Citations

3576 Background: Identification of a marker with strong prognostic impact on clinical outcomes would aid in the management of pts with primary, curatively resected CRC. We conducted a pooled individual pt data analysis to confirm the prognostic value of pTIMP-1 for overall survival (OS) in stage II and III CRC. Methods: pTIMP-1 was determined in plasma using immunoassays. Samples and clinical data were included from 5 studies with a total of 1,042 stage II/III pts resected between 1991 – 2006 (Holten-Andersen (HA), 2000 [n=378], HA, 2004 [n=226], Waas, 2005 [52], Birgisson, 2010 [225], Nielsen, 2011[n=161]). To provide comparable measures across studies, pTIMP-1 values were rescaled using fractional ranks (FR, based on ranks and sample size). Pts were grouped by the median of the pTIMP-1 FR score. Stratified log-rank tests and multivariate Cox models assessed the association between OS and pTIMP-1 FR. Results: Overall, 59% were male; median age 69 years; 56% had stage II disease and 50% had colon cancer. The median follow-up was 5.5 years with 457 deaths. Continuous pTIMP-1 FR as a linear effect was significantly associated with OS (p < .0001, HR=1.38 for a 0.3 unit increase in FR score, 95% CI 1.25 to 1.52).Pts with pTIMP-1 FR above vs below the median had significantly shorter OS (HR, 1.72; 95% confidence interval[CI], 1.42 to 2.08, p <0.0001), which remained significant after adjusting for age, gender, stage, disease location, and serological CEA (HR, 1.31; 95%CI, 1.07 to 1.60; p = 0.009). Results by stage are shown in the table. No evidence of an interaction between pTIMP-1 and stage or disease location (colon vs. rectal) was observed (all p > 0.6). Conclusions: pTIMP-1 is a strong prognostic marker for predicting OS independent of age, gender, CEA, stage and site of disease in curatively resected stage II and III CRC. [Table: see text]

3526 Background: Prior observational studies have suggested that smoking is associated with poorer overall survival after colon cancer (CC) diagnosis. Using data from N0147, a phase III randomized adjuvant trial of patients with stage III CC, we assessed the relationship between smoking status and cancer outcomes (disease-free survival (DFS) and time to recurrence (TTR)), accounting for possible heterogeneity by patient and tumor characteristics. Methods: Patients completed a risk factor questionnaire at baseline prior to randomization to FOLFOX or FOLFOX+cetuximab (N=1968). Information was collected on smoking and other lifestyle factors, including alcohol consumption, body mass index (BMI), and physical activity. Multivariate Cox models assessed the association between smoking history and the primary trial outcome of DFS, as well TTR, adjusting for other lifestyle and clinical factors. The median follow-up was 3.5 years. Results: Overall, 52% of patients were former or current smokers. Compared to never smokers, ever smokers experienced significantly shorter DFS [3-year DFS: 70% vs 74%, hazard ratio (HR)=1.21, 95% confidence interval (CI): 1.02-1.42]. This association persisted after adjusting for age, sex, tumor subsite, number of nodes involved, T-stage, mismatch repair deficiency, BRAF mutation status, performance score, physical activity, BMI, and alcohol consumption (HR=1.23, 95% CI: 1.02-1.49). There was significant interaction in this association by BRAF mutation status (p=0.02): smoking was associated with shorter DFS in BRAF wildtype CC patients (HR=1.25, 95% CI: 1.04-1.51) but not in BRAF mutant CC patients (HR=0.72, 95% CI: 0.47-1.10). Patterns of association with smoking, overall and by BRAF status, were identical in analyses of TTR. Conclusions: Overall, smoking was significantly associated with shorter DFS and TTR in CC patients. These adverse relationships were most evident in patients with BRAF wildtype CC.

3514 Background: Activating mutations in the BRAF oncogene (mutBRAF) are associated with deficient MMR (dMMR) in sporadic colon cancers and are mutually exclusive with KRAS mutations (mutKRAS). We evaluated BRAF and KRAS in relation to MMR and disease-free survival (DFS) in stage III pts treated with adjuvant FOLFOX +/-cetuximab; study was amended mid-way to require prospective KRAS testing (Alberts SR, ASCO 2010). Methods: Extracted DNA from archival tumors was analyzed for BRAF exon 15 (V600E) and KRAS exon 2 mutations by allele specific RT-PCR. Tumors with loss of any MMR protein (MLH1, MSH2, MSH6) were categorized as dMMR vs proficient (pMMR). DFS was censored at 4 yrs; median follow-up was 3.1 yrs. Results: Among 2,686 pts, 12% (314/2580) of tumors were dMMR; 28% (716/2579) had mutKRAS, and 14% (346/2515) had mutBRAF of which 43% (150/346) were dMMR. mutBRAF was associated with high grade, proximal site, T3-4, >4 LNs, and dMMR (all p<0.002); mutKRAS with low grade, proximal, <3 LNs, and pMMR (all p<0.02). Among dMMR tumors, mutKRAS was more frequent in distal vs proximal cancers (28% vs. 8%, p=0.0004). Both mutKRAS (HR=1.42, p=0.0001) and mutBRAF (HR=1.34, p=0.02) were associated with worse DFS and remained significant after adjustment for treatment, MMR, grade, site, T stage and LNs. mutBRAF (p=0.003) and mutKRAS (p=0.003) were each prognostic for DFS in pMMR, but not dMMR tumors (KRAS, p=0.79; BRAF, p=0.60). While MMR was not significant univariately, proximal dMMR tumors had favorable DFS [HR 0.8, 95% CI (0.6-1)] and distal dMMR tumors had poor DFS [HR 2 (1.1-3.4)] after adjusting for KRAS, BRAF, grade, T stage and LNs (pinteraction=0.02). pMMR/mutBRAF (p=0.027) and pMMR/mutKRAS (p<0.001) had worse DFS vs pMMR/bothWT. Conclusions: mutBRAF and mutKRAS were significantly associated with worse DFS independent of treatment, covariates, and each other, with the DFS impact restricted to pMMR tumors. Prognostic impact of MMR was dependent upon tumor site after adjusting for covariates. Support: NIH Grant CA 25224, NCI K05CA 142885, Bristol-Myers Squibb, ImClone, Sanofi-Aventis, and Pfizer.

2004 Background: With the approval of novel agents for glioblastoma (GBM) treatment, reliable historical outcome data for the common phase II endpoint of progression free survival (PFS) is not available. In particular, the association between progression and overall survival (OS) has not been evaluated after temozolomide (TMZ) became the standard therapy for newly diagnosed GBM. Methods: Datasets from 5 TMZ-including NCCTG clinical trials in newly diagnosed GBM (n=325, 2005-2011, 1 phase I only and 4 phase I and single arm phase II trials) were pooled and analyzed to evaluate the individual level correlation between PFS and OS using correlation coefficient, landmark analyses, and time-dependent variable analyses. A historical control group was constructed by pooling datasets from 12 pre-TMZ era NCCTG clinical trials in the same patient population (n=1359, 1980-2004). Each of the 5 newer trials was compared to the control group to estimate the treatment effect (hazard ratio) on PFS and OS, which were analyzed to evaluate the trial level correlation between PFS and OS using correlation coefficient and weighted linear regression. Results: The estimated correlation coefficient between PFS and OS at individual patient level is 0.75 (95%CI, 0.7 to 0.8). The estimated hazard ratios of death in the landmark analysis at 4 and 6 months are 2.15 (95%CI, 1.47 to 3.15) and 2.17 (95%CI, 1.57 to 3.12), respectively, and the estimated hazard ratio of death is 10.1 (95%CI, 6.6 to 15.4) when progression is treated as a time-dependent variable. For trial level correlation (against the historical control group), the estimated correlation coefficient between HRos and HRpfs is 0.51 (95%CI, -0.67 to 0.96), with moderate prediction: the predicted HRos is very close to the observed HRos in 4 of 5 cases, however the predicted HRos is significantly different from 1 in only 2 out of 5 cases due to small sample size in other 3 cases. Conclusions: PFS and OS are highly associated at the individual patient level but only moderately at the trial level, the latter possibly due to small sample size in some of the trials. Additional validation in other trials could support use of PFS as a primary endpoint for randomized phase II trials in GBM.

6520 Background: We previously reported that alternative cutpoints and alternate categorical metrics to RECIST standards provided no meaningful improvement in OS prediction. Here, we sought to identify and validate continuous, longitudinal TM-based metrics for predicting OS. Methods: Data from 13 trials (1739 patients with sufficient imaging-based measurements from breast, lung or colon cancer) were randomly split (60:40) into training and validation datasets and landmarked at 12-weeks (w). Multiple candidate metrics that capture tumor curve trajectory were considered: first slope (or % change) from baseline to 6 w, last slope (or % change) from 6 w to 12 w, and indicator of increase (<5% (includes no increase or a decrease) vs. ≥5%) from the prior assessment (baseline to 6 w, 6 w to 12 w). Multivariable Cox models stratified by study and number of baseline target lesions (<3 vs. >3) and adjusted for baseline tumor burden (ratio of the sum of lesions/number of lesions) were fit for each tumor type. Predictive ability was assessed via concordance (c)-index (0.5: no association to 1.0: perfect association), and hazard ratios/p-values (HR/p). Results: Patients whose tumor size increased by at least 5% from the prior assessment, as well as whose tumors showed a positive slope (or % change) from 6 to 12 w (indicating tumor growth) had worse OS. The first slope (or % change) metric was less consistent, with an initial decrease in slope suggesting worse OS in some cases. Predictive ability of these metrics varied across tumor type with poorest prediction in Breast (Table). Conclusions: Continuous TM-based metrics performed no better than categorical RECIST metrics for OS prediction. [Table: see text]

3639 Background: The first phase of the multi-center prospectively specified retrospective study Validating Indicators To Associate Recurrence (VITAR), assessing the relationship between GCC gene expression in formalin fixed (FFPE) LNs and time to recurrence (TTR) in stage II CC pts not treated with adjuvant chemotherapy (Sargent, Annals Surg Onc 2011), showed promising initial results. Here we report a validation set of 463 new stage II CC pts. Methods: GCC mRNA was quantified by RT-qPCR using FFPE LNs from untreated T3N0 CC pts diagnosed from 1999-2008 with at least 12 LNs examined , blinded to clinical outcomes. Patients were classified by GCC LN ratio (LNR) (high risk: LNR > 0.1; low risk: LNR ≤ 0.1), with LNR defined as ratio of GCC positive to GCC informative LNs. Cox regression models tested the relationship between GCC and the primary endpoint of TTR, adjusted for age, tumor grade, number of LN examined pathologically, and lymphovascular invasion. Mismatch repair (MMR) status was also assessed. All primary analyses and cut-points were pre-specified. Results: 46pts (10%) recurred (rec), median follow-up was 65 months, median LNs examined was 20, and 42% (195/463) were classified high risk. Overall, TTR was not significantly associated with binary GCC LNR risk class (HR=1.47, p=.208) or DFS (HR= 1.39, p=.097). One site’s (n=97) tissue grossing method precluded appropriate LN assessment with existing GCC qualification methods. Excluding this site resulted in a TTR HR=1.91, p=0.051 (multivariate). In a post-hocanalysis excluding this site and using a 3-level GCC risk group of high (LNR > 0.20), intermediate (0.10 < LNR < 0.20) and low (LNR < 0.10), high risk group pts had a 5-yr rec risk of 22% versus 8% in low risk (HR 2.72, p=0.006). MMR status was not significantly associated with TTR (multivariate p=0.30). Conclusions: GCC status is a promising prognostic factor in appropriately staged stage II CC pts not treated with adjuvant therapy independent of traditional histopathology risk factors, but GCC determination must be performed with methodology adapted to the tissue procurement and fixation technique. Outcome associations were strengthened when considering a 3-level GCC categorization.

3636 Background: In the CORRECT phase III trial, the multikinase inhibitor REG demonstrated significant improvement in overall survival (OS) and progression-free survival vs placebo (Pla) in patients (pts) with mCRC whose disease progressed on other standard therapies. The most frequent REG-related grade ≥3 adverse events (AEs) of interest were hand–foot skin reaction (HFSR), fatigue, diarrhea, hypertension, and rash/desquamation. We explored whether the impact of REG in pts aged ≥65 years differed from that in younger patients. Methods: Pts with mCRC progressing following all other available therapies were randomized 2:1 to receive REG 160 mg once daily (n=505) or Pla (n=255) for the first 3 weeks of each 4-week cycle. The dose could be modified to manage AEs. The primary endpoint was OS. We report efficacy, safety, and dosing data from REG recipients by age. Results: The REG treatment group included 309 pts <65 years (307 evaluable for safety) and 196 pts ≥65 years (193 evaluable for safety). The OS hazard ratio (REG/Pla) was 0.72 (95% confidence interval [CI] 0.56–0.91) in pts <65 years and 0.86 (95% CI 0.61–1.19) in pts ≥65 years (interaction p-value = 0.405). Median OS was 6.7 vs 5 months for REG vs Pla in pts <65 years, and 6.0 vs 5.6 months, respectively, in pts ≥65 years. Most pts experienced drug-related AEs (<65 years: 93.8%; ≥65 years: 91.7%). The rates of grade ≥3 REG-related AEs of interest and dose modifications are shown in the Table. In pts <65 years vs ≥65 years, median (interquartile range [IQR]) duration of REG was 7.6 weeks (6.6–15.4) vs 7.1 weeks (5.1–17.2), median (IQR) daily REG dose was 160.0 mg (134.6–160.0) vs 160.0 mg (137.5–160.0), and median (IQR) proportion of planned REG dose was 83.3% (65.7–100.0) vs 78.6% (66.7–100.0), respectively. Conclusions: In the CORRECT trial, REG demonstrated an OS benefit in pts <65 years and ≥65 years. Safety and tolerability of REG appeared to be similar in both age subgroups. Clinical trial information: NCT01103323. [Table: see text]

3523 Background: The association of KRAS or BRAFV600E mutations with prognosis in colon cancers has been inconsistent. Since primary tumor site may influence outcome, we analyzed KRAS and BRAFV600E stratified by tumor site in resected stage III colon cancers from a randomized adjuvant trial of mFOLFOX6 + cetuximab chemotherapy (N= 2,686) where no survival difference by treatment was found. Methods: 2,580 tumors were analyzed for mutations in BRAFV600E (exon 15) or KRAS (codons 12, 13), and for deficient DNA mismatch repair (dMMR). Cox models were used, adjusting for mutation status, age, sex, treatment, T-stage, histologic grade, nodal status, tumor site, and MMR. After study initiation, eligibility was restricted to patients (pts) with KRAS wild-type (WT) tumors. At a median follow-up 4.1 yrs, 83% of pts are alive. Results: KRAS and BRAFV600E mutations were detected in 716 (28%) and 346 (14%) tumors, respectively; dMMR was found in 314 (12%). Proximal (to splenic flexure) tumors (50%) were associated with older age and were more likely to be high grade (33 vs 18%), T-stage3,4 (88 vs 82%), mutated for KRAS (33 vs 23%) or BRAFV600E (23 vs 4%), and dMMR (21 vs 3%)[all p <0.02]. Pts with distal vs proximal tumors showed improved disease-free survival (DFS)[ HR 0.7, 95% CI (0.6-0.9); p<0.01 unadjusted for KRAS or BRAF;]. Mutant KRAS [HR=1.5 (1.2-1.7); p<.0001] or mutant BRAF [HR=1.4 (1.1-1.7); p=0.02] were each independently associated with worse DFS. The KRAS and tumor site interaction was significant (unadjusted pinteraction=0.02; adjusted pinteraction=0.07), with poorer DFS among distal tumors having mutant KRAS [HR 1.7 (1.3-2.2); p<.0001]. No interaction was observed for BRAFV600E and tumor site, despite worse DFS for mutant BRAFV600E among proximal tumors [HR 1.3 (1-1.8); p=0.04]. Conclusions: The adverse prognostic impact of KRAS, but not BRAFV600E, mutations was confined to the distal colon. This finding underscores the importance of tumor site in the interpretation of the prognostic impact of KRAS status, and motivates similar analyses in pts with advanced disease.

3520 Background: EOTR has been suggested as a potential surrogate for overall survival (OS) in patients (pts) with mCRC and allows early assessment of treatment efficacy, facilitating adaptive trial design. We assessed at the individual patient level, the correlation between EOTR (complete or partial response) at 6, 8 and 12 weeks (wk), OS and progression free survival (PFS) in pts with mCRC treated with 1stline chemotherapy with or without a targeted agent as a first step in a surrogacy demonstration. Methods: IPD from 13,949 pts enrolled on 15 randomized Phase III trials in 1st line mCRC were used; 8 trials included targeted (anti-angiogenic and anti-EGFR) agents. EOTR prognostic value was assessed by landmark analyses using Cox models stratified by treatment assignment. P-values <0.01 were considered statistically significant to account for multiple comparisons. Results: Of 13,949 pts, 11,987 had sufficient response data to be included in the analysis. Median OS was 21.7 months (mo) in pts with an EOTR vs. 16.5 mo without EOTR at 6 wk (p<.0001, Hazard Ratio [HR] 0.64, 95% confidence interval [CI] 0.58-0.70, c statistic [c] 0.55). HRs were similar whether pts were treated with targeted therapies (p<.0001, HR 0.68, 95% CI 0.58-0.80, x 0.54) or non-targeted therapies (p<.0001, HR 0.61, 95% CI 0.55-0.69, x 0.55). Median PFS was 8.4 mo in pts with EOTR at 6 wk vs. 7.0 mo in pts without EOTR (p<.0001, HR 0.79, 95% CI 0.73-0.85). EOTR at 8 and 12 wks were also significantly associated with longer OS and PFS. The prognostic value of EOTR at 6, 8 and 12 wks remained significant (p<0.0001) after adjusting for age, gender, performance statusand location of metastatic disease (lung or liver). Overall tumor response (to 26 wk) however provided superior OS prediction (p<.0001, HR 0.51, 95% CI, 0.47-0.56, CS 0.61) vs. EOTR. Conclusions: Early response measured at 6, 8 or 12 wk after starting 1st line treatment was a strong and independent predictor of both OS and PFS in patient with mCRC and warrants further consideration as a potential endpoint for future trials, particularly randomized phase II trials.

3510 Background: Prediction of 5-FU-related adverse events (5FU-AEs) continues to be problematic. Pharmacogenetic studies on the rate-limiting enzyme in 5-FU metabolism, dihydropyrimidine dehydrogenase (DPD), suggest a link between three variants and both decreased enzyme activity and increased toxicity: c.1905+1 G>A (DPYD*2A; rs3918290), c.1679 T>G (I560S; DPYD*13; rs55886062), and c.2846A>T (D949V; rs67376798). Since the adverse impact of DPYD variants on 5-FU toxicity remains controversial, we determined associations between the three known DPYD variants and 5FU-AEs in stage III CC patients receiving FOLFOX or FOLFIRI (+ cetuximab) after curative resection. Methods: 2886 patients were genotyped by multiplexed single-base extension assays using the IPLEX Gold Kit and analyzed on the Sequenom MassARRAY system. Grade 3+ AEs were recorded per CTCAE v3. Fisher’s exact test, unequal variance two-sample t-test, and Wilcoxon rank sum test were used to compare categorical variables, continuous variables, and counts between patients with wild-type and mutant status. Logistic regressions were used to assess univariate and multivariate associations. Results: Patients displayed the following characteristics: male gender 53.2%, median age 58 [19-86], proficient DNA mismatch repair status 88.6%, PS-0 76.6%, + irinotecan 8.1%, and + cetuximab 45.9%. A total of 27 (0.9%), 4 (0.1%), and 32 (1.1%) patients carried the DPYD*2A, I560S, and D949V variants, respectively. Analysis identified significant associations between DPYD*2A and D949V variants and toxicity, with grade 3+ 5FU-AEs identified in 22 DPYD*2A carriers (OR=11.9, 95% CI 4.0-32.7, p<0.0001) and 22 D949V carriers (OR=5.5, 95% CI 2.5-12.1, p< 0.0001). No interaction effect was found between DPYD*2Aand D949V on grade 3+ 5FU-AEs (p = 0.98), nor on overall grade 3+ AEs (p = 0.97). No significant association was identified between I560S and grade 3+ 5FU-AEs. Conclusions: In the largest study to date, statistically significant associations were found between DPYD*2A and D949V variants and increased incidence of grade 3+ 5FU-AEs, suggesting utility in 5-FU toxicity prediction.