Genotype-based selection of treatment of patients with advanced colorectal cancer (SETICC): a pharmacogenetic-based randomized phase II trial

Genotype-based selection of treatment of patients with advanced colorectal cancer (SETICC): a... Abstract Background There has been little progress toward personalized therapy for patients with metastatic colorectal cancer (mCRC). TYMS-3′ untranslated region (UTR) 6 bp ins/del and ERCC1-118C/T polymorphisms were previously reported to facilitate selecting patients for fluoropyrimidine-based treatment in combination with oxaliplatin as first-line therapy. We assessed the utility of these markers in selecting therapy for patients with mCRC. Patients and methods This randomized, open-label phase II trial compared bevacizumab plus XELOX (control) versus treatment tailored according to TYMS-3'UTR 6 bp ins/del and ERCC1-118C/T polymorphisms. Patients randomized to the experimental treatment received bevacizumab plus FUOX, FUIRI, XELIRI, or XELOX depending on their combination of favorable polymorphisms for FUOX treatment (TYMS-3′UTR ins/del or del/del; ERCC1-118T/T). Progression-free survival (PFS) was the primary end point. Results Overall, 195 patients were randomized (control n = 65; experimental n = 130). The primary objective was not met: median PFS was 9.4 months in the control group and 10.1 months in the experimental group (P = 0.745). Median overall survival was similar in both groups (16.5 versus 19.1 months, respectively; P = 0.797). Patients in the experimental group had a significantly higher overall response rate (ORR; 65% versus 47% in the control group; P = 0.042) and R0 resection rate (86% versus 44%, respectively; P = 0.018). Neuropathy, hand–foot syndrome, thrombocytopenia, and dysesthesia were significantly less common in the experimental group. Conclusions This study did not show survival benefits after treatment personalization based on polymorphisms in mCRC. However, the improved ORR and R0 resection rate and fewer disabling toxicities suggest that tailoring therapy by TYMS-3′UTR and ERCC1-118 polymorphisms warrants further investigation in patients with mCRC. ClinicalTrials.gov NCT01071655. colorectal cancer, polymorphisms, biomarkers, bevacizumab, oxaliplatin, irinotecan Key Message This study shows that prospective genotyping of TYMS-3′UTR 6 bp ins/del and ERCC1-118C/T polymorphisms is feasible and easy to perform. Overall response and R0 resection rates were significantly improved by genotype-driven treatment choice. Our findings support continued investigations into optimal treatment selection according to germline variations in patients with metastatic colorectal cancer. Introduction Treatment of patients with metastatic colorectal cancer (mCRC) is generally based around chemotherapy consisting of 5-fluorouracil (5-FU) or capecitabine, plus oxaliplatin or irinotecan, plus a targeted agent, e.g. bevacizumab, cetuximab, or panitumumab [1]. Physicians have many treatment options for patients with mCRC, and selecting appropriate regimens for individual patients can be challenging. To date, personalizing treatment applies only to selection of anti-epidermal growth factor receptor drugs based on existence of RAS oncogene mutations [2]. Recommended chemotherapy regimens have substantial activity in most patients; however, many patients experience resistance or excessive treatment-related toxicity. This variability may be related to interpatient differences in genes involved in drug transport, metabolism, signaling, and cellular response pathways [3]. Previous studies suggest that germline polymorphisms in genes involved in the mechanism of action or metabolism of chemotherapeutic agents may be responsible for variations in patients’ response or tolerability [4]. However, most previous studies were retrospective, with the exception of a few prospective studies in which patients were treated in a genotype-driven fashion [5, 6]. Our previous study suggested that patients harboring the TYMS-3′ untranslated region (UTR) 6 bp ins/ins and ERCC1-118C/T or C/C genotypes might benefit from combining oxaliplatin with capecitabine rather than 5-FU [7]. The aim of the present study was to examine whether assigning first-line chemotherapy based on relevant germline polymorphisms would improve outcomes compared with treating all patients with a standard regimen. Patients and methods Study design and patients This randomized, multicenter, open-label, phase II study was conducted in Spain by the Grupo Español de Tratamiento de Tumores Digestivos (Spanish Cooperative Group for the Treatment of Digestive Tumors). Patients had histologically confirmed colon or rectal adenocarcinoma with measurable metastatic disease (Response Evaluation Criteria in Solid Tumors [RECIST] Version 1.1), were aged ≥18 years, had Eastern Cooperative Oncology Group performance status 0–2, and adequate renal function. Key exclusion criteria were prior systemic treatment of metastatic disease; concomitant cardiovascular disease; history or signs of central nervous system disease, uncontrolled hypertension, bleeding diathesis, or coagulopathy. Exclusion criteria are described in detail in the supplementary Appendix, available at Annals of Oncology online. The study was conducted in accordance with the Declaration of Helsinki and Good Clinical Practice guidelines, and was approved by institutional review boards or independent ethics committees at each center. All patients provided written informed consent. Randomization Between February 2010 and April 2011, patients were randomized 1 : 2 to the control or experimental group. Treatment assignment was through an electronic system integrated in the electronic case report form, which automatically balanced groups. For the experimental group, treatment was assigned depending on TYMS-3′UTR 6 bp ins/del and ERCC1-118C/T polymorphisms. Investigators were informed of treatments by automatically generated e-mails. Polymorphisms were determined for patients assigned to the control group, but these results were not needed before treatment commenced. The investigator still received an automatically generated e-mail. Randomization is described in detail in the supplementary Appendix, available at Annals of Oncology online. Treatments Patients assigned to the control group received bevacizumab 7.5 mg/kg on day 1 with XELOX. Patients assigned to the experimental group received different schedules according to the number of TYMS-3′UTR 6 bp ins/del and ERCC1-118C/T favorable genotypes as previously described [7] (supplementary Table S1, available at Annals of Oncology online). Briefly, we considered TYMS-3′UTR ins/del and del/del to favor treament with 5FU, ins/ins genotype to favor capecitabine, ERCC1-118T/T to favor oxaliplatin and C/T or C/C genotypes to favor irinotecan. Accordingly, patients with favorable genotypes for FUOX received bevacizumab 5 mg/kg on days 1, 15, and 29 with FUOX [8]. Patients with no favorable polymorphisms received bevacizumab 7.5 mg/kg on day 1 with XELIRI. Patients with TYMS-3′UTR 6 bp ins/ins and ERCC1-118T/T genotype received the same treatment as the control arm. Patients with TYMS-3′UTR 6 bp ins/ins and ERCC1-118C/T or C/C genotype received bevacizumab 5 mg/kg on days 1, 15, and 29 with FUIRI [9]. Further detail can be found in the supplementary Appendix, available at Annals of Oncology online. Treatment continued until disease progression, unacceptable toxicity, or patient withdrawal. Outcomes The primary end point was progression-free survival (PFS). Secondary end points included overall survival (OS), overall response rate (ORR; assessed using RECIST Version 1.1), proportion of patients whose disease became resectable, adverse events, and evaluation of KRAS exon 2 mutation status as a molecular prognostic marker. Toxicity was assessed using National Cancer Institute Common Toxicity Criteria for Adverse Events (Version 3.0). Adverse events were assessed at each cycle and every 12 weeks during the study follow-up. In patients treated with FUOX and FUIRI, hematology, biochemistry, and urinalysis assessments were carried out every 2 weeks. Statistical analysis It was estimated that the difference between groups for PFS would be statistically significant if, when 50% of patients in the control group had disease progression, 65% of patients in the experimental group had no disease progression. The maximum permitted error was considered to be α = 0.05 (unilateral contrast, assuming that the efficacy was greater in the experimental group) and β = 0.2 (80% power). Based on these assumptions, 171 patients (57 in the control group and 114 in the experimental group) were needed, assuming a 5% drop-out rate. We planned to include 180 patients. Further detail is provided in the supplementary Appendix, available at Annals of Oncology online. The study was registered with ClinicalTrials.gov (NCT01071655). Results Baseline and genotype data Of 202 patients enrolled, 195 were randomized (Figure 1). Baseline characteristics were generally balanced between the two groups (Table 1); we found no statistically significant differences between the groups in baseline characteristics (data not shown). Supplementary Table S1, available at Annals of Oncology online, shows the distribution of genotypes and assigned treatments in the experimental arm. Frequencies of TYMS-3′UTR ins and del alleles were 0.68 and 0.32, respectively; frequencies of ERCC1-118T and C alleles were 0.57 and 0.43, respectively. Genotype frequencies were TYMS ins/ins 50%, ins/del 37%, and del/del 13%; ERCC1-118T/T 35%, C/T 45%, and C/C 21%. Table 1. Patient characteristics at baseline (intent to treat population, n = 195) Characteristic  Control group (n = 65)  Experimental group (n = 130)  Age, years (range)  64.8 (36.8–82.9)  63.6 (28.8–86.1)  Interquartile range  58.7–72.7  57.2–70.7  Male  35 (54)  82 (63)  ECOG performance status       0  29 (45)  50 (38)   1  34 (52)  79 (61)   2  2 (3)  1 (1)  Tumor location       Colon  18 (28)  38 (29)   Rectum  41 (63)  80 (62)   Both  6 (9)  12 (9)  No. of affected organs       1  25 (38)  55 (42)   2  21 (32)  52 (40)   >2  19 (29)  23 (18)  Location of metastases       Local/regional  17 (26)  34 (26)   Liver  51 (78)  106 (82)   Liver only  21 (32)  45 (35)   Lung  22 (34)  50 (38)   Other  20 (31)  29 (22)  Surgery for primary tumor  34 (52)  73 (56)  Prior adjuvant therapya  10 (15)  15 (12)   Chemotherapy  9 (14)  15 (12)   Radiotherapy  3 (5)  7 (5)  KRAS mutation  27 (42)  42 (32)  Characteristic  Control group (n = 65)  Experimental group (n = 130)  Age, years (range)  64.8 (36.8–82.9)  63.6 (28.8–86.1)  Interquartile range  58.7–72.7  57.2–70.7  Male  35 (54)  82 (63)  ECOG performance status       0  29 (45)  50 (38)   1  34 (52)  79 (61)   2  2 (3)  1 (1)  Tumor location       Colon  18 (28)  38 (29)   Rectum  41 (63)  80 (62)   Both  6 (9)  12 (9)  No. of affected organs       1  25 (38)  55 (42)   2  21 (32)  52 (40)   >2  19 (29)  23 (18)  Location of metastases       Local/regional  17 (26)  34 (26)   Liver  51 (78)  106 (82)   Liver only  21 (32)  45 (35)   Lung  22 (34)  50 (38)   Other  20 (31)  29 (22)  Surgery for primary tumor  34 (52)  73 (56)  Prior adjuvant therapya  10 (15)  15 (12)   Chemotherapy  9 (14)  15 (12)   Radiotherapy  3 (5)  7 (5)  KRAS mutation  27 (42)  42 (32)  Data are median (range) or n (%). No P values are shown as no statistically significant differences were found between the control and experimental groups for any of the characteristics studied. a With or without radiotherapy. ECOG, Eastern Cooperative Oncology Group. Table 2. Response to treatment (response-evaluable population, n = 161) and R0 surgery (intent to treat population, n = 195) Outcome  Control group (n = 55)  Experimental group (n = 106)  P-value  Complete response (CR)  2 (4)  4 (4)    Partial response (PR)  24 (44)  65 (61)    Stable disease  22 (40)  30 (28)    Progressive disease  7 (13)  7 (7)    Overall response rate (PR + CR)  26 (47)  69 (65)  0.042  Disease-control rate (CR + PR + SD)  48 (87)  99 (93)  0.240  R0 surgery  7 (44)a  18 (86)b  0.018  Outcome  Control group (n = 55)  Experimental group (n = 106)  P-value  Complete response (CR)  2 (4)  4 (4)    Partial response (PR)  24 (44)  65 (61)    Stable disease  22 (40)  30 (28)    Progressive disease  7 (13)  7 (7)    Overall response rate (PR + CR)  26 (47)  69 (65)  0.042  Disease-control rate (CR + PR + SD)  48 (87)  99 (93)  0.240  R0 surgery  7 (44)a  18 (86)b  0.018  Data are n (%). a Proportion of the 16 patients with surgical resection. b Proportion of 21 patients who underwent surgical resection. Figure 1. View largeDownload slide Patient disposition. AE, adverse event; Bev, bevacizumab; FUIRI, 5-fluorouracil + irinotecan; FUOX, 5-fluorouracil + oxaliplatin; ITT, intent to treat; XELIRI, capecitabine + irinotecan; XELOX, capecitabine + oxaliplatin. Figure 1. View largeDownload slide Patient disposition. AE, adverse event; Bev, bevacizumab; FUIRI, 5-fluorouracil + irinotecan; FUOX, 5-fluorouracil + oxaliplatin; ITT, intent to treat; XELIRI, capecitabine + irinotecan; XELOX, capecitabine + oxaliplatin. Overall, 40% of patients in the control group and 51% of patients in the experimental group withdrew from the study for reasons other than disease progression or adverse events (Figure 1). Patients in the control group received a median of nine treatment cycles; those in the experimental group received a median of six cycles (NS). Dose intensities were lowest for bevacizumab + FUOX and bevacizumab + FUIRI (supplementary Table S2, available at Annals of Oncology online). Fifty of 65 patients in the control group (77%) and 94 of 130 (72%) in the experimental group received second-line treatments. PFS and OS analysis The primary end point was not met; after a median follow-up of 18 months (range 1.1–41.7 months), median PFS was similar in both groups {9.4 months [95% confidence interval (CI) 6.8–12.1 months] in the control group versus 10.1 months (95% CI 8.5–11.6 months) in the experimental group; hazard ratio (HR) 0.942; 95% CI 0.657–1.351; P = 0.745; Figure 2A}. Median OS was numerically lower in the control group (16.5 months; 95% CI 13.7–19.4 months) than in the experimental group (19.1 months; 95% CI 15.5–22.7 months; HR 0.956; 95% CI 0.678–1.348; P = 0.798; Figure 2B). In line with our previous findings [7], we did not expect to observe a survival difference in the control group when patients were stratified according to genotype. Indeed, no statistically significant differences in PFS and OS according to genotype were found in the control group (supplementary Figure S1, available at Annals of Oncology online). Figure 2. View largeDownload slide Progression-free survival (A) and overall survival (B) in the control and experimental groups. Patients in the control group (group A) received bevacizumab + capecitabine + oxaliplatin and patients in the experimental group (group B) received bevacizumab + chemotherapy tailored according to the presence of specific genotypes. CI, confidence interval; HR, hazard ratio. Figure 2. View largeDownload slide Progression-free survival (A) and overall survival (B) in the control and experimental groups. Patients in the control group (group A) received bevacizumab + capecitabine + oxaliplatin and patients in the experimental group (group B) received bevacizumab + chemotherapy tailored according to the presence of specific genotypes. CI, confidence interval; HR, hazard ratio. Response rate and R0 surgery A total of 161 patients were evaluable for response (Table 2). The ORR was significantly higher in the experimental group than in the control group (P = 0.042). When confirmed responses alone were considered, the same trend was observed (51 of 106 patients [48%] versus 18 of 55 patients [33%] for experimental versus control groups, respectively; P = 0.07). Patients in whom tumor response could not be confirmed were those who had undergone surgery for metastases. Twenty-two of 130 patients (17%) in the experimental group and 16 of 65 (25%) in the control group had surgical resection of metastases during or after the study (P = 0.250). Eleven patients had rescue surgery twice; three patients had three rescue surgeries. The R0 resection rate was statistically significantly higher in the experimental group than in the control group (P = 0.018; Table 2). The most common resection sites were the liver and lung. Additional genetic analysis Ninety-seven tumor specimens were available for tumor and blood TYMS-3′UTR 6 bp ins/del genotyping. We found 100% concordance between tumor and blood results in the 62 homozygous samples. However, results for 23 of the 35 (66%) heterozygous blood samples did not agree with the corresponding tumor sample, as previously described [10]. This disagreement was total in four cases (17%) and partial (one of the alleles was partially lost but still present) in 19 cases (83%) (supplementary Figure S2, available at Annals of Oncology online). According to tumor genotyping, only two patients in the experimental group received the wrong treatment regimen. Information regarding KRAS mutations in exon 2 codons 12 and 13 was available for 164 patients (84%): 58 of 65 patients (89.2%) in the control group and 106 of 130 patients (81.5%) in the experimental group. KRAS mutations were present in 27 patients (47%) in the control group and 42 patients (40%) in the experimental group. KRAS mutation status had no impact on any of the clinical variables studied (data not shown). UGT1A1*28 polymorphism genotyping in the 75 irinotecan-treated patients was: UGT1A1*1/*1, n = 31 (41%), UGT1A1*1/*28, n = 33 (44%), and UGT1A1*28/*28, n = 11 (15%). Allelic frequencies for UGT1A1*1 and UGT1A1*28 alleles were 0.63 and 0.37, respectively, in agreement with Spanish population data [11]. The association between genotype and outcomes is summarized in supplementary Table S3, available at Annals of Oncology online. PFS and OS were worst in patients with the UGT1A1*28/*28 genotype according to codominant and recessive analyses. Repeating the PFS analysis without these patients did not affect the differences between the control and experimental groups (data not shown). An additional analysis of UGT1A1*28 polymorphism in patients who did not receive irinotecan was carried out. When patients were stratified by UGT1A1*28 polymorphism no statistically significant association with toxicity or efficacy outcomes was observed (data not shown). Toxicity Adverse events are summarized in supplementary Table S4, available at Annals of Oncology online. Neuropathy, hand–foot syndrome, thrombocytopenia, and dysesthesia were significantly less frequent in the experimental group (all P < 0.05 versus the control group); diarrhea was numerically more frequent in the experimental group (P = 0.074). Grade ≥3 diarrhea was significantly more frequent in the experimental group (P = 0.005); grade ≥3 neuropathy was significantly more frequent in the control group (P = 0.032). The UGT1A1*28/*28 genotype was significantly associated with a higher incidence of severe diarrhea in patients treated with irinotecan (supplementary Table S3, available at Annals of Oncology online). Discussion Although combinations of fluoropyrimidines with oxaliplatin or irinotecan remain the backbone of mCRC treatment, oncologists do not currently have effective tools to help choose the most suitable first-line schedule [12]. For the first time, to the best of our knowledge, standard treatment with bevacizumab plus XELOX has been compared with treatment tailored by TYMS-3′UTR 6 bp ins/del and ERCC1-118C/T polymorphisms in mCRC patients in a multicenter randomized phase II study. We demonstrated that this approach is feasible and applicable to clinical practice as patients were randomly assigned to treatment according to genotyping within 3 days of phlebotomy. The primary study end point—difference in PFS—was not met; although we observed a 3-month difference in OS in favor of tailored treatment, this was not statistically significant. There were no genotyping errors, and genotype and allelic frequencies were in Hardy–Weinberg equilibrium. Therefore, other possible explanations for these results should be considered. Importantly, 27% of patients in the experimental arm withdrew, citing patient or investigator’s decision as their reason. This may have affected outcomes in the experimental arm if subsequent treatment was not optimal for the patient. Although not all responses were confirmed as some patients underwent surgery, the trend toward better RR and higher rate of conversion surgery, together with statistically significant differences in unconfirmed treatment response, reinforces this hypothesis. The impact of bevacizumab should also be considered. Unlike our earlier study [7], the present study was designed after approval of bevacizumab for patients with mCRC; consequently, bevacizumab was included in treatment regimens. Reported interaction of genetic variants in VEGF and TYMS and influence of SHMT1 gene variants [13, 14] suggest that addition of bevacizumab could have affected the study results. This warrants further investigation. We also investigated whether loss of heterozygosity (LOH) or UGT1A1*28 polymorphism affected our results. TYMS LOH frequency was similar to previous reports [10, 15]. Tumor genotyping revealed that only two patients in the experimental group received the wrong treatment, suggesting that LOH was unlikely to have affected our results. Irinotecan-treated patients with UGT1A1*28/*28 genotype had worse PFS than those with UGT1A1*1/*28 or UGT1A1*1/*1 genotypes. Only 11 of 130 patients had the UGT1A1*28/*28 genotype, which appeared to have little effect on PFS, as suggested by the PFS analysis excluding patients with the UGT1A1*28/*28 genotype. A significant benefit for tailored treatment was observed in ORR and R0 surgery rates. R0 surgery is considered a good prognostic factor but no predictive markers are available to tailor neoadjuvant treatment before resection of metastases. Despite the reasonable caution needed in assessing results regarding secondary end points, patients whose liver metastases cannot be operated on at treatment outset may benefit from an approach such as we have described. Improving response rates and R0 surgery rates in these patients can dramatically improve their chance of surviving beyond 5years. Toxicities were as expected: the overall incidence of neurotoxicity in the experimental group was lower than in the control group, and the incidence of diarrhea was higher in the experimental group than in the control group. Neurotoxicity is chronic and causes varying degrees of functional impairment, whereas diarrhea is generally acute and has no sequelae. Therefore, tailoring treatment may reduce disabling toxicity while at least maintaining pharmacological activity, as demonstrated by the similar survival in both arms and better R0 resection rates in the experimental group. Furthermore, the higher incidence of severe diarrhea in patients with UGT1A1*28/*28 genotype who received irinotecan confirms previous results [16–18], highlighting the importance of assessing this polymorphism to avoid severe toxicity. Some limitations should be considered. Serum samples were not analyzed for cell-free tumor DNA as the study was designed before evidence for the prognostic value of such analyses became available. This technique may allow identification of variability in response due to the tumor genome, complementary to information provided by examination of germline polymorphisms. Conclusion In conclusion, the results of this study suggest that prospective first-line treatment selection according to TYMS-3′UTR 6 bp ins/del and ERCC1-118C/T polymorphisms in patients with mCRC does not improve TTP or OS. However, we showed this approach to be feasible and applicable to clinical practice. Promising ORR and R0 surgery rates in the experimental arm, without compromising toxicity, combined with possible logistical reasons for lack of significant difference in PFS, provide the rationale for continuing to investigate how best to select treatment according to germline variations and improve survival in patients with mCRC. Acknowledgements Spanish Cooperative Group for the Treatment of Digestive Tumors (TTD) Data Center: Inmaculada Ruiz de Mena and Susana Rodríguez. Data Management, Monitoring and Statistics (Dynamic): Laura Casas Herrero—Biostatistician, and Fátima González Hurtado—Monitoring. Editorial support was provided by Deirdre Carman, PhD, Miller Medical Communications Ltd, funded by Roche Farma, Spain. Funding Roche Farma (no grant number applies). Disclosure JMV received personal fees from Roche during the conduct of the study. VAO reports personal fees from Roche during the conduct of the study and personal fees, outside the submitted work, from Merck and Sanofi. MB received personal fees from Roche during the conduct of the study and personal fees from Merck, Amgen, Bayer, Sanofi, Lilly, and Servier that were outside the submitted work. MV received personal fees from Roche during the conduct of the study and personal fees, outside the submitted work, from Amgen and Merck-Serono. JS received personal fees from Roche during the conduct of the study and personal fees outside the submitted work from Bayer, Amgen, Boehringer Ingelheim, Merck, Sanofi, Ipsen, and Celgene. FR received personal fees from Roche during the conduct of the study and personal fees from Merck-Serono, Amgen, Celgene, Bayer, MSD, and Lilly outside the submitted work. EA received personal fees from Roche during the conduct of the study and personal fees from Amgen, Bayer, Celgene, Merck, and Sanofi outside the submitted work. AA, AC, VC, PGA, JG, GG, JLM, EM-B, BM, CM, MP, RS, and MZ have nothing to disclose. References 1 National Comprehensive Cancer Network Clinical Practice Guidelines in Oncology. Colon cancer. Version 2. 2017. http://www.nccn.org/professionals/physician_gls/pdf/colon.pdf (5 July 2017, date last accessed). 2 Van Cutsem E, Lenz HJ, Köhne CH et al.   Fluorouracil, leucovorin, and irinotecan plus cetuximab treatment and RAS mutations in colorectal cancer. J Clin Oncol  2015; 33( 7): 692– 700. Google Scholar CrossRef Search ADS PubMed  3 Unger FT, Witte I, David KA. Prediction of individual response to anticancer therapy: historical and future perspectives. Cell Mol Life Sci  2015; 72( 4): 729– 757. Google Scholar CrossRef Search ADS PubMed  4 Watson RG, McLeod HL. Pharmacogenomic contribution to drug response. Cancer J  2011; 17( 2): 80– 88. Google Scholar CrossRef Search ADS PubMed  5 Tan BR, Thomas F, Myerson RJ et al.   Thymidylate synthase genotype-directed neoadjuvant chemoradiation for patients with rectal adenocarcinoma. J Clin Oncol  2011; 29: 875– 883. Google Scholar CrossRef Search ADS PubMed  6 Toffoli G, Cecchin E, Gasparini G et al.   Genotype-driven phase I study of irinotecan administered in combination with fluorouracil/leucovorin in patients with metastatic colorectal cancer. J Clin Oncol  2010; 28: 866– 871. Google Scholar CrossRef Search ADS PubMed  7 Martinez-Balibrea E, Abad A, Aranda E, Spanish Group for the Treatment of Digestive Tumours et al.   Pharmacogenetic approach for capecitabine or 5-fluorouracil selection to be combined with oxaliplatin as first-line chemotherapy in advanced colorectal cancer. Eur J Cancer  2008; 44( 9): 1229– 1237. Google Scholar CrossRef Search ADS PubMed  8 Abad A, Carrato A, Navarro M et al.   Two consecutive phase II trials of biweekly oxaliplatin plus weekly 48-hour continuous infusion of nonmodulated high-dose 5-fluorouracil as first-line treatment for advanced colorectal cancer. Clin Colorectal Cancer  2005; 4( 6): 384– 389. Google Scholar CrossRef Search ADS PubMed  9 Aranda E, Carrato A, Cervantes A et al.   Phase I/II trial of irinotecan plus high-dose 5-fluorouracil (TTD regimen) as first-line chemotherapy in advanced colorectal cancer. Ann Oncol  2004; 15( 4): 559– 567. Google Scholar CrossRef Search ADS PubMed  10 Dotor E, Cuatrecases M, Martínez-Iniesta M et al.   Tumor thymidylate synthase 1494del6 genotype as a prognostic factor in colorectal cancer patients receiving fluorouracil-based adjuvant treatment. J Clin Oncol  2006; 24: 1603– 1611. Google Scholar CrossRef Search ADS PubMed  11 Marcuello E, Altes A, Menoyo A et al.   GT1A1 gene variations and irinotecan treatment in patients with metastatic colorectal cancer. Br J Cancer  2004; 91( 4): 678– 682. Google Scholar CrossRef Search ADS PubMed  12 Van Cutsem E, Cervantes A, Adam R et al.   ESMO consensus guidelines for the management of patients with metastatic colorectal cancer. Ann Oncol  2016; 27( 8): 1386– 1422. Google Scholar CrossRef Search ADS PubMed  13 Budai B, Komlósi V, Adleff V et al.   Impact of SHMT1 polymorphism on the clinical outcome of patients with metastatic colorectal cancer treated with first-line FOLFIRI+bevacizumab. Pharmacogenet Genomics  2012; 22( 1): 69– 72. Google Scholar CrossRef Search ADS PubMed  14 Pander J, Wessels JA, Gelderblom H et al.   Pharmacogenetic interaction analysis for the efficacy of systemic treatment in metastatic colorectal cancer. Ann Oncol  2011; 22( 5): 1147– 1153. Google Scholar CrossRef Search ADS PubMed  15 Uchida K, Hayashi K, Kawakami K et al.   Loss of heterozygosity at the thymidylate synthase (TS) locus on chromosome 18 affects tumor response and survival in individuals heterozygous for a 28-bp polymorphism in the TS gene. Clin Cancer Res  2004; 10( 2): 433– 439. Google Scholar CrossRef Search ADS PubMed  16 Martinez-Balibrea E, Abad A, Martínez-Cardús A et al.   UGT1A and TYMS genetic variants predict toxicity and response of colorectal cancer patients treated with first-line irinotecan and fluorouracil combination therapy. Br J Cancer  2010; 103( 4): 581– 589. Google Scholar CrossRef Search ADS PubMed  17 McLeod HL, Sargent DJ, Marsh S et al.   Pharmacogenetic predictors of adverse events and response to chemotherapy in metastatic colorectal cancer: results from North American Gastrointestinal Intergroup Trial N9741. J Clin Oncol  2010; 28: 3227– 3233. Google Scholar CrossRef Search ADS PubMed  18 Cecchin E, Innocenti F, D'Andrea M et al.   Predictive role of the UGT1A1, UGT1A7, and UGT1A9 genetic variants and their haplotypes on the outcome of metastatic colorectal cancer patients treated with fluorouracil, leucovorin, and irinotecan. J Clin Oncol  2009; 27: 2457– 2465. Google Scholar CrossRef Search ADS PubMed  © The Author 2017. Published by Oxford University Press on behalf of the European Society for Medical Oncology. All rights reserved. For Permissions, please email: journals.permissions@oup.com. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Annals of Oncology Oxford University Press

Genotype-based selection of treatment of patients with advanced colorectal cancer (SETICC): a pharmacogenetic-based randomized phase II trial

Loading next page...
 
/lp/ou_press/genotype-based-selection-of-treatment-of-patients-with-advanced-QF0MiOYtwL
Publisher
Oxford University Press
Copyright
© The Author 2017. Published by Oxford University Press on behalf of the European Society for Medical Oncology. All rights reserved. For Permissions, please email: journals.permissions@oup.com.
ISSN
0923-7534
eISSN
1569-8041
D.O.I.
10.1093/annonc/mdx737
Publisher site
See Article on Publisher Site

Abstract

Abstract Background There has been little progress toward personalized therapy for patients with metastatic colorectal cancer (mCRC). TYMS-3′ untranslated region (UTR) 6 bp ins/del and ERCC1-118C/T polymorphisms were previously reported to facilitate selecting patients for fluoropyrimidine-based treatment in combination with oxaliplatin as first-line therapy. We assessed the utility of these markers in selecting therapy for patients with mCRC. Patients and methods This randomized, open-label phase II trial compared bevacizumab plus XELOX (control) versus treatment tailored according to TYMS-3'UTR 6 bp ins/del and ERCC1-118C/T polymorphisms. Patients randomized to the experimental treatment received bevacizumab plus FUOX, FUIRI, XELIRI, or XELOX depending on their combination of favorable polymorphisms for FUOX treatment (TYMS-3′UTR ins/del or del/del; ERCC1-118T/T). Progression-free survival (PFS) was the primary end point. Results Overall, 195 patients were randomized (control n = 65; experimental n = 130). The primary objective was not met: median PFS was 9.4 months in the control group and 10.1 months in the experimental group (P = 0.745). Median overall survival was similar in both groups (16.5 versus 19.1 months, respectively; P = 0.797). Patients in the experimental group had a significantly higher overall response rate (ORR; 65% versus 47% in the control group; P = 0.042) and R0 resection rate (86% versus 44%, respectively; P = 0.018). Neuropathy, hand–foot syndrome, thrombocytopenia, and dysesthesia were significantly less common in the experimental group. Conclusions This study did not show survival benefits after treatment personalization based on polymorphisms in mCRC. However, the improved ORR and R0 resection rate and fewer disabling toxicities suggest that tailoring therapy by TYMS-3′UTR and ERCC1-118 polymorphisms warrants further investigation in patients with mCRC. ClinicalTrials.gov NCT01071655. colorectal cancer, polymorphisms, biomarkers, bevacizumab, oxaliplatin, irinotecan Key Message This study shows that prospective genotyping of TYMS-3′UTR 6 bp ins/del and ERCC1-118C/T polymorphisms is feasible and easy to perform. Overall response and R0 resection rates were significantly improved by genotype-driven treatment choice. Our findings support continued investigations into optimal treatment selection according to germline variations in patients with metastatic colorectal cancer. Introduction Treatment of patients with metastatic colorectal cancer (mCRC) is generally based around chemotherapy consisting of 5-fluorouracil (5-FU) or capecitabine, plus oxaliplatin or irinotecan, plus a targeted agent, e.g. bevacizumab, cetuximab, or panitumumab [1]. Physicians have many treatment options for patients with mCRC, and selecting appropriate regimens for individual patients can be challenging. To date, personalizing treatment applies only to selection of anti-epidermal growth factor receptor drugs based on existence of RAS oncogene mutations [2]. Recommended chemotherapy regimens have substantial activity in most patients; however, many patients experience resistance or excessive treatment-related toxicity. This variability may be related to interpatient differences in genes involved in drug transport, metabolism, signaling, and cellular response pathways [3]. Previous studies suggest that germline polymorphisms in genes involved in the mechanism of action or metabolism of chemotherapeutic agents may be responsible for variations in patients’ response or tolerability [4]. However, most previous studies were retrospective, with the exception of a few prospective studies in which patients were treated in a genotype-driven fashion [5, 6]. Our previous study suggested that patients harboring the TYMS-3′ untranslated region (UTR) 6 bp ins/ins and ERCC1-118C/T or C/C genotypes might benefit from combining oxaliplatin with capecitabine rather than 5-FU [7]. The aim of the present study was to examine whether assigning first-line chemotherapy based on relevant germline polymorphisms would improve outcomes compared with treating all patients with a standard regimen. Patients and methods Study design and patients This randomized, multicenter, open-label, phase II study was conducted in Spain by the Grupo Español de Tratamiento de Tumores Digestivos (Spanish Cooperative Group for the Treatment of Digestive Tumors). Patients had histologically confirmed colon or rectal adenocarcinoma with measurable metastatic disease (Response Evaluation Criteria in Solid Tumors [RECIST] Version 1.1), were aged ≥18 years, had Eastern Cooperative Oncology Group performance status 0–2, and adequate renal function. Key exclusion criteria were prior systemic treatment of metastatic disease; concomitant cardiovascular disease; history or signs of central nervous system disease, uncontrolled hypertension, bleeding diathesis, or coagulopathy. Exclusion criteria are described in detail in the supplementary Appendix, available at Annals of Oncology online. The study was conducted in accordance with the Declaration of Helsinki and Good Clinical Practice guidelines, and was approved by institutional review boards or independent ethics committees at each center. All patients provided written informed consent. Randomization Between February 2010 and April 2011, patients were randomized 1 : 2 to the control or experimental group. Treatment assignment was through an electronic system integrated in the electronic case report form, which automatically balanced groups. For the experimental group, treatment was assigned depending on TYMS-3′UTR 6 bp ins/del and ERCC1-118C/T polymorphisms. Investigators were informed of treatments by automatically generated e-mails. Polymorphisms were determined for patients assigned to the control group, but these results were not needed before treatment commenced. The investigator still received an automatically generated e-mail. Randomization is described in detail in the supplementary Appendix, available at Annals of Oncology online. Treatments Patients assigned to the control group received bevacizumab 7.5 mg/kg on day 1 with XELOX. Patients assigned to the experimental group received different schedules according to the number of TYMS-3′UTR 6 bp ins/del and ERCC1-118C/T favorable genotypes as previously described [7] (supplementary Table S1, available at Annals of Oncology online). Briefly, we considered TYMS-3′UTR ins/del and del/del to favor treament with 5FU, ins/ins genotype to favor capecitabine, ERCC1-118T/T to favor oxaliplatin and C/T or C/C genotypes to favor irinotecan. Accordingly, patients with favorable genotypes for FUOX received bevacizumab 5 mg/kg on days 1, 15, and 29 with FUOX [8]. Patients with no favorable polymorphisms received bevacizumab 7.5 mg/kg on day 1 with XELIRI. Patients with TYMS-3′UTR 6 bp ins/ins and ERCC1-118T/T genotype received the same treatment as the control arm. Patients with TYMS-3′UTR 6 bp ins/ins and ERCC1-118C/T or C/C genotype received bevacizumab 5 mg/kg on days 1, 15, and 29 with FUIRI [9]. Further detail can be found in the supplementary Appendix, available at Annals of Oncology online. Treatment continued until disease progression, unacceptable toxicity, or patient withdrawal. Outcomes The primary end point was progression-free survival (PFS). Secondary end points included overall survival (OS), overall response rate (ORR; assessed using RECIST Version 1.1), proportion of patients whose disease became resectable, adverse events, and evaluation of KRAS exon 2 mutation status as a molecular prognostic marker. Toxicity was assessed using National Cancer Institute Common Toxicity Criteria for Adverse Events (Version 3.0). Adverse events were assessed at each cycle and every 12 weeks during the study follow-up. In patients treated with FUOX and FUIRI, hematology, biochemistry, and urinalysis assessments were carried out every 2 weeks. Statistical analysis It was estimated that the difference between groups for PFS would be statistically significant if, when 50% of patients in the control group had disease progression, 65% of patients in the experimental group had no disease progression. The maximum permitted error was considered to be α = 0.05 (unilateral contrast, assuming that the efficacy was greater in the experimental group) and β = 0.2 (80% power). Based on these assumptions, 171 patients (57 in the control group and 114 in the experimental group) were needed, assuming a 5% drop-out rate. We planned to include 180 patients. Further detail is provided in the supplementary Appendix, available at Annals of Oncology online. The study was registered with ClinicalTrials.gov (NCT01071655). Results Baseline and genotype data Of 202 patients enrolled, 195 were randomized (Figure 1). Baseline characteristics were generally balanced between the two groups (Table 1); we found no statistically significant differences between the groups in baseline characteristics (data not shown). Supplementary Table S1, available at Annals of Oncology online, shows the distribution of genotypes and assigned treatments in the experimental arm. Frequencies of TYMS-3′UTR ins and del alleles were 0.68 and 0.32, respectively; frequencies of ERCC1-118T and C alleles were 0.57 and 0.43, respectively. Genotype frequencies were TYMS ins/ins 50%, ins/del 37%, and del/del 13%; ERCC1-118T/T 35%, C/T 45%, and C/C 21%. Table 1. Patient characteristics at baseline (intent to treat population, n = 195) Characteristic  Control group (n = 65)  Experimental group (n = 130)  Age, years (range)  64.8 (36.8–82.9)  63.6 (28.8–86.1)  Interquartile range  58.7–72.7  57.2–70.7  Male  35 (54)  82 (63)  ECOG performance status       0  29 (45)  50 (38)   1  34 (52)  79 (61)   2  2 (3)  1 (1)  Tumor location       Colon  18 (28)  38 (29)   Rectum  41 (63)  80 (62)   Both  6 (9)  12 (9)  No. of affected organs       1  25 (38)  55 (42)   2  21 (32)  52 (40)   >2  19 (29)  23 (18)  Location of metastases       Local/regional  17 (26)  34 (26)   Liver  51 (78)  106 (82)   Liver only  21 (32)  45 (35)   Lung  22 (34)  50 (38)   Other  20 (31)  29 (22)  Surgery for primary tumor  34 (52)  73 (56)  Prior adjuvant therapya  10 (15)  15 (12)   Chemotherapy  9 (14)  15 (12)   Radiotherapy  3 (5)  7 (5)  KRAS mutation  27 (42)  42 (32)  Characteristic  Control group (n = 65)  Experimental group (n = 130)  Age, years (range)  64.8 (36.8–82.9)  63.6 (28.8–86.1)  Interquartile range  58.7–72.7  57.2–70.7  Male  35 (54)  82 (63)  ECOG performance status       0  29 (45)  50 (38)   1  34 (52)  79 (61)   2  2 (3)  1 (1)  Tumor location       Colon  18 (28)  38 (29)   Rectum  41 (63)  80 (62)   Both  6 (9)  12 (9)  No. of affected organs       1  25 (38)  55 (42)   2  21 (32)  52 (40)   >2  19 (29)  23 (18)  Location of metastases       Local/regional  17 (26)  34 (26)   Liver  51 (78)  106 (82)   Liver only  21 (32)  45 (35)   Lung  22 (34)  50 (38)   Other  20 (31)  29 (22)  Surgery for primary tumor  34 (52)  73 (56)  Prior adjuvant therapya  10 (15)  15 (12)   Chemotherapy  9 (14)  15 (12)   Radiotherapy  3 (5)  7 (5)  KRAS mutation  27 (42)  42 (32)  Data are median (range) or n (%). No P values are shown as no statistically significant differences were found between the control and experimental groups for any of the characteristics studied. a With or without radiotherapy. ECOG, Eastern Cooperative Oncology Group. Table 2. Response to treatment (response-evaluable population, n = 161) and R0 surgery (intent to treat population, n = 195) Outcome  Control group (n = 55)  Experimental group (n = 106)  P-value  Complete response (CR)  2 (4)  4 (4)    Partial response (PR)  24 (44)  65 (61)    Stable disease  22 (40)  30 (28)    Progressive disease  7 (13)  7 (7)    Overall response rate (PR + CR)  26 (47)  69 (65)  0.042  Disease-control rate (CR + PR + SD)  48 (87)  99 (93)  0.240  R0 surgery  7 (44)a  18 (86)b  0.018  Outcome  Control group (n = 55)  Experimental group (n = 106)  P-value  Complete response (CR)  2 (4)  4 (4)    Partial response (PR)  24 (44)  65 (61)    Stable disease  22 (40)  30 (28)    Progressive disease  7 (13)  7 (7)    Overall response rate (PR + CR)  26 (47)  69 (65)  0.042  Disease-control rate (CR + PR + SD)  48 (87)  99 (93)  0.240  R0 surgery  7 (44)a  18 (86)b  0.018  Data are n (%). a Proportion of the 16 patients with surgical resection. b Proportion of 21 patients who underwent surgical resection. Figure 1. View largeDownload slide Patient disposition. AE, adverse event; Bev, bevacizumab; FUIRI, 5-fluorouracil + irinotecan; FUOX, 5-fluorouracil + oxaliplatin; ITT, intent to treat; XELIRI, capecitabine + irinotecan; XELOX, capecitabine + oxaliplatin. Figure 1. View largeDownload slide Patient disposition. AE, adverse event; Bev, bevacizumab; FUIRI, 5-fluorouracil + irinotecan; FUOX, 5-fluorouracil + oxaliplatin; ITT, intent to treat; XELIRI, capecitabine + irinotecan; XELOX, capecitabine + oxaliplatin. Overall, 40% of patients in the control group and 51% of patients in the experimental group withdrew from the study for reasons other than disease progression or adverse events (Figure 1). Patients in the control group received a median of nine treatment cycles; those in the experimental group received a median of six cycles (NS). Dose intensities were lowest for bevacizumab + FUOX and bevacizumab + FUIRI (supplementary Table S2, available at Annals of Oncology online). Fifty of 65 patients in the control group (77%) and 94 of 130 (72%) in the experimental group received second-line treatments. PFS and OS analysis The primary end point was not met; after a median follow-up of 18 months (range 1.1–41.7 months), median PFS was similar in both groups {9.4 months [95% confidence interval (CI) 6.8–12.1 months] in the control group versus 10.1 months (95% CI 8.5–11.6 months) in the experimental group; hazard ratio (HR) 0.942; 95% CI 0.657–1.351; P = 0.745; Figure 2A}. Median OS was numerically lower in the control group (16.5 months; 95% CI 13.7–19.4 months) than in the experimental group (19.1 months; 95% CI 15.5–22.7 months; HR 0.956; 95% CI 0.678–1.348; P = 0.798; Figure 2B). In line with our previous findings [7], we did not expect to observe a survival difference in the control group when patients were stratified according to genotype. Indeed, no statistically significant differences in PFS and OS according to genotype were found in the control group (supplementary Figure S1, available at Annals of Oncology online). Figure 2. View largeDownload slide Progression-free survival (A) and overall survival (B) in the control and experimental groups. Patients in the control group (group A) received bevacizumab + capecitabine + oxaliplatin and patients in the experimental group (group B) received bevacizumab + chemotherapy tailored according to the presence of specific genotypes. CI, confidence interval; HR, hazard ratio. Figure 2. View largeDownload slide Progression-free survival (A) and overall survival (B) in the control and experimental groups. Patients in the control group (group A) received bevacizumab + capecitabine + oxaliplatin and patients in the experimental group (group B) received bevacizumab + chemotherapy tailored according to the presence of specific genotypes. CI, confidence interval; HR, hazard ratio. Response rate and R0 surgery A total of 161 patients were evaluable for response (Table 2). The ORR was significantly higher in the experimental group than in the control group (P = 0.042). When confirmed responses alone were considered, the same trend was observed (51 of 106 patients [48%] versus 18 of 55 patients [33%] for experimental versus control groups, respectively; P = 0.07). Patients in whom tumor response could not be confirmed were those who had undergone surgery for metastases. Twenty-two of 130 patients (17%) in the experimental group and 16 of 65 (25%) in the control group had surgical resection of metastases during or after the study (P = 0.250). Eleven patients had rescue surgery twice; three patients had three rescue surgeries. The R0 resection rate was statistically significantly higher in the experimental group than in the control group (P = 0.018; Table 2). The most common resection sites were the liver and lung. Additional genetic analysis Ninety-seven tumor specimens were available for tumor and blood TYMS-3′UTR 6 bp ins/del genotyping. We found 100% concordance between tumor and blood results in the 62 homozygous samples. However, results for 23 of the 35 (66%) heterozygous blood samples did not agree with the corresponding tumor sample, as previously described [10]. This disagreement was total in four cases (17%) and partial (one of the alleles was partially lost but still present) in 19 cases (83%) (supplementary Figure S2, available at Annals of Oncology online). According to tumor genotyping, only two patients in the experimental group received the wrong treatment regimen. Information regarding KRAS mutations in exon 2 codons 12 and 13 was available for 164 patients (84%): 58 of 65 patients (89.2%) in the control group and 106 of 130 patients (81.5%) in the experimental group. KRAS mutations were present in 27 patients (47%) in the control group and 42 patients (40%) in the experimental group. KRAS mutation status had no impact on any of the clinical variables studied (data not shown). UGT1A1*28 polymorphism genotyping in the 75 irinotecan-treated patients was: UGT1A1*1/*1, n = 31 (41%), UGT1A1*1/*28, n = 33 (44%), and UGT1A1*28/*28, n = 11 (15%). Allelic frequencies for UGT1A1*1 and UGT1A1*28 alleles were 0.63 and 0.37, respectively, in agreement with Spanish population data [11]. The association between genotype and outcomes is summarized in supplementary Table S3, available at Annals of Oncology online. PFS and OS were worst in patients with the UGT1A1*28/*28 genotype according to codominant and recessive analyses. Repeating the PFS analysis without these patients did not affect the differences between the control and experimental groups (data not shown). An additional analysis of UGT1A1*28 polymorphism in patients who did not receive irinotecan was carried out. When patients were stratified by UGT1A1*28 polymorphism no statistically significant association with toxicity or efficacy outcomes was observed (data not shown). Toxicity Adverse events are summarized in supplementary Table S4, available at Annals of Oncology online. Neuropathy, hand–foot syndrome, thrombocytopenia, and dysesthesia were significantly less frequent in the experimental group (all P < 0.05 versus the control group); diarrhea was numerically more frequent in the experimental group (P = 0.074). Grade ≥3 diarrhea was significantly more frequent in the experimental group (P = 0.005); grade ≥3 neuropathy was significantly more frequent in the control group (P = 0.032). The UGT1A1*28/*28 genotype was significantly associated with a higher incidence of severe diarrhea in patients treated with irinotecan (supplementary Table S3, available at Annals of Oncology online). Discussion Although combinations of fluoropyrimidines with oxaliplatin or irinotecan remain the backbone of mCRC treatment, oncologists do not currently have effective tools to help choose the most suitable first-line schedule [12]. For the first time, to the best of our knowledge, standard treatment with bevacizumab plus XELOX has been compared with treatment tailored by TYMS-3′UTR 6 bp ins/del and ERCC1-118C/T polymorphisms in mCRC patients in a multicenter randomized phase II study. We demonstrated that this approach is feasible and applicable to clinical practice as patients were randomly assigned to treatment according to genotyping within 3 days of phlebotomy. The primary study end point—difference in PFS—was not met; although we observed a 3-month difference in OS in favor of tailored treatment, this was not statistically significant. There were no genotyping errors, and genotype and allelic frequencies were in Hardy–Weinberg equilibrium. Therefore, other possible explanations for these results should be considered. Importantly, 27% of patients in the experimental arm withdrew, citing patient or investigator’s decision as their reason. This may have affected outcomes in the experimental arm if subsequent treatment was not optimal for the patient. Although not all responses were confirmed as some patients underwent surgery, the trend toward better RR and higher rate of conversion surgery, together with statistically significant differences in unconfirmed treatment response, reinforces this hypothesis. The impact of bevacizumab should also be considered. Unlike our earlier study [7], the present study was designed after approval of bevacizumab for patients with mCRC; consequently, bevacizumab was included in treatment regimens. Reported interaction of genetic variants in VEGF and TYMS and influence of SHMT1 gene variants [13, 14] suggest that addition of bevacizumab could have affected the study results. This warrants further investigation. We also investigated whether loss of heterozygosity (LOH) or UGT1A1*28 polymorphism affected our results. TYMS LOH frequency was similar to previous reports [10, 15]. Tumor genotyping revealed that only two patients in the experimental group received the wrong treatment, suggesting that LOH was unlikely to have affected our results. Irinotecan-treated patients with UGT1A1*28/*28 genotype had worse PFS than those with UGT1A1*1/*28 or UGT1A1*1/*1 genotypes. Only 11 of 130 patients had the UGT1A1*28/*28 genotype, which appeared to have little effect on PFS, as suggested by the PFS analysis excluding patients with the UGT1A1*28/*28 genotype. A significant benefit for tailored treatment was observed in ORR and R0 surgery rates. R0 surgery is considered a good prognostic factor but no predictive markers are available to tailor neoadjuvant treatment before resection of metastases. Despite the reasonable caution needed in assessing results regarding secondary end points, patients whose liver metastases cannot be operated on at treatment outset may benefit from an approach such as we have described. Improving response rates and R0 surgery rates in these patients can dramatically improve their chance of surviving beyond 5years. Toxicities were as expected: the overall incidence of neurotoxicity in the experimental group was lower than in the control group, and the incidence of diarrhea was higher in the experimental group than in the control group. Neurotoxicity is chronic and causes varying degrees of functional impairment, whereas diarrhea is generally acute and has no sequelae. Therefore, tailoring treatment may reduce disabling toxicity while at least maintaining pharmacological activity, as demonstrated by the similar survival in both arms and better R0 resection rates in the experimental group. Furthermore, the higher incidence of severe diarrhea in patients with UGT1A1*28/*28 genotype who received irinotecan confirms previous results [16–18], highlighting the importance of assessing this polymorphism to avoid severe toxicity. Some limitations should be considered. Serum samples were not analyzed for cell-free tumor DNA as the study was designed before evidence for the prognostic value of such analyses became available. This technique may allow identification of variability in response due to the tumor genome, complementary to information provided by examination of germline polymorphisms. Conclusion In conclusion, the results of this study suggest that prospective first-line treatment selection according to TYMS-3′UTR 6 bp ins/del and ERCC1-118C/T polymorphisms in patients with mCRC does not improve TTP or OS. However, we showed this approach to be feasible and applicable to clinical practice. Promising ORR and R0 surgery rates in the experimental arm, without compromising toxicity, combined with possible logistical reasons for lack of significant difference in PFS, provide the rationale for continuing to investigate how best to select treatment according to germline variations and improve survival in patients with mCRC. Acknowledgements Spanish Cooperative Group for the Treatment of Digestive Tumors (TTD) Data Center: Inmaculada Ruiz de Mena and Susana Rodríguez. Data Management, Monitoring and Statistics (Dynamic): Laura Casas Herrero—Biostatistician, and Fátima González Hurtado—Monitoring. Editorial support was provided by Deirdre Carman, PhD, Miller Medical Communications Ltd, funded by Roche Farma, Spain. Funding Roche Farma (no grant number applies). Disclosure JMV received personal fees from Roche during the conduct of the study. VAO reports personal fees from Roche during the conduct of the study and personal fees, outside the submitted work, from Merck and Sanofi. MB received personal fees from Roche during the conduct of the study and personal fees from Merck, Amgen, Bayer, Sanofi, Lilly, and Servier that were outside the submitted work. MV received personal fees from Roche during the conduct of the study and personal fees, outside the submitted work, from Amgen and Merck-Serono. JS received personal fees from Roche during the conduct of the study and personal fees outside the submitted work from Bayer, Amgen, Boehringer Ingelheim, Merck, Sanofi, Ipsen, and Celgene. FR received personal fees from Roche during the conduct of the study and personal fees from Merck-Serono, Amgen, Celgene, Bayer, MSD, and Lilly outside the submitted work. EA received personal fees from Roche during the conduct of the study and personal fees from Amgen, Bayer, Celgene, Merck, and Sanofi outside the submitted work. AA, AC, VC, PGA, JG, GG, JLM, EM-B, BM, CM, MP, RS, and MZ have nothing to disclose. References 1 National Comprehensive Cancer Network Clinical Practice Guidelines in Oncology. Colon cancer. Version 2. 2017. http://www.nccn.org/professionals/physician_gls/pdf/colon.pdf (5 July 2017, date last accessed). 2 Van Cutsem E, Lenz HJ, Köhne CH et al.   Fluorouracil, leucovorin, and irinotecan plus cetuximab treatment and RAS mutations in colorectal cancer. J Clin Oncol  2015; 33( 7): 692– 700. Google Scholar CrossRef Search ADS PubMed  3 Unger FT, Witte I, David KA. Prediction of individual response to anticancer therapy: historical and future perspectives. Cell Mol Life Sci  2015; 72( 4): 729– 757. Google Scholar CrossRef Search ADS PubMed  4 Watson RG, McLeod HL. Pharmacogenomic contribution to drug response. Cancer J  2011; 17( 2): 80– 88. Google Scholar CrossRef Search ADS PubMed  5 Tan BR, Thomas F, Myerson RJ et al.   Thymidylate synthase genotype-directed neoadjuvant chemoradiation for patients with rectal adenocarcinoma. J Clin Oncol  2011; 29: 875– 883. Google Scholar CrossRef Search ADS PubMed  6 Toffoli G, Cecchin E, Gasparini G et al.   Genotype-driven phase I study of irinotecan administered in combination with fluorouracil/leucovorin in patients with metastatic colorectal cancer. J Clin Oncol  2010; 28: 866– 871. Google Scholar CrossRef Search ADS PubMed  7 Martinez-Balibrea E, Abad A, Aranda E, Spanish Group for the Treatment of Digestive Tumours et al.   Pharmacogenetic approach for capecitabine or 5-fluorouracil selection to be combined with oxaliplatin as first-line chemotherapy in advanced colorectal cancer. Eur J Cancer  2008; 44( 9): 1229– 1237. Google Scholar CrossRef Search ADS PubMed  8 Abad A, Carrato A, Navarro M et al.   Two consecutive phase II trials of biweekly oxaliplatin plus weekly 48-hour continuous infusion of nonmodulated high-dose 5-fluorouracil as first-line treatment for advanced colorectal cancer. Clin Colorectal Cancer  2005; 4( 6): 384– 389. Google Scholar CrossRef Search ADS PubMed  9 Aranda E, Carrato A, Cervantes A et al.   Phase I/II trial of irinotecan plus high-dose 5-fluorouracil (TTD regimen) as first-line chemotherapy in advanced colorectal cancer. Ann Oncol  2004; 15( 4): 559– 567. Google Scholar CrossRef Search ADS PubMed  10 Dotor E, Cuatrecases M, Martínez-Iniesta M et al.   Tumor thymidylate synthase 1494del6 genotype as a prognostic factor in colorectal cancer patients receiving fluorouracil-based adjuvant treatment. J Clin Oncol  2006; 24: 1603– 1611. Google Scholar CrossRef Search ADS PubMed  11 Marcuello E, Altes A, Menoyo A et al.   GT1A1 gene variations and irinotecan treatment in patients with metastatic colorectal cancer. Br J Cancer  2004; 91( 4): 678– 682. Google Scholar CrossRef Search ADS PubMed  12 Van Cutsem E, Cervantes A, Adam R et al.   ESMO consensus guidelines for the management of patients with metastatic colorectal cancer. Ann Oncol  2016; 27( 8): 1386– 1422. Google Scholar CrossRef Search ADS PubMed  13 Budai B, Komlósi V, Adleff V et al.   Impact of SHMT1 polymorphism on the clinical outcome of patients with metastatic colorectal cancer treated with first-line FOLFIRI+bevacizumab. Pharmacogenet Genomics  2012; 22( 1): 69– 72. Google Scholar CrossRef Search ADS PubMed  14 Pander J, Wessels JA, Gelderblom H et al.   Pharmacogenetic interaction analysis for the efficacy of systemic treatment in metastatic colorectal cancer. Ann Oncol  2011; 22( 5): 1147– 1153. Google Scholar CrossRef Search ADS PubMed  15 Uchida K, Hayashi K, Kawakami K et al.   Loss of heterozygosity at the thymidylate synthase (TS) locus on chromosome 18 affects tumor response and survival in individuals heterozygous for a 28-bp polymorphism in the TS gene. Clin Cancer Res  2004; 10( 2): 433– 439. Google Scholar CrossRef Search ADS PubMed  16 Martinez-Balibrea E, Abad A, Martínez-Cardús A et al.   UGT1A and TYMS genetic variants predict toxicity and response of colorectal cancer patients treated with first-line irinotecan and fluorouracil combination therapy. Br J Cancer  2010; 103( 4): 581– 589. Google Scholar CrossRef Search ADS PubMed  17 McLeod HL, Sargent DJ, Marsh S et al.   Pharmacogenetic predictors of adverse events and response to chemotherapy in metastatic colorectal cancer: results from North American Gastrointestinal Intergroup Trial N9741. J Clin Oncol  2010; 28: 3227– 3233. Google Scholar CrossRef Search ADS PubMed  18 Cecchin E, Innocenti F, D'Andrea M et al.   Predictive role of the UGT1A1, UGT1A7, and UGT1A9 genetic variants and their haplotypes on the outcome of metastatic colorectal cancer patients treated with fluorouracil, leucovorin, and irinotecan. J Clin Oncol  2009; 27: 2457– 2465. Google Scholar CrossRef Search ADS PubMed  © The Author 2017. Published by Oxford University Press on behalf of the European Society for Medical Oncology. All rights reserved. For Permissions, please email: journals.permissions@oup.com.

Journal

Annals of OncologyOxford University Press

Published: Feb 1, 2018

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 12 million articles from more than
10,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Unlimited reading

Read as many articles as you need. Full articles with original layout, charts and figures. Read online, from anywhere.

Stay up to date

Keep up with your field with Personalized Recommendations and Follow Journals to get automatic updates.

Organize your research

It’s easy to organize your research with our built-in tools.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

Monthly Plan

  • Read unlimited articles
  • Personalized recommendations
  • No expiration
  • Print 20 pages per month
  • 20% off on PDF purchases
  • Organize your research
  • Get updates on your journals and topic searches

$49/month

Start Free Trial

14-day Free Trial

Best Deal — 39% off

Annual Plan

  • All the features of the Professional Plan, but for 39% off!
  • Billed annually
  • No expiration
  • For the normal price of 10 articles elsewhere, you get one full year of unlimited access to articles.

$588

$360/year

billed annually
Start Free Trial

14-day Free Trial