Relationship Between18F-FDG PET/CT Scans andKRAS Mutations in Metastatic Colorectal Cancer
Abstract
Several studies have shown thatKRAS mutations in colorectal cancer (CRC) result in the lack of response to anti-epidermal growth factor receptor–based therapy; thus,KRAS mutational testing has been incorporated into routine clinical practice. However, 1 limitation of this test is the heterogeneity ofKRAS status, which can be either intratumoral heterogeneity within an individual primary CRC or discordantKRAS status between a primary CRC and its corresponding metastases. We previously reported that18F-FDG accumulation was significantly higher in primary CRCs with mutatedKRAS than in those with wild-typeKRAS. However, the clinical utility of the previous report has been limited because endoscopic biopsy for testingKRAS status is safe and feasible only in primary CRC. The purpose of this study was to investigate whetherKRAS status is associated with18F-FDG accumulation in metastatic CRC and whether18F-FDG PET/CT scans can be used to predict theKRAS status of metastatic CRC.Methods: A retrospective analysis was performed on 55 metastatic CRC tumors that were identified by18F-FDG PET/CT before surgical resection. Maximum standardized uptake value (SUVmax) of the respective metastatic tumor was calculated from18F-FDG accumulation.Results: From the analysis with the 55 tumors, no significant correlation was found between SUVmax andKRAS status. We next analyzed only tumors larger than 10 mm to minimize the bias of partial-volume effect and found that SUVmax was significantly higher in theKRAS-mutated group than in the wild-type group (8.3 ± 4.1 vs. 5.7 ± 2.4, respectively;P = 0.03). Multivariate analysis indicated that SUVmax remained significantly associated withKRAS mutations (P = 0.04).KRAS status could be predicted with an accuracy of 71.4% when an SUVmax cutoff value of 6.0 was used.Conclusion:18F-FDG accumulation into metastatic CRC was associated withKRAS status.18F-FDG PET/CT scans may be useful for predicting theKRAS status of metastatic CRC and help in determining the therapeutic strategies against metastatic CRC.
Colorectal cancer (CRC) develops through accumulation of genetic alterations in oncogenes and tumor suppressors. Mutations in theKRAS gene occur in approximately 40% of CRCs and involve codons 12 and 13 in more than 90% cases. Several studies have shown thatKRAS mutations predict a lack of response to therapies targeted to the epidermal growth factor receptor (EGFR) (1,2). The anti-EGFR monoclonal antibodies cetuximab and panitumumab are currently recommended to use only for CRC tumors with wild-typeKRAS, although a wild-typeKRAS does not guarantee a response to either antibody.KRAS mutational testing of primary CRC samples has been incorporated into routine clinical practice for the purpose of treatment algorithms. However, 1 limitation ofKRAS mutational testing is the heterogeneity ofKRAS status, which can be either intratumoral heterogeneity within an individual primary CRC (3) or discordantKRAS status between a primary CRC and its corresponding metastases (4,5). Another limitation is failure to determineKRAS mutational status due to poor DNA quality of biopsy samples. In addition, mutational testing requires tumor tissue samples resected by biopsy or surgery, but the samples from metastatic tumors are usually difficult to access and may need invasive procedures.
PET/CT with18F-FDG is used to evaluate glucose metabolism by measuring uptake of18F-FDG, a glucose analog. This is a less invasive tool for diagnosis, treatment response monitoring, surveillance, and prognostication of CRC.18F-FDG is transported into cells via glucose transporters (GLUTs) and then phosphorylated by hexokinases to FDG-6-phosphate, which becomes trapped within the cells. In most types of cancers,18F-FDG accumulation depends largely on the glucose transporter-1 (GLUT1) and the rate-limiting glycolytic enzyme hexokinase type 2 (6). For CRC, several recent studies have suggested that GLUT1-mediated18F-FDG accumulation is more essential than hexokinase activity (6). It was previously reported that in CRC cell lines, under normoxic conditions, the increase in GLUT1 expression and glucose uptake is critically dependent onKRAS mutations (7). Using human clinical samples, we previously reported thatKRAS mutations significantly increased18F-FDG accumulation into primary CRC possibly through upregulation of GLUT1 expression but not hexokinase type 2 expression (8). Hypoxia-inducible factor-1α (HIF-1α) is a transcriptional factor that mediates cellular response to hypoxia, including angiogenesis and glucose metabolism. In hypoxic cells, HIF-1α enhances glycolysis by inducing glucose transporter and several enzymes involved in glycolysis (9). In both in vitro and in vivo animal experiments, we recently showed that mutatedKRAS caused higher18F-FDG accumulation possibly by upregulation of GLUT1 and at least partially by upregulating HIF-1α induction under hypoxic conditions (10).
In a retrospective analysis of 51 primary CRCs, we previously reported that maximum standardized uptake value (SUVmax) was significantly higher in primary CRCs with mutatedKRAS than in those with wild-typeKRAS and thatKRAS status could be predicted by18F-FDG PET/CT scans with an accuracy of 75% (8). The study by Kawada et al. (8) was the first clinical study showing the causal relationship betweenKRAS mutations and18F-FDG accumulation using18F-FDG PET/CT scans in a variety of cancers. There is also emerging evidence from other groups that18F-FDG accumulation reflectsKRAS mutational status of CRC and non–small cell lung cancer (11–13). However, the clinical utility of these findings has been limited because endoscopic biopsy forKRAS mutational testing is safe and feasible only in primary CRC. It has not been investigated whether the similar relationship betweenKRAS mutations and18F-FDG accumulation exists in metastatic CRC. In particular,KRAS mutational testing derived from metastatic CRC samples is usually difficult because of limitations in sample availability. Therefore, the purpose of this study was to assess whetherKRAS mutations are associated with18F-FDG accumulation in metastatic CRC and whether18F-FDG PET/CT scans can be used to predict theKRAS status of metastatic CRC. To our knowledge, this is the first clinical study showing a causal relationship betweenKRAS mutations and18F-FDG accumulation in metastatic CRC. Our study suggests that18F-FDG PET/CT scans may be useful to determine therapeutic strategies for CRC by predicting tumor response to anti-EGFR antibody therapy.
MATERIALS AND METHODS
Study Population
Sixty distant metastases were obtained from 38 CRC patients undergoing18F-FDG PET/CT scans before surgical resection at Kyoto University Hospital between April 2009 and March 2014. The diagnosis of metastatic CRC was confirmed by pathologic examination of surgical specimens. No patients received chemotherapy or radiation therapy 6 mo before18F-FDG PET/CT scans. Five distant metastases were excluded because they had the following non–tumor-related factors that can affect18F-FDG accumulation: uncontrolled diabetes mellitus—that is, a blood glucose level of 150 mg/dL or greater (n = 4)—and severe inflammation with C-reactive protein of 5.0 mg/dL or greater (n = 1). Finally, fifty-five distant metastases obtained from 35 CRC patients were included in this retrospective study. This study protocol was approved by the institutional review board of Kyoto University, Kyoto, Japan, and all patients provided their consent for data handling.
PET Imaging and Analysis
The methods for PET/CT imaging and quantitative analysis were detailed in our previous report (8). PET/CT scans were performed using a combined PET/CT scanner (Discovery ST Elite; GE Healthcare). This system integrates a PET scanner with a multidetector-row CT (16 detectors) scanner and permits the acquisition of coregistered CT and PET images in a single examination. Patients fasted for at least 4 h before18F-FDG administration. We checked patients’ plasma glucose levels just before injecting18F-FDG, and there were no patients whose blood glucose level exceeded 150 mg/dL in this study. Data acquisition started approximately 60 min after the injection of a standard dose of 3.7 MBq/kg of18F-FDG. Initially, starting at the level of the upper thigh, the low-dose CT scans were obtained with the following parameters: 40–60 mA, 120 kV, 0.6-s tube rotation, and 3.75-mm section thickness. The CT images were acquired during shallow breathing, and scanning included the area from the upper thigh to the skull. Immediately after CT, a PET emission scan was obtained with an acquisition time of 2–3 min per bed position. The total acquisition time was approximately 20 min. The CT data were used for attenuation correction, and images were reconstructed using the 3-dimensional iterative reconstruction algorithm called VUE Point Plus. For quantitative analysis, a board-certified radiologist/nuclear medicine physician assessed18F-FDG accumulation on a workstation (Advantage Workstation 4.4; GE Healthcare) by calculating the standardized uptake value (SUV) in the regions of interest placed over the suspected lesions and the normal liver. The SUV was calculated using the following formula: SUV = Cdc/(Di/W), where Cdc is the decay-corrected tracer tissue concentration (in Becquerel per gram); Di, the injected dose (in Becquerel); and W, the patient’s body weight (in grams). For evaluating metastatic CRC, the highest SUV in a metastatic tumor was taken as SUVmax.
KRAS Mutational Analysis
DNA was extracted from formalin-fixed, paraffin-embedded tumor tissue sections using the NucleoSpin DNA FFPE XS (Macherey-Nagel).KRAS exon 2 was amplified by polymerase chain reaction. The polymerase chain reaction products were directly sequenced using an ABI 3130 Genetic Analyzer (Applied Biosystems) according to the manufacturer’s instruction.
Statistical Analysis
All values are expressed as mean ± SD. Differences in SUVmax between mutated and wild-typeKRAS were tested by a Mann–WhitneyU test. The statistical significance of differences inTables 1 andTable 2 was determined by the χ2 test or Mann–WhitneyU test. All analyses were 2-sided, and aP value of less than 0.05 was considered statistically significant. To determine the factors associated withKRAS mutational status inTable 3, multivariate logistic regression analysis was used, and factors with aP value of 0.10 or less were included in the model. The relationship between SUVmax and tumor size was determined by Pearson correlation coefficients. Statistical analyses were performed using SPSS software (version 11.50; SPSS Inc.).
Univariate Analysis of Factors Associated withKRAS Status (n = 55)
Univariate Analysis of Factors Associated withKRAS Status in Tumors Larger Than 10 mm (n = 42)
Multivariate Analysis ofKRAS Status in Metastatic CRC (n = 42)
RESULTS
Patient Population
The characteristics of patients and their metastatic tumors are presented inTable 4. The study group consisted of 55 distant metastases (liver,n = 38; lung,n = 11; distant lymph nodes,n = 4; peritoneal dissemination,n = 2) obtained from 35 CRC patients. All metastatic tumors were surgically resected within 30 d after18F-FDG PET/CT scans.KRAS mutations at codons 12 and 13 were found in 21 and 9 (38% and 16%, respectively) of the 55 metastatic tumors, whereasKRAS was wild-type in the remaining 25 samples (46%). SUVmax in the metastatic tumors ranged from 1.2 to 19.7 (5.9 ± 3.6).
Tumor Characteristics (55 Metastases Obtained from 35 Patients)
Correlation Between SUVmax andKRAS Mutations
On the basis ofKRAS mutational status, distant metastatic tumors were classified into 2 groups: tumors with wild-typeKRAS (n = 25) and those with mutatedKRAS (n = 30).Table 1 shows the results of the univariate analysis for each factor. SUVmax in the mutatedKRAS group was not significantly different from that of wild-typeKRAS group (6.3 ± 4.2 vs. 5.4 ± 2.6, respectively;P = 0.84;Fig. 1C). However, the tumor size of the mutatedKRAS group was smaller than that of the wild-typeKRAS group, although not significantly different (P = 0.06). SUVmax can be underestimated because of partial-volume effect, particularly when tumor size is small (14). In fact, we found that SUVmax was significantly correlated with tumor size (Pearson correlation coefficient,P = 0.006; Supplemental Fig. 1A [supplemental materials are available athttp://jnm.snmjournals.org]).
(A) A 78-y-old man had 1 liver metastasis (diameter, 23 mm) with mutatedKRAS.18F-FDG PET/CT scans showed intense accumulation of18F-FDG in liver tumor (arrow; SUV, 8.3). (B) A 61-y-old man had 1 liver metastasis (diameter, 27 mm) with wild-typeKRAS.18F-FDG PET/CT scans showed modest accumulation of18F-FDG in tumor (arrow; SUV, 4.5). (C) Analysis of SUVmax according to status ofKRAS. With all liver tumors (n = 55), SUVmax of mutatedKRAS group was not significantly different from that of wild-typeKRAS group (6.3 ± 4.2 and 5.4 ± 2.6, respectively;P = 0.84; exact Mann–WhitneyU test). Means = bars. (D) With metastatic tumors larger than 10 mm (n = 42), SUVmax was significantly higher in mutatedKRAS group than in wild-typeKRAS group (8.3 ± 4.1 and 5.7 ± 2.4, respectively;P = 0.03; exact Mann–WhitneyU test). Means = bars.
Therefore, we next examined the tumors larger than 10 mm to minimize bias produced by partial-volume effect. On the basis ofKRAS status, tumors were classified into 2 groups: tumors with wild-typeKRAS (n = 23) and those with mutatedKRAS (n = 19).Table 2 shows the results of the univariate analysis for each factor. No significant differences were found between the 2 groups in terms of sex, blood glucose level, serum C-reactive protein level, serum carcinoembryonic antigen level, and tumor size. However, a significant difference in18F-FDG accumulation into the metastatic tumors was found between these 2 groups. Namely, SUVmax was significantly higher in the mutatedKRAS group than in the wild-typeKRAS group (8.3 ± 4.1 vs. 5.7 ± 2.4, respectively;P = 0.03;Fig. 1D).Figure 1 shows typical18F-FDG PET/CT scans of the patients with mutatedKRAS (Fig. 1A) and wild-typeKRAS (Fig. 1B). In the multivariate analysis including factors with aP value of 0.1 or less, only SUVmax remained to be significantly correlated withKRAS mutations (Table 3; odds ratio, 0.78; 95% confidence interval, 0.61–0.99;P = 0.044). We also confirmed that SUVmax was not correlated with tumor size in this setting (Pearson correlation coefficient,P = 0.29; Supplemental Fig. 1B), indicating that these results were independent of tumor size.
We then sought to determine the threshold for optimal differentiation between these 2 groups. Receiver-operating-characteristic curve analysis revealed that the highest accuracy (71.4%) was obtained with an SUVmax cutoff value of 6.0 and that the area under the curve was 0.70 (Supplemental Fig. 2). Sensitivity and specificity for the prediction ofKRAS mutations were 68% (13/19) and 74% (17/23), respectively (positive predictive value, 68%, 13/19; negative predictive value, 74%, 17/23; accuracy, 71.4%, 30/42). These results suggested that18F-FDG PET/CT scans can be predictive of theKRAS status of metastatic CRC.
Concordance ofKRAS Status Between Primary Tumor and Its Corresponding Metastatic Tumor
Of the 55 distant metastases in this study, 49 samples (89%) could be used to assess the association ofKRAS status between paired primary and metastatic CRC samples. The aim was to investigate whether theKRAS status of primary CRC could be used as a surrogate for its corresponding metastatic CRC. Heterogeneity ofKRAS status between a primary CRC and its corresponding metastases was found in 7 samples (14%; 7/49), which is consistent with the frequencies reported in previous studies (15). Namely, 3 metastatic CRCs had mutatedKRAS in codon 13, whereas paired primary CRCs had wild-typeKRAS; 2 metastatic CRCs had wild-typeKRAS, whereas paired primary CRCs had mutatedKRAS in codon 12; and 2 metastatic CRCs had mutatedKRAS in codon 13, whereas paired primary CRCs had mutatedKRAS in codon 12. In addition, discordantKRAS status also existed among metastatic CRCs from the same patient; 1 patient simultaneously had both codon 12–mutated and codon 13–mutated metastases.
DISCUSSION
The American Society of Clinical Oncology suggests that patients with metastatic CRC, having aKRAS mutation in codon 12 or 13, should not receive anti-EGFR antibody treatment (16). Although anti-EGFR antibody therapy has been established in CRC patients with wild-typeKRAS, up to 50% of these patients do not respond to this therapy (17). Failure of EGFR antibody against CRC patients with wild-typeKRAS may result from the intratumoral heterogeneity ofKRAS status (3) and the discordantKRAS status between a primary CRC and its corresponding metastases (4,5). In fact, it remains unclear whether mutational testing of a primary CRC is sufficient to characterize its corresponding metastases. Some studies have found a high (>95%) concordance ofKRAS mutations between primary CRCs and corresponding metastases (18,19), although others have reported a relatively low number (∼70%) (3,4); the most commonly reported rate is approximately 90% (15). In addition, tumor tissue samples obtained by biopsy or surgery are necessary for mutational testing, but samples from metastatic tumors are usually difficult to access and may need invasive procedures. Therefore, alternative noninvasive strategies, such as18F-FDG PET/CT scans, to predict mutation profile could be of value to overcome these limitations. We previously reported that18F-FDG PET/CT scans can predict theKRAS status of primary CRC with an accuracy of 75% (8). In the present study, we have also shown that18F-FDG PET/CT scans can predict theKRAS status of metastatic CRC with an accuracy of 71.4%, particularly in tumors larger than 10 mm. Although18F-FDG PET/CT scans may not be enough for predictingKRAS status determined by mutational testing, they may reflect the macroscopic status ofKRAS mutations. On the other hand, mutational testing of resected specimens may not reflect the macroscopic status of the whole tumor. Miles et al. recently reported that a combination of SUVmax, CT texture, and blood perfusion could potentially improve the accuracy for the prediction ofKRAS status of primary CRC (11). To optimize the clinical application of18F-FDG PET/CT scans, future prospective studies should include a larger number of patients and use standardized protocols for18F-FDG PET/CT acquisition and correction of partial-volume effect. In addition, together with more comprehensive genomic information, it is imperative to investigate whether18F-FDG PET/CT scans can predict the actual response to anti-EGFR–based therapy and survival rates.
18F-FDG PET/CT scans are used to evaluate glucose metabolism by measuring uptake of18F-FDG, a glucose analog. It was reported that metastatic liver CRC tumors more than 10 mm could be detected by18F-FDG PET/CT scans with a sensitivity of approximately 97%, whereas those with a diameter of 10 mm or smaller could be detected with a sensitivity of approximately 45% (20). The molecular mechanisms causing upregulation of glucose metabolism in CRC have not yet been investigated. Yun et al. previously reported that, under normoxic condition, the increase in GLUT1 expression and glucose uptake was critically dependent onKRAS mutations in CRC cell lines (7). In vitro assays using CRC cell lines indicated thatKRAS mutations caused about a 2.0-fold increase in glucose uptake by upregulation of GLUT1 expression (7). We previously conducted a retrospective analysis of 51 primary CRCs and found thatKRAS mutations significantly increased18F-FDG accumulation possibly through upregulation of GLUT1 expression (8), which indicates that18F-FDG accumulation may reflect a genetic mutation—that is,KRAS. Primary CRCs with mutatedKRAS showed about a 1.5-fold increase in SUVmax when compared with those with wild-typeKRAS (P < 0.01). There is also emerging evidence from other groups that18F-FDG accumulation reflects theKRAS mutational status of CRC and non–small cell lung cancer (11–13). In this clinical study, we have shown that, in metastatic CRC tumors larger than 10 mm, mutatedKRAS showed about a 1.45-fold increase in SUVmax compared with wild-typeKRAS (P < 0.05;Fig. 1D;Table 2). To our knowledge, this is the first study to analyze the association betweenKRAS status and18F-FDG accumulation in metastatic CRC.
The mechanisms underlying18F-FDG accumulation into cancer tissues are more complex. These factors include tumor-related (e.g., tumor size and hypoxia) and non–tumor-related components (e.g., diabetes mellitus, inflammation, and chemotherapy) (21–23). It was reported that SUVs of liver metastases in CRC patients who received chemotherapy within 3 mo of hepatic surgery were lower than those who did not receive chemotherapy within 3 mo of surgery (24). In this study, patients with uncontrolled diabetes mellitus and severe inflammation were not included. Moreover, patients who received chemotherapy 6 mo before18F-FDG PET/CT scans were also excluded. HIF-1α has been shown to regulate transcription of GLUT1 in hypoxic conditions (25). When CRC cells were treated under hypoxic conditions, mutatedKRAS enhanced the translation of HIF-1α through the phosphoinositide 3-kinase pathway (26). We recently reported that CRC cells with mutatedKRAS increased18F-FDG accumulation by upregulating GLUT1 and at least partially by upregulating HIF-1α induction under hypoxic conditions (10). In this study, we investigated a possible association betweenKRAS status and HIF-1α expression by immunohistochemical analysis and found that HIF-1α did not correlate withKRAS status in metastatic CRC (data not shown). Our previous studies on primary CRC showed a significant correlation between HIF-1α andKRAS status (10). One possible reason for this discrepancy may be the difference in tumor size (primary CRC, 47.9 ± 20 vs. metastatic CRC, 20.9 ± 14.2 mm;P < 0.01). Another reason could be the difference of microenvironment between the colon and its metastatic sites.
CONCLUSION
This study is a relatively small, retrospective analysis, but it highlights the fact that18F-FDG accumulation in metastatic CRC with mutatedKRAS is significantly higher than that with wild-typeKRAS, when the tumors are larger than 10 mm. Although a larger number of patients are needed to confirm our findings, these results indicate that18F-FDG PET/CT scans could be useful in the prediction ofKRAS mutational status.
DISCLOSURE
The costs of publication of this article were defrayed in part by the payment of page charges. Therefore, and solely to indicate this fact, this article is hereby marked “advertisement” in accordance with 18 USC section 1734. This work was supported by grants from the Ministry of Education, Culture, Sports, Science and Technology of Japan. No other potential conflict of interest relevant to this article was reported.
Footnotes
↵* Contributed equally to this work.
Published online Jul. 1, 2015.
- © 2015 by the Society of Nuclear Medicine and Molecular Imaging, Inc.
REFERENCES
- 1.↵
- KarapetisCS,
- Khambata-FordS,
- JonkerDJ,
- et al
- 2.↵
- AmadoRG,
- WolfM,
- PeetersM,
- et al
- 3.↵
- BaldusSE,
- SchaeferK-L,
- EngersR,
- HartlebD,
- StoeckleinNH,
- GabbertHE
- 4.↵
- AlbaneseI,
- ScibettaAG,
- MigliavaccaM,
- et al
- 5.↵
- MolinariF,
- MartinV,
- SalettiP,
- et al
- 6.↵
- JadvarH,
- AlaviA,
- GambhirSS
- 7.↵
- YunJ,
- RagoC,
- CheongI,
- et al
- 8.↵
- KawadaK,
- NakamotoY,
- KawadaM,
- et al
- 9.↵
- SemenzaGL
- 10.↵
- IwamotoM,
- KawadaK,
- NakamotoY,
- et al
- 11.↵
- MilesKA,
- GaneshanB,
- Rodriguez-JustoM,
- et al
- 12.
- ChenSW,
- ChiangHC,
- ChenWT,
- et al
- 13.↵
- CaicedoC,
- Garcia-VellosoMJ,
- LozanoMD,
- et al
- 14.↵
- 15.↵
- BaasJM,
- KrensLL,
- GuchelaarHJ,
- MorreauH,
- GelderblomH
- 16.↵
- AllegraCJ,
- JessupJM,
- SomerfieldMR,
- et al
- 17.↵
- LièvreA,
- BachetJB,
- BoigeV,
- et al
- 18.↵
- Etienne-GrimaldiMC,
- FormentoJL,
- FrancoualM,
- et al
- 19.↵
- SantiniD,
- LoupakisF,
- VincenziB,
- et al
- 20.↵
- SuenagaY,
- KitajimaK,
- AokiH,
- et al
- 21.↵
- PauwelsEK,
- RibeiroMJ,
- StootJH,
- McCreadyVR,
- BourguignonM,
- MazièreB
- 22.
- GilliesRJ,
- RobeyI,
- GatenbyRA
- 23.↵
- 24.↵
- AkhurstT,
- KatesTJ,
- MazumdarM,
- et al
- 25.↵
- ChenC,
- PoreN,
- BehroozA,
- Ismail-BeigiF,
- MaityA
- 26.↵
- KikuchiH,
- PinoMS,
- ZengM,
- ShirasawaS,
- ChungDC
- Received for publicationMay 8, 2015.
- Accepted for publicationJune 18, 2015.
In this issue
