~ 51 S P E C I F I C A T I 0 M page 1 1 Background of the Inven_ion
2 Thls invention relates to automatic scanning
3 apparatus, which in rapid sequence performs a series of
4 related densitometric or optical density tests on samples contained in a large number of wells in a tray. It also 6 relates to measurement o~ the susceptibility of bacteria 7 to different antimicrobic drugs, with automatic quan-ti-8 fication of the susceptibility to each drug, so that a 9 physician may select a drug that will most effectively treat an infecting bacterlum and choose the appropriate 11 dosage for effective treatment. It further relates to 12 the identification of microorganisms that have been 13 isolated from patients.
14 In the clinical laboratory, the -bacteriology department has two major functions: 1) the identifica-16 tion of organims that are isolated from patients and 17 2) the determination of the susceptibility of these 18 organisms to antimicrobic medication. Both of these 19 are involved here.
~l Identification of rliicroorganisms 22 Organism identification has generally been 23 accomplished by noting both the microscopic appearance 24 of the bacteria and their gross appearance (colonial 25 morphology) as they grow on a solid medium. In addi-26 tion to morphologic examination, a technologist sometimes 27 tested the organism with immunological techniques and 28 special stains to gain further information on the micro-29 organism's identity.
However, the most important technique for 31 bacterial identification relates to that organism's bio-32 chemical properties.
33 Each organism possesses a set of enzymes that 34 act as chemical catalysts or fermentors. By performing a 35 series of chemical reactions in a medium where an organism 36 is growing, a technologlst is able to identify a combi-37 nation of positive and negative reactions that effectively .,,-, ~
. - , . ~ , .. .... ~ , . - . -9L:13~71~
1 provide a chemical fingerprint for that organism. Typi~
2 cally, these reactions include fermentation of a wide 3 range of carbohydrates, citrate utilization, malonate 4 utilization, phenylalanine deaminase production, beta
5 galactosidase production, indole production, hydrogen
6 sulfide production, lysine decarbonxylase production,
7 ornithine decarboxylase production, urease production,
8 sucrose utilization, and arginine dehydroxylase pro-
9 duction. A reaction result is determined by a visual 1~ color change in the medium. The color reagent in most 11 cases is pH indicator which measures the alkalinity or 12 acidity resulting from the chemical reactions.
13 variety of indicators such as bromphenol bllle and phenol 14 red may be used to measure pH changes over a wide range 15 of the pH scale. Another mechanism for chemical color 16 development is the enzymatic splitting of a chromogen 17 (color producing chemical) off the original substrate, 18 thus signalling a positive chemical reaction.
19 A combination of color reactions as just 20 described forms a profile that may be used to identify 21 the organism. For this purpose, identification can pro-22 ceed in either a parallel mode in which a large number 23 of tests are performed at one time, or in a serial mode, 24 also known as a sequential or "branching" mode, in which subsequent tests are chosen on the basis of previous 26 results. The serial mode saves reagents since only those 27 tests are performed which will directly affect the final 28 results; however, it is quite time consuming, since each 29 subsequent test cannot be performed until the results from the preceding test have been obtained. In a medical 31 setting, each test may take 24 hours to obtain a result, 32 and where time is of the essence, the system may be far 33 too slow for practical application. Additionally, should 34 the technologist misread one result early in the decision tree, then all subsequent results could possibly be 36 misleading~
37 For these reasons, the parallel mode employing 1 the performance of a large number of biochemical tests on ~ bacteria is presently preferred by most workers. In the 3 parallel mode, large numbers of known organisms are tested 4 with a battery of biochemical substrates, and the proba-5 bilities of each of these bacterial taxa having a positive 6 reaction are tabulated. With this information, the 7 probability of each organism occuring for each combi-8 nation of chemical reactions can be computed by standard 9 statistical methods.
Systems presently available comTnercially pro-11 vide a convenient battery of biochemical substrates and 12 indicators to test a given organism. A "book" (typically 13 a computer printout) is provided to enable a technologist 14 to translate combinations of chemical reactions into the correspondingly most probable organism. These commercial 16 systems provide a plurality of microtubes (e.g., 15 to 17 30) which contain substrate indicators. The microtubes 18 are inoculated with the bacteria and, after an appropriate 19 time for the organism to grow and elaborate its enzymes, 20 the reactions can be read as a color change. Combinations 21 of these reactions can then be transposed into a unique 22 numerical code.
23 A standard way of transposing these reactions 24 into numbers involves reading the reactions in groups 25 of threes and e~pressing the reactions as an octal code.
26 This octal code ranges from 0-7 with 0 representing no 27 positive reaction and 7 indicating that all 3 reactions 28 were positive. This octal number or "biotype" can be 29 found in a book where the most probable organisms are 30 listed for each biotype.
31 With the present art it is necessary for a 32 technologist to read each reaction visually, record 33 each of the results, compute a biotype number, and then 34 find this biotype number in a computer printout, in order 35 to make the organism identification. Thus, although they 36 are much more convenient than the original serial branch-37 ing technique the present manual multi-test battery ~ 7~ ~
1 methods are still laborious and time-consuming.
3 Minimum inhibitory concentration 4 The physician usually has a choice of about 5 twelve to fifteen types of an~imicrobial agents for 6 treating the forty to sixty groups of pathogenic bacteria.
7 Many of these agents are ineffective against a given 8 bacterial strain, but normally some of them will be 9 appropriate for treatment. In order for the physician
10 to choose the best antimicrobic, it is necessary to iso-
11 late the pathogenic organism in the laboratory and then
12 test it against a panel of drugs to determine which drugs
13 inhibit growth and which do not. Ideally, the doctor
14 should receive susceptibility information the same day the
15 culture is taken, since it is usually necessary to
16 initiate therapy immediately. Unfortunately, it currently
17 takes one day to isolate an organism, and it has required
18 another day to test the susceptibility of the organism to
19 the antimicrobics. Therefore, it has been customary for
20 the physician to institute therapy based on an educated
21 guess at the time the patient is first seen. If the
22 sensitivity studies completed two days later indicate that
23 the guess was incorrect, therapy is changed to the proper
24 drug.
Clearly an important goal in automating anti-26 microbic testing would be to diminish the time lag between 27 the initial culture and the obtaining of sensitivity infor- ;
28 mation. An estimated 30 million antimicrobic susceptibil-29 ity tests are performed annually in the United States by 30 labor intensive manual methods. In addition to the 31 potential economic advantages of automation and obvious 32 advantages to the patient in receiving only the proper 33 treatment, one could also anticipate better precision, 34 quality control and objectivity.
The most frequently used technique to measure 36 antimicrobial susceptibility has been the standardized 37 disc-diffusion method described by Kirby and Bauer 38 (Bauer, Kirby et al, "~ntibiotic Susceptibility Testing . , , - : :. , ': ' '.: ' ' ' ' .. , ;: ' ' ' ', : :' ;: ' ~
~L3L3 ~ 5 --1 by a Standardized Single Disk Method", America _Journal of 2 Clinical Pathology, 1966, Vol. 45, No. 4, p. 493). By 3 this method, isolates of bacteria are grown in suspension 4 to a standardized concentration (usually determined by 5 visual turbidity) and streaked onto nutrient agar (culture 6 medium) in a f~at glass Petri dish. Paper discs impreg-7 nated with different anti-microbial materials are placed 8 upon the agar streaked with bacteria, and the drug is 9 allowed to diffuse through the agar, forming a gradient 10 halo around the disc. ~s the bacteria replicate, they 11 form a visible film on the surface of the agar 7 but in the 12 zones surrounding the antibiotic-impregnated discs, 13 growth is inhibited if the organism is susceptible to that 14 particular antimicrobial agent. Since a concentration gradient has been established, the zone of inhibition 16 around the disc is roughly proportional to the degree o~
17 susceptibility. Typically, the laboratory classifies an 18 organism as "sensitive", "intermediate", or "resistant"
19 to each drug in the test panel. Thus the results estab-20 lish a characteristic profile or "antibiogram" for that 21 organism.
22 The Kirby-Bauer disc-diffusion method has the 23 advantage of simplicity, but is suffers from several 24 drawbacks. One problem is that of time efficiency. In
25 order that the initial inoculum become visible on the
26 Petri dish so that zones of growth can be distinguished
27 from zones of inhibition, the bacteria numbers must
28 increase by several orders of magnitude over the original
29 number. ~owever, for determination of whether or not the
30 organism is growing in theantimicrobial milieu, which is
31 the only information required, a period that would allow
32 doubling of all the organisms should be theoretically
33 sufficient with suitable detection equipment. For most
34 Gram-negati~e organisms, the doubling period is between
35 twenty and thirty minutes, following a lag phase. There-
36 fore, an automated system should be able to distinguish
37 growth within a thirty-minute period.
38 ~L~3'~7~J6 1 Another difficulty with the Kirby-Bauer disc 2 method is that of standardization. If an organism is 3 "resistantl', does that mean that it cannot be treated 4 with higher than normal doses of the microbial agent?
5 Also, how does this information relate to a site in the 6 body where the antimicrobic is concentrated (such as 7 bile) or decreased in amount (such as cerebrospinal 8 fluid)?
9 To answer these questions, quantitative data 10 are necessary. To obtain quantitative results, it m~lst 11 be determined what minimum concentration o a drug will 12 inhibit the organism's growth. This quantitation of 13 susceptibility is known as minimum inhibitory concentra-14 tion or MIC. The MIC may be determined by making serial 15 dilutions o~ the drug in agar or broth, and then inocu-16 lating each dilution of each drug with a standardized 17 suspension of bacteria. Since the test procedure may 18 involve as many as 70 to 80 individual tubes, it can 19 become a formidable task if the test is performed in 20 individual test tubes on a macro scale. Systems are 21 available in which the individual dilutions of anti-22 microbics are made in plastic trays containing small 23 micro-tubes. (March and MacLowry, "Semiautomatic 24 Serial-Dilution Test for Antibiotic Susceptibility", 25 Automation and Data Processing the the Clinical Labora-26 ~2~ Springfield, Illinois, C. C. Thomas 1970). Orga-27 nisms can be inoculated in a single step using a multi-28 pronged template. Thus, setting up the test is simpli-29 fied, and it takes slightly less time to provide 30 quantitative data than qualitative Kirby-Bauer infor-31 mation. There are now semiautomated devices that dispense 32 antimicrobial solutions into the microtubes. Trays of 33 microtubes are also commercially available with frozen 34 solutions in the tubes, and the Gram-negative anti-35 microbial panels have been combined with biochemical tests 36 to identify enteric bacteria as well as to determine 37 their antimicrobic susceptibility.
7E~;
1 Although MIC results give quanti~ative infor-2 mation which allows consideration of multiple doses and 3 multiple sites, the MIC numbers in themselves can be 4 confusing to the clinician. To use MIC data correctly, 5 a physician must refer to tables of achievable anti-6 microbic levels as a function of dosage and body site.
7 Therapy will be effective if the achievable drug level 8 for a particular dose and site in the body is two to 9 four times the MIC. With the present invention 10 described ~elow, such interpretation of MIC data is 11 accomplished by a computer, which compares the MIC
12 with a table of achievable drug levels at different 13 body sites and different doses.
15 Optical testing methods and ~pparatus 16 Optical detection methods have been suggested 17 and have proven to be powerful tools to measure bacterial 18 growth. A laser light-scattering system can have the 19 sensitivity to detect a single bacterium. Optical 20 methods measure the presence of bacteria either by 21 nephelometry or turbidity measurements. Nephelometry 22 measures the ability of the bacteria particles to scatter 23 light, and the detector is aligned at an angle to the 24 axis of the light source. Turbidity measures the net 25 effect of absorbance and scatter, and the transducer 26 is placed on the axis of the radiation source. Nephelo-27 metry measurements are significantly more sensitive than 28 turbidity measurements, but since the nephelometer 29 measures only that fraction of light scattered by 30 bacteria, the signal to the detector is small, and both 31 light source and transducer amplification must be corres-32 pondingly large.
33 Some apparatus heretofore relied direct inspec-34 tion by the human eye, as did Astle in U.S. Patent No.
3,713,985. Aware of inaccuracies involved, Astle sug-36 gested, but did not disclose details of automatic equip-37 ment for reading and recording the results of densitometric 3~
.; . ~ : :
1 tests and suggested tl~at his tur'bidity data could be fed 2 to a device that would translate the data to machine 3 language for ~ecording on computer punch cards. Astle 4 does not teach how to do that. His own device is a strip 5 having a series of wells, all in a single line which 6 involve mechanical position shifts.
7 Automation in microbio'Logy has lagged far behind 8 chemistry and hematology in the clinical laboratory.
9 However, there is presently an intensive e~ort by indus-10 try to develop this field. The best publicized devices 11 for performing automated antimicrobic susceptibility 12 testing use optical detecti.on methods. A continuous 13 flow device for detectin~ particles 0.5 micron or less 14 has been commercially available since 1970; however, 15 probably due to its great expense, it has not been widely 16 used in the laboratory. Other devices using laser li~ht 17 sources have been suggested but have not proven com-18 mercially practicable. Recently, the most attention has 19 been directed to three devices discussed below.
The Pfizer ~utobac 1 system (U~S. Patent No.
21 RE. 28,801) measures relative bacterial growth by light 22 scatter at a fixed 35 angle. It includes twelve test 23 chambers and one control chamber in a plastic device 24 that forms multiple contiguous cuvettes. Antibiotics are 25 introduced to the chambers via impregnated paper discs.
26 The antimicrobîc sensitivity reader comes with an 27 incubator, shaker, and disc dispenser. Results are 28 expressed as a light scattering index (LSI~, and these 29 numbers are related to the Kirby-~auer "sensitive, 30 intermediate and resistant''. MIC measurements are not 31 available routinely with this instrument. In a compari-32 son with susceptibilities of clinical isolates measured 33 by the Kirby-Bauer method, there was 91~/~ agreement.
34 However, with this system some bacteria strain-drug ~-combinations have been found to produce a resistant 36 Kirby-Bauer zone dlameter and at the same time a sensi- '-' 37 tive'LSI.
~'7~
g 1 The Auto Microbic System has been developed by 2 McDonnell-Douglas to perform identification, enumeration 3 and susceptibility studies on nine urinary tract patho-4 gens using a plastic plate containing a 4 x 5 array of 5 wells. See &ibson et al~ U.S. Patent No. 3,957,583;
6 Charles et al, U.S. Patent No. 4,118,280, and Charles 7 et al, U.S. Patent No. 4,116,775. The specimen is 8 drawn into the small wells by negative pressure and the 9 instrument monitors the change in optical absorbance and 10 scatter with light-emitting diodes and an array of 11 optical sensors. A mechanical device moves each plate 12 into a sensing slot in a continuous succession so that 13 each plate is scanned once an hour, and an onboard digi-14 tal computer stores the optical data. The system will 15 process either 120 or 240 specimens at a time. One can 16 query the status of each test via a CRT-keyboard console, 17 and hard copy can be made from any display. When the 18 system detects sufficient bacterial growth to permit a 19 valid result, it automatically triggers a print-out.
20 Following identification in four to thirteen hours, a 21 technologist transfers positive cultures to another 22 system which tests for antimicrobic susceptibility.
23 The results are expressed as "R" (resistant) and "S"
24 (susceptible); no quantitative MIC data are provided.
It should be noted that Gibson et al, U.S.
~6 Patent No. 3,957,583 do not include automation, but 27 use naked-eye inspection or a manually-operated color-28 imeter. Scanning is therefore a hand or a mechanical 29 operation. Charles et al, Patents Nos. 4,116,775 and 30 4,118,280 also require mechanical movement of their 31 cassette for reading different rows.
32 The Abbott MS-2 system consists of chambers 33 composed of eleven contiguous cuvettes. Similar to the 34 Pfzier Autobac 1, the antimicrobial compounds are intro-35 duced by way of impregnated paper discs. An inoculum 36 consisting of a suspension of organisms from several 37 colonies is introduced into the culture medium, and the . .- -. . -. - . - ~ . . , . .
, : ~
L3~
1 cuvette cartridge is filled with this suspension. The 2 operator inserts the cuvette cartridge into an analysis 3 module which will handle eight cartridges (additional 4 modules can be added to the system). Following agitation 5 of the cartridge, the instrument monitors the growth rate 6 by turbidimentry. When the log growth phase occurs, the 7 system automatically transfers the broth solution to the 8 eleven cuvette chambers; ten of these chambers contain 9 antimicrobial discs, and the eleventh is a growth control.
10 The device performs readings at five-minute intervals, 11 and stores the data in a microprocessor. Following a 12 pre-set increase of turbidity of the growth control, 13 the processor establishes a growth rate constant for each 14 chamber. A comparison of the antimicrobic growth rate 15 constant and control growth rate constant forms the basis 16 of susceptibility calculations. The print,out presents 17 results as either resistant or susceptible; if inter-18 mediate, susceptibility information is expressed as an 19 MIC.
Non-optical methods have also been used or 21 suggested for measuring antimicrobic sensiti~ity in 22 susceptibility testing. These have included radio-23 respirometry, electrical impedance, bioluminescence and 24 microcalorimetry. Radiorespirometry, based on the 25 principLe that bacteria metabolized carbohydrate and 26 the carbohydrate carbon may be detected following its 27 release as C02, involves the incorporati~n of the isotope 28 C14 into carbohydrates. Released C1402 gas is trapped 29 and beta counting techniques are used to detect the 30 isotope. The major difficulty in applying the isotope 31 detection system to susceptibility testing, however, is 32 that an antimicrobic agent may be able to stop growth of 33 a species of bacteria, yet metabolism of carbohydrate 34 may continue. Less likely, a given drug may turn off 35 the metabolic machinery that metabolizes certain carbo-36 hydrates, but growth may continue. This dissociation 37 between metabolism and cell growth emphasizes the fact ~L~3~
1 that measurements for detecting antimicrobic susceptibility 2 should depend upon a determination of cell mass or cell 3 number rather than metabolism, 4 The electrical impedance system is based on the fact that bacterial cells have a low net charge 6 and higher electrical impedance than the surrounding 7 electrolytic bacterial growth media. A pulse impedance 8 cell-counting device can be used to count the cells;
9 however, available counting devices are not designed 10 to handle batches of samples automatically, and generally 11 do not have the capacity ~o distinguish between live 12 and dead bacterial cells. Another approach with electri~
13 cal impedance has been to monitor the change in the 14 conductivity of the media during the growth phase of 15 bacteria. As bacteria utilize the nutrients, they pro-16 duce metabolites which have a greater degree of electri-17 cal conductance than the native broth, so that as 18 metabolism occurs, impedance decreases. However, since 19 this technique measures cell metabolism rather than 20 cell mass, its applicability to antimicrobic suscepti-21 bility detection suffers from the same drawback as 22 radiorespirometry.
23 Bioluminescence has also been sugges~ed for the 24 detection of microorganisms. It is based on the princi-25 ple that a nearly universal property of living organisms 26 is the storage of energy in the form of high energy 27 phosphates (adenosine triphosphate, ATP), which can be 28 detected through reaction with firefly luciferase. The 29 reaction results in the emission of light energy which 30 can be detected with great sensitivity by electronic 31 light ~ransducers. Although a clinical laboratory may 32 obtain a bioluminescence system to detect the presence 33 of bacteria in urine, the technique is expensive due to 34 the limited availability of firefly luciferase, and 35 problems have been enco~mtered in standardizing the 36 system.
37 Microcalorimetry is the measurement of minute "~ ;, "" ~ , "
` ~37'~
amounts of heat generated by bacterial metabolism. The principle exhibits certain advantages, but laboratories have not adopted such a system, one serious drawback being that the system measures metabolic activity rather than bacterial mass or number.
Summary of the Invention The present invention employs optical methods and apparatus for automatically identifying microorganisms and auto-matically determining bacterial susceptibility to a number of different antimicrobic drugs, utilizing turbidimetry.
The invention provides apparatus for identifying a micro-organism, employing a sample tray having a series of wells for con-taining uniform samples of microorganism culture and a reagent, said wells having translucent bottoms, comprising:
tray holding means for holding said tray accurately in a predetermined position without blocking off light paths -through said wells, a single diffused light source means positioned above the sample tray, for sending light down through all said wells at approximately uniform intensity, collimation means beneath said tray holding means for collimating the light from each well after it has passed through the wells, light filter means below said tray holding means for filtering the color values of the light passing through the wells, an array of light-intensity-detecting photocells on the opposite side of the filter means from the tray holding means, one adjacent to each well and positioned to receive light from the light source which has been transmitted through the well and its .~ - 12 -~IL3~7&3~
contents, a reference detecting photocell for receiving light directly from said single diffused light source means without pass-ing through a said tray, sequential signal receivi.ng means connected to all the photocells for receiving sequentially a signal from each said photocell in a prescribed order, each signal corresponding to the intensity of light transmitted through the ad~acent well and thus to the opacity of the contents of t.he well, electronic sequencing means connected to said signal receiving means for electronically causing it to receive its signals in order, first comparator means connected to said signal receiviny means, for sequentially comparing the signal from each said photo-cell of said array with the signal from said reference detecting photocell and developing a difference signal therefrom, data storage means for holding data values corresponding to inhibited growth and for holding data relating to various organl sms, second comparator means connected to said first com-parator means and to said data storage means for sequentially making a comparison of each said difference signal value with a value corresponding to inhibited growth and developing a resultant value from that comparison, third comparator means connected to said second com-parator means and to said data storage means for sequentially comparing said resultant values with a large number of stored values and for determining the probability values for the presence .. ~.. j ~ . . , .- - ~, ~ : . . ;:
~3'7~7~
of selected orcJanisms in the sample, and output means connected to said third comparator means for giving the results obtained by said third comparator means.
The invention, from another aspect, provides a method of identifying microorganisms, comprising:
placiny a series of different reagents in series of wells in a light-transmissive sample tray, establishing a known uniform concentration o~ a culture of the microorganism and placing the uniform concentration in equal volumes in the wells, following a predetermined period for bacterial growth, passing light from a single light source in substantially equal intensity through all said wells and through a color filter and collimator, according to the opacity value for each well, automatically sensing the intensity of the collimated light transmitted through each well by photodetector means adjacent to the wells and filter and opposite the light source, automatically and sequentially comparing the opacity values for each well with an opacity for light from the same source not passing through any well but passing through the filter, and for generating a signal from such comparisonl automatically and sequentially comparing that signal with a value corresponding to inhibited reaction for each well, and automatically and sequential~y comparing the opacity values from different tests to obtain probability values for various suspected organisms.
The invention also provides a method for performing optical density tests, employing a sample tray having a series of - 13a -~3~'7~6i wells, said wells having translucent bottoms, comprising:
holding said tray, with the wells empty, accurately in a single predetermined reading poisition without blocking off light paths through said wells, sending light from a single light source down through all said wells at roughly the same intensity to an array of light-intensity-detecting photocells, there being one photocell adjacent to each well, sending light directly from said light source means to a reference detecting photocell without passing the light through the tray, electronically sequentially transmitting the signal from all said photocells in a prescribed order, each signal correspond-ing to the intensity of light received by a said photocell, sequentially comparing the signal from each said photo-cell of said array with the signal from said reference detecting photocell and developing a first related signal therefrom for each well, storing said first related signals, filling the wells with liquid samples, culturing said liquid for a predetermined length of time, holding said filled tray, after culture, accuratsly in said single predetermined reading position without blocking off light paths through said wells, sending light from said single light source down through all said filled wells at roughly the same intensity to said array of light-intensity-detecting photocells, there being one photocell adjacent to each well, - 13b -7~i6 sending light directly from said light source means to said reference detecting photocell without passing the light through a said sample, electronically sequentially transmitting the signals for filled wells from all said photocells in a prescribed order, each signal corresponding to the intensi.ty of light received by a said photocell, sequentially comparing the signal for the filled wells from each said photocell of said array with the signal from said reference detecting photocell and developing a second related signal therefrom for each well, sequentially making a comparison of each said second related signal value with the corresponding Eirst related signal value for the same well, and developing a resultant value from that comparison, se~uentially comparing said resultant values with other stored values and for determining a desired result therefrom, and reading out the desired results thereby obtained.
The invention also provides a method for determining susceptibility of a bacteria culture to various antimicrobic drugs and of determining the minimum inhibitory concentration of the bacteria culture to those drugs to which it is susceptible, com-: prising:
providing a sample tray having a series of light-trans-missive wells, and a series of photodetectors, including a reference photodetector not associated with a tray well, each photodetector being adapted to provide a signal corresponding to the sensed light intensity, - 13c - .`
7~
initially calibrating the photodetectors by passing light from a source of generally uniform intensity over the photo-detectors and electronically sequencing the photodetectors to read a signal from each photodetector, comparing the values of the signals obtained for each well-associated photodetector sequential-ly with the value of the reference signal obtained for the refer-ence photodetector, and providing an initial calibration value for each well-associated photodetector which is a function of the well photodetector signal and the reference signal, and storing and retaining the calibration value for each well-associated photo-detector, placing in the wells a plurality of different antimicrobic drugs, each drug being included in a series of wells in serially diluted known concentrations, and samples of equal volumes of a known uniform concentration of the bacteria, with the wells adjacent to the well-associated photodetectors, following a period for bacterial growth, passing light of generally uniform intensity simultaneouialy through each well and to the reference photodetector and reading the intensity of the trans-; 20 mitted light with the photodetectors by electronically sequencing the photodetectors by electronically sequencing the photodetectors to read an after culture signal from each, comparing the values of the after culture signals obtained from the well-associated photodetectors with the value of the after culture signal from the reference photodetector and pro-viding an after culture value for each well-associated photo-detector which is a function of the after culture well photo-detector signal and the after culture reference signal, 13d -comparing, for each well, the after culture value with the initial calibration value and providing a comparison signal for each well which allows for variations in the intensity of the light directed from the source onto the clifferent wells and for varia-tions in the sensitivities of the photocells, automatically and sequentially comparing each comparison signal value with a limit comparison signal value which represents a cutoff between inhibition and growth, correlating the comparisons with stored data identi~ying the antimicrobic drug and concentra-tion in each well, and obtaining therefrom an indication of which antimicrobic drugs inhibit growth of the bacteria, and automatical-ly selecting the minimum inhibitory concentration of each inhibitory drug by selecting the minimum concentration of each drug which produced a comparison signal value on the inhibition side of the limit comparison signal value, and automatically displaying the minimum inhibitory concen-tration for each inhibitory drug, and for each drug that does not inhibit growth, displaying that the bacteria is resistant to that drug.
The in~ention further provi~es a method of identifying microorganisms, comprising:
placing a series of different reagents, including one for determining whether an organism is a dextrose fermenter in a series of we.~ls in a light-transmissive sample tray, establishing a known uniform concentration of a culture of the microorganism and placing the uniform concentration in equal volumes in the wells, following a period for bacterial growth, passing light ~.~
~;~ - 13e -~l3~ 6 in substantially equal intensity through each well and through appropriate color filter means to determine an opacity value for each well by automatically sensing the intensity of the light transmitted through each well with photodetector means adjacent to the wells and to the filter opposite the light source, comparing the obtained opacity values with a value corre-sponding to zero reaction, determining from the opacity ~alue from the dextrose reagent well whether the microorganism is a dextrose fermenter, depending on whether the microorganism is or is not so found to be a dextrose fermenter, comparing the opacity values from appropriate wells to tables of values indicative of microorganism to obtain probability values for various suspected microorganisms, cumulatively multiplying each such probability value by other probabilities for a given taxon to give the non-normalized frequency for each taxon, adding the non-normalized frequencies for all such taxa, determining the three most probable taxa fxom said non-normalized frequencies, determining from the values for the three most probable taxa whether any of them have a frequency greater than 1 x 10 6 and, if they do, normalizing the frequencies o the three most pro~able organisms, and indicating the normalized frequencies of the most probable microorganism and the identity of that microorganism and the probability percentage, if the probability is greater than 75%.
The pre.sent invention makes it possible to use an optical-~à~ - 13f -electrical method for automatically .readinc3 the color changes of a plurality of bicchemical reactions in small microtubes and for calculating and printing out the most probable organism by means of an inboard computer and probability data stored in the computer memory.
The microtubes are, preferably, all part of a unitary sample tray, made of suitable translucent material. Each micro-tube is a well of this tray. In each well and in a standardized manner, is placed a suitable chemical reagent ~' - 13g -~:3~t71~
1 or reagents; then each well is inoculated with the sample.
2 Photodetection o~ color changes is accomplished by the 3 passage of uni~orm intensity light through each of the 4 wells and through the translucent well bottoms following 5 an incubation period. At the opposite side of the tray, 6 preferably below the tray, is an optical filter designed 7 to pass only certain wavelengths of light. Beneath the 8 filter is an array of sequentially-scanned transd~tcers 9 such as photoelectric cells, one associated with each 10 well. The optical filter is designed so that a shift in 11 color in the wells will result in a predictably greater 12 or lesser amo~nt of light passing through to the photo-13 electric cells.
1~ Previously, the reading of an identiEication system required a technologist manua].ly to record visual 16 impressions oE color changes indicating either positive or 17 negative biochemical reactions generated by the enzymes 18 contained in the bacteria to be tested. The apparatus 19 in the present invention provides this reading auto-20 matically and objectively. With present manual methods 21 and apparatus, after the reactions had been determined, 22 it was necessary for the technologist to calculate a 23 numerical summation of these reactions and to express 24 them as an octal number or "biotype". This biotype 25 number was then searched out in a large book containing 26 various biotypes and corresponding organism probabilities.
27 Once a biotype was found, the most probable organism 28 was noted and reported. With the present invention, the 29 computer which is an integral part of the instrument, 30 computes the probability for each organism and prints 31 out the identification on a laboratory form.
32 Signals from the transducers (photoelectric 33 cells) are transmitted to a computer which contains an 34 algorithm that transforms the reaction results to 35 organism identification. The following description 36 presents in detail the algorithm used by the computer 37 to convert the reaction colors to organism identification.
~9 3~
:....... , - . , , .::~ .~:
- i . . ...
~37'~
1 This algorithm is also summarized in the accompanying 2 flow diagram.
3 For each biochemical reaction, a voltage value, 4 which discriminates between a positive and negative 5 result, is or has already been determined by experi-6 mentation. Each of these "cutoff" points is stored in 7 the computer's memory together with a module that indi-8 cates if a given value above that point is negative or 9 positive.
The computer is programmed to compile a table 11 of the probability of occurrence for each biochemical 12 reaction with each of the organisms (taxa) in the data 13 base. This probability assumes a positive reaction.
14 If, in act, a negative reaction occurs, then the proba-15 bility of the observa~ion would be l.000-P. For example, 16 i a given biochemical with a given organism has a proba-17 bility of 0.005 of occurring, and the reaction was found 18 to be negative, then the probability would be .995 (1.000-19 .005) that this reaction would not occur. So the program 20 at this point calls for converting all the negative 21 probabilities to l.000-P values for the table. The 22 positive reactions are left unchanged, and are manipulated 23 exactly as they occur in the table.
24 In addition to printing out the most probable ;
25 organism, the instrument provides the operator wlth 26 several indices of reliability. First, the overall non- ;
27 normalized probabllity of the reaction is computed. If 28 this probability is very low, this may mean that there 29 was an error in reading or that the suspension of test 30 bacteria contained more than one taxon. Second, the 31 relative normalized probabilities of three most likely 32 organisms are computed and displayed to the technologist.
33 Clearly, if several organisms have equal probability of 3~ occurring with a given set of biochemical reactions, 35 further testing is necessary to discriminate between 3~ them. Third, the instrument measures the susceptibility 37 of the test organism to several antimicrobics. If the ~'~31~78 1 known susceptibility is in conflict with the identification 2 by biochemicals, a warning is given to the operator.
3 Thus, after the individually observed proba-4 bilities have been determined, each of the biochemical 5 probabilities is cumulatively multiplies by the other 6 probabilities for a given taxon. For example, the 7 observed probability (P) for the organism, E. Coli, 8 with dextrose is multiplied by the P for sucrose, and 9 this product is multiplied by the P of sorbitol, and 10 so on. This continues until a p:roduct of, for example, 11 twenty-one multiplica~ions is obtained for each organism.
12 Each of these products is the non-normalized frequency Eor 13 each taxon. As these non-normalized frequencies are 14 being computed, they are added to each other, so that a sum of all of the non-normalized frequencies for each 16 organism is obtained.
17 Rare combinations of biochemical reactions can 18 occur with organisms, but more commonly, a very low 19 frequency will indicate a technical error. The most 20 common technical errors are due either to a mixed culture ~1 or to areading error. The instrument software is 22 designed so that an organism frequency (non-normalized) 23 of less than 1 x 10- 6 Will be read out as unacceptable.
24 If the organism with the greatest frequency is computed 25 to have a frequency of less than this value, the display 26 indicates: "VERY RARE BIOTYPE", and the program goes 27 back to the beginning. If the first organism frequency 28 is greater than 1 x 10- 6 but less than or equal to 1 x 10- 5, 29 the display says: "RARE BIOTYPE~PRINT? (1 or 0)". If the 30 operator wishes to go ahead and print, then he presses 31 "1" on the keyboard; if he wishes to go back to the main 32 program, then he presses "0". The instrument waits for 33 either of these keys to be pressed.
34 Normalization is accomplished by dividing each 35 of the three highest frequencies by the sum of all of the 36 frequencies. If the most probable organism has a normal-37 ized frequency between .950 and .999, then the display ... . . . . . . .
.. ~ .
1 shows "MOST PROBABLE--XX.~%". In this case, the proba-2 bility is converted to a percent figure. The program 3 then returns to check the dextrose fermenter flag. If 4 the organism is a dextrose fermenter, then the program 5 goes on to print the name o~ the most probable organism 6 and the biotype in appropriate spaces on the form. If 7 it is a non-fermenter it is compared with Colisti.n and 8 nitrofurantoin (Furadantin) results, as outlined below.
9 If the most probable organism has a normalized frequency 10 between .850 and .950, then the display indicates 11 "VERY PRORABLE--XX.X%". The program again checks for 12 fermenter or non-fermenter status as above. If the 13 relative (normalized) frequency is between .750 and 14 .850, the display indicates: "PROBABLE--X~.X%" and 15 loops through the fermenter/non-fermenter check as 16 above. If the relative probability is less than .750, 17 the display outputs three messages in sequence at one-18 second intervals: "LOW SEJ.ECTIVITY-RECHECK"; followed 19 by "000000000000000000000--XX.X%" where 000 is the 20 organism name, and XX.X is the percentage as above.
2~ The third display is "STILL WAMT TO PRINT? (1 or 0)".
22 As stated above, if the organism is a non-23 fermenter, the instrument also measures the suscepti-24 bility of the test organism to several antimicrobics.
25 Thus, identification may be evaluated for its sensitivity 26 to the two antibiotics Colistin and Nitrofurantoin. If 27 there is growth in these wells (hex voltage less than 28 threshold), then this means the organisms are resistant 29 ("R"). If there is no growth (hex voltage greater than 30 threshold~, then the organism is sensitive or "S".
31 Once the sensititvity or resistance for Colistin 32 has been determined, the program looks up a table to see 33 if the result is correct; if not, t~en it displays on 3~ the visual display: "REC~ECK I.D. & COLISTIN DISAGREE";
35 the most probable organism is then printed out, and the 36 routine returns to the main program. If the table and 37 results agree with Colistin, then a similar procedure is ' .' , . ' , ' ':
3-~t~
1 performed with Nitrofurantoin. If there is disagreement, 2 the display says: "RRCHECK-ID & FURAMTOIN DISAGREE", If 3 there is agreement, then the result is printed out as 4 above.
In the method for determining bacterial suscep-6 tibility to various antimicrobic drugs, the system o the 7 invention uses broth-dilution to determine susceptibility.
8 Serial dilutions of the antimicrobic agent are inoculated 9 with the organism and incubated for a period sufficient lO to allow detectable growth. The apparatus of the inven-11 tion determines minimum inhibitory concentration (MIC) 12 of a particular antimicrobic drug, which is the lowest 13 concentration of that drug that results in no detectable 14 bacterial growth. Typically, ten antimicrobic drugs are evaluated, with seven different dilutions of each 16 drug being tested. Therefore, to obtain an MIC deter-17 mination for ten drugs, seventy tubes or wells must be 18 inoculated and examined. In contrast to previous methods 19 using individual full-sized test tubes, which were 20 cumbersome and expensive, the present system utilizes 21 "micro-tubes", which are presently available as dis-22 posable, molded plastic trays, each well of which holds23 approximately 0.5 milliliter.
24 For measurement of the MIC values in these 25 trays, appropriate dilutions of each antibiotic must be 26 placed in the wells or micro-tubes. Semiautomated 27 devices for making the dilutions and filling the trays 28 in large batches are available commercially. Alterna-29 tively, a laboratory may obtain trays that are already 30 filled with antibiotic dilutions and kept frozen until 31 use. To prepare the bacteria cultures for inoculation 32 into the wells, a suspension of the bacterial organisms 33 in water is made in a container. By means of a multiple-34 pronged device, a technician is able to inoculate a 35 uniform drop of bacterial suspension into each of the 36 large plurality (e.g., seventy) of wells with a single 37 motion. The bacteria and the various dilutions of the " ~
~ 3~
~3~7~i l antimicrobic agents are incubated for Q time period suf-2 ficient to produce detectable bacterial growth, and the 3 MIC may then be determined as the lowest concentration 4 of the effective antimicrobic agents in which -there is 5 no evidence of growth.
6 Previously the reading of such an ~IIC tray was 7 done by manual viewing performed by a technician, and was 8 a laborious procedure. An overnight incubation period was 9 generally required in order to produce visually detectable 10 patterns of growth. However, the apparatus and method of 11 the present invention provided for the performance of the 12 reading and interpretive task automatically. Moreover, 13 the device has the capability of interpolating the MIC
14 between twofold dilutions, whereas by visual reading a technician can only detect the difference between growth 16 and no growth and thus can only read MIC to the nearest 17 twofold dilution. With the sensltive photoelectric appa-18 ratus described herein, together with the capabilities of 19 a microcomputer the different gradations of growth can be 20 measured even after a relatively short incubation period, 21 and a precise MIC can be calculated and displayed on a 22 screen or printed out. Thus, the device makes available 23 contînuous numerical data that improves accuracy and allows 24 quantitative quality-control techniques.
Photodetection of bacterial growth is accom-26 plished by passage of uniform intensity light throu~h each 27 of the wells and through the translucent well bottoms 28 following the incubation period. The uniform light may 29 be obtained from plural uniform sources, one at each well, 30 or by a single source of uniform, diffused light over the 31 entire tray. At the opposite side of the tray, preferably 32 below the tray, are an array of sequentially-scanned photo-33 electric cells, one associated with each well. The sensed 34 light intensity level at each well is compared by computer 35 with a light level corresponding to zero bacteria growth 36 to determine a relative value of turbidity. The reference 37 value may be obtained by the reading of a sterile control 38 well.
- . :, ~.~ 3~7~6 1 In addition to the quantitative MIC data, 2 the apparatus and method of the invention provide a 3 graphic interpretive printout to guide the physician's 4 therapy. The computer is programmed to translate the 5 MIC value into dosage ranges that would be necessary to 6 achieve blood levels of the antimicrobic drug effective 7 to inhibit growth of the organism at a particular site.
8 For example, a printout of "-" might be used to indicate 9 that the organism is resistant and no dosage of a drug 10 can effect the organism. A printout of "+" may be used 11 to mean that the organism is resistant and may respond 12 to high intramuscular or intravenous doses, with "-~t"
13 indicating intermediate sensitivity and that the 14 organism may respond to higher than recommended doses.
15 ~ printout of "+++" would indicate that the organism 16 may be sensitive to the usual doses o~ an antibiotic, 17 and "++++" would indicate a high degree of sensitivity 18 and thus an optimal drug with which to treat the infec-19 tious agent.
In one embodiment, a method according to the 21 invention for determining susceptibility of a bacteria 22 culture to various antimicrobic drugs and of determining 23 the minimum inhibitory concentration (MIC) of the bacteria 24 culture to those drugs to which it is susceptible com-25 prises the steps of placing the plurality of different 26 antimicrobic drugs in a plurality of wells in a light-27 transmissive tray, each drug being included in a series 28 of wells in serially-diluted known concentrations;
29 establishing a known uniform concentration of the bacteria 30 and placing the uniform concentration in equal volumes of 31 the wells: following an incubation period, passing light 32 in substantially equal intensity through each well and 33 determining a turbidity value for the bacterial suspension 34 of each well by sequentially sensing the intensity of 35 light transmitted through the bacterial suspensions of the 36 wells by means of photodetectors adjacent to the wells 37 opposite the light source; and in a computer, comparing ~L~3'~71~
1 turbidity values with a turbidity value corresponding 2 to zero bacterial growth, thereby determining which anti-3 microbic drugs have inhibited bacterial growth and the 4 minimum concentration of each inhibitory drug required 5 to inhibit growth, and displaying the determined infor-6 mation. The concentration of the bacteria culture may 7 itself be initially determined by turbidimetric measure-8 ment utilizing a light source and at least one photo-9 detector. The antimicrobic drugs may be placed in the 10 tray in a rectangular matrix of wells, with each column 11 of wells containing incrementally varying concentrations 12 of a single drug. Of course, any arrangement of the wells 13 or of the antimicrobics in the wells is suitable, so long lh as the computer has the proper in~ormation as to what ls 15 being tested in each well. Control wells containing only 16 the bacterial suspension, as well as sterile control 17 wells, may be included for self-checking of the system 18 and/or providing a transmitted light value corresponding 19 to zero bacteria growth. The system may, as explained 20 above, provide for translation of the MIC values to 21 dosage ranges necessary to establish the required anti-22 microbic concentration at the body sites involved.
23 There are other applications for the instru-24 ment. For example, heretofore, very sensitive techniques 25 have used bacteria as biological indicators for detecting 26 trace amounts of chemicals. Strains of bacteria are ~7 obtained by mutation that manifest growth that is ~ -28 directly proportional to the quantity of a given sub-29 stance so that calibration curves are easily made. Such 30 a substance and the amount thereof can therefore be 31 detected in a cultured sample by using the apparatus of 32 the lnvention, preferably using optical filters. Suitable 33 programs can, of course, be preferred.
34 Another example is the instrument's applica-35 bility to the technique of enzyme-linked immuno absorbent 36 assay (E~ISA) to detect the presence of a specific 37 species of protein molecules (e.g. bacteria, virus, or ..... . . . . .
;. ~ .~ ., ,. . - ' `,' :
.. . . ... .. .
,................... . . ... . . . . .
.:
. ~ - .. . ..
L3~ 6 l hormone) which is detected by its combina~ion with an 2 antibody. An antigen-antibody reaction is detected 3 (in this technique) by a color change caused by an 4 enzyme or enzymes and detected by the instrument.
S
6 Brief Description of the Drawings ; 7 In the drawings:
9 Fig. 1 is a perspective view showing an auto-10 mated apparatus embodying the principles of the invention.
12 Fig. 2 is an exploded perspective view showing 13 a sample tray and optical detection equipment forming a 14 part o:E the apparatus of Fig. 1.
16 Fig. 3 is a schematic sectional elevational 17 view showing a portion of the apparatus of Fig. 1.
19 Fig. 4 is a block diagram of the apparatus of 20 Fig. 1.
22 Fig. 5 is a block diagram of an analog-to-23 digital converter subsystem usable in the apparatus 24 of Fîgs. 1-4.
26 Fig. 6 is a block diagram of a microcomputer 27 portion of the apparatus.
29 Fig. 7A, 7B, and 7C are flow charts of 30 operational steps involved in the method of determining 31 minimum inhibitory concentration.
33 Fig. 8 is a schematic elevation view showing an 34 alternati~e~form of optical detection apparatus which may 35 be included in the apparatus of the i.nvention.
37 Fig. 9 shows a form of printout which may be ~`' 38 ,. ., . . .. . ,. ~
' ~ ' , " . ~ I . '. . . ' " ., ' `, ' , ., ' `' . ' " ., ` ' ' ' ' ' ", '` ' ~ .
3~
1 utilized in connection with the apparatus of the invention 2 when determining minimum inhibitory concentration.
4 Figs. lOA, lOB, and lOC (which are on the same S sheet as Fig. 1) are a set of three spectral absorbent 6 responses for three respective optical filters that may 7 be used in the invention.
9 Figs. llA, llB, and llC are flow charts of 10 operational steps involved in the method for identifying 11 microorganisms according to the principles of the 12 invention.
14 Description of the Preferred Embodiments 15 The apparatus of Figs. 1-6:
16 Fig. 1 shows one example of an external con-17 figuration which the susceptibility testing apparatus 10 18 of the invention may take. The unit 10 comprises a photo 19 unit or optical detection unit 11 and a processor unit 12.
20 The optical detection unit 11 preferably includes a 21 drawer 13 for receiving, supporting, and correctly posi-22 tioning a sample tray 14 which is examined ~y detection 23 apparatus of the unit 11 when the drawer is closed and 24 the testing operation is begun. The detection unit 11 may 25 also include a patient identification input switch 16, 26 a run switch 17 and a calibrate switch 18. The processor 27 unit 12 may include a readout display 19 J an on/off 28 power switch 21, printer control buttons 22, and a 29 printout exit 23 which dispenses a printed "ticke-t" 24 30 bearing the desired susceptibility information.
31 Fig. 2 somewhat schematically represents the 32 configuration of the detection apparatus associated 33 with the optical detection unit 11 of the apparatus 10.
34 Within the detection unit 11 above the drawer 13 is a 35 source of uniform, diffuse light which may comprise, 36 for example, a fluorescent light bulb 26, a parabolic 37 reflector 27 positioned thereabove such that the lamp 26 3~
~ .. .. .. . .
... : ~ , . . . . .
1 is at the focal point of the reflector 27, and a diffuser 2 28 just below the lamp and reflector. The arrangement 3 of the lamp 26 and the reflector 27 provides a nearly 4 uniform distribution of light over the surface of the 5 diffuser 28, and the diffuser împroves uniformity and 6 reduces intensity to the desired level.
7 Within the drawer 13 are a sample tray holder 8 29 having a matrix of openings 3:L, and an array of photo-9 cells 32 therebelow in a matrix conforming to the posi-10 tion of the openings 31 above. The openings 31 and the 11 photocells 32 also correspond precisely to the position 12 of sample testing wells 33 of a sample tray 14 which is 13 received in registry above the tray holder 29 when a test 14 is to be conducted. The sample tray 14, or at least the lS bottom of each well 33, is translucent so that light 16 passing through the diffuser 28 penetrates the wells and 17 their contents, passes through the openings 31 in the 18 tray holder 29 (and usually through a collimator~ and 19 reaches the photocells 32 below, which individually 20 sense the intensity of the light passing through each 21 well. The photocells may be of the type manufactured 22 by Clairex Electronics of Mt. Vernon, N.Y. as Model 23 CL702L. The tray holder 29 is preferably of a dark, 24 light-absorbing color such as black to reduce light 25 transmission between the wells and reflection of dif-26 fracted light within any one well. The tray holder 27 arrangement assures that all light passing through the 28 openings 31 is from the wells 33 rather than through other 29 areas of the translucent sample tray 14. The embodiment 30 as described is particularly suitable for determining 31 MIC.
32 In another embodiment adapted for bacterial 33 identification, a collimator 34 and an optical filter 35 34 are placed between the sample tray 14 and the photocells 35 32. The exact filter 35 used depends on the test con-36 cerned. The filter 35 is made to be easily removable 37 and replaceable. For example, a large number of tests , , , .,. . ~ . . .. .. .. . . .
71~
1 may be run using only three filters one at a time; these 2 three being (for example) ~ilters numbers 809, 863, and 3 878, of Edmund Scientific Co., 785 Edscorp Building, 4 Barrington, New Jersey 08007. Fig. 14 shows the spec-5 tral absorber responses of these three filters, 809 at 6 A, 863 at B, and 878 at C.
7 The sample tray 14 is preferably a disposable, 8 molded plastic tray, each well of which holds approxi-9 mately 0.5 milliliter. Trays of this type are commer-10 cially available and have been used previously for 11 simple visual type "reading" techniques as discussed 12 above. The wells 33 are often referred to as "micro 13 tubes", since they replace cumbersome full-sized test 14 tubes which were used in the past for this type testing.
16 Fig. 3 shows a preferred arrangement for ~he 17 light source, the sample tray, and the means of holding 18 the sa~ple tray and collimating the light through the wells 19 to the photocells. A partially broken-away, schematic 20 sectional view in Fig. 3 shows the lamp or bulb 26 21 with the reflector 27 above and the diffuser 28 below.
22 The drawer arranement slides in and out of the appratus 23 10 above and independen-tly of the array of photocells 32.
24 As indicated, the photocells 32 are mounted fixedly below 25 the drawer 13 and of course positioned to receive light 26 passing through each well 33 when the drawer is fully 27 inserted, in the testing position.
28 The tray 14 rests on a tray block 40 secured 29 within the drawer 13 and having a matrix of openings 30 31 for receiving the depending sample wells 33. Below 31 the tray block 40 is a drawer plate 44, also bored at 32 each location of a well as indicated, the drawer plate 33 bores 45 being directly in registry with the openings 34 31 above. The drawer plate 44 with its bores 45 serves 35 as the collimator 34.
36 Alternatively, the matrix of photocells 32 may be 37 slidab~y mounted with a tray holder 29 having a matrix of ., ~ . . .. . .
., . , , . -. ~
.
13~'7~6 1 openings 31 into which the wells 33 extend, with little 2 side-to-side tolerance so that the registry of the ~ample 3 wells with the photocells is assured. In this embodiment, 4 which may be used particularly for determining MIC, the 5 tray holder and photocell assembly slides in and out of 6 the apparatus 10 with the drawer 13. The collimator 34 7 may be omitted.
8 As indicated in Fig. 3, there is a space left 9 between the tray block 40 and the drawer plate 44 for an 10 optical filter 35. The filter 35 preferably is slidably 11 received between the two drawer-attached components 40 12 and 44 above and below, and different filters can be 13 used.
14 Another feature illustrated in Fig. 3 is the use 15 of a microswitch 46 which is tripped by the back end of 16 the drawer 13 as it is fully inserted into the testing 17 position. This starts the test automatically, and the 18 testing cycle proceeds to completion.
19 The light source illustrated is a convenient and 20 preferred form; however, any light source or a plurality 21 of light sources which will provide light of equal inten-22 sity directed into each well 33 of the sample tray 14 is 23 sufficient. In this regard, an alternative form of light 24 source and detection system is described below in con- -~
25 nection with Fig. 8.
26 The single diffuse source 26, 27, 28 need not put 27 out a uniform ligh~. The light need only be roughly even.
28 Also, the photodetectors 32 may be inexpensive ones, 29 providing signals of different strengths for the same 30 light intensity, 90 long as the invention is practiced 31 with an initial calibration step. In this step, the 32 light source 26, 27, 28 directs light over all the photo-33 detectors 32 either without a tray ~4 positioned above `
34 them or with an empty tray 14, to take any variations in 35 the plastic material of the tray into account in the 36 calibration. As another alternative, the -tray wells 33 37 may be filled and then run through before any culture, '71~i 1 at zero time relat:ive to growth. In the calibration, 2 a scan is made and all values, i.e. photodetector out-3 put signal values, are stored. When each actual test 4 is run, a difference or ratio signal is created for each 5 photocell, so that only the difference in sensed light 6 intensity is used, disregarding effects of localized 7 differences and intensit~ and differences in the photo-8 cells themselves.
9 The reference photocell 32n is preferably 10 outside the area of the tray (as shown in Fig. 4), 11 although it may be beneath a sterile or empty well 33n 12 of the tray.
13 As discussed above, the sample tray 14 is 14 preferably laid out in a rectangular matrix, which may comprise for example eight rows and ten columns. Other 16 arrangements would be adequate, but a rectangular matrix 17 is space-efficient and convenient. The wells 33 may, as 18 for obtaining MIC values, contain various dilutions of 19 different antlbiotics, and these may be arranged such 20 that each of the ten columns of wells contains a single 21 antibiotic in a series of different dilutions. There 22 may be seven different concentrations of each anti-23 biotic, with theeighth well of a~ least several of the 24 columns used for control purposes. For example, one 25 control well might be used for unrestricted growth of 26 bacteriaj, and another well used to represent no growth, 27 with no bacteria inoculated into the well.
28 Into the wells containing the various dilutions 29 of diferent antibiotics ~for determining ~IIC values) is 30 introduced the patient bacteria sample borne within a 31 culture medium. This bacteria culture is uniformly 32 inoculated into each well, and this may be accomplished 33 by commercially available devices having a matrix of 34 prongs (not shown) arranged to register with each well 35 to be inoculated in the commercially available sample 36 tray 14. Of course, the antibiotics and the bacteria 37 culture may be introduced to the wells in the reverse :
1 order, but for convenience, efficiency and reliability 2 it is preferred that the antibiotic be introduced first.
3 Fig. 4 indicates diagrammatically the operation 4 of the susceptibility testing apparatus 10. The lamp 26, 5 reflector 27 and diffuser 28 are shown transmitting 6 uniform diffuse light through a sample well 33a of the 7 matrix of wells of the sample tray 14. The well 33a 8 contains one dilution of one of the antibiotics being 9 tested, inoculated with a controlled volume and known 10 concentration of the bacteria in a culture sample. The 11 same uniform difEuse light may be transmitted through a 12 well 33n containing no bacteria or providing a light 13 intensity reading corresponding to zero bacteria growth.
14 Alternatively, the sterile control well 33n may be eliminated, as shown, the diffuse light passing through 16 a collimator 34 directly to the reference photocell 32n.
17 An optional color filter 35 may also be added for bacterial 18 identification purposes.
19 After an incubation period sufficient to allow 20 some detectible growth of the bacteria in the well 33a 21 in the event that growth is not prevented by the particular 22 antibiotic in the particular concentration being tested, 23 a growth culture 36 results therein. The light from the 24 diffuser passes through this culture 36 and through the 25 bottom of the well to a photocell 32a of the photocell 26 matrix. Here the intensity of the light is sensed and 27 converted into an electrical analog value corresponding 28 to the opacity of the culture 36. This opacity value 29 represents the turbidity of the culture, stemming from 30 the net effect of light absorption and scatter in the 31 well 33a. At the same time, ~he diffuse light passes 32 through the steri~e control well 33n to a photocell 32n 33 of the photocell matrix. Again, the sensed light 34 intensity is converted into an electrical analog 35 reference value.
36 The photocell 32_ is connected to a plus input 37 of a differential amplifier 37 through a noise filter l 38 and a multiplexer 39 which functions to select each 2 photocell 33 of the matrix of photocells in a prescribed 3 sequence under direction of a microcomputer 41. Alterna-4 tively, the differential amplifier 37 may be replaced by 5 a log ratio module 237, Thus, the "difference" is made 6 into a quotient, giving more sensitivity. The sequencing, 7 being automatic, is very fast, going through 80 or 96 8 wells of a tray 14 in about five seconds or less. The 9 automatic electronic scan has no moving parts--an lO important feature.
11 Electronic sequencing :is much more reliable than 12 mechanical movement of a tray or other mechanical seque~c-13 ing. Multiplexing has the advantages of speed, accuracy, 14 reliability and maintainability, i.e. easy maintenance.
15 For at least these reasons, the invention is a signifi-16 cant improvement over mechanical scanning. Thus, the 17 photocell 32a shown in Fig. 4 is connected to the dif-18 ferential amplifier 37 only when the multiplexer 39 19 momentarily selects that particular photocell. The 20 reference photocell 32n is connected to the minus input 21 of the differential amplifier 37 and provides a reference 22 voltage which is subtracted from the plus input to provide 23 an analog differential output. Thus, the light intensity 24 or turbîdity value signal emanating from the differential 25 amplifier 37 is in the form of a reference voltage which 26 varies according to turbidity of the sample being sensed, 27 representing the increase in turbidity of that sample 28 since inoculation. Each analog signal is transmitted in 29 its turn to an analog-to-digital converter 42 which 30 converts the analog to a digital signal and sends it to 31 the microcomputer 41.
32 The microcomputer 41 (See Figs. 5 and 6) func-33 tions to correlate differential digital values (from the 34 ADC 42) representing, for example, bacterial growth for 35 the various wells witll the particular drug and its con-36 centration in the subject well. From such correlation, 37 the microcomputer selects, for example, the zero growth ,: . :.1: . . .~ :
1 indication ste~ming from the weakest concentration of 2 each drug, and this concentration becomes the MIC for 3 that particular drug. If none of the wells containing 4 a particular drug indicates inhibition of growth, the 5 microcomputer prints out the fact that the infectious 6 organism is resistant to that particular drug.
7 The remaining apparatus indicated in Fig. 4 is 8 described below with reference to the other figures.
9 The analog circuitry associated with this 10 system 10, including the analog-to-digital converter, 11 is set forth in the detailed block diagram of Fig. 5.
12 Fig. 5 includes the nolse filter circuit 38, the multi-13 plexer circuits 39 which are included within the dashed 14 line box, the di~ferential amplifier 37, and the analog-to-digital converter 42 along with its related supporting 16 circuitry. A ten by eight photocell matrix is also shown 17 in Fig. 5 for clarity of understanding of this part of 18 the system 10. A twelve-b~-eight system (or other such 19 system) may be used instead.
The multiplexer 39 includes a binary coded 21 decimal (BCD) to decimal decoder 101 driving column 22 drivers 103, and another BCD to decimal decoder 113 con-23 trolling FET switches 111. A four bit digital line 100 24 from the microcomputer 41 is connected to the binary 25 coded decimal input of the binary coded decimal to deci-26 mal decoder circuit 101 (which may be pre~ereably imple-27 mented as a type 7442 TTI. integrated circuit or equiva-28 lent). Ten output lines 102 from the decoder 101 are 29 connected to ten driver circuits 103. The driver circuits 30 are preferably implemented as operational amplifiers type 31 LM 324 or equivalent.
32 As already explained above, the photocell 33 matrix is arranged as a rectangle with ten columns and 34 eight rows. Thus, the outputs from the ten driver 35 circuits 103 are applied to the ten columns respectively 36 via a bus 104 such that when one driver îs excited by 37 operation of the decoder 101, an excitation voltage is ~3~7~
1 provided to one of the column drive lines corresponding 2 to the binary coded decimal column select information 3 input to the decoder 101 via the data line 100 from the 4 microcomputer 41. An eleventh of the drivers 103 applies 5 voltage continuously through a drive line 105 to the 6 reference cell 33n.
7 Eight row lines 106 and one line 107 from the 8 reference cell 32n are applied as inputs to nine active 9 filter circuits within the filter 38. Each filter cir-10 cuit is preferably implemented by an operational ampli-11 fier, type LM 324 or equivalent. The filters 38 function 12 to remove power line ripple so that the eight row output 13 lines 108 and a reerence output line 109 carry DC voltage 14 levels only. The eight output lines 108 are applied to 15 eight field effect transistor switches 111, respectively.
16 The switches are preferably implemented as integrated 17 circuits type CD4016 CMOS quad bilateral switch gate 18 chips or equivalent. An output line 110 from the switches 19 111 is connected directly to the plus input of the dif-20 ferential amplifier 37. The ninth line 109 is applied 21 directly to the minus input of the logarithmic differen-22 tial amplifier 37.
23 A three bit digital line 112 from the micro-24 computer 41 is connected to the input of a second binary 25 coded decimal to decimal decoder 113 which is also 26 preferably implemented as a type 7442 TTL integrated 27 circuit or equivalent. The decoder 113 functions to 28 select one of eight output control lines 114 which in 29 turn select one of the eight field effect transistor 30 switches 111 to connect one of the filtered row lines to -31 the plus input of the logarithmic differential ampli-32 fier 37, in accordance with digital row select information 33 received from the microcomputer 41. .
34 The logarithmic differential amplifier 37 is 35 preferably implemented as an Analog Devices type 757 or 36 equivalent, and the purpose of the amplifier 37 is to 37 correct for variations in light intensity from the light 3~
~ . , . , .. " , - , ~ , ~, ; ., . ,-, .
1..13~7'~
1 source 26. The light-variation-corrected analog voltage 2 output from the amplifier 37 is supplied as an input to 3 an operational amplifier 116 which is provided with 4 external potentiometers to control gain and DC offset 5 of the incoming signal from the amplifier 37.
6 An output line 117 from the amplifier 116 is 7 supplied as an analog input to the analog-to-digital 8 converter 42 which is preferably implemented with a 9 National Semiconductor MM5357 integrated circuit or 10 equivalent. A digital control line 118 from the micro-11 computer 41 is connected as a trigger input to a mono-12 stable multivibrator one shot 119, preferably implemented 13 as a type 74121 TTL integrated circuit or the equivalent.
~ An output pulse from the one shot 119 of appropriate 15 amplitude and duration is supplied via a line 121 to the 16 analog-to-digital converter 42 to start the conversion 17 process. A timing genera-tor (e.g. type 555) 122 applies 18 timing pulses via a line 123 to the analog-to-digital 19 converter 42 to control the sequence of operations thereof.
20 The analog to-digital converter 42 utilizes the timing 21 pulses supplied on the line 123 during a conversion cycle 22 to digitize the analog information on the line 117 and 23 provide an eight bit digital output via an eight bit 24 output bus 124 which is supplied to an input port of the 25 microcomputer 41.
26 The microcomputer 41 forms the central portion 27 of the system 10. The microcomputer includes a single 28 chip monolithic microprocessing unit (MPU) 140, which is 29 preferably implemented as a type 6800 manufactured by 30 Motorola Semiconductor, American Microsystems, and other 31 suppliers. Although this particular microprocessor 32 was chosen for the described preferred embodiment of 33 the present invention, other types of microprocessors 34 would function equally as well, for example the Intel 35 8080, the Mostec 6502, the Zilog Z80, the Fairchild F-8, 36 etc. A suitablle two-phase clock 141 provides the neces-37 sary clock signals to the microprocessing unit 140.
, ~ . :,- ,., , ~ . ., 3~7~7~6 - 33 ~
1 A main system program like that which is set 2 forth in hexadecimal code in the table following the 3 specification of the present invention may be loaded 4 into one and a half kilobytes of programmable read only S memory 142. The read only memory 142 is preferably 6 implemented with 2708 programmable read only memories 7 produced by Intel and other suppliers. Other PROMs 8 would be well suited for the program memory 142. The 9 microcomputer 41 also includes one kilobyte of random 10 access memory (RAM) 143 which provides volatile storage 11 of data to be processed as well as a stack for the 12 microprocessing unit 140. The microprocessing unit 140, 13 the clock 141 through the microprocessing unit 140, the 14 program memory 142 and the data storage memory 143 are 15 connected in parallel to the system bus 144 which 16 includes an eight bit data bus, an eight bit control 17 bus, and a sixteen bit address bus.
18 Input output interface is accomplished with 19 three peripheral interface adapters ~PIA) 146, 147 and 20 148 which are connected to the sys~em bus 144. The 21 interface adapters 146, 147 and 148 are preferably 22 implemented as type 6820 integrated circuits produced 23 by Motorola Semiconductor, ~merican Microsystems, and 24 other suppliers. These integrated circuits contain two 25 ports apiece. Each port may be used either to input 26 data to the microprocessing unit 140 or to output data 27 to output devices, as will be explained hereinafter.
28 The first interface adapter 146 has its first 29 port connected to receive the eight bit digiti~ed 30 information via the bus 124 from the analog-to-digital 31 converter 42, as shown in Fig. 5. The first port of the 32 interface adapter 146 also provides the control signal 33 line 118 which is connected to the one shot 119 which 34 functions to start the analog-to-digital conversion 35 process of the converter 42. The line 118 will be 36 further explained hereinafter. The second port of the .
37 interface adapter 146 is connected to the multiplexer 39 , ~ . , . . ; ! ' ' I ' ' . ' ' ;
'7~
1 with four bits provided for the column select control 2 signal via the bus 100, and the three remaining bits 3 provide for the row select control signal via the bus 4 112.
The second peripheral :interface adapter 147 6 includes a first port which controls the printer 20.
7 Two bits of data are input from status indicators in 8 the printer 20 via a line 149. One of these bit posi-9 tions is from a microswitch which indicates that the 10 paper form has been properly inserted and tha~ a print-11 out can be made. The other bit is a signal from the 12 printer electronics which indicates that the printer is 13 either in a "print" or a "wait" operational mode. Four 14 bits of the first port of the interface adapker 147 are 15 also used to control the printer and shift data to be 16 printed into the printer 20. The data is entered seri-17 ally via a line 151 from the first port of the adapter 18 147 to the printer 20. Other control functions carried 19 out by the four bits on the line 151 include line feed 20 (advance the paper one line), print (cause the print 21 solenoid to make an impression on the paper), and shift 22 (move the next data bit into position for printing~. The 23 second port of the interface adapter 147 is not used in 24 the present embodiment.
The third peripheral interface adapter 1~8 in-26 cludes a first port which reads the thumbwheel switch 16 27 for patient identification information via a four bit 28 line 153. The upper our bit positions of this first 29 port of the adapter 148 are used to select and enable 30 one of the four thumbwheel positions via a four bit line 31 152. One bit position of the line 152 is low to enable 32 one of the four switching positions. The lower four bits 33 of the first port of ~he adap~er 148 are used to read 34 data via a bus 153 from the switch position selected by 35 the upper four bits. The data from the switch repre-36 sent a binary number between zero and nine. The second 37 port of the in~er:Eace adapter 148 is used to supply data 38 to the alpha-numeric display readout 19. The display 1 is , , :.
- , ::
..
37"~3~;
l the Burroughs model SSD0132-0070 self-scan display unit 2 with built-in electronics. As explained, it is controlled 3 via a line 154 from the second port of the third peripheral 4 interface adapter 148. Data to be displayed on the 5 display 19 are entered into the lmit via a line 156 in a 6 six bit code for all alpha-numeric characters as well 7 as some special symbols. The data are read in from left 8 to right and appear on the display until new data are 9 entered. Thus, the upper two bits are provided via the 10 line 154 to control the display, with one of the bits 11 being a clear line and the other being an enable line.
12 The lower six bits are provided via the line 156 for the 13 purpose of sending parallel data to the display presented 14 to the user in accordance with the operation of the system 10.
16 In addition to the characteristics of the 17 interface adapters 146, 147 and 148 described hereinabove, 18 each adapter also has an interrupt function. The 19 interrupt is an additional line which is available for 20 monitoring the status of external devices. In the pres-21 ently described system 10, the interrupts a~e used to 22 monitor operator actions of several types. Interrupt 23 capability which results in an output rather than an 24 input is termed a strobe. Strobes are utilized in the 25 system 10 as well as interrupts. Thus, the first peri-26 pheral interface adapter 146 controls the conversion of 27 data ~rom analog--to-digital format via the analog-~o 28 digital converter 42 by utilizing a strobe line 118 which 29 is connected to the one shot 119 (Fig. 5) to start the 30 analog-to-digital conversion operation.
31 The second peripheral intsrface adapter utilizes 32 an interrupt from the printer 20 via a line 155 and 33 utilizes one interrupt each from the run switch 17 via a 34 line 158 and calibrate switch 1& via a line 159. The ;-35 second port of the second adapter 147 utilizes the output 36 strobes via a line 157 to cause the printer 20 to execute 37 a print cycle.
:: , ,, ~ -3~'7~i 1 A third peripheral interface adapter 148 has 2 two interrupt lnputs: one from a microswitch indicating 3 that the photo unit drawer is open via a line 161 and one 4 indicating that the drawer is closed via a line 162.
The printer 20 may be implemented as an MFE
6 model TKllE or Practical Automation DMPT-9, both with 7 electronics package. Data is fed from the microcomputer 8 41 via the line 151 which generates the proper control 9 signals to enable the printer electronics to cause the 10 printer 20 to print, line feed or shift data into internal 11 registers. The data is fed to the printer 20 in serial 12 format, stored in buffers in the printer electronics, and 13 is then printed in parallel. The command to print is 14 generated as a strobe output of the second port of the second peripheral interface adapted 147 via the line 157.
16 The printer is a commercially available unit presently 17 being sold for the original equipment manufacturer (OEM~
18 market.
20 Determining minimum inhibitory concentration (MI~) (Fi~. 7):
21 One method using the system 10 is explicated by 22 the flow chart set forth in Fig. 7. Therein, at a power 23 on step 166, the operator turns the power on to the system 24 10. At that point, the display 19 informs the operator 25 to insert the calibration tray. At insertion step 168, 26 the operator inserts the tray, and at step 169, the 27 operator closes the drawer. At a logical step 170, the 28 system checks the identification of the tray in the 29 drawer. For this purpose a binary code may be implemented 30 using the uppermost right two wells of the tray, either 31 of these wells being Qither opaque or transparent, thus 32 providing identification of four possible types of trays.
33 This code is made to correspond to the combination 34 antibiotics which the tray contains.
In the event that the type of tray is not 36 identified at step 171, the system asks whether the tray 37 is inserted backwards at step 172. If so, the readout 19 '' ' ' ~ ' '` ' 'i` ' ' .` '', ;;~ ' ' " ` . ' ' .;: '.' ~ `"` " '' ' .. . .
3~ 6 1 displays a ~ray backwards indication at step 173, and 2 the operator opens the drawer at a step 174 and removes 3 the tray, orients it correctly, and reinserts it, then 4 repeats steps 168, 169, 170 and 171.
Once the tray is identified at step 171, the 6 readout 19 displays the tray type at step 175, and 7 directs the operator to press the calibration switch 18 8 at a step 176. At step 177, the operator presses the 9 calibration switch 18 whereupon the system tells the 10 operator to wait at step 178. The wait signal remains 11 until the system informs the operator to remove the 12 tray at step 179. The operator opens the drawer at step 13 180. In the event that the tray is not in backwards, and 14 yet the tray re~lains unidentified at step 181, the 15 operator is then instructed to open the drawer to manually 16 inspect the tray to find out why the system 10 is unable 17 to identify it.
18 At step 182, the readout 19 tells the operator 19 to close the drawer, and at step 183 the operator removes 20 the tray and closes the drawer. The readout 19 then tells 21 the operator that if a next test is desired, he should 22 press the run or calibrate button at step 184. At a step 23 185, the operator actually presses the run or the cali-24 brate switch. If the system has been previously cali-25 brated at step 186, then the readout l9 directs the 26 operator to insert the test tray at step 187. However, 27 if the system 10 has not been calibrated at step 186, the 28 program returns to step 167 and the calibration procedure 29 is carried out as set forth in steps 167 through 185.
30 At step 188, the operator opens the drawer and ~-31 inserts the test tray. The display 19 then tells the 32 operator to close the drawer at step 189. The operator 33 closes the drawer at step 190 and the tray identification 34 is determined at step 191. In the event that the tray is 35 not identified, the system then determines whether the 36 tray is in backwards at step 192. If so, the system 37 informs the operator that the tray is in backwards by a 3~
1 readout display at step 193. In the event that the tray 2 remains unidentified and it is not in backwards, then at 3 step 194, the operator is informed that the tray is 4 unidentified and the program loops back to step 180 5 whereupon the operator opens the drawer and repeats 6 steps 180 through 191.
7 Once the identification of the tray has been 8 determined at logical step 191, the system 10 displays 9 the type of tray at the readout with step 195. Then the 10 operator is informed to set the patient identification 11 information i.nto the identification switch 16 and insert 12 the form to be printed into the printer 20 at step 196.
13 The operator performs these operations at step 197 and 14 when they are completed, the display 19 tells the operator to press the run switch 17 at step 198. The operator 16 presses the run switch 17 at step 199 and the patient 17 identification information is displayed at step 200.
18 Then, the patient identification is printed on the form 19 at a step 201 and then the MIC values and lnterpretive 20 information are printed on the form in step 202 to produce 21 the form 203.
22 Once the form is printed with the patient 23 identification MIC values and interpretive information 24 the display tells the operator to remove the tray at 25 step 204. The operator opens the drawer and actually 26 removes the tray at step 205 whereupon the display 19 27 tells the operator to close the drawer at step 206. The 28 operator closes the drawer at step 207 and the apparatus 29 10 then instructs the operator to perform the next ~-30 operation of either "run" or "calibrate" at step 208 31 whereupon the program loops back to step 185 where the 32 run or calibration switches are operated and the program 33 is repeated as heretofore described until all of the 34 samples have been evaluated by the system 10.
36 Comparisons to reduce errors due to the tray and to light 37 intensity and_photodetector differences:
38 It will be apparent that the tray 14 itself might 31~l3~ 9
- 39 -1 be a source of error. That is, its own light trans-2 missivity and opaqueness and flaws can substantially 3 affect the light transmissivities received by the photo-4 cells 32, in addition to the light transmissivity of the 5 liquid in the wells. The trays 14 can vary from tray to 6 ~ray, and they can also vary in a tray from well to well.
7 This could, of course, lead to substantial errors that 8 would give false impressions and false results if not 9 compensated or corrected.
The present invention accomplishes the needed 11 correction by two different types of comparison stages.
12 First, for each reading in any sequence of wells 13 33 in the tray 14, each well 33 is immediately compared 14 with the value obtained by direct light transmission to 15 the reference photocell 32_. While this may be done 16 through a sterile control well, as shown in Fig. 4, it 17 is preferably done directly, completely outside the tray 18 14, as shown in Figs. 10 and 13, with the light to the l9 reference photocell not passing through any portion of the 20 tray. F'rom this comparison, the device provides an after 21 culture value for each well, which is a function of the 22 after culture signal values (or amplification thereof) for 23 the tested well and for the reference photocell. This, 24 of course, represents a comparison of the light received 25 at each photocell in the main array and the intensity of 26 the light received at the reference photocell. The signal 27 may be amplified and is used as the operative signal, as 28 shown in Fig. 4. The after culture value for each well 29 may be called a "difference" signal value, regardless 30 of the type of function which is used in comparing the 31 two values (test well vs. reference photocell) to produce 32 this value. In the embodiment of Fig. 4 the subtractive 33 difference preferably is taken between the two values, 34 and the differential amplifier 37 amplifies the difference 35 signal. However, the signal value produced in the other 36 embodiment of Fig. 4 is a ratio, and the signal from each 37 well is compared with the reference photocell signal by : . . : . ~ ~: ~ .
. .. :. ; ~ ; , r - ~
- 40 -1 means of a log ratio module 237. In other words, there 2 is again a "difference" signal, but it is a difference 3 in logarithms, so that the suhtraction is really a 4 division, and a quotient or ratio is obtained instead 5 of a difference expressed as a logarithm.
6 Thus a first comparator may incorporate a log 7 ratio module and send out its related signal as an ampli-8 tude ratio between each signal Sw obtained through a well 9 and its photocell and a signal SR obtained from the 10 reference photocell. This related signal Sx =
12 where kl is a constant. This first comparator may also 13 incorporate a log ratio module and sends out its out its 14 resultant value Sv as a ratio k2 SX , where DV is the DV
16 data reference value and k2 a constant.
17 The first and second comparators may use the 18 same log ratio module. The second comparator may utilize 19 as its data reference value Dv, stored ratios read earlier 20 from an empty tray, so that DV = kl SWE for each well, 22 where SwE is the signal coming from an empty well.
23 The second comparator may utilize as its 24 data reference value Dv, stored ratios read earlier from 25 a tray containing the same liquid from which the signal 26 Sw are generated, but read at a time when there has been 27 zero growth, so that DV = 1 SWO for each well, where 29 sWO is the signal coming from a well containing the liquid 30 at zero growth time.
31 Also, the preliminary comparing means may include 32 a log ratio module for sending as its derived signal a 33 signal based on the ratio of the two signals it compares.
34 Thus, in the invention, each reading of each 35 well, at each stage where readings are taken, is compared 36 by a first comparator means with the reading at the refer-37 ence photocell, and a difference or ratio signal developed 38 ~`~
~L~L3~
1 from it. By this procedure, variations in light intensity 2 from the source over time, as would be induced by supply 3 voltage fluctuations, have no effect on the readings.
Such variations will vary the reference and well photo-5 cells proportionately, so that a ratio will cancel the 6 errors out. This is the purpose of the reference photo-7 cell.
8 Second, to further reduce the possibility of 9 error particularly due to flaws in the tray, and in view 10 of the fact that each tray 14 is positively identified 11 in the apparatus, as has already been descrlbed, a prior 12 reading may be taken through the tray before the reading 13 after bacterial culture; this prior reading is stored 14 and is later compared with the sample reading.
One way of taking the prior reading is to take 16 a reading of the tray 14 in its empty state, before it 17 is filled with fluid, to compare the reading through each 18 empty well with the reading of the reference photocell, 19 as above, and to store the resulting difference signal or 20 ratio signal in the data storage portion of the micro-21 computer 41. Then the ratio signal (or difference signal) 22 derived from the liquid at the time of the after culture 23 reading is compared with the ratio signal (or difference 24 signal) of the empty wells. Thereby, each well is compared 25 with itself when full and when empty, and errors due to 26 the wells are substantially eliminated.
27 Ano~her way of taking this prior reading is to 28 take the prior reading, not of the empty tray but of the 29 tray just after its wells have been filled with the solu-30 tion and prior to the culture; in other words, at sub-31 stantially zero time so far as growth or culture is 32 concerned. This means that the reading is taken through 33 the actual solution, and the ratio of that reading to 34 the reference electrode is stored in the data storage 35 bank for the later use.
36 With the zero based signal (however obtained) 37 in the data bank, and with the ratio or difference signal 38 provided for each well for the liquid after culture, then, ,.3~t~6 1 before proceeding further, the ne.~t step is to compare by 2 a second comparator rneans the two ratio (or difference) 3 values, that is, to compare the ratio of the signal 4 derived from the light transmissivity of the specimen 5 after culture to the direct light recep-tion by the refer-6 ence cell, with the ratio of the empty tray or tray with 7 the same liquid at zero time to the signal from the ref-8 erence cell. This second comparison may also be made by 9 calculating a ratio of the two ratios, which is prefer-10 ably accomplished by taking the difference in logarithms 11 of the two ratios, resulting in another logarithm which i9 12 the log of the comparison ratio, or of what may be called 13 the comparison signal.
14 In the next step, a third comparison depends 15 upon what test is being run. Basically, it is a com~
16 parison of the ratio signal obtained from the second 17 comparator means, which preferably is the logarithm of 18 the comparison signal, with values that are stored in the 19 data storage means to determine the final asked-for result.
For good results in this last step, especially 21 when applied to MIC procedure, a distinction is made 22 between a growth state and a no-growth state. The 23 instrument determines at the output from the second 24 comparator means, a voltage level or logarithm value that 25 represents the extent of bacterial growth, when that 26 voltage level is compared to voltages that are obtained 27 from known sterile and growth controls, these voltage 28 values being stored in the data bank of the microcomputer 29 41. A first step here is to determine whether there is 30 an adequate voltage (logarithm value) difference between 31 the readings obtained from the sterile and the growth -~
32 control wells. This is done preferably by comparing the 33 ratios for the two wells, i.e. the products of the first 34 comparator means for the two wells, which are logarithms 35 of ratios of well readings vs. reference readings. The 36 comparison of the two control values is done by taking 37 a difference between the two logarithms. The resulting 38 difference is compared to a predetermined, stored value - ~3 -1 representing adequate growth-sterile difference for the 2 test. If there is an inadequate difference, this means 3 either one of two things, either that there had not been 4 sufficient growth to provide an adequate difference, 5 or that the sterile well had been contaminated and that 6 there had been growth there. In either case, the instru-7 ment will display a reading such as "insufficient growth-8 sterile difference", and the computer returns to the 9 beginning of the program. The operator then checks to lO see which of these two possibilities is the one that is 11 present. If ~here is insufficient growth, it may be due 12 to a lack of time or because there was nothing to grow.
13 If there were contamination, that would show and be 14 readily detectable, and the test must be re-done.
Once the computer has established that there is 16 an adequate difference between the sterile condition and 17 the expected growth condition from one well to another, 18 the calculated logarithm values and their difference are 19 used for computation of a break point, or a limit com-20 parison signal value. Preferably, the break point is 21 biased toward the sterile value to achieve more sensitivity 22 to growth detection, via a preselected fraction of the 23 sterile-growth logarithm difference. The break point may, 24 for example, be placed at 25% of the determined sterile-25 growth difference (preferably a logarithm value as above), 26 added to the log value for sterility. For all wells 27 where there has been less growth than that r~presented 28 by 25% of the determined growth-sterile difference for 29 the test being conducted, then the concentration of 30 those wells is considered as inhibitory. For each drug 31 being tested, the concentration closest to the break 32 point, but on the inhibitory side, is selected as the 33 minimum inhibitory concentration value. Thus, supposing 34 that there are a series of wells of different dilutions 35 and that the operation is moving from wells of greater 36 growth towards those of lesser growth and toward the 37 sterile condition, then the minimum inhibitory concentration ,...... " . .. .. .. . . .
3~
; 1 is not found until the first well is reached which shows 2 less than 25~/o of the determined difference between the 3 sterile and growth control wells. In this way, a 4 "floating threshold" is utilized, i.e. one which is calcu-5 lated from controls in the very test being conducted and 6 with the same organism being tested, rather than a fixed 7 threshold which has been calculated based on prior infor-8 mation and stored.
9 Another important comparison which should be 10 performed preferably at least once a day, before series 11 of tests are performed, is an initial calibration step.
12 This initial calibration is in lieu of the empty tray 13 (or just filled tray) reading procedure described above.
14 Like that procedure, this calibration procedure is important in that it enables the use of a light source ]6 which is not totally uniform Eor each photodetector, 17 but only generally uniform, and also the use of inexpen-18 sive photodetectors which may not be uniform or totally 19 constant, over a long period of time, in their sensi-20 tivity. By this procedure the light is first passed 21 directly (no tray) to all photodetectors) including 22 the reference photocell, and ratio readings (preferably 23 their logarithms) are taken as above and recorded. These 24 values are stored and give a relative base line or initial 25 calibration value for each photocell. All subsequent 26 after culture values (which are pre~erably logarithms 27 of ratios as above) are compared to these base line 28 readings, and expressed as "difference" (or log ratio) 29 readings. Thus, any differences in sensitivities of the 30 various photocells, or differences in light intensity due 31 to position, are "zeroed out" by comparison of after 32 culture ratios with initial calibration ratios, the 33 comparisons being separate for each well.
35 An alternate type of light source (Fi~. 8~:
36 Fig. 8 shows schematically an alternative 37 arrangement for passing light th~ough the wells 33 of the 38 sample tray 14 and detecting the resultant light intensity 7~36 - ~5 -1 passing through each well. The apparatus of Fig. 8, which 2 utilizes fiber optics to transmit light,would replace the 3 form of light source and diffuser 26, 27 and 28 shown in 4 Fig.s. 2, 3 and ~. It would also eliminate the need for 5 a large plurality of photocells 32 in a matrix as shown 6 in Fig. 2, and would replace the multiplexing unit 39 7 (Fig. 4) with a substitute arrangement which selects one 8 cell at a time for receipt of a penetrating quantum of 9 light.
The apparatus of Fig. 8 includes a light source 11 221 and a reflector 222, directing light through a lens 12 223 toward a rotatable selector plate 224 driven by a 13 stepper motor 225. The selector plate 224 has a single 14 opening 226 (dashed lines) which sequentially directs 15 light to different fiber optic fibers 228 o~ a fiber 16 optic bundle 229. The stepper motor 225 is under the 17 control of the microcomputer 41 via the lines 100 and 18 112 (Figs. 4 and 6), in lieu of an to perform the same 19 function as the multiplexer 39 indicated in Figs. 4 and 20 6. The fiber optic fibers 228 of the bundle 229 each go 21 to individual testing wells 33 of the tray 14. The 22 fibers are indicated only schematically, as is the bundle 23 229.
24 Below the wells 33 are a second plurality of 25 fiber optic fibers 230 of a second bundle 231. Trans-26 mitted light from each well is collected by a fiber 230 27 Of the bundle 231 and fed via a lens 232 to a single 28 photocell detector 233. A value corresponding to the 29 intensity of incident light is then fed to the filter 38, 30 then to the plus input of the differential amplifier 37, 31 as in the apparatus of the other embodiment described 32 above.
33 In order to provide a control or reference 34 value which may be fed into the minus input of the dif-35 ferential amplifier 37 to represent a base light inten-36 sity corresponding to zero bacterial growth, there must 37 be an optical fiber which always carries light through 1 a reference sterile control well, i.e. the well 33n of 2 Fig. 4, also indicated in the schematic representation 3 of Fig. 8. Accordingly, a single optical fiber 228n is 4 positioned at the lens 223 in such a way that it receives S and carries llght continuously whenever the lamp 221 is 6 energized, i.e. whenever any of the wells 33 is being 7 tested. The fiber 228n extends to a position adjacent to 8 the sterile control well 33n as shown, and a receiving 9 fiber 230n carriesthe transmitted light to second lens 10 232n. The resultant analog light intensity value or 11 the control well is fed through the filter 38 to the 12 minus input of the differential amplifier 37, so that 13 the di~ferential ampliier yields a differential analog 14 signal corresponding to increaæed turbidity in the tested 15 well from bacterial growth.
16 The remainder of the system remains the same as 17 described above. The principal advantage of the form 18 illustrated in Fig. 8 is the use of a single light source ;
19 focused on the fiber optic bundle and a single detector 20 for all test wells of the sample tray, providing a more 21 uniform measurement over the matrix of test wells in ~he 22 tray. Light is transmitted through only two wells of 23 the tray at any given time: the well currently being 24 tested for turbidity, and the sterile reference well 33n.
25 The subsystem of Fig. 8 allows for close standardization 26 and easy calibration and checking.
28 A printout ticket for MIC (Fig. 9):
29 Fig. 9 shows a form of printout ticket 24 which 30 may be used in conjunction with the present invention, 31 with exemplary MIC susceptibility information and therapy 32 information. As discussed above, the apparatus of the 33 invention provides a graphic interpretive printout to 34 guide the physician's therapy, an example of this type 35 information being located in the right column of the 36 ticket 24. The computer algorithm translates the MIC
37 values (left column) to dosage ranges that would be 7~7~t~
1 necessary to achieve blood levels of each antimicrobic 2 drug to effectively inhibit growth of the organism. Fig.
3 9 indicates one form that the "therapy guide" information 4 may take. With this format, "-" indicates that the 5 organism tested is resistant to that particular anti-~ microbic, and that no dosage of the antimicrobic can 7 affect the organism. "~" indicates resistance but that 8 the organism may respond to high intra-muscular intra-9 venous doses. "++" indicates that the organism is 10 intermediate in sensitivity to the particular antimicrobic, 11 and may respond to higher than recommended doses. A
12 printout of "~" indicates sensitivity to the usual 13 recommended doses of the antibiotic, and "~ " means that 14 the organism exhibits a high degree o:~ sensitivity and thus is an optimum drug with which to treat the infectious 16 organism. A printout of "****" tells the physician that 17 a dosage of that particular antibiotic necessary for 18 therapy may be toxic to the patient.
20 Use of the apparatus of Figs. 2-4 in bacterial identifi-21 cation:
22 As discussed above, the sample tray14is prefer-23 ably laid out in a rectangular matrix, which may comprise, 24 for example, eight rows and ten columns. Other arrange-25 ments would be adequate, but a rectangular matrix is 26 space-efficient and convenient. In this method, the wells 27 33 contain various reagents.
28 Into the wells containing the various reagents 29 is introduced the patient bacteria sample borne within 30 a culture medium. This bacteria culture is uniformly 31 inoculated into each well, and this may be accomplished 32 by commercially available devices having a matrix of 33 prongs (not shown) arranged to register with each well 34 to be inoculated in the commercially available sample 35 tray 14 Of course, the reagents and the bacteria 36 culture may be introduced to the wells in the reverse 37 order, but for convenience, efficiency and reliability ~: ~ .. . . - , . .
: , . - , , . .: : . ::
~ ~ ~ 377~i - 4~3 -1 it is preferred that the reagents be introduced first.
2 Fig. 4 indicates diagrammatically the operation.
3 The lamp 26, reflector 27, and diffuser 28 are shown 4 transmitting uniform diffuse light through a sample 5 well 33a of the matrig of wells of the sample tr~y 14.
6 The well 33a contains one reagent, or group of reagents, 7 and one sensor (e.g.~ one reagent and one pH indicator), 8 inoculated with a controlled volume and known concentra-9 tion of the microorganisms in a culture sample. The same 10 uniform dif~use light is also directly transmitted through 11 a filter 35 to a reference photocell 32n.
12 After an incubation period sufficient,to allow 13 some detectible reaction of the microorganism with the 14 reagent in the well 33a, in the event that there is a 15 reaction, a reaction product 36 results therein. The 16 light from the diffuser passes through this reaction 17 product 36 and through the bottom of the well and through 18 the filter 35 to a photocell 32a of the photocell matrix.
19 Here the intensity of the light is sensed and converted 20 into an electrical analog value corresponding to the 21 opacity of the reaction product 36. This opacity value 22 represents the intensity of color of the culture and its 23 reaction with the reagent, stemming from the net effect 24 of light absorption and scatter in the well 33_. At the 25 same time, the diffuse light passes to a photocell 32n 26 of the photocell matrix. Again, ~he sensed light intensity 27 is converted into an electrical analog reference value.
28 The photocell 32a is connected to a plus input 29 of a log ratio module 37 through a noise filter 38 and a 30 multiplexer 39 which functions to select each photocell 31 33 of the matrix of photocells in a prescribed sequence 32 under direction of a microcomputer 41. Thus, the pho~ocell 33 32a shown in Fig. 4 is connected to the log ratio module 34 37 only when the multiplexer 39 momentarily selects that 35 particular photocell. The reference photocell 32n is 36 connected to the minus input of the log ratio module 37 -~ 37 and provides a reference voltage which is subtracted from ~13~
_ ~9 _ 1 the plus input to provide an analog differential output.
2 Thus, the light intensity signal emanating from the log 3 ratio module 37 is in the form of a reference voltage 4 which varies according to the opacity of the sample being 5 sensed, representing the increase in opacity of that 6 sample since inoculation. Each analog signal is trans-7 mitted in its turn to an analog-to~digital converter 42 8 which converts the analog to a digital ~ignal and sends 9 it to the microcomputer 41.
The microcomputer 41 functions to correlate 11 differential digital values (from the ADC 42) represent-12 ing bacterial reaction with the reagent for the various 13 wells.
14 The analog circuitry associated with the system 15 of Fig. 4 includes the noise filter circuit 38, the multi-16 plexer circuits 39 which are included within the dashed-17 line box, the log ratio module 237, and the analog-to-18 digital converter 42 along with its related supporting 19 circuitry, all as described before in connection with 20 Figs. 1-6.
21 The logarithmic log ratio module 237 is prefer-22 ably implemented as an Analog Devices type 756 or equiva-23 lent, and the purpose of the log ratio module 37 is to 24 correct for variations in light intensity from the light 25 source 26. The light-variation-corrected analog voltage 26 output from the log ratio module 37 is supplied as an 27 output line 117.
29 Example of bacteria identification:
Bacteria that must be identified in the clinical 31 laboratory may be taken from a large number of body sites.
32 Wounds suspected of being infected, nose and throat cul-33 tures, aspirates from abscesses, feces, and sputum speci-34 mens are some of the more common sites from which bacteria 35 may be cultured. Normally sterile body fluids are also 36 frequently investigated for the presence of bacteria. In 37 suspected cases of septicemia, blood may be sent to the ~ ~ ~.3~f'7~i l laboratory, and urine is frequently cultured ~or the 2 diagnosis of urinary tract infections. Additionally, 3 cerebrospinal fluid for suspected meningitis, pleural 4 fluid for suspected pleuritis, pericardial fluid for 5 pericarditis, and ascitic fluid Eor suspected peritonitis 6 may be sent to the laboratory for the culture, identi-7 fication, and susceptibility testing of bacteria.
8 A nurse, physician, or technologist may obtain 9 cultures by either using cotton swabs, or directly 10 inoculating the specimen into a :Liquid medium. In the ll case of a liquid specimen, such as blood or a body fluid, 12 the original material is introduced into a bottle or 13 tube of sterile nutrient broth media. For cultures of 14 solid structures, such as a wound, skin, eye, etc., it is 15 necessary to collect a specimen with a swab. Once 16 transported to the laboratory, the swab is streaked over 17 the surface of nutrient agar. Agar is a seaweed deriva-18 tive that forms a gel. Molten agar may be poured into 19 a shallow, cylindrical glass dish (Petri dish) to form 20 a layer appro~imately five millimeters deep. When 21 bacteria grow on the surface of the agar, individual 22 organisms which are originally invisible to the naked 23 eye multiply to become colonies that are easily visible.
24 Each isolated colony is the aggregate "off-spring" of 25 a single bacterial progenitor. Thus, by utilizing a 26 single colony or a group of similar colonies, a pure 27 culture of bacteria may be obtained for susceptibility 28 testing or identification. In the case of bacteria 29 that were originally isolated in a liquid nutrient 30 broth, it is necessary to subculture these organisms 31 on agar plates in order to obtain pure cultures. It is 32 these isolated colonies that are subjected to identifi-33 cation testing by the present invention. A detailed 34 discussion of specimen collection and preparation is 35 described in the American Soc~e~y o~ Microbiology Manual 36 of Clinical Microbiology, 2nd Ed. Lennette, E.H., Spaulding, 37 E.H., and Truant, J.P., Editors, ASM, Washington, D.C. 1974.
8~3 l In time sequence, once a culture is obtained, 2 brought to the laboratory, and plated on agar plates 3 for isolation, it is usually necessary to wait twelve 4 to eighteen hours for there to be sufficient colonial 5 growth for further testing. Once pure colonies have been 6 isolated, the medical technologist makes a preliminary 7 identification based on the colonial morphology and the 8 microscopic appearance of the bacteria. To assist in 9 this classification, the bacteria are stained with a 10 dye and iodine mordant together with a red counterstain.
11 If the bacterial walls have affinity for the stain, they 12 will appear blue and are referred to as "Gram Positive".
13 If the bacteria do not stain positively, the red counter-14 stain will prevail; and these organisms are classified as 15 being "Gram Negative". A well-isolated colony is -trans-16 ferred into eight ml. sterile saline which is supple-17 mented with 0.02% Tween 80.
18 The saline suspension of bacteria is trans-19 ferred into a plastic seed tray, and a transfer lid is 20 placed over the tray. The transfer lid contains plastic 21 prongs that are spaced in such a way that each prong will 22 pick up a small but uniform drop of bacteria suspension 23 and mate with the wells 33 in another plastic tray 14 24 that contain biochemical reagents. After the bacteria 25 have been introduced to the biochemical microtubes 33, 26 certain tubes (H S, lysine, arginine, and ornithine) 27 are overlayed with mineral oil to seal the reaction 28 mixtures from a~mospheric oxygen. The biochemicals 29 containing bacteria are incubated in a non-CO incubator 3Q at 35C. for 18-24 hours. The reactions may then be 31 read by the instrument 10, presently described.
32 The biochemical tests which are read and 33 interpreted by this invention span a wide range of 34 fermentative reactions. The following list expands 35 in detail these bioc~emical reactions.
. : . ,,: : ~ , . , . - ,.- :,. .. ... - .. . ~ : : , i``
3~
5~
1 Carbohyd ate Fermentation 2 The carbohydrates used with this invention are 3 dextrose, sucrose, raffinose, rhamnose, arabinose, 4 inositol, adonitol, and cellobiose. The fermentation of S a specific carbohydrate results in acid formation. The 6 resulting drop in pH is detected by a phenol red indicator 7 changing the color from red to yellow.
9 Urea Bacteria which produce urease split urea forming 11 two molecules of ammonia. Since ammonia is basic, the 12 resulting rise in pH can be detected by a phenol red pH
13 indicator changing the color from orange to red.
Indole 16 The metabolism of the amino acid tryptophane 17 results in the formation of indole which is detected by the 18 addition of Kovac's reagent. If indole is present, a red 19 color develops.
21 Lysine, Arginine, Ornithine 22 Decarboxylation of these compounds results in 23 an alkalization of the media which is detected by the pH
24 indicator bromscresol purple. A positive reaction is brown and a negative reaction is colorless to gray.
27 Tryptophane Deaminase 28 Bacteria capable of deaminating tryptophane 29 produce phenyl pyruvic acid. In the presence of ferric ammonium citrate, this reaction product produces a brown 31 color, whereas a negative reaction is clear.
33 Esculin Hydrolysis 34 The ability of an organism to hydrolyze esculin 35 is de~ected by ferric ammonium citrate in the medium, which -36 reacts with the hydrolysis products to form a black pre-37 cipitate.
,':' ' , ~ ' , ,` ' ., . '.' ' ! ' ' ' ' . ' ' ,' ' ', ,' ' ' ~L~L3~7~;
1 Vo~es Proskauer 2 Acetoin is produced from sodium pyruvate and 3 indicated by the formation of a red color after addition 4 of KOH and alpha-naphthol.
6 O.N.P.G.
7 Beta galactosidase hydrolizes orthonitrophenyl-8 beta-galactose, which liberates the yellow colored 9 orthonitrophenyl.
;'' 11 Citrate, Malonate, Acetamide, Tartrate 12 The utilization of these substrates as the sole 13 source of car-bon for metabolism results in a rise in pH
14 that is detec-ted as a s~ift of green to blue by the pH
15 indicator bromthymol blue.
17 O.F. Carbohydrates 18 Oxidation or fermentation of a carbohydrate 19 results in acid formation. The consequent drop in pH
20 is detected as a shiftfrom blue or dark green to yellow 21 or light green by the pH indicator bromthymol blue.
23 Nitrate 24 The ability of an organism to reduce nitrate 25 to nitrite is detected by the addition of alpha-26 naphthylamine and sulfanilic acid, which produce a red 27 color in the presence of nitrite. To confirm that nitrate 28 has not been reduced to nitrogen gas, zinc powder is added 29 to all negative tests to detec~ the pr~sence of unreduced 30 nitrate. This test is performed before the plate is read 31 by the instrument, and -the results are manually entered 32 when the instrument's display ~ueries the operator.
34 Starch Hydrolysis Starch reacts with Gram's iodine to produce a 36 blue-black color. If an organism hydrolyzes starch, the 37 absence of starch is detected by the iodine yielding a 38 brown rather than blue-black color.
~3~36 1 Oxidase 2 Like nitrate, this test is performed "off-line"
3 and manually entered into the instrument on command. The 4 recommended oxidase test is the tetramethyl-p-phenylene-5 diaminedihydro-chloride procedure described on page 679 6 of the second edition of the ASM Manual of Clinical 7 Microbiology cited above 9 MacConkey MacConkey's agar is a selective medium that is 11 used to differentiate major groups of gram negative micro-12 organisms from one another. Growth or no-growth on this 13 medium is manually entered into the instrument on command.
14 The organisms presently identified by this 15 system are gram negative bacilli. These fall into two 16 major classifications: enteric, or dex~rose fermentors;
17 and non-enteric, or dextrose non-fermentors. The follow-18 ing list includes a number of organisms that are identi-19 fied by the present system.
~7~;
1 D~ OS~ 9~ K~S DE~TROSE NON-EF,~ TEP~S
3 Escherichia coli Pseudomonas aeruginosa 4 E. coli indole neg. Ps. fluorescens E. coli H2S pos. Ps. putida 6 E. coli urea pos. Ps. cepacia 7 E. coli adecarboxylata Ps. maltophilia 8 Shi~ella dysenteriae Ps. stutzeri g Sh. flexneri Ps. putrefaciens Sh. boydii Ps. pickettii 11 Sh. sonnei Flavobacterium meningosept.
12 Ed~ardsie~la tarda Flavo. species 13 Sal~onella enteriditis Acinetobacter anitratus 14 Sal. typhi Ac. lwofi Sal. cholera-suis Achromobacter sp.
16 Sal. parat~Jphi A A. xyloso~idans 17 Arizona hinshawii Moraxella 18 Citrobacter freundii B. bronchiseptica 19 Ci. diversus Alkaligenes sp.
Ci. amalonaticus Eikenella corrodens 21 Klebsiella pneumoniae CDC Group II F
22 Kl. oxytoca CDC Group II J
23 Kl. ozaenae CDC Group II K-l 24 Kl. rhinoscleromatis CDC Group II K-2 Enterobacter aerogenes CDC Group IV C-2 26 Ent. cloacae CDC Group VE-l 27 Ent. agglomerans CDC Group VE-2 28 Ent. gergoviae 29 Ent. sakazakii Haniae alviae 31 Serratîa marcescens 32 Ser. liquefaciens 33 Ser- rubidea - ~ -34 Proteus w lgaris Prot. mirabilis .
36 Morganella morganii 37 Providencia rettgeri J 38 Prov. stuartii ~ .
, : .
,. . ~ ;, . . "
, :.:~ .
: ~ .. , , , ,, ~, , , , , , :
- 5~ -1 DE~TROSE FERMENTERS (continued) 3 Prov. alcalifaciens 4 Yersinia enterocolitica Y- pestis 6 Y- pseudotuberculosis 7 Chromobacter violacium 8 Pasteurella sp.
g Past. multocida Aeromonas hydrophilia 11 Vibrio cholera 12 Vibrio par-ahemolyticus 13 V. alginolyticus 14 Plesiomonas shigelloides 16 The data base in the present invention thus may 17 be the frequency of occurrence of twenty-one chemical 18 reac~ions with seventy-seven different organisms, or a total 19 cor.~bination of 1,617 probabilities.
21 The following three cases will be considered: -22 1) A Pseudomonas aeruginosa infec~ion of the kidneys o~ a 23 patient with chronic pyelonephritis, 2) A case of Klebsiella 24 pneumoniae infection of the lungs in an alcoholic, 3) A case o~ Salmonella enteritis in a patient who owns a 26 pet turtle.
28 The first patient was a middle aged woman who has 29 had chronic urinary tract infections with flank pain and fever for many years. She was sent to a laboratory where 31 a urine specimen was passed for bacterial analysis. The 32 urine was streaked onto an agar plate with a calibrated 33 platinum Ioop so that a quantitative estimate of bacteria 34 growth could be obtained. A growth of similar appearing colonies were obtained that numbered over 100,000 per 36 milliliter. The organism grew on McConkey's agar, and was 37 both ni~rate and oxidase positive. Several simi.lar colonies ~ ~ .
~L~377 l were suspended in saline (salt solution) for evaluation b~
2 the ins~rument being described.
4 The second patient was an alcoholic who was found unconscious and brought to a co~nty hospital. He subse-6 quen~ly suffered pne~lmonia, and a sputum specimen was 7 obtained for bacterial evaluation. When the sputum cup 8 was sent to the laboratory, representative portions were 9 streaked onto various types of agar plates, and the next day, colonies were noted that to the technologist did not 11 appear to be normal flora. A gram stain of these organisms 12 revealed gram negative bacilli, so the tec~nologist made the 13 decision for further evaluation. A suspension of the 14 bacteria was made in saline for further testing.
16 The third case was that of a grade school pupil 17 who had a sudden onset of diarrhea. Upon questioning, the 18 physician learned that the child had recently been given 19 a pet turtle. The mother was asked to send the child and tur~le to a local laboratory so they could obtain stool 21 specimen from both the child and the turtle. The technolo-22 gist plated representative parts of the stool on several 23 types of selective agar, and there were some suspicious 24 colonies that warranted further evaluation. These colonies were placed into saline for further identification by the 26 present invention.
28 All of the three specimens had bacteria isolated 29 from them that contained bacteria suspicious for disease.
These bacterial colonies were placed into a saline suspension 31 and thoroughly agitated to obtain optimum dispersion. The 32 saline suspension was poured into a sterile plastic dish, 33 and a transfer lid containing a matrix o~ 96 prongs was 34 used to inoculate a drop of bacterial suspension into each of 96 wells of a plastic tray. The wells contained anti-36 microbic dilutions as well as biochemical substrates and 37 indicators. The trays were allowed to incubate o~ernight;
,A
' . ' .' ' ;: , " ' . ' ' ~ `.
- 5~ -1 and the next day, following proper calibration, were placed 2 into the instrument for automatic identification. The 3 instr~mlent made a reading of each of the wells through 4 appropriate filters,and a table residing in computer memory interpreted these digitized voltages as either a positive 6 reaction or a negative reaction,, The computer then went 7 through each of the seventy-seven possible organisms and 8 computed the probability of occurrence. With the three 9 present organisms, the frequencies and results obtained by the instrument are summarized in the following table:
13.
'19 ..
3~ , .., 38 ., .
.377~
1 PS.AERUGI~IOS~ K.PNE~IONIAE S EIITERITIDIS
2 BI0C~MICAIJ Fl~Q. R~SULT ~ ~ . RESULT
3 De~trose 00.1 neg 99.9 pos 99.9 pos 4 Sucrose -- 99.0 pos 00.6 neg 5 Sorbitol -- 99.4 pos 95.0 pos 6 Raffinose -- 99.2 pos 3.0 neg 7 Rhamnose -- 99.3 pos 90.0 pos 8 Arabinose -- 99.9 pos 99.9 pos 9 Inositol -- 98.0 pos 30.0 pos 10 Adonitol -- 90.0 pos 00.1 neg 11 Cellobiose -- 99.O pos 5.0 neg 12 Urea 50.0 pos 90.0 pos 00.1 neg 13 H2S 00.1 neg O0.1 neg 95.0 pos 14 Indol 00.1 neg 6.0 neg 1.0 neg 15 Lysine 00.1 neg 98.0 pos 95.0 pos 16 Arginine 95.0 pos 1.0 neg 50.0 pos 17 Ornithine 00.1 neg 1.0 neg 97.0 pos 18 Tryptophane 00.1 neg 00.1 neg 00.1 neg 19 Esculin 00.1 neg 99.O pos 1.0 neg 20 V.P. 00.1 neg 90.0 pos 00.1 neg 21 O.N.P.G. 00.1 neg 99.0 pos 1.0 neg 22 Citrate 95.0 pos 98.0 pos 90.0 pos 23 Malonate 90.0 pos 94.0 pos 00.6 neg 24 OF Glucose 95.0 pos -- __ 25 OF Maltose 00.1 neg 26 OF Xylose 85.0 pos -- __ 27 Acetamide 90.0 pos 28 Tartrate 00.1 neg _ __ 29 Starch 00.1 neg -- --30 Nitrate 75.0 pos -- --31 MacConkeY 85.0 pos 32 Oxidase 99.9 pos --rJ
. .
~L3t^~7~6 1 The table o~ probabilities stated a~ove are 2 relevant only to positive reactions. If, in fact, the 3 reactions were negative, the result would be 1 0 minus this 4 probability. For instance, with the sucrose reaction of S.enteritidis, the probability for a positive reac~ion would 6 be 00.6%, but since the reaction was negative, the actual 7 probability is 99.4%.
g Each of the actual probabilities of each biochemical reaction are cumulatively multiplied for each of the se~enty-11 seven organisms in the data base to obtain the net probability 12 for each organism. The organism with the highest net 13 probability is the most likely organism. If the net prob-14 ability of the most likely organism is less than 1 x 10 6, then the instr~ment flashes a warning to the operator that 16 the frequency is low, and possible technical errors should 17 be checked out. If the net probability is greater than this 18 value, then the instrument proceeds to normalize. This is 19 done by dividing each of the organisms' net probabilities by the sum of all of the net probabilities. Thus, an 21 estimate of the probabilities relative to each of the 22 organisms is obtained. In the case of the three examples 23 described above, these normalized probabilities are greater 24 than 95%, so the instrument proceeds to display the most likely organism's probability on a thirty-two character 26 Burroughs display and print the most probable organism's 27 genus and species on a Practical Automa~ion Model DMTP-9 28 alpha-~umeric ticket printer.
A specific example of procedure 31 (Figs. llA, llB, and llC):
32 The flow sheets llA, llB, and llC illustrate 33 procedure according to the present invention, after the 34 type of tray-checking etc. shown in Figs. 7A, 7B, and 7C.
~.
.3~'7~
1 Thus, when the device is instructed to commence, 2 at start 300, the apparatus determines at 301 whether there 3 is a tray 14 containing biochemicals in place or not. If 4 there is no such tray 14 the light transmission will be the same for all wells and such a known transmission will 6 give the answer "No"; then the computer returns at 302 to 7 the main program. If the light transmissions result in the 8 answer "Yes", then the dextrose voltage is read at 303 and 9 the value compared at 304 with t:he stored positive-negative table in the computer. If the answer is positive, the 11 organism is a dextrose fermenter and each of the biochemicals 12 in receptacles 1 through 21 are read in at 305.
14 If the answer is "No", the organism is a dextrose non-fermenter, and the nex~ stage is for the operator to 16 enter manua~ly whether the organism grew or did not grow on 17 MacConkey's agar, at 306. This is an "off-line" test. If 18 it grew, the operator enters "1"; if not, he enters "0".
19 The processor waits for this input and stores the data at 307. Next, the operator manually enters "1" if oxidase is 21 present or "0" if it is not present, as determined by another 22 "off-line" test, queried and entered at 308; the processor 23 again waits for the data and stores it at 309. A third 24 "off-line" test used when the organism is not a dextrose fermenter is the nitrate test, and the operator is queried 26 and enters the nitrate at 310 as either "1" or "0", depending 27 on whether it is present or not, and at 311 the processor 28 again waits for the data and stores it. After that, the 29 biochemicals receptacles 10-27 are read in at step 312.
31 Thus, if the dextrose test is positive, the results 32 for biochemicals 1 to 21 are read into the program at 305;
33 if the dextrose results are negative, the presence or absence 34 of MacConkey's growth, oxidase, and nitrate are determined and ~hen the results for biochemicals 10 to 27 are read in.
. . , ~ , , , r --~37~
- 62 ~
either event, the next step (after either 305 2 or 312, whichever is applicable) i5 the step 3~3 7 where the 3 stored data in the tables is used to determine for each 4 biochemical still pertinent (l to 21 or 10 to 27) whether each is posi~ive or negative. Step 314 then stores the 6 positive and negative indications in a table or as packed 7 data, in terms of probabilities. If any reaction is negative, g then the probability used is 1.000 minus the actual negative 9 probability, as box 315 shows. E.g., a negative probability of 99.5% is stored as 0.005. At this point, in box 316, 11 the biotype is calculated and stored.
13 The next step 317 (Fig. 11B) multiplies the prob-14 abilities of each taxon and accumulates the sum, and then at step 318 sets up a table of non~normalized probabilities 16 for each ta~on. From this, the computer then sorts at 319 17 the three organisms with the highest probabilities.
19 If at step 320 the most frequent probability found is less than 1 x 10~6, step 321 displays to the operator 21 "VERY RARE BIOTYPE" and instructs the operator to call the 22 company that provides the trays, and then the device returns 23 at 322 to the main program for this answer is unacceptable.
24 If the answer if "No" then the comparator determines at s~ep 323 whether the most frequent organism is greater than 26 1 x 10- 6 but less or equal to 1 x 10- 5 ~ If the answer is 27 "Yes" the display at 324 says "RARE BIOTYPE-PRINT?
28 (1 or 0)". If the operator wishes to go ahead and print 29 this information he presses "1" on the keyboard, steps 325 and 326. If he presses "O", the computer returns to the 31 main program at 327.
~- 32 33 If he presses "1"~ or if the most probable 34 organism has a probability greater than 1 x 10- 5 ~ then the computer normalizes the three most probable organisms at 36 step 328, by dividing the three highest frequencies by the 37 sum of all the frequencies.
-:ll13 77~ !r ~ 63 l In Fig. llC, the outpllt from step 328 is dealt2 with. If the most probable organism has a normalized 3 frequency between .950 and .999 (as asked at step 329), then 4 the machine prints that one (or ones) out at step 329 in terms of probability percentage (e.g., 98.21%) and returns 6 to the dextrose positive flag at 331.
8 If the most probable organism has a normali~ed 9 frequency between 0.850 and 0.950 (step 332), that organism and its percentage are printed out at step 333, and the 11 program goes to step 331 to determine again w~ether the 12 organism is a dextrose fermenter or not.
14 If the response to both steps 329 and 332 is negative and if the most probable organism has a probability 16 bet~een .75-.85 at step 334, then it is printed out at the 17 step 335 and the percentage indicated, and the program at 18 that point goes to step 331 for dextrose determination.
l9 If the relative probability of the most probable organism is less than 75%, the display first says "LOW SELECTIVITY
21 RECHECK" at step 336, followed one second later by 22 "OOOOOOOOOOOOO--XX.X%" at step 337, followed in ~urn one 23 more second later by "STILL WANT TO PRINT? (1 OR O)" at 24 step 338. If at step 33~ the operator does want to print, he presses l'l't at step 240 and the information is printed, 26 followed by sending the program to step 331 for the question 27 of whether the dextrose is posi~ive or negative. If the 28 answer is "no", he presses "O" and returns to the main 29 program at 341.
31 The step 342 asks whether the dextrose is positive, 32 and if the answer is "yes", then the program goes via an 33 output line 342. If the answer is "no" then the org~anism 34 is evaluated for its sensitivity to the antibody Colistin.
Comparison of Colis~in with expected result for the organism 36 is made at step 343 and then at step 344. The program looks 37 up a ta~le to see whether the result is correct. If not, - 6~ -1 it displays "R~CHECK ID & COLISTIN DISAGl~E" and goes to 2 the output llne 342. If the answer is "Yes", a similar 3 procedure is performed at steps 346, 347, and 348 with the 4 antibiotic Nitrofurantoin.
6 A "Yes" result leads to step 349 where the most 7 probable organism and biotype number are printed. The output 8 from de~trose positive along line 342 and from the two 9 recheck steps 345 and 348, also go to this step 349.
After that has been printed, the computer returns to the main ll program at 350.
2g ~3~7 ]. TABIIE 0~ PROGP~M FOR ANTIEIOTIC SUSCEPTIBILIT~ TESTING
_ 3 The following l.isting constitutes the program ~or 4 antibiotic susceptibility testing in hexadecimal code ~or direct loading into the programmable read only memory 6 142. -8 Address _ro~ram Instructions 12 .: ol oo ~C,E :~O.CO'; Gr 'C5 61- ,C4 85 ,2C ~.7,.C5''6F, O,7,,~6 .?F ~.73 ,~0110 06, C6'O4- E7 0.7- 6r 0~ -85.. OF...A7:'O0 8~, 0D.^.,7.,C;7 ',~
13 i 01?.0 2r)'A7 OV '.,r OD ~6' FO A7 OC .86 OD A7~ O"a,5F C~r' '3G I
14 , Ol3O~,Fr A7 CE .'C5 2C E7~ OF .7F~ 8? E;~ 7F ;~7 E9 Cl C~; 01l ..0l~:O 7E Fl 3,~ Cl Cl 01 Gl CI 01 01 Oi' OI 01 01 01 ~CI~
150l5~ bl Ol Ol Ol Ol Cl i~ 37,Dl '7F. ~7 EO 7F 87 ~l 4r-' 16 ~, 0150 ~? 3D ~:3 ~4 FO,,26 b5 .CE F5;,i,4 43 2C .CC ~5 ~,. '.~-0l70 2~ 43'g4, FO 2? 07 C~ l~5.'.6l ?. ~? Dl,:39, r76 47 8D, 17 ' 0l:3O Ir, 43 B4~,F.O. 27,.0~ CE rG .IO'.. B7.. 5?.EO ,20 07lCE~I'6' 18 , Cl9O IF,~I3,,~,7~:3r E1,~ D,D F3 36 BD .F2 FS B.D.. ,r2, cr. 39 '37 ,~ 51~0 rJO 06. C6, 5O'ED F2 V9 36 50 C4 BD. F2 F5.B5 ,C7O C4' 19 . 013,0 ,39 01 Cl C~ D .F,Q~ 56 36 37.~.DI,,2.G' 5C CE F5 D4 ~D
c I cn F2 4f~ 20 .C3 t;D F2 50 V6 57 D5 27 03 7E F2 90 j~36 OIDO ~7 DG 2; FO r;D r3 35 rV~D ~2.FS ,C~. F6' 2E V~ E2 CF
21 ; OIEO ,~6, Of~ BD r2 r3~ CE rS4 00 36. 57 EO 25, 03 C_ 8~ .0 . .'OIFO.36 j'7'r)5 27 b3 7E F2 .98 .3D F3 60 E;5 87 D5 27 IS'.:
23 :0200 ~0 ~7 EO 27 03 7F.:~7.. E8 B6 .3.?. ~ 1 .'2.7.03 7F j'37 E9 0210 7E F2 9~ 01 Ol,.E6i7 ~0 .2? Q3 B?: 5? E3 36 ~7 El 24 . 0220 27.. 03 ~7,:,87 E9 CE F5 ,'3E~'~33 r''2. ~1'7 'v~j. g7 '3:5 .27 .,.
2t 0230 CE.'F5. 9V,3D,F,2,,,47.:36; ~:7.D4 ~7` F8 ,C,E ,F5 29 Vj~ F2, ~:~.G24O 47 B~ 87 ~D6 27- 03 ~7E ~FO ~B.1 ~5 ~37 ~D7~26 03 ~7E Fl ~
26 l 02.50 3v, 01 01 .01. ~3D FO 43 æ6 57 :Dl 2? 03.7r Fl 28,B6.
!.~0260 !37 EO ,21,7,OOS v6 87 E8 26' lO CE FS JjI~ ~7~ F 1 ~ n8 E'~6 27 1 027 0 ~r37 .EI 27.. ''F5 v6 87 ~9 27 FO 86 C~3 P.D F2 vD Cl. F5 28 L ~O-n30 . A9 . '3D F2 47 B6 87 D5 27 b3 7E r 2 .9~ ~6 ~37 D~ 27.
029b FO CE~r5 C9 BD F.;2 47''.~6 ~37 D5 2.7; 03 7E F;2 .9S B6, 29 ~;. 0 2.~ 0 ~,7., D7 . ',27;.: F.0 ~ 6 ' FFr.~7 ~ i, AO, B7~8,7 . ,A3 ~7., ~ 7 A ! ~ n :~ 02~30 F,2 Fl~ :~.D:~F3. J-17 7E F6 .FE. bi 0101: 01 01. 01 01 01 02CO Oi OI ;oi ol, oi o:i ol: bl: ol oi oi oi :ol :!Oi.;OI `01 31 C2DO of~:o~ ol .ol .oi :o.l ol :oi bl:,ol -oi OD.ol; bl oi ol .
32 02` 0`01 Oi 01^ Ol, OL. ~;1 01 jO 01 i I 1 O g 3~ .
' ' ' ' `' ' " ~' ' . , ' :
" . ' , ~ ' .'' . . :
:; ' ' ' ` : , ~ ' ' ' . '. ' , .
~ " ' ' ,' . ' : ' ' ' - ~13~'7~6 3 ~.~,,r~,~ ,.q~.~a~s~ ~r~ }-~r;-~s~ rq~ ~t~ f~
4 ~06CO:.,03 lO.`~i9 4E'.53,.45' S2 5'4'?0,.5il Z15,,,,53'5,4 2G 54'52' oSlo~ili 5$.:.05..'.i'5,,.49,4E~53 45-52''54 20,~i3;41.'/1C 43.42 1.0520 52 ~1l 5~1 ~15 20 5~1 52 ~il 59 00 20 4~ ~1~ 5~ 5~1 20 ~,06.30;54.45,53'54 20,50.52 45 53 53 20 52 55 4~ 20,4F
6 ;.C6'lO.S~ ~0.. ~13 ~ IC ~9,42 5~ 5~ ~15 O".'l~
7 ,,.065Q 4? 4~ 52, ~! ~ 54 49 4F 4F. 20 52 45 5.1 55 49 52 45 ~.06~0 44 Q7 11 55 4E 4? 44'45 4E'54',~i~ 4G 49 45 44 2a 8 r:CG7,0 54 52 4:1 59 07 11 54'52 41'59 20'~19 4~ 20 42 41 9' 106~0 43 4~.57 ~1l 52 4~1l53 OD OS 45,:'52 52 4F 52 OA 0 ,0~90,52,,,45:4D'4F:.55.il5 20' 54'52 41'59,0A OC 43 4C 41 ~,'C5AO''53 ~S 20;~ 52'41 S7~ 45 52.01.IT' 53 45 54'20 50, 1.06~0 4~1 54'49,45.'4E 54 2U 49 44 20 41'4~ ~14 20 . 4~
11 ~OGCO.53 45 52-54 20 46 i1~ 5~ 4D'0~ 0~'50 52'45'5?,.53 12 ~.,C5nO'20.. 52:55.4~.,02.. 0~,.. 50.. 52-.45'53,,.53,20.. ~13.4;1 4C 49 -0,~'0 ~12 52',41`5~~'45 06 '13 4~ , 53.,55 4~'46 ~ 3'~
13 ~0~10 4~ 4T 54.,2,0.47.52 4~ 57 5~1 4~ Of~ 1~; 53~54 45_.52 1~ '0700.4~ 4C.45 20.~i3 4F .~ 54 41, 4D 4~ 4 r iii 5iJ ~;5 ~0710 00 OD..~i.7..52....41.4D..20 50:4F,53 20 54 52",.4,1 59.CC
L0720-OD.47.52.41 4D ~'0 4T~. 45.4i 20 54 52 41.,59,, 0!; Q4 16 ~.0730 57 ~ i9 5~1 51 2~ 25 6A 12 ~,AL 64 AOl32.~LC 1~ r~C;
1.0740 RA 00 4^ ob 2A OG'lA oo ~A, 50 pA 25 OA 12 CP 05 17 r,o750 60 ~ 30,4A 15 2A.7G..... A0 35 AO:,19:A0 9A 5C 2~j.31 ~75C 3A 43:4C 54' 5C .05 1~ 2i 29',32 3~ 44.,4C 05 23 31 18 1 0770 3A.42 ilB 5ij 5C 05 2~ 30 3.q 43 4~ 54,5C 05 00 0~
19 07~0'11 19:22'2~ 0~ 2C'29'32 :3~ 43 ~IC 5~1 OS,I~ 20 07~0,29 '3~'3~ '3'4'C 05'2r3 30' 39"4~ '5~'5C Q5 CC'CS
loi.~o 10 Ir.23 2~ 33 'C5 70 73 31 ~A ~3 9C'A4 05 2'i 30 21 1'0730 3~ ~12 4~ 5'4 5C 05 2r~ 3'1 3~ 42'4'~ 54 5C 05..... 2J ~l 107co 3.4 43'4Y..54 5C 05 39 42, 4B 53,5C 64 GC C~ .21 22 ~07~,0 2~ 32'3~, 4'h.'`4C.05 2D'31'3A 42 43 54 5C 05 2~.30 ~Ci_C 3R 43,42 54 5C~5:0C~O~ .22~2i~33 05 OC'03 23 iO7F0 10~ 3.2B'33 05 70 '~ P. ~3,,?6 'A~l C5 C~ ~:
iO300 50 P6 ~7.E0 25 03 CE,.54'FO F~ ~7 D~ 9D F3 60 3G
!oZ10 ~7 ri5 27 03 7E F2 9~.CE ~4 00 D~ 8'7 EO 2G C3.C~
¦'0020.84 AO ~ oi.;.r.o.s7.21~ o i: ~o 31 0.~ 2~ 05 C~ F5 FA-26 iC~qO 7~ Fl 2~ Cr AO 5F 2A-O'1 i~o CE F5 ~5 ~1 o.r~ 2D
27 ~OS40 EF:~ F3 ~C.7E Fi 25 01 01 C! 01 31 Cl OI 01 01 , I G~5C 01 Cl 01 01 01 Ol.Ol pl 0l 01,01 oi ol bl ol ol 28 ! 0~63 ! ol ~ol Ql .bl; o.l 01. ai .ol ol~ 31 ol Ql ol .ol pl.;
29 1'.0~7').01 01 0! C1.01'01.`01~01... 01 01 01 01 ~ 1.-C1 OI
'O~'.O 0,I,'C1,01"0.1,.01,;''01''01j'~01.01 01i''01 .01"'0,r 01 01,01 .G~?C 01.0,1. 01 01 OL Oi 0.1'01 01 01 .0l 01,GI, 01 01 'Cl.
O~RG 01,bl OI~O'i 01 bl.Ql 01 01'01 01 01 01~01 01 .01 . . . . O ~ i ~ O ~ O I O I , O I O 1 . O I O 1 0 1 0 1 . p 1 0 ,1 , o, i ~ ,o i i C) I O i ~ O I
32 ~.C~. 01.,.'01'0.1''01'01,-01' 01 01:01'.01.C1.01 ,0,1~CI Cl OI -, '-o.~Dn .oi ~o~i ol o.I :o.i..~i ol ol oi ol` ol; c! ol~ oi 'OI'''C,1 33 -OSEO 01'01,01 0'1'.Oi.b1 01:'01.... 0'i Oi.01:01~01 .,01 01"01 34 i'o~o Ol':Ol`'Oi':n'1.""01''01' 01 0'1:'01'01 nl 0.1-01,01'Gl 01 ' 33 ,, ' , : :i ~ . : . : .: .. . .
1 Concerning Comparisons 2 As indicated above, the apparatus of ~his inven-3 tion is capable oE performing various types of compari-4 sons. Any specific comparison cLepends upon what method is 5 being used and which types of comparison are appropriate.
6 In some instances, there may be only one compari-7 son, the specific type of comparison depending on the 8 particular apparatus or particular method being used. For 9 example, it is advisable to relate or every sample the si~nal received from each well with a signal of a reference 11 transducer. Such comparison negates the e~fect of the 12 variation of light intensity or power fluctuation with 13 time.
14 A second type of comparison is often made, in 15 addition to the first one. This may be considered as 16 a type of calibration procedure aimed at negating the 17 variation of response of the different photosensors. The 18 comparison may be achieved by storing the signals from 19 each of the photosensors before the tray is introduced, 20 and then subtracting from this the corresponding signals 21 generated by each of the wellsafter the filled tray with 22 its cultured samples has been read. This technique of 23 eliminating transducer-to-transducer variation is impor-24 tant.
25 Further refinements, which are not necessarily ~!:
26 crucial, may be added to eliminate further well-to-well 27 variation. For e~ample, variations in the plastic trays 28 or their contents may affect the accuracy of a reading.
29 One way to eliminate this problem is to calibrate with 30 an empty tray instead of calibrating without any tray in 31 the holder. A single empty tray may be used, assuming 32 that all the trays to be used are substantially identical.
33 Another approach is to compare the wells of each individual 34 tray when empty with the results obtained after filling 35 them with liquid and culturing the liquid. This is more 36 time consuming and not usually necessary, but it is more 37 accurate. With suitable multiplexing wired into the '.' ~3~B~
1 device, however, this becomes quite practical. rrhus, 2 it is possible to eliminate the variations in the 3 signal fluctuating with time, to eliminate the vari-4 ation of one sensor versus another, and also to compensate 5 for tray-to-tray and well-to-welL variations.
6 A third type of compar:ison may be used for 7 certain tests, such as the MIC test, where the signal 8 level indicating bacterial growth is differentiated 9 from the signal level indicating no growth. This may be 10 accomplished by comparison between various wells on the 11 tray; that is, some wells may be control wells or sterile 12 no-growth wells, in which there is no growth or which 13 are inoculated with suitable inhibitors. There is a 14 possible interpolation between the values of growth and 15 no-growth, as discussed above. Alternatively, by experi-16 mentation, one can determine a signal value that dif-17 ferentiates between growth and no-growth, and this decision 18 point may be used instead of one derived by controls on 19 board each tray.
Some of the claims which follow specifically 21 identify the types of comparisons made, while others 22 merely call for suitable comparisons to be made or for 23 apparatus which make these comparisons possible.
24 The above described preferred embodiments provide 25 apparatus and a method for automatically determining the 26 minimum inhibitory concentration of a plurality of 27 different antibiotics necessary to stop growth of an 28 infective organism being tested. Minimum inhibitory 29 concentration information is also transferred to dosage 30 information by the apparatus and method of the invention.
31 The required time to perform such a test is greatly reduced 32 in comparison to other methods, a great deal more infor-33 mation is provided, and accuracy is improved. ~arious other 34 embodiments and variations to the preferred embodiments 35 will be apparent to those skilled in the art and may be 36 made without departing from the spirit and scope of the 37 following claims.
38 We claim: