Grothey, Axel

3558 Background: Anti-programmed death-1 (PD-1) antibodies have been shown to be effective in the treatment of dMMR CRC. We describe an updated analysis from a cohort of 19 patients (pts) with dMMR CRC treated with pembrolizumab. Methods: Pts were identified through review of the Mayo Clinic electronic medical record (EMR) and chemotherapy administration records from May 2015 through January 2017. All pts with dMMR CRC who received treatment with pembrolizumab were included. The EMR was reviewed to identify demographic, clinical, pathologic and treatment details. Overall survival (OS), progression free survival (PFS) and disease control rate (DCR = CR+PR+SD) are reported. Time to event analysis was calculated using the Kaplan-Meier method. Results: Nineteen pts were included in this analysis; median age at diagnosis was 48.6 years (range 25-93); 53% were female. Most primary tumors were right sided (n = 12; 63%), 6 (32%) were left sided and 1 (5%) was a tumor of unknown primary within the bowel. Twelve (63%) pts received ≥2 lines of therapy prior to pembrolizumab (range 1-4). Five pts received 2 mg/kg every 3 weeks, 11 received 200 mg/day every 3 weeks and 2 pts received 10 mg/kg (1 every 2 weeks, the other every 3 weeks). The most common alterations identified were loss of MLH1 (11/15) and PMS2 (12/15); 4 pts had germline mutations (mut) identified. KRAS mut was identified in 6/16 pts and 2 pts had BRAF mut. Three pts had MLH1 hypermethylation. Median number of cycles of pembrolizumab was 8 (range 1-35+), with 13 (68%) pts receiving ≥6 cycles. DCR at first assessment was 68%, with 5% CR, 47% PR and 16% SD. Median follow-up from diagnosis was 29 months (95% CI 18-42). Median OS was 103 months (95% CI 85-103); 12-month OS was 89%. Median OS from PD-1 therapy was 16.1 months (95% CI 16-NR); 12-month OS from PD-1 therapy was 79%. Median PFS was NR (95% CI 5-NR); 12-month PFS was 54%. At time of analysis, 9 pts remain on PD-1 therapy; 5 pts have died; 3 have received subsequent therapy. Conclusions: Anti-PD1 blockade with pembrolizumab can provide long lasting benefit in dMMR mCRC, even in heavily pretreated pts.

3556 Background: To define and update survival rates and relapse patterns in patients (pts) with isolated advanced abdominal nodal metastasis secondary to colorectal cancer (CRC), treated with curative intent using aggressive trimodality therapy. Methods: Fifty-seven pts with isolated advanced abdominal lymph node metastasis (retroperitoneal and mesenteric) secondary to colorectal cancer received trimodality therapy defined as chemotherapy delivered in conjunction with external beam radiotherapy (EBRT) followed by lymphadenectomy and intraoperative radiotherapy (IORT). Infusional 5-FU was the most common radiosensitizer used (66%, 38 pts). The median dose of EBRT was 50 Gy & the median dose of intraoperative radiotherapy was 12.5 Gy. End points included distant metastasis, toxicities, local failure within EBRT field, recurrence within the intraoperative radiotherapy field, and survival. Results: 49% of pts were male, median age 50.5 yrs. All patients had ECOG ≤ 1. 27 pts had primary right sided colon cancer, 16 left sided colon cancer and 14 rectal primaries. Median time from initial CRC diagnosis to development of abdominal lymph node metastatic disease was 24 months (95% CI, 23.5-45.1 months). 84% (48 pts) had paraaortic nodal metastases, 12% (7 pts) had mesenteric nodal metastases, and 3% (2 pts) had both. With a median follow up of 89.4 months, the median overall survival and 5-year estimated survival rate were 53.2 months (95% CI, 46.4-78.8 months) and 42%, respectively. Median progression free survival was 19.3 months (95% CI, 15.6-32.8 months). 21 (37%) pts never developed distant disease. Outcome was not affected by disease sidedness, rectal primary, or mutational profile. Treatment was well tolerated without any grade 3/4 toxicities. Conclusions: The use of trimodality therapy including EBRT with radiosensitizing chemotherapy, lymphadenectomy and IORT produces sustainable long-term survival in selected metastatic CRC pts presenting with isolated retroperitoneal/mesenteric nodal relapse.

LBA1 Background: Since 2004, 6 months (m) of oxaliplatin (oxali)-based treatment has been the standard of care as adjuvant therapy for stage III CC. Since oxali is associated with cumulative neurotoxicity, shorter duration of adjuv therapy could spare pts toxicity and lead to substantial saving in health expenditures. Methods: A prospective, pre-planned pooled analysis of 6 concurrently conducted randomized phase III trials (SCOT, TOSCA, Alliance/SWOG 80702, IDEA France (GERCOR/PRODIGE), ACHIEVE, HORG) was performed to evaluate the non-inferiority (NI) of 3m compared with 6m (ref) of adjuv FOLFOX/XELOX. The primary endpoint was disease-free survival (DFS), defined as time from enrolment to relapse, 2nd CRC, and death (all causes). NI was to be declared if the 2-sided 95% confidence interval (CI) for DFS hazard ratio (HR 3m v 6m) estimated by a stratified Cox model was below 1.12. NI was examined within regimen and stage subgroups as pre-planned. Results: The analysis included 12,834 pts from 12 countries, accrued from 6/07 to 12/15. Stage distribution: 13% T1-2, 66% T3, 21% T4; 28% N2; 40% received XELOX. G3+ neurotoxicity was higher in the 6m v 3m arm (16 v 3% FOLFOX, 9 v 3% XELOX, p<0.0001). With a median follow-up of 39 mos, 3263 DFS events were observed. Overall, the 3 year DFS rate was 74.6% (3m) and 75.5% (6m), with estimated DFS HR of 1.07 (95%CI, 1.00-1.15). The 3m v 6m DFS HRs were 1.16 (95%CI, 1.06-1.26) and 0.95 (95% CI, 0.85-1.06) for FOLFOX and XELOX treated pts, respectively. The 3m v 6m DFS HRs were 1.01 (95%CI, 0.90-1.12) in T1-3 N1, and 1.12 (95%CI, 1.03-1.23) for T4 or N2 pts. Conclusions: While NI was not established for the overall cohort, NI of 3m v 6m oxali-based adjuv therapy was supported for XELOX. As each IDEA trial treated varying proportions of pts with XELOX (0 to 75%), the regimen interaction likely produced the differential outcomes observed between individual studies. Certain substages (T1-3 N1) also showed NI for 3m v 6m. These data provide a framework for discussions on risks and benefits of individualized adjuv therapy approaches. Support: U10CA180821, U10CA180882, U10CA180888, U10CA180835, INCA, PHRC2009, EME 09/800/34, HTA 14/140/84, CRUK C1348/A15960, JFMC. Clinical trial information: NCT01150045.

757 Background: Progression free survival (PFS) and overall survival (OS) appear to differ across trials, even when pts receive the same regimen (REG). This analysis aimed to benchmark the clinical outcomes in the 2nd line treatment of metastatic colorectal cancer (mCRC). Methods: Individual patient (pt) data was available on 6,462 pts with mCRC enrolled in 7 2ndline clinical trials (8 REGs) in the ARCAD database. Regimens used in at least two trials included FOLFIRI +/- an EGFR inhibitor (EGFRi), FOLFOX +/- a VEGF inhibitor (VEGFi), and irinotecan (IRI). Descriptive statistics were used to describe pt demographics. Multivariable Cox models were used to assess differences (DIFFs) in PFS and OS, while p-value <0.05 were considered statistically significant (SS). Results: 500-1400 pts received each REG (see Table). Median OS and PFS differed by 0.2 to 4.6 months (mo) and 0.8 to 2.5 mo across trials within the same REG. These DIFFs were SS except for PFS in REGs containing FOLFIRI +/- EGFRi. After multivariable adjustment, all DIFFs in OS were attenuated and non-significant. DIFFs in PFS attenuated for most REGs but remained SS for FOLFOX and IRI. Conclusions: After multivariable adjustment, there were no SS DIFFs in outcomes across trials within the same REG, with the exception of PFS in FOLFOX and IRI. The initial observed DIFFs in outcomes could be due to confounding factors and the remaining SS DIFFs in PFS in FOLFOX and IRI may be due to variation in definition of progression across trials. Therefore, caution is warranted when using historical data to design future trials.[Table: see text]

714 Background: Mismatch repair deficient (dMMR) colorectal cancer (CRC) has been shown to be responsive to blockade with anti-programmed death-1 (PD-1) antibodies. We describe a cohort of 17 patients with dMMR mCRC receiving treatment with pembrolizumab. Methods: Patients were identified through review of the Mayo Clinic electronic medical record (EMR) and chemotherapy administration records from May 2015 through September 2016. All patients with dMMR mCRC who received treatment with pembrolizumab were included. The EMR was reviewed to identify demographic, clinical, pathologic and treatment details. Overall survival (OS), progression free survival (PFS) and disease control rate (DCR=CR+PR+SD) are reported. Time to event analysis was calculated using the Kaplan-Meier method. Results: A total of 17 patients were included in this analysis; median age at diagnosis was 48.5 years (range 25-93); 59% were female. Most primary tumors were right sided (n=12; 71%), 4 (24%) were left sided and 1 (5%) was a tumor of unknown primary within the bowel. Thirteen (75%) patients received 2 or more lines of therapy prior to pembrolizumab (range 1-4). All patients received 2 mg/kg every 3 weeks, with the exception of two patients on a clinical trial; one who received 200 mg every 3 weeks, the other received 10 mg/kg every 2 weeks. The most common mutations (mut) identified were loss of MLH1 (9/12) and PMS2 (10/12), 4 patients had germline mut. KRAS mut was identified in 6/15 patients and only 1 patient had BRAF mut. Number of cycles of pembrolizumab ranged from 1 to 35, with 10 (59%) patients receiving 6 or more cycles. DCR at first assessment was 57%, with 7% CR and 50% PR. Median follow-up was 28 months (95% CI 18-55). Median OS was 103 months (95% CI NR-NR); 12-month OS was 88% and 24-month OS was 81%. Median OS post-PD-1 therapy was NR (95% CI 2-NR). Median PFS was 19 months (95% CI 5-19); 12-month PFS was 64% and 24-month PFS was 0%. At time of analysis, 11 patients remain on PD-1 therapy; 4 patients have died. Conclusions: Anti-PD1 blockade with pembrolizumab can provide long lasting benefit in dMMR mCRC, even in heavily pretreated patients.

713 Background: Exploratory analyses from RAISE (NCT01183780), FOLFIRI with ramucirumab or placebo in previously treated mCRC pts, suggested pts with treatment-emergent neutropenia may have improved overall survival (OS) compared to pts who did not experience neutropenia on study (Ciuleanu T, et al. abst. 2475, ESMO 2016). Subsequent analyses suggested an even stronger association between baseline ANC and OS, with high baseline ANC associated with shorter OS. A retrospective observational study was conducted to test this association as a pre-specified hypothesis. Methods: The IMS Oncology Database (electronic medical record data from US community practices) was used to identify pts ≥ 18 yrs with mCRC (ICD-9 code 153.x, 154.0 or 154.1) who initiated second-line fluoropyrimidine and irinotecan (+/- bevacizumab; bev) in 2007-2013, following progression during or after first-line fluoropyrimidine and oxaliplatin (+/- biologic). Eligible pts had ≥ 1 ANC in the 60 days prior to second-line therapy. Pts were stratified by second-line bev use and high/low baseline ANC ( ≥ or < 5.5x109/L) and matched by propensity scores. OS based on a validated proxy was analyzed using the Kaplan-Meier method and a Cox proportional hazards model, adjusted for age, gender, ECOG PS, stage, length of first-line treatment, and BMI. Results: 747 eligible pts were identified, of whom 617 were matched. Pts with baseline characteristic data had a median age of 59 yr, 57% were male, 67% were Caucasian, 74% had stage IV disease at diagnosis, and 63% had ECOG PS of 0/1 at start of second-line. Bev-treated pts with high baseline ANC (n = 122) had a median OS (mOS) of 7.7 mo vs 14.7 mo in pts with low baseline ANC (n = 323); adjusted HR 1.69 (95% CI, 1.35, 2.12; p < 0.0001). Pts with high baseline ANC not treated with bev (n = 59) had a mOS 5.4 mo vs 14.4 mo in pts with low baseline ANC (n = 113); adjusted HR 2.73 (95% CI, 1.89, 3.93; p < 0.0001). There was a modest interaction effect between bev treatment and baseline ANC (p = 0.012), indicating greater bev benefit in the high ANC group. Conclusions: Baseline ANC appears to be a strong prognostic factor and potentially a weak predictive factor for antiangiogenic therapy for OS in pts with previously treated mCRC.