Open-label, randomized, comparative, phase III study on effects of reducing steroid use in combination with Palonosetron

The purpose of this study is to compare the efficacy of a single administration of dexamethasone (DEX) on day 1 against DEX administration on days 1–3 in combination with palonosetron (PALO), a second-generation 5-HT3 receptor antagonist, for chemotherapy-induced nausea and vomiting (CINV) in non-anthracycline and cyclophosphamide (AC) moderately-emetogenic chemotherapy (MEC). This phase III trial was conducted with a multi-center, randomized, open-label, non-inferiority design. Patients who received non-AC MEC as an initial chemotherapy were randomly assigned to either a group administered PALO (0.75 mg, i.v.) and DEX (9.9 mg, i.v.) prior to chemotherapy (study treatment group), or a group administered additional DEX (8 mg, i.v. or p.o.) on days 2–3 (control group). The primary endpoint was complete response (CR) rate. The CR rate difference was estimated by logistic regression with allocation factors as covariates. The non-inferiority margin was set at −15% (study treatment group − control group). From April 2011 to March 2013, 305 patients who received non-AC MEC were randomly allocated to one of two study groups. Overall, the CR rate was 66.2% in the study treatment group (N = 151) and 63.6% in the control group (N = 154). PALO plus DEX day 1 was non-inferior to PALO plus DEX days 1–3 (difference, 2.5%; 95% confidence interval [CI]: −7.8%–12.8%; P-value for non-inferiority test = 0.0004). There were no differences between the two groups in terms of complete control rate (64.9 vs 61.7%) and total control rate (49.7% vs 47.4%). Anti-emetic DEX administration on days 2–3 may be eliminated when used in combination with PALO in patients receiving non-AC MEC.

Association of morphologic response with progression free survival in patients with metastatic colorectal cancer treated with bevacizumab-based chemotherapy: HGCSG0802

743

Background: It was reported that an optimal morphologic response to preoperative chemotherapy was associated with better overall survival (OS) in patients (pts) with colorectal liver metastases (CLM). We investigated association of morphologic response with progression free survival (PFS) in pts with unresectable CLM from HGCSG0802 observational cohort study in pts with mCRC treated with first-line bevacizumab (BV)-based chemotherapy. Methods: The objective of HGCSG0802 was to evaluate PFS, OS, time to treatment failure (TTF), response rate (RR), safety, etc. The key eligibility criteria were evaluable lesions, older than 20 years old, ECOG PS 0-2. Pts with CLM underwent contrast-enhanced CT at the start and every 8-weeks of BV-based chemotherapy. In this analysis, three blinded, independent radiologists evaluated images for morphologic response, based on metastases changing from heterogeneous masses with ill-defined margins into homogeneous hypoattenuating lesions with sharp borders. Association of morphologic response and pts characteristics, RR, and PFS were evaluated. PFS was analyzed with Kaplan-Meier method, log-rank test, and Cox proportional hazards model. Results: Of 108 pts (the full analysis set), 73 pts with CLM were evaluable for morphologic criteria. Eighteen pts (24.7%) had optimal morphologic response (OR), 31 (42.5%) had incomplete (IR), and 24 (32.9%) had no response (NR). The pts characteristics between those with OR, IR and NR were generally balanced. The median TTF was 7.2 months in NR versus 7.2 months in IR versus 6.8 months in OR (HR (OR/NR) = 0.91, HR (OR/IR) = 0.90; p = 0.93). RR was 77.8% in OR versus 64.5% in IR and 58.3% in NR (p = 0.528). The median PFS was 8.3 months in NR versus 8.5 months in IR versus 9.1 months in OR (HR (OR/NR) = 0.72, HR (OR/IR) = 1.04; p = 0.420). Conclusions: In this analysis, morphologic response might not be a prognostic marker in first-line BV-based chemotherapy in pts with CLM.

Association of early tumor shrinkage with progression-free survival in patients with metastatic colorectal cancer treated with bevacizumab-based chemotherapy: HGCSG0802

749

Background: It was reported that early tumor shrinkage (ETS) was associated with better overall survival (OS) in patients (pts) with metastatic colorectal cancer (mCRC) receiving cetuximab. We investigated association of ETS with progression free survival (PFS) in pts with unresectable colorectal liver metastases (CLM) from HGCSG0802 observational cohort study in pts with mCRC treated with first-line bevacizumab (BV)-based chemotherapy. Methods: The objective of HGCSG0802 was to evaluate PFS, OS, time to treatment failure (TTF), response rate (RR), safety and so on. The key eligibility criteria were evaluable lesions, older than 20 years old, ECOG PS 0-2. In this analysis, association of ETS at 8 weeks from the start of chemotherapy with pts characteristics, PFS and TTF was evaluated. Pts characteristics were compared using Student-t test, chi-square test and Fisher’s exact test. PFS and TTF were analyzed with Kaplan-Meier method and compared using log-rank test. Univariate analysis for the association of pts characteristics with PFS and TTF was performed using log-rank test, and multivariate analysis was performed using Cox proportional hazards model. Results: Of 108 pts (the full analysis set), 74 pts with CLM were evaluable for ETS. Forty-nine pts (66.2%) had ETS ≥20%. The pts characteristics between ETS ≥20% and <20% were well balanced. The median PFS was 7.3 months in ETS <20% versus 10.0 months in ETS ≥20% (HR 0.55; p=0.025). In multivariate analysis for PFS, there was no significant difference between ETS ≥20% and <20% (HR 0.585; p=0.066). The median TTF (ETS <20% v ≥20%) was 5,1 months vs. 7.7 months (HR 0.46; p=0.003). In multivariate analysis for TTF, there was significant difference between ETS ≥20% and <20% (HR 0.509; p=0.017). Conclusions: In this analysis, ETS ≥20% might be positive predictive marker for PFS and TTF in pts with CLM receiving first-line BV-based chemotherapy.

Observational cohort study of first-line bevacizumab combined with chemotherapy in metastatic colorectal cancer (HGCSG0802): Sub-group analysis by KRAS Exon2 status

782

Background: A few reports have shown the efficacy of bevacizumab (BV) independent of the KRAS Exon2 mutational status (KRAS). We performed a sub-group analysis by KRAS from the HGCSG0802 observational cohort study with investigated 115 patients (pts) treated with first line BV for metastatic colorectal cancer (mCRC). Methods: The objective of HGCSG0802 was to evaluate progression-free survival (PFS), overall survival (OS), time to treatment-failure (TTF), response rate (RR), safety, etc. The key eligibility criteria were with evaluable lesions, older than 20 years, ECOG PS 0-2. In this analysis, pts characteristics, RR and safety were compared using Fisher’s exact test. Univariate and multivariate analysis for PFS were performed using patient characteristics. Survival analyses were performed with Kaplan-Meier method, log-rank test, and Cox proportional hazards model. Results: Of 108 pts (the full analysis set), 98 pts were evaluable for KRAS. Sixty-one pts (62.2%) had KRAS wild-type (wt) and 37 pts (37.8%) had mutation (mt). The pts characteristics between those with wt and with mt were generally balanced except for PS 0 (91.8% in wt, 75.7% in mt; p=0.037) and lung metastasis (34.4% in wt, 62.2% in mt; p=0.012).The median TTF was 7.4 months in wt versus 6.6 months in mt, and RR was 69.5% in wt versus 65.7% in mt. Adverse events related to BV were almost balanced except for bleeding (any grade) (31.1% in wt, 13.5% in mt; p=0.056).The median PFS was 9.9 months in wt versus 7.9 months in mt (HR=1.506; p=0.071). Although KRAS mt showed shorter PFS, there were not significant difference regardless of KRAS in Cox multivariate analysis (HR 1.314, 95% CI, 0.820-2.107, p=0.257). Conclusions: The HGCSG0802 could be a database to investigate first line BV for mCRC in clinical practice. Depending on the KRAS Exon2 mutational status, efficacy, and adverse events were no significant difference. We plan to conduct further follow up for survival, and to perform this analysis again.

Observational cohort study of first-line bevacizumab combined with chemotherapy in metastatic colorectal cancer (HGCSG0802): Comparison of infusional FU/oxaliplatin(OX)+BV and oral FU/OX+BV

527

Background: A few reports have shown no difference between the efficacy of infusional FU and that of oral FU (Capecitabine/S-1) for colorectal cancer, and some studies have reported the non-inferiority between infusional FU/Oxaliplatin (OX) and oral FU/OX for metastatic colorectal cancer (mCRC). We performed a sub-group comparison between infusional FU/OX (mFOLFOX6 + BV: iFU) and oral FU/OX (CapeOX/SOX + BV: oFU) from the HGCSG0802 observational cohort study with investigated Japanese patients (pts) treated with first line BV for mCRC. Methods: The objective of HGCSG0802 was to evaluate progression-free survival (PFS), overall survival (OS), time to treatment-failure (TTF), response rate (RR), safety and so on. The key eligibility criteria of HGCSG0802 were with evaluable lesions, older than 20 years, ECOG PS 0-2, and this analysis used the cohort treated with OX-based regimens.In this analysis, pts characteristics, RR and safety were compared using Fisher’s exact test. PFS and TTF were compared using log-rank test. Results: Of 108 pts (the full analysis set), 95 pts were evaluable for treated with OX-based regimens. Forty-eight pts (50.5%) were treated with iFU and 47 pts (49.5%) were treated with oFU (CapeOX + BV 42 pts/SOX + BV 5 pts). The pts characteristics between those were generally balanced except for PS 0-1 (72.9% in iFU/93.6% in oFU; p=0.012) and synchronous liver metastases (mets) (93.8% in iFU/78.8% in oFU; p=0.040). Adverse events ≥grade 3 were balanced except for leucopenia (25.0% in iFU versus 2.1% in oFU; p=0.002) and neutropenia (43.5% in iFU and 10.9% in oFU; p=0.001). Hand-foot skin reaction was not different between two cohorts. RR was 62.5% in iFU versus 71.1% in oFU (p=0.835). The median PFS was 8.3 months in iFU versus 8.2 months in oFU (p=0.835). Conclusions: The HGCSG0802 could be a database to investigate first line BV for mCRC in clinical practice. As a result of this analysis, in Japanese daily practice, efficacy was no significant difference between infusional FU/OX and oral FU/OX, and the profiles of adverse events varied from each regimens.

Safety analysis of FOLFOX as adjuvant chemotherapy for stage III colon cancer in phase II study (NORTH/HGCSG1003): Detailed analysis of peripheral sensory neuropathy

701

Background: Oxaliplatin-containing regimen is a standard adjuvant chemotherapy for resected stage III colon cancer. Oxaliplatin-containing regimens were investigated for their efficacy in patients with resected stage III colon cancer in MOSAIC and XELOXA studies. Since these two international randomized studies were performed outside of Japan, we conducted a phase II study (NORTH/HGCSG1003) to assess the efficacy and safety of FOLFOX as adjuvant chemotherapy in Japanese patients (pts) with resected stage III colon cancer (UMIN ID: 000004590). Methods: This phase II study enrolled patients with resected stage III colon cancer. Patients received 12 biweekly cycles of FOLFOX4 or mFOLFOX6. Sample size was determined to be 243 pts. Primary endpoint was DFS. Secondary endpoints included overall survival (OS) and safety. Results: From September 2010 to March 2013, 273 pts were enrolled at 28 institutions. Safety analysis included 265 patients who received FOLFOX. Patients characteristics were as follows: median age, 65 (33-84); male/female: 131/134; PS 0/1:258/7; stage IIIA/IIIB/IIIC: 37/197/31; colon/rectosigmoid: 214/51. The most common grade 3-4 adverse events were neutrophil count decreased (48.1%), platelet count decreased (2.3%), and allergic reaction (1.5%). The incidence of peripheral sensory neuropathy (PSN) was 41.9% (grade 1), 38.1% (grade 2), and 6.4% (grade 3). PSN tended to be serious depending on the cumulative dose of oxaliplatin (table 1). Median cumulative dose of oxaliplatin at which PSN occurred were as follows: grade 1; 170 mg/m2, ≥ grade 2; 850 mg/m2, ≥ grade 3; (-). The median number of cycles of chemotherapy was 12, and the completion treatment rate was 80.4%. There was no treatment-related death. Conclusions: In Japanese patients with stage III colon cancer, FOLFOX is a well-tolerable regimen as adjuvant chemotherapy. Clinical trial information: 000004590. [Table: see text]

The efficacy of first-line IRIS with or without bevacizumab in patients with metastatic colorectal cancer: Including multivariate analysis of two phase II studies

603

Background: The safety and efficacy of first-line IRIS (S-1 in combination with irinotecan) and IRIS/Bev (IRIS in combination with bevacizumab [Bev]) have been evaluated in patients with metastatic colorectal cancer (mCRC). To date, no randomized studies comparing these regimens have been performed. This retrospective analysis compared efficacy data for the two regimens from separate phase II studies performed at Hokkaido Gastrointestinal Cancer Study Group (HGCSG). Methods: Patients with histologically confirmed unresectable metastatic or recurrent CRC and received no prior chemotherapy were enrolled. In the first trial, patients received irinotecan 100 mg/m2 on day 1,15 and oral S-1 40 mg/m2 twice daily on days 1-14 every 4 weeks (IRIS study: HGCSG0302). In the second trial, patients received the same regimen plus Bev 5 mg/kg on day 1,15 (IRIS/Bev study). Results: A total of 40 and 52 patients were enrolled the IRIS and IRIS/Bev studies, respectively. Patient characteristics were generally similar in both groups, whereas there were more cases of good performance status and less number of metastatic organ in IRIS/Bev group. The median overall survival was 39.6 months in IRIS/Bev, as compared with 23.4 months in IRIS, corresponding to a hazard ratio for death of 0.418 (p<0.001). The median progression-free survival was 17.0 months in IRIS/Bev, as compared with 8.6 months in IRIS (hazard ratio for disease progression, 0.402; p < 0.001); the corresponding response rate were 63.5 percent and 52.5 percent (p = 0.393). In a multivariate analysis of PFS and OS, IRIS/Bev was significantly associated with longer PFS and OS compared with IRIS alone. Conclusions: In this retrospective comparison of two studies, the addition of Bev to IRIS appeared to improve outcome compared with IRIS alone in the first-line treatment of patients with mCRC.

Analysis of Kohne's prognostic index in KRAS wild-type patients with metastatic colorectal cancer (mCRC) treated with salvage-line cetuximab-based regimen: HGCSG0901

634

Background: In the treatment for mCRC, it is essential for understanding the prognosis of each individual patient. Köhne’s index (KI) based on performance status, white blood cell count, alkaline phosphatase and number of metastatic sites has been previously proposed. However, in the salvage setting, the validity of KI has not been reported in patients treated by cetuximab-based chemotherapy. Methods: 269 patients with mCRC treated by cetuximab contained chemotherapy were retrospectively registered from 27 centers in Japan. This analysis was included in the KRAS wild-type patients who were refractory to or intolerant for 5-FU/irinotecan/oxaliplatin and were never administered anti-EGFR-antibody. Univariate and multivariate analysis for overall survival were performed using patient characteristics. Survival analyses were performed with Kaplan-Meier method, log-rank test and Cox proportional hazards model. The analysis was also designed to determine whether the Köhne’s classification could be extended to other endpoints such as progression-free survival. Results: All data were available for prognostic categorization in 127 patients. Median overall and progression-free survival was 9.8 and 4.2 months. The distribution and median survival / progression-free survival for KI were as follows: low risk (L) (n = 40; 13.1/5.1 months), intermediate risk (I) (n = 17; 9.6/3.5 months), and high risk (H) (n = 70; 7.6/4.1 months). For overall survival, there was significant difference between L and H (p = 0.004), but not between L and I (p = 0.213), and between I and H (p = 0.321). For progression-free survival, there was tended to difference between L and H (p = 0.083), but not between L and I (p = 0.392), and between I and H (p = 0.630). In Cox multivariate analysis, KI showed an independent prognostic impact (HR 1.370, p = 0.010), but not predictive impact (HR 1.147, p = 0.212). Conclusions: In this analysis, KI might be a prognostic factor in salvage treatment with cetuximab-based regimen, but no effect predicted impact. Moreover, the prospective evaluation is needed for the further validation.

Evaluation of usefulness of Royal Marsden Hospital prognostic index in second-line chemotherapy of advanced gastric cancer

163

Background: Royal Marsden Hospital prognostic Index (RMH-I), which was based on performance status, ALP, liver metastasis and peritoneal metastasis, was reported as prognostic factor of advanced esophago-gastric cancer before first line chemotherapy (Chau I, et al. J Clin Oncol 22:2395-2403, 2004). Usefulness of RMH-I in second line chemotherapy is not elucidated. Methods: Advanced gastric cancer patients who started second line chemotherapy in Hokkaido University Hospital from July 2001 to May 2013 with prior fluoropyrimidine plus platinum administration were retrospectively analyzed. Univariate and multivariate analysis for overall survival were performed using patient characteristics (RMH-I, hemoglobin, CRP, CEA, Alb, TTP in first line, primary lesion resection, and bone metastasis). Survival analyses were performed with Kaplan-Meier method, log-rank test and Cox proportional hazards model. Results: There were 77 eligible patients. Male/Female were 52/25, median age was 60 years (range 31-80) and unresectable/recurrent were 70/7. Median survival was 7.1 months. The distribution and median survival for RMH-I groups were as follows: good risk (n = 8), 8.2 months; moderate risk (n = 57), 9.6 months; and poor risk (n = 12), 4.4 months. Although poor risk group showed shorter survival time, there were not significant difference regardless of RMH-I in Cox multivariate analysis (HR 1.12, 95%CI 0.61-2.09, P=0.72). Conclusions: In this retrospective analysis, RMH-I was not independent prognostic factor in second line chemotherapy of advanced gastric cancer. Prognostic factors in this population need to be investigated further.

Phase II trial of S-1 plus split cisplatin (SSP) in patients with advanced gastric cancer (HGCSG0702): Final report

113

Background: On the basis of SPIRITS trial, S-1 plus cisplatin has been regarded as standard first-line chemotherapy for patients with advanced gastric cancer (AGC) in Japan (Koizumi W, et al. Lancet Oncol, 2008). However, conventional S-1 plus cisplatin (60mg/m2) regimen requires hospitalization for hydration. Therefore, we conducted phase II trial of S-1 plus split Cisplatin (SSP) for outpatient chemotherapy. Methods: Eligibility criteria included pathologically confirmed AGC; no prior chemotherapy; Age 20 to 75, ECOG performance status (PS) of 0 to 1; adequate organ function; and written informed consent. S-1 (40 mg/m2) was given orally, twice daily for 21 days, and cisplatin (30 mg/m2) was given intravenously on day 1 and 15, followed by 2-week rest period, within a 5-week cycle. Primary endpoint was the response rate (RR), and secondary endpoints were progression-free survival, overall survival, safety profile, and non-hospitalized survival. Results: Between Mar 2008 and Mar 2012, 40 pts were enrolled. Patients characteristics were as follows: median age 63 years (range 41-75), Male: female 30:10, PS 0:1 33:7, diffuse: intestinal 23:17. Median number of cycles was 3. The main grade 3-4 AE were neutropenia (37.5%), anemia (30%), anorexia (30%) and fatigue (15%). These toxicities were safely managed. The median relative dose intensity of S-1 was 0.782, and cisplatin was 0.824. Response rate was 57.5% (95%CI 42.2-72.8%) and disease control rate was 90.0%. Median progression-free survival was 6.1 months (95%C.I. 3.0-9.1 months) and median survival time was 15.8 months (95%C.I. 12.7-18.8 months). Conclusions: SSP showed comparable tolerability and efficacy to SPIRITS trial. In addition, most patients underwent the treatment without hospitalization. SSP may be one of practical alternatives for AGC.

Exploratory randomized trial to evaluate the effect of indisetron tablets for preventing chemotherapy-induced nausea and vomiting (CINV)/acute-onset diarrhea induced by IRIS/FOLFIRI: HGCSG0704

624

Background: Indisetron is a 5-HT3 receptor antagonist that also antagonizes 5-HT4receptors. Indisetron tablets showed the non-inferiority to ondansetron tablets in terms of efficacy for preventing CINV. Moreover, preclinical data administered with irinotecan showed indisetron significantly reduced stool frequency in mice and inhibited the colonic peristalsis in dogs. We designed a pilot study compared with granisetron in the efficacy and tolerability of indisetron for irinotecan-induced nausea, vomiting and especially in diarrhea. Methods: Advanced colorectal cancer patients treated with FOLFIRI or IRIS (Irinotecan + S-1) with or without bevacizumab were enrolled in this study. Treatment: Arm A: indisetron tablets 8mg po day1. Arm B: granisetron 3mg iv day 1. The primary endpoints were the incidence of acute-onset diarrhea and complete protection from vomiting. Secondary endpoints were tolerability, complete protection from nausea and rate of no rescue medication. Results: Between May 2008 and July 2012, 33 patients (pts) were randomized. The study was closed prematurely due to poor accrual. Arm A: 16 pts, arm B 17 pts. Median age A: 68 y.o., B: 66 y.o.: ECOG PS 0/1: A: 12/4, B: 14/3pts. There was no significant difference of the incidence of acute-onset diarrhea between both groups (18.8% in A vs. 35.3% in B, p = 0.438). The proportion of complete protection from vomiting was 87.5% in A and 88.2% in B (p = 1.000). Similarly, complete protection from nausea, rate of no rescue medication, proportion of patients with a complete response (defined as no emetic episodes and no rescue medication) did not have a significant difference. Severity of nausea/vomiting and AE induced by 5-HT3 receptor antagonist were also similar between two groups. Conclusions: Compared with granisetron, indisetron showed effective and feasible results for preventing CINV induced by FOLFIRI or IRIS. Indisetron had also not improved the incidence of acute-onset diarrhea induced by irinotecan. Clinical trial information: UMIN000010162.

[New molecular targeting drugs for metastatic colorectal cancer]

I explain an outline of the molecular target of new drugs for colorectal cancer. First, regorafenib is an orally active, multikinase inhibitor. Second, as angiogenesis inhibitors, aflibercept and ramucirumab. Aflibercept is the second anti-angiogenic agent after bevacizumab with a FDA approval for metastatic colorectal cancer. Ramucirumab is a fully human mAb that binds human VEGFR-2, thus blocking VEGF binding and inhibiting angiogenesis. Finally, I introduce a new combination chemotherapy regimen with molecular targetting drugs.

Successful lung transplantation in a case with diffuse pulmonary arteriovenous malformations and hereditary hemorrhagic telangiectasia

Diffuse pulmonary arteriovenous malformations (AVMs) are associated with a poor prognosis and the therapeutic strategy remains controversial. We describe a pediatric patient with diffuse pulmonary AVMs associated with hereditary hemorrhagic telangiectasia (HHT), who presented with two cerebral AVMs in the parietal and occipital lobes as well. Of note, successful bilateral lung transplantation not only improved the hypoxemia but also resulted in size reduction of the cerebral AVMs. Although it is essential to consider involvements other than pulmonary AVMs, especially brain AVMs, to decide the indication, lung transplantation can be a viable therapeutic option for patients with diffuse pulmonary AVMs and HHT.

Randomized phase III clinical study comparing postoperative UFT/LV,UFT+LV/UFT and UFT+LV+PSK/UFT+PSK as adjuvant therapy for curatively resected stage III colorectal cancer HGCSG-CAD study

3638

Background: Study showed that the oral anticancer agent UFT/LV is useful as postoperative adjuvant chemotherapy for stage III colorectal cancer. PSK, a protein-bound polysaccharide extracted from the mycelia of Coriolus versicolor, is an immunomodulator widely used in gastric, colorectal and lung cancers. Methods: Patients aged 20-80 years with stage III colorectal cancer registered in 35 facilities were randomized to: group A (UFT/LV 28 days/5 weeks for 6 months); group B (UFT+LV 28 days/5 weeks for 6 months, then UFT for 12 months); and group C (UFT+LV+PSK 28 days/5 weeks for 6 months, then UFT+PSK for 12 months). Treatment was started within 6 months after curative resection. Outcome measures were relapse-free survival (RFS), overall survival (OS), incidence and severity of adverse events, and QOL. Results: Of 342 patients registered, 340 eligible patients were analyzed (84 in group A, 85 in group B, and 171 in group C). At baseline, variation in QOL score was observed but histopathological parameters were not different among 3 groups. Median observation period was 36 months. 3-year RFS was 73.8%, 77.6% and 73.9% in groups A, B, and C [A vs C: hazard ratio (HR) 0.960, 95% confidence interval (CI) 0.575-1.601; B vs C: HR 0.837, CI 0.488-1.433; A vs B: HR 1.151, CI 0.623-2.126]. 3-year OS was 95.2%, 91.8% and 89.9% in groups A, B, and C (A vs C: HR 0.460, CI 0.155-1.367; B vs C: HR 0.814, CI 0.338-1.963 A vs B: HR 0.570, CI 0.167-1.947). Adverse events ≥grade 3 included gastrointestinal symptoms and general status. There was no treatment-related death. Excluding high fatigue score in QOL scale that showed pretreatment variation, stratification analysis showed interaction between family score and group, and efficacy was suggested especially in group C with high score (3-yesr RFS: 66.7%, 68.2% and 88.1% in groups A, B, and C. A vs C: HR 3.289, CI 0.951-11.375, B vs C: HR 3.070, CI 0.973-9.685, A vs B: HR 1.084, CI 0.344-3.417). Conclusions: A significant difference in primary endpoint was not detected. Variation in QOL at treatment initiation probably greatly affected outcome. Clinical trial information: NCT00209742.