Analysis of Mutational Spectra in Metastatic Colorectal Carcinoma: KRAS as an Indicator of Oxaliplatin-Based Chemotherapy
Mutation spectra in colorectal cancer with metastasis and its response to chemotherapy. No molecular markers are available for selecting the optimal chemotherapeutic regimen (irinotecan or oxaliplatin) for metastatic colorectal cancer (mCRC). We enrolled 161 mCRC patients who underwent surgery for their primary tumors at Taipei Veterans General Hospital from 2004 to 2010. The prevalence of gene mutations was measured and correlated with responses to different cytotoxic agents. We detected 1,836 mutations in 12 genes. KRAS mutants affected 44.3% of the tumors. The rate of good response was insignificantly higher for patients with KRAS mutant tumors who received oxaliplatin-based chemotherapy compared with patients with KRAS wild-type tumors (65.6% versus 47.0%; P = 0.15). For patients who received irinotecan-based chemotherapy, the rate of good response was similar in patients with wild-type (55.0%; n = 11) and those with KRAS mutant tumors (54.5%; n = 12; P = 1). In patients with KRAS mutant tumors treated with an oxaliplatin-based regimen, the overall survival was 38.5 months (95% CI: 26.6–50.5 months), which was insignificantly better than that for patients treated with an irinotecan-based regimen (30.4 months; 95% CI: 15.8–45.1 months; P = 0.206). Our data could not come to the conclusion that patient with KRAS mutation mCRC may have better response with oxaliplatin-based first-line chemotherapy. Further study is needed to confirm the relationship between gene mutation and chemotherapy response.Objective:
Summary of Background Data:
Methods:
Results:
Conclusions:
Colorectal carcinoma (CRC) is the most common cancer in Taiwan.1 Half of the patients with CRC eventually develop distant metastasis, resulting in poor outcomes. With the introduction of cytotoxic and biologic agents to treat metastatic colorectal cancer (mCRC), the median overall survival (OS) increased from 12 to 30 months, and nearly up to 70% of patients could have a chance to receive at least 2 lines of treatment.2–6 In addition to tumor aggressiveness, therapeutic responses impact patient outcomes significantly. In particular, rapid and deep tumor shrinkage could convert unresectable metastatic lesions into resectable lesions, thus prolonging patient life and improving quality of life.7,8 Increased molecular knowledge and randomized clinical trial results have identified the association between mutations in RAS genes, specifically KRAS and NRAS, and anti-EGFR monoclonal antibody (mAb) effectiveness. This has led to selected use of this class of drugs in patients with RAS-wild-type CRC.9
On the other hand, no ideal molecular marker has been identified to aid in selection of cytotoxic agents. Because irinotecan and oxaliplatin are regarded as equally effective agents, the use of particular chemotherapeutic combinations often depends on patients' comorbidities as well as institutional or personal preferences.10–13
Recently, some in vitro studies showed that manipulation of mutant KRAS could alter expression of ERCC1 and affect sensitivity to oxaliplatin,14,15 but this was not reproduced in clinical analysis. Previously, we established a high throughput MassARRAY platform (Sequenom, San Diego, California) that included 12 genes and 138 hotspots, and we detected 1,836 mutations in those 12 genes in 997 (79.8%) nonmetastatic CRCs.16 In this study, we analyzed clinical data from patients with mCRC who received at least 6 administrations of first-line oxaliplatin-based or irinotecan-based chemotherapy. We determined carcinoembryonic antigen (CEA) ratios and overall survival (OS) as indicators for treatment responses. Here, we present the relationships between first-line chemotherapy regimens and tumor response as well as OS.
Materials and Methods
Clinical data
We enrolled 161 patients with mCRC who underwent surgery for their primary tumors at Taipei Veterans General Hospital from 2004 to 2010. The exclusion criteria were preoperative radiochemotherapy, emergency operations, or death within 30 days after surgery. Clinical information that was prospectively obtained and stored in a database included age, sex, personal and family medical history, location, tumor-node-metastasis (TNM) stage, differentiation, pathologic prognostic features and follow-up conditions. Chemotherapy regimens (FOLFOX or FOLFIRI) were chosen by clinical physicians. Following surgery, patients were monitored quarterly for the first 2 years and semi-annually thereafter. The follow-up protocol included physical examination, digital rectal examination, CEA analysis, chest X-ray, abdominal sonogram, and computed tomography (CT). If needed, proton emission tomography (PET) or magnetic resonance imaging (MRI) was arranged.
Tumor tissues
Before sample collection, written informed consent for tissue collection was obtained from all patients. Samples were meticulously dissected and collected from different quadrants of the tumors. Samples were immediately frozen in liquid nitrogen and stored in the Taipei Veterans General Hospital Biobank. Sections of cancerous tissue and corresponding normal tissue were reviewed by a senior gastrointestinal pathologist.
DNA isolation and quantification
After approval from the Institutional Review Board of the Taipei Veterans General Hospital, (2013-04-042B), samples for this study were obtained from the Biobank. DNA from tissue specimens was extracted using the QIAamp DNA Tissue Kit (Qiagen, Valencia, California), according to the manufacturer's recommendations. DNA quality and quantity were confirmed using the Nanodrop 1000 Spectrophotometer (Scientific, Waltham, Massachusettes).
MassARRAY-based mutation characterization
The MassDetect CRC panel (v1.0), which enables the identification of 139 mutations in 12 genes (Supplementary Table 1), was designed as previously described.16 Polymerase chain reaction (PCR) and extension primers for the mutations were designed using MassARRAY Assay Design 3.1 software (Sequenom, San Diego, California). The mutation alleles were manually designed by extension in either the forward or reverse direction to have lower masses than the reference alleles. After analyzing the primer designs with BLAST, any necessary modifications were made to avoid pseudogene amplification. MassARRAY-based mutation detection methods are described in Supplementary Table 1. PCR products of multiplexed reactions were spotted onto SpectroCHIP II arrays, and DNA fragments were resolved by on the MassARRAY Analyzer 4 System (Sequenom). Each spectrum was then analyzed using Typer 4.0 software (Sequenom) to identify mutations. Putative mutations were further filtered by manual review.
Microsatellite instability (MSI) analysis
Five reference microsatellite markers were used according to international criteria for determination of MSI: D5S345, D2S123, BAT25, BAT26, and D17S250. Primer sequences were obtained from GenBank (www.gdb.org). MSI detection was performed as previously described.17,18 Briefly, DNA was amplified using fluorescent PCR. PCR products were denatured and analyzed by electrophoresis on 5% denaturing polyacrylamide gels, and results were analyzed using GeneScan Analysis software (Applied Biosystems, Carlsbad, California). Tumor samples that exhibited allele peaks different from the corresponding normal sample(s) were classified as having MSI for that particular marker. Samples with ≥2 MSI markers were defined as having MSI and those with 0–1 MSI markers were classified as having MSS.
CEA measurement and CEA ratio assessment
CEA levels were measured before chemotherapy and every 3 months after initial chemotherapy. The CEA ratio was defined as the CEA level 3 months after chemotherapy divided by the pretreatment CEA level (post-CEA/pre-CEA).
The serum CEA level was measured in the Department of Nuclear Medicine at the Taipei Veteran General Hospital. Thirty-nine patients whose CEA levels were within normal limits (<5 μg/L) throughout the treatment were excluded from the study. Thus, a total of 122 patients were enrolled in this study.
Similar to the Response Evaluation Criteria In Solid Tumors (RECIST) criteria, the tumor response to chemotherapy was defined according to the CEA ratio before analysis: complete response, having CEA level return to normal value; partial response, having CEA ratio < 50% but CEA level above the normal value; stable disease, having 50% CEA ratio < 100%; and progressive disease, having CEA ratio > 100%.
Response assessment via imaging
CT or MRI were collected before chemotherapy and 3 months after treatment. Image changes were evaluated according to RECIST criteria. The RECIST criteria have limitations. Patients with disseminated metastasis or immeasurable metastasis could not be evaluated. Patients that were unable to be evaluated using RECIST criteria were excluded. Therefore, 77 patients were excluded. The imaging results for 85 patients were evaluable and considered in this study.
Statistical analysis
OS, which was defined as the date from surgery until death, was the statistical endpoint for analyses in this study. Patients not known to have died were censored at the date of last follow-up. Kaplan–Meier survival curves were plotted and compared using log-rank tests. The impact of chemotherapeutic regimens, clinicopathologic features, and genetic mutations on OS were assessed using Cox regression univariate and multivariate analyses. The chi-squared test and 2-tailed Fisher's exact test were used to compare genotype frequency according to clinicopathologic features. Numerical values were compared using Student's t-test. Data were expressed as mean ± standard deviation. Statistical significance was defined as P < 0.05. Statistical analyses were performed using SPSS for Windows (version 16.0).
Results
There were 77 men (63.1%) and 45 women (36.9%) included in this study. The mean age at tumor resection was 66.2 ± 12.4 years (range, 35–92 years; median, 67.6 years). Patients were diagnosed with 39 proximal colon cancer (32.0%), 42 distal colon cancer (34.4%), and 41 rectal cancer (33.6%). The patients were also diagnosed with metastatic lesions in the liver (N = 88; 72.1 %), lungs (N = 22; 18.2 %), and other sites (N = 23; 18.9%) (Table 1).
Seventy-five patients were administered a median 8.6 cycles of FOLFOX as first-line treatment (range, 1–19 cycles). Forty-five patients received a median 9.4 cycles of FOLFIRI as first-line treatment (range, 1–32 cycles). The OS for all patients was 37.7 months (95% confidence interval [CI]: 29.5–45.8 months) in the FOLFOX group and 36.7 months (95% CI: 25.8–47.7 months) in the FOLFIRI group (P = 0.635) (Supplementary Fig. 1).
The most frequently mutation was loss of 18q (45.9%, N = 56), followed by KRAS (44.3%, N = 54), TP53 (26.2%, N = 32), APC (26.2%, N = 32), PIK3CA (8.2%, N = 10), and NRAS (6.6%, N = 8) (Supplementary Fig. 2).Detailed mutation information for individual genes is described in Table 2.
Our previous study showed that CEA ratios (post-CEA/pre-CEA) and imaging changes according to RECIST criteria both correlated with OS and treatment responses.19 In this study, we used CEA ratio (3 months posttreatment CEA/pretreatment CEA) as a parameter for evaluating chemotherapy response. Overall good responses were achieved in 53.5% (N = 23) of patients treated with FOLFIRI and 55.9% (N = 38) of patients treated with FOLFOX, when using CEA as an evaluation parameter (P = 0.85). Responses to oxaliplatin-based and irinotecan-based chemotherapy were not associated individual gene mutations, with the exception of a marginal effect for KRAS mutation (Supplementary Table 2). Detailed mutation information for individual genes response is described in Supplementary Table 3.
For patients who received oxaliplatin-based chemotherapy, the rate of good response was 47.0% (N = 16) for patients with wild-type KRAS tumors. This was insignificantly lower than that for patients with KRAS mutant tumors (65.6%, N = 21; P = 0.15). For patients who received irinotecan-based chemotherapy, the rates of good responses was similar for patients with wild-type KRAS tumors (55.0%; N = 11) and those with KRAS mutant tumors (54.5%; N = 12; P = 1) (Table 3).
The OS for patients with KRAS wild-type tumors who were treated with oxaliplatin-based first-line chemotherapy was 36.3 months (95% CI: 25.6–47.1 months), and OS was 42.3 months (95% CI: 25.6–59.0 months) for patients treated with an irinotecan-based regimen (P = 0.702).
In patients with KRAS mutant tumors who were treated with an oxaliplatin-based regimen, the overall survival was 38.5 months (95% CI: 26.6–50.5 months), which was insignificantly better than that for patients treated with an irinotecan-based regimen (30.4 months; 95% CI: 15.8–45.1 months; P = 0.206) (Fig. 1).
Citation: International Surgery 103, 1-2; 10.9738/INTSURG-D-17-00050.1
KRAS as a predictor of response to oxaliplatin-based chemotherapy
The National Cancer Institute (NCI) database containing data from the 60 NCI60 cell lines was used as the source of cytotoxicity data for oxaliplatin and irinotecan (SN-38). The GI50, which is the concentration required to inhibit growth by 50%, was used as a cytotoxicity parameter. Of these 60 cell lines, 53 had KRAS mutations status data, which was, identified from the COSMIC (Catalogue of Somatic Mutations in Cancer) database. As shown in Fig. 2, cell lines with KRAS mutations were more sensitive to oxaliplatin than wild-type KRAS cells. The GI50 values for mutant and wild-type KRAS cell lines were 10-6.21 and 10-5.66, respectively. This revealed that wild-type KRAS cells required 10 times more drug to inhibit growth than mutant KRAS cells. In contrast, SN-38 had a similar median GI50 in cell lines with/without KRAS mutation (10-4.16 versus 10-4.05).
Citation: International Surgery 103, 1-2; 10.9738/INTSURG-D-17-00050.1
Discussion
Our study implies that using first-line oxaliplatin-based chemotherapy might result in better survival benefits for patients with KRAS mutant mCRC. Although not statistically significant, patients with KRAS mutant tumors who were treated with FOLFOX had a better response rate than patients treated with FOLFIRI (65.6% versus 54.5%, P = 0.15). This better response rate might lead to a higher chance of surgical resectability.8,10
The mutational profiling in our study revealed that loss of 18q (45.9%, N = 56), followed by KRAS (44.3%, N = 54), P53 (26.2%, N = 32), APC (26.2%, N = 32), PIK3CA (8.2%, N = 10), and NRAS mutations (6.6%, N = 8) were most common among our patients with mCRC. The previous study by Russo et al, who reviewed 222 American patients with mCRC(20), showed a prevalence of patients with KRAS mutations (36%), followed by TP53 mutations (21%), APC mutations (8%), NRAS mutations (4%), BRAF mutations (10%), and PIK3CA mutations (13%). The APC mutations were much more frequent in our database.
The OS rates for patients with mutant KRAS tumors who underwent oxaliplatin-based or irinotecan-based regimens were 38.5 months and 30.4 months, respectively, in our study. In the PRIME study, the OS for patients with KRAS mutant tumors who were treated with FOLFOX4 alone was 19.3 months.5 In the CRYSTAL study, the OS for patients with KRAS mutant tumors who were treated with FOLFIRI was 17.7 months.21 Overall survival was higher in our study. This was mainly because the patient in our study underwent surgical resections for their primary tumors. The response rate of patients with KRAS mutant tumors in the CRYSTAL study was 40.2% (FOLFIRI), and it was 40% in the PRIME study (FOLFOX4). The response rates were also higher in our study (65.6% for oxaliplatin-based and 55% for irinotecan-based). Our data was collected from 2004 to 2010, when targeted agent therapy had not yet been reimbursed by the National Health Insurance System in Taiwan.
In recent years, CRC molecular and cytogenetic characteristics have allowed to increasingly understand CRC. Several biomarkers have been identified and confirmed to predict progression and survival.16–18,22 Our study tried to compare different biomarkers to oxaliplatin-based or irinotecan-based chemotherapy alone in response rate for patients with mCRC. We revealed an approximately 65% response rate in patients with KRAS mutant tumors who were treated with FOLFOX compared to the 42.1–54.5% for patients with KRAS mutant tumors who were treated with irinotecan-based chemotherapy. Thus FOLFOX elicited a 35.3% better response rate. Although the OS for patients with KRAS mutant tumors who were treated with first-line oxaliplatin or irinotecan based chemotherapy was not significantly different (P = 0.20), oxaliplatin-based first-line chemotherapy tended to result in a more favorable outcome.
According to Huang et al, CEA ratio and imaging change according to RECIST criteria both correlated with overall survival.19 In addition to using CEA ratio as a parameter, we have also used RECIST criteria to evaluate treatment response rates. The overall response rate was 39.4% (N = 13) for patients treated with FOLFIRI and 57.7% (N = 30) for patients treated with FOLFOX (P = 0.12) (Supplementary Table 2). When comparing the patients with tumors that have wild-type or mutant KRAS, oxaliplatin-based regimens elicited a slightly better response rate than irinotecan-based regimens. In patient who were administered oxaliplatin-based chemotherapy, the rates of good responses were 55.6% (N = 15) for those with KRAS wild-type tumors and 65.2% (N = 15) for those with KRAS mutant tumors (P = 0.56). For patients who were administered irinotecan-based chemotherapy, the rate of good responses were 38.5% (N = 5) for those with KRAS wild-type tumors, which was similar to that for patients with KRAS mutant tumors (42.1%; N = 8; P = 1). This result also suggested more favorable outcomes for patients with KRAS mutant tumors who receive oxaliplatin-based first-line chemotherapy. We've also used RECIST criteria to assess other mutational factors. The rates of good response for patients with wild-type loss of 18q were 15.4% (N = 2) when treated with FOLFIRI and 46.7% (N = 14) when treated FOLFOX (P = 0.032). The response rates for patient with mutant loss of 18q treated with FOLFIRI or FOLFOX were 55% (N = 11) and 72.7% (N = 16) (P = 0.089), respectively. For patients with tumors expressing wild-type APC, the rates of good responses were 28.6% (N = 6) when treated FOLFIRI and 51.9% (N = 27) with FOLFOX (P = 0.31). In patients with tumors expressing mutant APC, the rate of good response were 46.7% (N = 7) when treated with FOLFIRI and 68.4% (N = 13) with FOLFOX (P = 0.283). For patients with tumors expressing wild-type P53, the rates of good responses were 33.3% (N = 9) when treated with FOLFIRI, and 61.9% (N = 27) with FOLFOX (P = 0.693). For patients with tumors expressing mutant P53, the rates of good responses were 44.4% (N = 4) when treated with FOLFIRI, and 68.4% (N = 13) with FOLFOX (P = 0.283) (Supplementary Table 4).
The superior results after oxaliplatin treatment in patients with KRAS mutant tumor are comparable to the result of Lin et al.14,15 For mCRC patients treated with oxaliplatin-based chemotherapy, the median progression-free survival (PFS) was 8.5 months for patients with KRAS mutant tumors versus 5.8 months for patients with KRAS wild-type tumors (P = 0.008). However, for patients treated with irinotecan-based chemotherapy, the median PFS was 3.9 months for patients with KRAS mutant tumors versus 6.0 months for patients with KRAS wild-type tumors (P = 0.23).
Furthermore, according to an recent in vitro study by Lin et al,15 KRAS mutation is a predictor of oxaliplatin sensitivity in colon cancer cells. In the in vitro study, KRAS mutant CRC cells were more sensitive to oxaliplatin than KRAS wild-type CRC cells. Overexpression of excision repair cross-complementation group 1 (ERCC1) is associated to resistance to platinum-based chemotherapy. Mutant KRAS cells might downregulate ERCC1 through hypermethylation of ERCC1 gene. In the in vitro study, using the apoptosis rate 48 hours after platinum-based treatment, the response rate was 22.5% for KRAS wild-type CRC wells versus 39.1% for KRAS mutant cells. The results implied that KRAS mutated CRC cells are more sensitive to oxaliplatin-based treatment. The role of ERCCI was discussed in a previous study. Chua et al23 showed that ERCC1 is associated with worse progression-free survival in patients treated with FOLFOX. H Baba et al18 also reported higher ERCC1 levels lead to resistance to oxaliplatin. Furthermore, KRAS mutated cell lines were more sensitive to oxaliplatin in vitro.24
The National Cancer Institute (NCI) database containing data from the NCI60 cell lines was used as the source of cytotoxicity data for oxaliplatin. The GI50, which is the concentration required to inhibit growth by 50%, was used as a parameter for cytotoxicity. Of these 60 cell lines, there was information about the KRAS mutation status for 53 lines. The median log (GI50) was -5.33 in 39 wild-type KRAS cell lines when treated with oxaliplatin. This result was higher than that for 14 mutant KRAS cell lines (-6.21). The cell line data revealed that a higher concentration of oxaliplatin was needed to inhibit tumor growth in wild-type KRAS cells. In the contrast, when cell lines were treated with SN-38, the result was similar for both groups (-4.05 in KRAS wild-type cell lines versus -4.16 in KRAS mutant cell lines). This result further strengthened our results in this study.
There are limitations to our study. First, it was a retrospective study with a relatively small sample size. Limited power might result in statistical insignificance in our study. A larger prospective study is needed to confirm the result. However, it is hard to design this kind of study. Target therapy has been the protocol for first-line chemotherapy for patients with mCRC. However, in vitro studies might help confirm the results of oxaliplatin resistance using different molecular markers. Second, we lack data on progression-free survival, which is an indicator of treatment response.
Our data could not come to the conclusion that patient with KRAS mutation mCRC may have better response with oxaliplatin-based first-line chemotherapy. Further study is needed to confirm the relationship between gene mutation and chemotherapy response.
Overall survival stratified by KRAS status. (A) Patients with mutant KRAS tumor who were treated with different chemotherapy regimens (P = 0.702). (B) Patients with wild-type KRAS tumors who were treated with different chemotherapy regimens (P = 0.206).
(A) The log (GI50) of SN-38 in 53 NCI60 panel cell lines. The median log(GI50) is -4.05 for 39 wild-type KRAS cell lines (red), which was similar to that for the 14 mutant KRAS cell lines (-4.16; blue). (B) The log (GI50) of oxaliplatin in 53 NCI60 panel cell lines. The median log(GI50) is -5.33 for 39 wild-type KRAS cell lines (red), which was higher than that for the 14 mutant KRAS cell lines (-6.21; blue).
Contributor Notes