Search

77 Background: Colorectal Cancer (CRC) is the second leading cause of cancer mortality worldwide although highly curable if detected early. An accurate liquid biopsy test for early cancer and precancerous lesion, detection is key to promote prevention and reduce mortality. We previously developed a liquid biopsy test based on immune cells response to the tumor, commercialised under the name of Colox, with best-in-class clinical performances for AA detection (52% sensitivity and 92% specificity) (Ciarloni L et al, Clin Cancer Res, 2016). An improved second generation of the test, leveraging transcriptome profiling and advanced machine learning tools, is under development and first results are presented. Methods: Prospective peripheral whole-blood (PAXgene) samples from subjects diagnosed with CRC, advanced adenoma (AA), non-advanced adenoma, other types of cancer as well as controls without colorectal neoplasia (CON) were divided into discovery and validation sets. Transcriptome profiles were generated by RNA-seq and gene expression signatures identified leveraging Novigenix’s proprietary LITOseek platform, which integrates several advanced Machine Learning (ML) methods into a ranking systems. Biological functional analysis was performed by over-representation (ORA), gene set enrichment (GSEA) and gene network analyses (STRINGdb). Results: Significant changes in the whole blood transcriptome profile of AA compared to CON were identified. Functional analysis highlighted upregulation of fatty acid derivative biosynthesis, interferon signaling, tryptophan catabolism to kynurenine, and downregulation of DNA repair in AA blood samples compared to CON. A novel gene classifier was generated, showing for AA detection 71% sensitivity, 94% specificity and an AUC of 74% by cross-validation on the discovery set. Validation of the gene classifier in an independent set is currently ongoing and will be presented at the congress. Conclusions: Capturing immune-related information using whole blood trascriptome profiling and applying cutting-edge machine learning technologies demonstrated to be valuable for identifying gene signatures for advanced adenoma detection. This differentiated solution has demonstrated best-in-class potential to significantly improve patient outcomes by detecting precancerous lesions with a simple non-invasive blood test.

321 Background: Cetuximab is an epidermal growth factor receptor monoclonal antibody indicated for the treatment of squamous cell carcinoma of the head and neck and RAS wild-type metastatic colorectal cancer. Hypomagnesemia is a reversible adverse event associated with cetuximab treatment. Here, we provide the key considerations in prevention and management of cetuximab-related hypomagnesemia events in clinical practice. Methods: A PubMed literature search was conducted for articles published between 18 March 2012 and 18 March 2022 using “cetuximab” and “hypomagnesemia” as keywords. Of the 75 articles retrieved, 20 were considered relevant and are included here.Articles were excluded if they focused on hypomagnesemia-associated clinical benefits. Data were summarized and grouped into the following categories: risk and risk factors; monitoring; prevention; and management. Results: The relative risk of grade 3/4 hypomagnesemia following cetuximab therapy appears to be approximately 1.5 times higher for colorectal cancer compared with head and neck cancer. Several risk factors have been identified including age, comorbidities, treatment duration, baseline electrolyte levels, concomitant platinum-based chemotherapy, and polypharmacy. Although the frequencies of hypomagnesemia events vary between the studies identified, data from the included meta-analyses suggest that the incidence of grade 3/4 hypomagnesemia is around 4% among patients receiving cetuximab therapy. Early and regular monitoring of magnesium levels is recommended ideally every 4 to 8 weeks, especially in patients at high risk of serious complications, and up to 8 weeks following treatment cessation. Proposed proactive prophylactic interventions include magnesium supplementation for patients with baseline magnesium levels < 1.8 mg/dL, to prevent severe hypomagnesemia. The management approach depends on the degree of hypomagnesemia, the symptoms and risk factors, and may include supplementation with oral magnesium salts and/or parenteral forms; intravenous forms may be preferred due to low bioavailability and gastrointestinal adverse events related to oral supplements. While recommendations for initiating magnesium replacement in grade 1 (> 1.2 mg/dL) and grade 2 (0.9–1.2 mg/dL) hypomagnesemia events vary, magnesium supplementation should be given to patients with grade 3 (0.7–0.9 mg/dL) or grade 4 (< 0.7 mg/dL) hypomagnesemia, due to high risk of serious adverse cardiac events. If magnesium levels do not improve or continue to decrease despite supplementation, a pause in cetuximab administration may be considered. Conclusions: Hypomagnesemia is a reversible and manageable cetuximab-related adverse event and should not interfere with treatment outcomes. A periodic check of magnesium levels during treatment with cetuximab is recommended.

TPS157 Background: Patritumab deruxtecan (HER3-DXd; U3-1402) is a novel, investigational antibody drug conjugate comprising an anti-HER3 monoclonal antibody, a tetrapeptide-based linker, and a topoisomerase I inhibitor payload. Ongoing clinical trials of HER3-DXd in pts with metastatic breast cancer or non-small cell lung cancer have shown promising clinical activity and acceptable safety. HER3 (human epidermal growth receptor 3), a member of the tyrosine kinase receptor family, is overexpressed in most CRC tumors and associated with an adverse prognosis. Significant tumor regression with HER3-DXd has been observed in CRC murine xenograft models, regardless of KRAS mutation status. Here we introduce the design of a phase 2 study (U31402-A-U202) that is evaluating HER3-DXd in previously treated pts with advanced/metastatic CRC. Methods: U31402-A-U202 (NCT04479436) is an open-label, multicenter phase 2 study that will enroll 80 pts in the USA, Europe and Asia. Pts are enrolled who are aged ≥ 18 years with advanced/metastatic colorectal adenocarcinoma that is resistant/refractory/intolerant to ≥ 2 prior lines of therapy including a fluoropyrimidine, irinotecan, a platinum agent, an anti-EGFR agent (if clinically indicated), an anti-VEGF agent (unless contraindicated [CI]), and an immune checkpoint inhibitor (unless CI) for microsatellite instability-high CRC. Pts with current/previous interstitial lung disease or clinically severe pulmonary compromise are excluded. Archival tumor biopsy and pre-treatment tumor biopsy are collected from all pts at screening, with HER3 protein expression measured by immunohistochemistry (IHC). In part 1, results of the HER3 IHC assay from the pre-treatment tumor biopsy are used to assign pts into 1 of 2 cohorts (C). C1: HER3 high (IHC 3+, 2+), n = 24; C2: HER3 low/negative (IHC 1+, 0), n = 12. Pts receive 5.6 mg/kg HER3-DXd IV every 3 weeks. An interim futility analysis will be conducted separately for C1 and C2 and will determine enrollment in part 2, with 2 potential scenarios: enrollment continues irrespective of HER3 IHC status, or enrollment continues in HER3 high pts only. The primary objective is the evaluation of the antitumor activity of HER3-DXd as measured by objective response rate (ORR) (assessed by BICR according to RECIST v1.1). ORR will be summarized with the 2-sided 95% confidence interval. Secondary objectives include the evaluation of efficacy as measured by ORR (assessed by investigator according to RECIST v1.1), duration of response, time to tumor response, disease control rate, progression-free survival (assessed by investigator and BICR according to RECIST v1.1), overall survival, safety and tolerability, HER3 protein expression in tumor tissue and relationship with efficacy, and pharmacokinetic properties. Clinical trial information: NCT04479436.

TPS266 Background: Patients (pts) with MSI-H/dMMR mCRC treated with CT have poorer outcomes than pts with microsatellite stable/MMR proficient mCRC. NIVO (anti–programmed death [PD]-1) and IPI (anti–cytotoxic T lymphocyte antigen-4 [CTLA-4]) are immune checkpoint inhibitors that act synergistically and promote antitumor immune response by complementary mechanisms. NIVO±IPI received accelerated US FDA approval for MSI-H/dMMR mCRC that progressed after fluoropyrimidine, oxaliplatin, and irinotecan treatment based on the phase 2, non-randomized, multicohort CheckMate 142 study. Indirect comparisons suggest that NIVO (3 mg/kg) + low-dose IPI (1 mg/kg) provides improved clinical benefit vs NIVO (investigator-assessed [INV] objective response rate [ORR] 55% vs 31%; 12-month [mo] INV progression-free survival [PFS] rate 71% vs 50%; 12-mo overall survival [OS] rate 85% vs 73%) with a favorable benefit-risk profile for previously treated MSI-H/dMMR mCRC (Overman et al. JCO 2018). NIVO+low-dose IPI also demonstrated robust and durable clinical benefit in first-line MSI-H/dMMR mCRC (INV ORR 64%; 12-mo INV PFS rate 77%; 12-mo OS rate 84%; Lenz et al. ASCO 2019, #3521). To date, no prospective phase 3 studies have reported results for anti–PD-1, anti–PD-1 + anti–CTLA-4, or CT in MSI-H/dMMR mCRC; these treatments will be evaluated in the international, multicenter, open-label, randomized, phase 3 CheckMate 8HW (NCT04008030) study. Methods: Approximately 494 pts aged ≥18 years with histologically confirmed recurrent or mCRC, irrespective of prior treatment with CT and/or targeted agents, not amenable to surgery, and with known tumor MSI-H or dMMR status, and Eastern Cooperative Oncology Group performance status ≤1 will be randomized to receive NIVO, NIVO+IPI, or investigator’s choice CT (pts in the CT arm can receive NIVO+IPI upon progression). The primary endpoint is PFS, assessed by blinded independent central review (BICR). Secondary endpoints include PFS by INV, ORR and disease control rate by BICR and INV, OS, time to and duration of response. Exploratory endpoints include safety. Clinical trial information: NCT04008030.

247 Background: This analysis was undertaken to forecast survival and enhance treatment decisions for patients (pts) with colorectal cancer (CRC) with liver metastases sensitive to folinic acid, fluorouracil and irinotecan (FOLFIRI) alone [F] or in combination with cetuximab [FC] using simple quantitative radiomic changes between CT scans at baseline and 8 weeks. Methods: We retrospectively analyzed 667 pts with KRAS-unselected metastatic CRC in NCT00154102 treated with F and FC. CT quality was classified as high (HQ) or standard (SQ), and four data sets were created and named by treatment quality. Pts were randomly assigned 1:2 to training or validation sets: FCHQ, 38/78 pts; FCSQ, 62/124 pts; FHQ, 51/78 pts; FSQ, 78/158 pts. A machine-learning signature was trained using data set FCHQ to classify pts as treatment-sensitive or treatment-insensitive using just 4 of 3,499 potential radiomic imaging features. Performance was calibrated/validated using ROC curves. Hazard ratios (HRs) and Cox regression models were used to evaluate association with overall survival (OS). Results: The signature used decrease in tumor heterogeneity plus boundary infiltration to successfully predict sensitivity to FC (FCHQ: AUC, 0.80; FCSQ: AUC, 0.72) but failed with non-cetuximab regimens (FHQ: AUC, 0.59; FSQ: AUC, 0.55). The radiomic signature outperformed existing biomarkers ( KRAS mutational status and tumor shrinkage by RECIST 1.1) for sensitivity to cetuximab-based therapy and was strongly associated with OS in the cetuximab-containing sets FCHQ (HR, 44.3; p = 0.0001) and FCSQ (HR, 6.5; p = 0.005). Conclusions: This signature, derived from simple radiomic analysis of tumor imaging phenotype using only standard-of-care CT scans, appeared to be treatment-specific and was superior to all tested prognostic biomarkers. The signature provided early prediction of sensitivity and survival and could be used to guide treatment continuation decisions.

3601 Background: High stromal content (CMS4) and immune infiltration (CMS1) are important features of the tumor micro-environment (TME) in early stage colorectal cancer, with poor and good RFS prognosis, respectively. However, it remains unclear how these TME features are influenced by intrinsic tumor epithelial expression profiles. We examined how the former can be predicted by the latter, and assessed their respective prognostic contributions. Methods: Eighteen public gene-expression datasets (n=3511) were used for discovery, and an adjuvant trial dataset (PETACC3 n=688) was used for validation. Whole-transcriptome hierarchical clustering was applied to identify intrinsic versus extrinsic co-expression modules, corroborated by single-cell profiling data. The epithelial gene subset were used to predict previously assigned CMS labels, producing new CRC classifier that we refer to as epithelial-CMS (eCMS). Results: Despite lacking stromal and immune response genes, eCMS predicted CMS with an accuracy of 82% [CI 78-85%] in the test dataset. Not only were the RFS prognostic values retained (eCMS1 vs non-eCMS1: HR=0.59 [CI 0.35–0.95, p=0.029], and eCMS4 vs non-e-CMS4: HR=1.82 [CI 1.38–2.41, p=0.0003]), but they were also significantly improved when compared to the original CMS classification (CMS1 vs non-CMS1: HR=0.76 [CI 0.49–1.16, p=0.2] and CMS4 vs non-CMS4: HR=1.43 [CI 1.07–1.91, p=0.017]. In a multivariable regression comparison, eCMS was superior to CMS (p=0.001 vs p=0.36). Each eCMS group was associated with specific combinations of up- and down-regulation of CRC-relevant pathways, such as CDX2, map-kinase, mucin secretion, DNA damage repair and cell proliferation. Conclusions: Stromal and immune infiltration patterns in the TME are strongly associated with tumor epithelial expression profiles, as captured by eCMS. This finding sheds new light on the interplay between intrinsic molecular features of CRCs and their TME, and illuminates new paths for potential combination therapies.

Purpose We investigated in the first-line setting our previous finding that patients with chemorefractory KRAS G13D–mutated metastatic colorectal cancer (mCRC) benefit from cetuximab treatment. Methods Associations between tumor KRAS mutation status (wild-type, G13D, G12V, or other mutations) and progression-free survival (PFS), survival, and response were investigated in pooled data from 1,378 evaluable patients from the CRYSTAL and OPUS studies. Multivariate analysis correcting for differences in baseline prognostic factors was performed. Results Of 533 patients (39%) with KRAS-mutant tumors, 83 (16%) had G13D, 125 (23%) had G12V, and 325 (61%) had other mutations. Significant variations in treatment effects were found for tumor response (P = .005) and PFS (P = .046) in patients with G13D-mutant tumors versus all other mutations (including G12V). Within KRAS mutation subgroups, cetuximab plus chemotherapy versus chemotherapy alone significantly improved PFS (median, 7.4 v 6.0 months; hazard ratio [HR], 0.47; P = .039) and tumor response (40.5% v 22.0%; odds ratio, 3.38; P = .042) but not survival (median, 15.4 v 14.7 months; HR, 0.89; P = .68) in patients with G13D-mutant tumors. Patients with G12V and other mutations did not benefit from this treatment combination. Patients with KRAS G13D–mutated tumors receiving chemotherapy alone experienced worse outcomes (response, 22.0% v 43.2%; odds ratio, 0.40; P = .032) than those with other mutations. Effects were similar in the separate CRYSTAL and OPUS studies. Conclusion The addition of cetuximab to first-line chemotherapy seems to benefit patients with KRAS G13D–mutant tumors. Relative treatment effects were similar to those in patients with KRAS wild-type tumors but with lower absolute values.

Purpose Our purpose was development and assessment of a BRAF-mutant gene expression signature for colon cancer (CC) and the study of its prognostic implications. Materials and Methods A set of 668 stage II and III CC samples from the PETACC-3 (Pan-European Trails in Alimentary Tract Cancers) clinical trial were used to assess differential gene expression between c.1799T>A (p.V600E) BRAF mutant and non-BRAF, non-KRAS mutant cancers (double wild type) and to construct a gene expression–based classifier for detecting BRAF mutant samples with high sensitivity. The classifier was validated in independent data sets, and survival rates were compared between classifier positive and negative tumors. Results A 64 gene-based classifier was developed with 96% sensitivity and 86% specificity for detecting BRAF mutant tumors in PETACC-3 and independent samples. A subpopulation of BRAF wild-type patients (30% of KRAS mutants, 13% of double wild type) showed a gene expression pattern and had poor overall survival and survival after relapse, similar to those observed in BRAF-mutant patients. Thus they form a distinct prognostic subgroup within their mutation class. Conclusion A characteristic pattern of gene expression is associated with and accurately predicts BRAF mutation status and, in addition, identifies a population of BRAF mutated-like KRAS mutants and double wild-type patients with similarly poor prognosis. This suggests a common biology between these tumors and provides a novel classification tool for cancers, adding prognostic and biologic information that is not captured by the mutation status alone. These results may guide therapeutic strategies for this patient segment and may help in population stratification for clinical trials.

Purpose To study the power of the epidermal growth factor receptor (EGFR) epiregulin (EREG) and amphiregulin (AREG) ligands' expression in primary tumors to predict the outcome in patients with chemorefractory metastatic colorectal cancer (cmCRC) treated with the combination of cetuximab and irinotecan. Patients and Methods Gene expression measurements and KRAS mutation analysis were performed on archival formalin-fixed paraffin-embedded primary tumors of 220 cmCRC patients. Response was measured using RECIST (Response Evaluation Criteria in Solid Tumors) criteria. The relation between ligand expression levels and outcome was evaluated using logistic regression for response and Cox regression for survival data. Receiver operating characteristics analysis was performed for response and survival data. CIs for the performance indices were obtained with a nonparametric bootstrap procedure. Findings were externally validated on a series of 67 samples treated in a similar setting. Results In KRAS wild type (WT) patients, there was a significant association between log-transformed ligand expression and response for EREG (odds ratio for objective response, 1.90; 95% CI, 1.27 to 2.83; P = .0005; concordance index [c-index], 0.681) and for AREG (odds ratio for objective response, 1.862; 95% CI, 1.22 to 2.72; P = .0017; c-index, 0.673). In a Cox regression model, dichotomized ligand expression was significantly associated with progression-free survival (PFS) and overall survival (OS). EREG PFS hazard ratio (HR) was 0.41 (95% CI, 0.274 to 0.609; P < .001; time-dependent c-index [Cτ index], 0.640), and AREG PFS HR was 0.43 (95% CI, 0.29 to 0.64; P < .001; Cτ index, 0.627). EREG OS HR was 0.42 (95% CI, 0.28 to 0.63; P < .0001; Cτ index, 0.639), and AREG OS HR was 0.40 (95% CI, 0.27 to 0.64; P < .0001; Cτ index, 0.625). There was no predictive power of ligand expression in patients with KRAS mutation. Conclusion Expression of EGFR ligands in primary tumors significantly predicts outcome in KRAS WT cmCRC treated with cetuximab and irinotecan.

3572 Background: Cetuximab (cet) increases survival in RAS wild-type (WT) metastatic colorectal cancer (mCRC) in first-line and chemorefractory patients (pts). Adding irinotecan (iri) to cet in refractory pts who had progressed on iri increased response and delayed progression in the BOND trial, which was conducted prior to recognition that RAS mutations are predictive of EGFR-inhibitor (EGFR-I) resistance. Subsequent trials in chemorefractory pts used single agent EGFR-I as standard. In the era of biomarker selection, the benefit of continuing iri versus single agent EGFR-I had not been resolved. Methods: ICECREAM is a randomised phase II trial evaluating cet v cet + iri in chemotherapy-refractory mCRC, stratified for KRAS G13D mutation (previously reported) or no mutations in KRAS, NRAS, BRAF and PI3KCA (Quad WT ). EGFR-I naïve, ECOG PS 0-1 pts, progressing within 6 months of iri and refractory (or intolerant) to fluoropyrimidine and oxaliplatin were randomised to cet 400mg/m2 IV loading then 250mg/m2 weekly +/- iri 180mg/m2q2 wks. The primary endpoint was progression-free survival at 6 months (6m PFS); secondary endpoints were response rate (RR), overall survival (OS), toxicity and quality of life (QOL). Results: From Nov 2012 to June 2016, 48 Quad WT pts were recruited: 2 ineligible (not iri-refractory, BRAF mutation) wre not included for analyses and a further 2 not evaluable for response. Characteristics were balanced between cet (n = 21) v cet-iri (n = 25), except for sex (male 62 v 72%) and primary site (left 95 v 68%). 6m PFS rate was 14% v 41%; HR 0.39 (95% CI: 0.20–0.78, p = 0.008). RR was 10% (2 PR) v 36% (1 CR, 8 PR); p = 0.04. Toxicities were higher with cet-iri; at least one grade 3/4 adverse event in 50 v 23%. No differences in global or specific QOL were seen. Conclusions: The AGITG ICECREAM trial confirms a significant synergy for the addition of iri to cet and improved PFS in Quad WT chemorefractory mCRC, echoing data in molecularly unselected pts. Clinical trial information: ACTRN12612000901808.

TPS3633 Background: For patients with initially unresectable colorectal liver metastases (CRLM) with good clinical response to chemotherapy, the presence of disappearing liver metastases (DLMs) diagnosed by CT is a major independent prognostic factor. DW-MRI as well as contrast enhanced (CE)-MRI is recommended to detect and characterize CRLM. However, the correlation between radiological and pathological complete response has not been fully investigated using these latest imaging and pathology techniques. Our main aim is to demonstrate the added value of DW-MRI, CE-MRI to that of CT alone to provide precise assessment of the viability of DLMs. In addition, we aim to optimize the therapeutic management of CRLM patients. No prospective study has been conducted to determine the predictive value of DW-MRI combined with CE-MRI in confirming sites of DLMs and assessing their true status. Methods: This is the first collaborative study between EORTC, ESSO and JCOG with an integrated quality assurance program for imaging, surgery and pathology. Patients with unresectable CRLM will receive standard systemic chemotherapy and liver resection if resectable. Both CT and MRI (DW-MRI, CE-MRI and T1/T2) will be used to identify confirmed DLMs (cDLMs). cDLMs will be either resected or, if resection is not possible, followed-up without resection until 2 years after surgery to evaluate the true status of the cDLMs. The primary endpoint is negative predictive value (NPV) of DW-MRI, CE-MRI, T1/T2 and CT in confirming the status of cDLMs using as reference either the histopathological complete response or the absence of a local recurrence at the site of cDLMs during the follow up period of 2 years. The study aims at excluding a NPV≤0.85 and is powered under the alternative that the NPV≥0.95. The planned sample size is 92 evaluable (resected or left behind) cDLMs, with a 1-sided alpha of 5% and a power of 90% adjusting for within-patient correlation between cDLMs of 0.2 and an average number of 2 cDLMs per patient. Approximately 400 patients will be registered from European, Japanese and US sites over 3 years. As of February 2017, 2 patients have been enrolled. Clinical trial information: NCT02781935.

3510 Background: FIRE-3 compared 1st-line therapy with FOLFIRI plus either cetuximab or bevacizumab in 592 KRAS exon 2 wt mCRC patients. CMS is grouping CRCs according to their gene-signature in 4 different types. Relevance of CMS for the treatment of mCRC remains unclear. Methods: Patients were grouped according to tumor CRC-CMSs. Using ALMAC’s Xcel tissue array, gene signatures of FIRE-3 tumor samples were analyzed. Survival was compared using Kaplan-Meier estimation and log-rank tests. Hazard ratios (HR) were estimated according to the Cox proportional hazard method. Results: CMS classification could be determined in 385 specimens available from the ITT population (n = 592). In this KRAS exon 2 wt population (n = 385), frequencies were: CMS1 (10.4%), CMS2 (36.6%), CMS3 (11.7%), CMS4 (29.1%), non-consensus (12.2%). In RAS wt (n = 315), frequencies were: CMS1 (11.1%), CMS2 (38.1%), CMS3 (9.5%), CMS4 (29.5%), non-consensus (11.7%). Independent of the treatment, CMS was a strong prognosticator for ORR (p = 0.023), PFS (p < 0.001) and OS (p < 0.001). For data on CMS and treatment efficacy in the RAS wt population see the following table. Conclusions: CMS classification is prognostic for mCRC. The survival benefit in RAS wt previously observed for FOLFIRI cetuximab vs. FOLFIRI bevacizumab is not significantly different across CMS groups, although there are trends when comparing OS HR between categories with CMS4 showing the best HR. [Table: see text]

3538 Background: Addition of (ziv)-aflibercept (A) to FOLFIRI in second-line therapy for metastatic colorectal cancer (CRC) has been shown to be beneficial in phase III VELOUR trial (NCT00561470). A follow-up study (NCT01754272) was undertaken to acquire tumor samples for biomarker analyses and identify subgroups of patients with differential treatment effects. The primary results assessing efficacy according to well-established CRC subgroups defined by RAS, BRAF status and sidedness are reported here. Methods: Tissue specimens were collected for 666 patients from 1226 ITT pts. Suitable specimens were assayed for somatic mutation using NGS targeting extended RAS and BRAF genes. NGS assays with no missing values were obtained for 482 pts. Affymetrix gene chip technology was used for whole-transcriptome profiling; sidedness was extracted from available pathological reports. Differences between subgroups were assessed by interaction analysis. Results: The treatment effects on overall survival (OS) for the 482 pts is still significant HR=0.80 (CI 0.65-0.99), and similar to the ITT (n=1226) results (HR=0.82, CI 0.71-0.93). Two established ways of defining mutations (traditional KRAS exon 2 and extended RAS using NGS) show a trend for a differential effect across mutation groups.(see table for OS). Interestingly, BRAF mutants (which are all RAS wild type) show a trend for better outcome Same is seen for PFS and RR. Sidedness did not affect efficacy (HR: 0.83 (0.63- 1.1) for left and HR: 0.83 (0.54-1.3) for right. Conclusions: None of the mutations subgroup results shows significant interaction, although the ratios of treatment HR favor RAS wild types. Similar trends were observed in published trials with bevacizumab or ramucirumab. Sanofi supported this ISS. Clinical trial information: NCT01754272. [Table: see text]

611 Background: Intratumour heterogeneity is a key challenge in colorectal cancer management. We have recently shown high levels of morphologic heterogeneity in microsatellite unstable (MSI) colorectal tumours (CRC). In this study we verified whether morphologic intratumour heterogeneity can be used as a guidance to test for molecular heterogeneity. Methods: 15 MSI CRCs tumour, characterized by a heterogenous morphology, were selected together with the associated metastasis. Macrodissection of the distinct morphologic tumour components was followed by MSI testing according to the Bethesda guidelines and Sanger sequencing for BRAF V600E mutations. In addition, 2 MSI CRC underwent macrodissection of the distinct morphologic components, followed by nanostring ncounter analysis. Results: MSI status varies between MSI-H and MSI-L in the primary tumours and 35% of the metastasis of MSI-H primary tumours presented as MSI-L or MSS. BRAF mutations status was constant throughout the primary tumours, but two BRAF mutated tumours presented with a BRAF wild type metastasis. Nanostring analysis revealed that ùorphologically identical clones clustered separately from tumour clones with a different morphologic display. Conclusions: Morphologically heterogeneous tumour compartments have less related gene expression profiles, linking morphologic homogeneity to molecular homogeneity. Our findings support the hypothesis that morphologic heterogeneity reflects molecular intratumour heterogeneity and can be used as a guidance for molecular testing of colorectal cancers.

589 Background: ALK, ROS1, and NTRKs fusions occur in 0.2-2.4% of CRCs, and represent therapeutic targets, as indicated by promising reports of individual patients treated with tyrosine kinase inhibitors. Clinical and molecular features of mCRCs harboring these fusions have not been elucidated. Methods: mCRC harboring ALK, ROS1 and NTRKs fusions were identified taking advantage of screening programs worldwide (Ignyta STARTRK trials; Foundation Medicine Database; Memorial Sloan Kettering-IMPACT; Australian MAX trial post-hoc analysis). Clinicopathological and molecular characteristics of ALK, ROS1, NTRKs rearranged cases were compared with those of non-rearranged cases, included in Ignyta’s STARTRK trials program. Results: 27 ALK, ROS1 or NTRKs rearranged (including a novel SCYL3-NTRK1 fusion) and 319 not rearranged mCRCs patients were included. Rearrangements were more frequent in older patients (p = 0.024), and tumors that were right-sided (80% vs 30%; p < 0.001), RAS wild-type (93% vs 52%; p < 0.001), MSI-high (48% vs 8%; p < 0.001) and spread more frequently to lymph nodes (46% vs 25%; p = 0.03) and less frequently to the liver (42% vs 66%; p = 0.026). At a median follow-up of 28.5 months, patients bearing rearranged tumors had a shorter overall survival (OS) when compared to non-rearranged (15.6 vs 33.7 mos; HR: 2.17, 95% CI 1.03-4.57; p < 0.001). In the multivariable model including significant prognostic variables (primary site, primary resection, BRAF and MMR status), rearrangements were still associated with shorter OS (HR: 2.78, 95% CI 1.27-6.07; p = 0.011). All of the 4 patients with rearranged tumors evaluable for benefit from anti-EGFRs experienced disease progression as best response. One patient with an MSI-high EML4-ALK rearranged tumor experienced durable response to PD-1 blockade. Conclusions: ALK, ROS1 and NTRKs rearrangements define a new molecular subtype of mCRCs associated with unfavorable prognosis, and specific clinicopathological and molecular features. Since sensitivity to available treatment options, including anti-EGFRs, is very limited, targeted approaches with or without immunotherapy should be investigated in these patients.

3519Background: AJCC v.7 staging and clpath covariates do not accurately predict outcome of CC patients (pts) eligible for adjuvant chemotherapy (adjCh). Mol markers, microsatellite instability (MS...

3035Background: We previously described a molecular subtype of tumors that presents an innate immune response to intrinsic DNA damage. The associated gene expression pattern is mediated by the STIN...

575 Background: Molecular diagnostics can identify subgroups of colorectal and other cancers that are relevant for the mode of action of new anti-cancer agents. The efficient, GCP-conform and quality assured molecular screening to identify potential study patients is one of the major challenges for targeted drug development. The EORTC has implemented a new collaborative molecular screening platform in order to facilitate these next generation trials. Methods: SPECTAcolor (Screening Patients for Efficient Clinical Trial Access in advanced colorectal cancer) is a pan European network of institutions collaborating in centralized, high-throughput screening of tumor material from patients with colorectal cancer. After informed written consent, patients are screened for molecular alterations. Molecularly annotated patients can be offered so called “downstream” clinical trials. Results: Launched in September 2013, SPECTAcolor is now initiated in 19 clinical centers in 9 countries. As of 12 September 2014, 406 patients were enrolled exceeding the expected accrual target of 300 patients in the first year. Of these 406 patients, 293 patients had their tumor block shipped to the central biobank at the Dresden University Hospital for central quality and pathological review, and core analyses. Tumor samples are being additionally analyzed using next generation sequencing for 360 key cancer alterations in cooperation with the Sanger Institute (Cambridge, UK). In the first year, the blocks were analyzed for 5 baseline biomarkers. KRAS was wild type for exon 2, 3 and 4 in 151/284 pts (53%) and mutated 133 pts (47%; 114 pts in exon 2 (40%), 8 pts in exon 3, and 11 pts in exon 4). NRAS was tested in KRAS wild type pts only; mutations were found in 14 pts (4.9%; 6 pts in exon 2 and 8 pts in exon 3). BRAF mutations, all in exon 15, were found in 18 pts (7%). PI3K mutations occurred in 41 pts (15%; 13 in exon 20 and 28 in exon 9). IHC staining was showing deficient mismatch repair in 16 patients. Conclusions: SPECTAcolor is the first pan-European and EORTC screening platform. Its successful implementation proves that a logistically complex infrastructure to run next generation trials in a multinational setting is feasible.

3515 Background: BRAF V600 mutations occur in 5–10% of CRC and confer poor prognosis. Unlike in BRAF mutant (BRAFm) melanoma, BRAF and MEK inhibitors show limited activity in BRAFm CRC. Preclinical...

3507 Background: MMR status has been established as an important prognostic factor in CC pts. We determined the association of dMMR status with clinical/pathological features and prognosis using th...

3511 Background: Recently, a number of independent groups reported novel molecular subtypes in colorectal cancer (CRC). A formal comparison across these classifiers is needed to reconcile findings ...

3520 Background: RNA expression of epiregulin (EREG) ligands EREG and/or amphiregulin (AREG) has shown correlation with the efficacy of EGFR-targeted therapy in advanced colorectal cancer (aCRC). T...

e14510 Background: In a phase 1 study investigating ziv-aflibercept (Z, known as aflibercept outside the United States) with FOLFIRI, an analysis was performed to evaluate safety and activity in colorectal cancer (CRC) patients previously treated with irinotecan (I). Methods: The first part of the study was an open-label, dose-escalation stage in which patients with advanced solid tumors received Z at doses of 2, 4, 5, or 6 mg/kg on day 1, followed by FOLFIRI q2wks. The second part of the study was an expansion stage at the recommended dose of Z (4 mg/kg). This analysis is limited to CRC patients previously treated with I and receiving a dose of Z ≥4 mg/kg in either part of the study. Results: Of the 61 patients in the overall study with a planned dose of Z ≥4 mg/kg, 37 CRC patients previously treated with I received an actual dose of Z ≥4 mg/kg (56.8% men; mean age 57.6 years; 94.6% with ECOG performance status 0-1). Table shows response rates for the prior I efficacy population (n=36) and overall study ...

LBA443 Background: The phase III CRYSTAL study showed patients with unresectable KRAS exon 2, codon 12 and 13, wild-type (wt) mCRC (n = 666) to benefit from the addition of cetuximab to FOLFIRI first-line. Progression-free survival (PFS), overall survival (OS) times and response rate (RR) were significantly better in patients who received FOLFIRI plus cetuximab, compared with FOLFIRI alone. Recent retrospective analyses of studies involving EGFR-targeted monoclonal antibody therapy have suggested that patients with KRAS and NRAS mutations in other codons were unlikely to benefit from the addition of an EGFR-targeted agent. Thus, the impact of these additional KRAS and NRAS mutations on the efficacy and safety of FOLFIRI plus cetuximab treatment is under investigation in patients from the CRYSTAL study. Methods: KRAS (eight mutations on exons 3 [codons 59 and 61] and 4 [codons 117 and 146]) and NRAS (18 mutations on exons 2 [codons 12 and 13], 3 [codons 59 and 61] and 4 [codons 117 and 146]) mutation detection is ongoing using BEAMing (Beads, Emulsions, Amplification, and Magnetics) with a mutation detection cut-off of 0.1%. Tumor genomic DNA samples from 541 randomized patients with KRAS exon 2 wt disease (81% of KRAS wt patient population) and tumor material available are currently being screened. Treatment arms will be compared using two-sided log-rank tests for PFS and OS, and the Cochran–Mantel–Haenszel (CMH) test for response. Results: RAS mutation status is currently available from 192 patients (28.8% ascertainment) with further RAS assessment ongoing. Currently 63.5% of the samples are RAS wt and 36.5% are new RAS mutations. Conclusions: Updated information on the full RAS mutation dataset of up to 400 samples combined with the clinical data will be presented and implications for patient care discussed.

495 Background: MSI is a well-documented pathway in the pathogenesis of colorectal cancer. A specific cphenotype has been associated with MSI tumors, as they tend to be more right-sided, have a better prognosis, mostly display a poor histologic differentiation and are marked by extensive inflammation. However, the biological basis and of all these features is still largely unknown. We hypothesize that MSI results in an accumulation of intratumoral mutations. Consequencely, lots of aberrant proteins are formed, probably resulting in a poor histologic differentiation but also evoking a strong immune response that may have the ability to limit tumor expansion (hence explaining a better prognosis) and to stimulate angiogenesis. Methods: To sustain this hypothesis we compared 35 MSI and 35 MSS tumors (matching for TNM staging) with respect to their morphology, inflammation and vessel density. Morphology was analyzed on HE stains. Immunohistochemical stainings for CD 3, CD4, CD8, CD20 and CD68 were performed to document the number of T-lymphocytes, B-lymphocytes and macrophages in relation to tumor location (intraepithelial, intratumoral, peritumoral). CD 31 staining was performed to highlight vessel density in the tumor and assess angiogenesis. Conclusions: (1) In accordance with literature, MSI tumors occured at younger age, with preponderance of the right colon and with a smaller likelihood of metastasis. (2) The most striking observation was the higher number of intra-epithelial T-lymphocytes in MSI tumors (not present in their microsatellite stable counterparts). No difference was found between MSI and MSS tumors in terms of intra/peritumoral B-cells, macrophages, and/or T-cells. (3) Half of the MSI tumors and none of the MSS tumors displayed different histological subtypes/patterns within one tumor. As such, we suggest that in daily routine practice, presence of morphologic tumor heterogeneity should initiate analysis for microsatellite instability. (4) MSI tumors had a statistically much higher (micro)vessel density as compared to their MSS counterparts, hence hinting at a potential application of anti-VEGF therapy in this subgroup of colorectal cancers.

20 Background: Poorly differentiated cancer of the upper GI tract is a challenging diagnosis for the pathologist, who will rely on a positive cytokeratin immunostaining to confirm the epithelial origin of the tumor. Being confronted with a highly unusual case (a poorly differentiated esophageal “carcinoma” with immunophenotypic and genetic sarcoma-like features, arising in the background of high grade Barrett dysplasia), we retrospectively investigated whether other cancer cases displayed the same unusual features. Methods: 34 poorly differentiated GI cancers were selected based on availability of FFPE. All cases were investigated by (1) immunohistochemistry for selected carcinoma and sarcoma markers (prekeratin, NCAM, vimentin, CD99) and by (2) FISH for presence/absence of abnormalities of 2 genes, SYT and EWS, that are frequently rearranged in synovial and Ewing’s sarcoma respectively. Results: 6/34 cases displayed an aberrant immunophenotype, with expression of NCAM, vimentin, and/or CD99. FISH revealed a substantial number (9/34) of SYT and EWS related anomalies: 1 EWS rearrangements, 1 SYT rearrangements, 1 SYT and 1 EWS focal amplification, and 4 aneuploidies. While EWS and SYT are (mostly) fused to ERG or SSX1/2 in Ewing’s and synovial sarcoma respectively, these genes were found to be intact in our EWS and SYT rearranged cases. Conclusions: Our data show that SYT & EWS abnormalities are recurrently occurring in poorly differentiated GI cancers that are (mis?)diagnosed as carcinomas. While EWS rearrangement has previously been described in cancer types other than Ewing’s sarcoma, SYT rearrangement was so far thought to be exclusive to the diagnosis of synovial sarcoma. In addition, EWS/SYT abnormalities in our series seemed to target cancers that are at risk of not responding well to the conventional chemoradiotherapy regimen. Of interest, the one “carcinoma” patient that started this study went into complete remission after being administered a sarcoma targeted therapy. We also infer from our results that good histologic assessment of GI cancers is essential for clinical trials aiming at targeted therapy, as not all carcinomas may present true carcinomas.

LBA444 Background: Patients with KRAS exon 2 codon 12/13 wild-type (wt) mCRC benefit significantly from the addition of cetuximab to first-line FOLFOX4 in relation to response and progression-free survival. Patients with KRAS exon 2 codon 12/13 mutations show no benefit with a trend to worse clinical outcomes. Methods: Tumors from OPUS study patients previously defined as KRAS codon 12/13 wt (n=179) were screened for specific mutations in KRAS exons 3 and 4 (8) and NRAS exons 2, 3, and 4 (18) using BEAMing technology. Treatment outcomes were assessed according to mutation status. Results: RAS tumor mutation status was evaluable for 118/179 (66%) patients. Mutations at the screened loci were detected in 36 (31%) patients. In the RAS wt population, there was benefit associated with the addition of cetuximab to FOLFOX4 (Table). In the overall RAS-mutant population, there was less favorable clinical outcome and no benefit from the addition of cetuximab to FOLFOX4. Conclusions: Patients with mCRC harboring any activating mutation of KRAS or NRAS are unlikely to benefit from the addition of cetuximab to FOLFOX4. Definitive conclusions for patients with new tumor mutations cannot be drawn due to low patient numbers. Restricting cetuximab administration to patients with tumors wt at all such loci might enable the further tailoring of therapy to maximize patient benefit. [Table: see text]

542 Background: There is increasing evidence that evaluation of tumor shrinkage on a continuous scale early on during treatment in metastatic colorectal cancer (mCRC) contains predictive information on long term treatment benefit. These findings were more specifically observed in trials evaluating addition of cetuximab to chemotherapy in first line mCRC (Piessevaux, JCO, 2013). Precise tumor measurements are necessary to implement these findings in clinical trials and practice. Independent central review has been advocated to improve the validity of the radiological measurements in clinical trials. In the phase II mCRC OPUS trial, we compared tumor shrinkage assessments given by an independent review committee (IRC) to results of a semi-automated software algorithm (SOFT). Methods: IRC measurements were performed and reported previously. SOFT requires a seed placement in a metastatic deposit. Once the seed placement is made, the computer automatically generates the boundary of the lesion and calculates the maximal diameter. The original digital radiological data from 217/337 ITT patients was available for review. Early tumor shrinkage, defined as the relative change at week 8 of the sum of the longest diameters was calculated with both measures and compared (ETS-IRC vs. ETS-SOFT). Results: The intra-class coefficient of correlation between ETS-IRC and ETS-SOFT was 0.689 (95%CI 0.604-0.759). When the population was split according to treatment group we observed the following ICC: 0.874 (95%CI: 0.815-0.914) in the chemotherapy with cetuximab arm and 0.486 (0.307-0.632) in the chemotherapy alone arm. Conclusions: Secondary independent review of radiological data from the OPUS trial showed discrepancies between IRC and SOFT. These discrepancies are potentially due to target lesion number and selection. Treatment group seems to differentiate cases more prone to discrepant review. This finding may be related to enhanced reliability when treatment effect is more pronounced. Further comparisons of traditional IRC reviews to semi-automated algorithm-based reviews with regard to outcome evaluation are warranted.

3577 Background: Studies on the association between colorectal cancer (CC) outcome and thymidylate synthase expression have provided inconsistent results. In this study we attempted to resolve the issue by assessing the associations between TS expression and outcome in a population of primary CC patients (pts), who after resection were randomized to 5-FU/FA vs. FOLFIRI adjuvant therapy. Methods: Immunohistochemical staining for TS protein was successfully performed for 1211 pts in the PETACC3 trial. TS immunoreactivity was scored as high expression (≥75% positive) and low expression (

e14561 Background: The RESPECT trial (n = 198) tested the addition of Sorafenib to standard mFOLFOX6 treatment in first line metastatic colorectal cancer (mCRC) patients but resulted in no statistical significant improvement in progression free survival, and no evidence for overall benefit. Sorafenib inhibits several Raf kinases (including B-Raf). Samples with high BRAF-mutant-like score were previously shown to identify a subset of colon tumors with a similar biology and outcome to BRAF mutant patients (Popovici et al. J. Clin. Oncol., 30(12):1288–95, 2012). Methods: A subset of 125 patients from the trial was available for gene expression analysis from their primary tumor FFPE samples, using the Colon DSA gene expression arrays from Almac. Mutation status for KRAS and BRAF was previously assessed. The potential prognostic and/or predictive effect of a high BRAF-mutant-like score was assessed. The analyses were performed using Cox proportional hazards regression models and Kaplan-Meier curves. The logrank test was used to compare the survival distributions (significance level: 5%). Results: Molecular profiling was performed on FFPE tissue samples from primary tumors of 125 mCRC patients, 95 samples (47 in the combined arm; 3 BRAF mutants) were successfully processed. Limitations in amount of material available in this retrospective analysis led to failure to reach the required RNA amount for amplification in 30 samples (dropout rate = 24%). The collected gene expression data was of good quality: all 95 array profiles could be used (10 were flagged for slightly inferior quality). The BRAF-mutant-like score was recognized as marker of poor OS in a Cox regression model (HR = 1.55 [95% CI: 1.12 - 2.13], 1 unit = 1 IQR, P = 0.007). The HR difference between the two arms (combined arm: HR = 1.36; reference arm: HR = 1.41) is not significant. Conclusions: The poor survival of metastatic colorectal cancer patients with a BRAF-mutated-like tumor is confirmed, but no predictive effect was found. FFPE tissues are well-suited for this kind of study and allow to accurately test the hypotheses of interest. Further analyses are planned to generate hypotheses about markers of sensitivity to mFOLFOX6 or Sorafenib.

3526 Background: It has been shown that tumors arising in the proximal and distal colon, defined by the embryological midgut and hindgut, have distinctive clinical and molecular features, but very little is known concerning the differences in the mechanism of tumorigenesis and the effect that this could have on therapy. Methods: The distribution of clinico-pathological and molecular features was evaluated between proximal (N = 1110, Caecum to hepatic flexure) and distal colon (N = 1728, splenic flexure down to sigmoid) in patients included in the PETACC3 trial. Gene expression profile was also available from 783 tumors and 32 normal colon. A further set of 473 metastatic patients treated with cetuximab combined with chemotherapy (De Roock Lancet Oncol. 2010) was used to test tumor location with response. Results: Pathological features, such as tumor differentiation, and mucinous histology as well as molecular characteristics, such as MSI status, BRAF, PIK3Ca mutations and LOH18q loss show higher frequency in proximal compared to distal colon (N = 1214; Fisher test, P

3626 Background: Although cetuximab treatment has been successful for the treatment of KRAS wild-type colorectal cancers, complete remissions are rarely seen in patients, leading ultimately to resistance. We hypothesized that in cetuximab-sensitive patients, the ErbB network is insufficiently targeted since network plasticity may occur. Methods: We have used EGFR-sensitive colorectal cancer cells and investigated ErbB network activity and adaptations by RTK arrays, western blot, and Collaborative Enzyme Enhance Reactive immunoassay (CEER). These were complemented by ErbB heterodimerization (HER2:3, HER1:2, HER1:3, HER3:PI3K) assays using CEER. Effects on cell survival were measured using colony formation assay. In addition to the EGFR sensitive cell line, 200 clinical colorectal cancer (CRC) samples were profiled utilizing CEER for RTKs, downstream signaling, and heterodimerization. Results: EGFR and downstream signaling proteins AKT, ERK, and RSK were potently inhibited by cetuximab or gefitinib at 24h of treatment in EGFR sensitive colorectal cancer cell line. At 24h of treatment, we observed approximately 2 folds increase in total HER2 and HER3 protein levels, 2.7 folds in phosphorylated HER3, and 5.4 folds in HER2:HER3 heterodimer formation. Concurrently, increased in ErbB heterodimer formation was accompanied by 5 folds increase in PI3K binding to HER3, resulting in enhanced HER3 signaling, with increase in AKT, ERK, and RSK. Co-treatment of these cells with cetuximab and HER2 inhibitor Trastuzumab or by treatment with lapatinib blocked the induction of HER2:HER3 heterodimer, HER3 phosphorylation, and PI3K binding to HER3. In 30% of the 200 clinical colorectal cancer samples profiled, we observed an increased in phosphorylated HER3, formation of HER2:HER3 and HER3:PI3K heterodimers along with HER1 activation (KRAS WT). Conclusions: Combination of HER2 inhibitor with an EGFR inhibitor could potentially increase the therapeutic index in cetuximab sensitive patients, and suppress activation of feedback mechanisms upon EGFR inhibition. Current findings suggest that CRC patients with similar profile would benefit from these combination therapies.

3588 Background: Peripheral blood monocytes (PBM) represent a reservoir of inflammatory cells that contribute to cancer progression. When recruited into tumors, monocytes give rise to either macrophages or dendritic cells. Tumor-derived soluble factors can influence the differentiation of tumor-associated macrophages (TAMs) and drive their phenotype to either tumoricidal (M1) or pro-tumorigenic macrophages (M2). We hypothesized that PBM might change their phenotype already when in circulation and can be used as potential markers for early diagnosis of colorectal cancer (CRC) and its progression to metastatic disease. Methods: Monocytes were isolated by using magnetic microbeads conjugated to antibodies against CD14. In a monocentric training set we included 55 (27 non metastatic (NM), 38 metastatic (M)) untreated CRC patients. The third group included age and gender matched healthy volunteers (HV). Following RNA extraction, gene expression profiles of PBM were investigated by high throughput screening using Illumina. Statistical analysis was done on the microarray expression data to delineate a disease specific signature. A cut off was set to select the genes showing a minimum fold change of 1,5. A validation study was performed in three additional major oncological centers throughout Europe. Results: There was a broad overlap in the gene signature of PBM from NM and M patients, indicating that this signature already exists in an early stage. Then we assessed the performance of the gene signature of cancer versus HV as diagnostic tool, in particular the sensitivity and specificity, which were remarkably high in the ROC analysis. Out of 40 differentially expressed genes, 21 genes were confirmed by qRT-PCR analysis on samples processed independently. In the validation study the number of genes was downsized to a 12 gene signature without loss of specificity and sensitivity. Conclusions: This first biomarker study of PBM in CRC suggests an ‘imprinting’ of PBM by the tumor in CRC patients. Since the imprinting profile might also be reversed, the PBM signature might not only be useful in CRC diagnosis, but also for the follow-up of treated CRC patients. Overall, our data offer new opportunities to develop a non-invasive test for CRC detection and monitoring.

3531 Background: Previously reported results of this randomized study demonstrated that the addition of dalotuzumab (Dz) worsened the PFS and OS of chemofractory mCRC patients receiving Cx and Ir (Watkins et al; ASCO 2011). Comprehensive molecular analysis has been undertaken retrospectively to identify possible predictors of Cx resistance and Dz response. Methods: Quantitative RT-PCR for IGF-1, IGF-2, immunohistochemistry for IGF-1R and microarray expression profiling was conducted on archival tumor tissue. All patients had received Cx and Ir with either placebo (n=107), weekly Dz (n=112), or 2 weekly Dz (n=110). Results: Data from 292 and 206 patients was successfully obtained by RT-PCR or microarray respectively. Within the placebo arm, high IGF-1 expression was found to be associated with resistance to Cx (IGF-1-/IGF-1+; PFS 6.7/3.7 months, OS 15.5/9.6 months). High IGF-1 expression was associated with benefit from the addition of weekly Dz (placebo/weekly Dz; PFS 3.7/5.7 months, OS 9.6/18.2 months). By contrast the addition of Dz was not effective in tumors with high IGF-2 expression (placebo/weekly Dz; PFS 8.4/2.7 months). Microarray analysis revealed distinct populations that differentially correlated with Cx and Dz response. An epithelial phenotype appeared more associated with Cx response, whereas a mesenchymal phenotype more associated with Dz response. Rectal cancers showed greater association with increased IGF-1 expression, EMT gene signature and Dz response. Conclusions: These data support IGF-1 and IGF-2 as potential biomarkers for response to Dz therapy and high IGF-1 as a marker of resistance to Cx therapy. Based on these data Dz is being further evaluated in a molecularly selected population of mCRC. [Table: see text]

3587 Background: KRAS mutated metastatic colorectal cancer (mCRC) is inherently resistant to EGFR targeted therapy and carries an inferior prognosis to wild-type disease with a median survival of approximately 16 months. Methods: The activity of dalotuzumab (Dz) has been investigated in KRAS mutant pre-clinical colon cancer models. In addition efficacy data relating to a sub-set of KRAS mutant mCRC patients enrolled in a randomized phase II/III study (initiated prior to the introduction of KRAS genotyping) of Dz in combination with cetuximab (Cx) and irinotecan (Ir) are reported. All patients had received Cx and Ir with either placebo (n=20), weekly Dz (n=18), or 2 weekly Dz (n=31). Results: In pre-clinical models, Dz was found to be effective in inhibiting IGF-1 mediated cellular growth. Furthermore, in colorectal cancer xenograft models high IGF-1 was found to identify a sub-set of Dz responsive tumours. Combination studies demonstrated that Irinotecan exposure resulted in the activation of IGF-1R and PI3K signaling pathways, representing a possible cellular survival mechanism. In xenograft models the combination of Dz with irinotecan produced lasting tumor growth inhibition that persisted even after treatment withdrawal. Outcome data are reported for 69 patients with KRAS mutant mCRC enrolled in the randomized study. Efficacy data demonstrated differential activity between colon versus rectal tumours with evidence of activity in the rectal subgroup. Molecular analysis suggests that this observation may be mediated by differential IGF-1 expression. Conclusions: These data suggest that Dz in combination with irinotecan may have utility in the treatment of KRAS mutated colorectal cancer patients. Based on these data Dz is being further evaluated in a molecularly selected population of mCRC. [Table: see text]

3509 Background: Prognosis prediction for resected primary colon cancer is currently based on the tumor, nodes, metastasis (TNM) staging system. Gene expression based risk scores have been proposed, but need to be validated and integrated with clinical and TNM variables. We performed an independent assessment of the individual recurrence signatures developed for commercial use by Veridex and Genomic Health. Methods: The Veridex (aVDS) and Genomic Health (aGHS) risk scores were applied to existing gene expression data of 580 stage III and 108 stage II tumors from the PETACC-3 trial. Association of the scores with relapse-free (RFS) and overall survival (OS) of the patients was assessed singularly, and in a combination with TNM and other clinico-pathological variables, by univariate and multivariate Cox regression models and logrank methods. Results: Both risk scores were significantly associated with RFS in univariate and multivariate models (Table) for the stage III cohort. The scores contributed different and additive prognostic information, and reached highest effect sizes (approximate p-value 0.001 and HR per interquartile range 1.4 each) in a combined model (Table) that includes both scores as well as T-stage, N-stage and MSI status as significant factors. Analysis in the stage II cohort gave similar effect estimates. Conclusions: This study confirms in an independent colon cancer patient cohort that gene expression based risk scores improve current prognostic models. The best prognostic model was obtained by using clinico-pathological variables and both gene-expression risk-scores. [Table: see text]

3511 Background: From a clinical perspective colorectal cancer (CRC) is a heterogeneous disease whose biological background is insufficiently understood. In order to adapt targeted treatment to biologically different categories of CRC patients, in depth understanding of the molecular mechanisms involved in CRC heterogeneity is urgently needed. Methods: Consensus cluster analysis of 1113 stage II-III CRC gene expression profiles from PETACC3 and 4 public datasets defined a set of subtypes which were confirmed in an independent set of 720 samples. The similarity between tumors was based on a set of 54 meta-genes. This approach improves the robustness to measurement noise and gives equal chances to each biological process to be accounted for in the subtype definition. We tested the association of the subtypes with clinical variables, molecular markers and patient survival. Results: We identified and validated 5 major subtypes A-E, with different levels of expression in 6 main molecular processes: epithelial-mesenchymal transition (EMT), immune response, colon crypt differentiation, proliferation, Wnt signaling and chromosome 20q. Significant differences in survival and an enrichment for markers such as MSI, BRAF mutation, site or p53 status were found between the subtypes. In addition, the new subtype classification uncovered heterogeneity within groups defined by these commonly used markers. Survival analysis showed significant prognostic value for meta-genes connected to EMT, proliferation and differentiation. Identical data processing and clustering applied to the validation set on the same meta-genes resulted in subtypes with almost identical expression patterns. Conclusions: We identified and validated robust molecular subtypes in the largest set of stage II-III CRC samples as a combination of multiple molecular processes, that complement current disease stratification based on clinico-pathological variables and molecular markers. The biological relevance of these subtypes was reflected in significant differences in survival. These insights open new perspectives for improving prognostic models and rationalize the prediction of tumor-specific drug sensitivity.

474 Background: KRAS mutation represents the only validated biomarker used in clinical practice for selection of metastatic colorectal cancer (mCRC) candidates for therapy with the anti-Epidermal Growth Factor Receptor (EGFR) monoclonal antibodies. Previous studies conducted in small cohorts of patients suggested that HER2, the major EGFR partner, could modify the sensitivity to anti-EGFR agents. Aim of the present study was to investigate the role of HER2 gene status in a cohort of mCRC patients treated with cetuximab or panitumumab. Methods: 266 chemorefractory mCRC patients treated with cetuximab or panitumumab alone or in combination with chemotherapy were collected in an international consortium effort. HER2 gene status was analyzed using the dual-color FISH assay LSI HER2/neu-CEP17 (PATHVYSION, Abbott) in one central lab, whereas KRAS/BRAF mutations were investigated locally. HER2 gene amplification was defined as the presence of a ratio between HER2 gene and chromosome 17 centromere > 2. Results: An objective response to anti-EGFR therapy (complete or partial response according to RECIST criteria) was observed in 79/266 (29%) of patients. Twelve cases (4.5%) showed the classical pattern of HER2 gene amplification (R>2) in the whole tissue (>80% of tumor cells). By matching HER2 gene status with clinical response we observed that HER2 gene amplification was significantly related to therapy resistance (p Conclusions: Data from this large retrospective study suggest that HER2 gene status evaluated by FISH may represent an additional marker useful for the prediction of mCRC patients’ response to EGFR-targeted therapies, in line with very recent evidence. In particular, patients with HER2 gene amplification seem to be resistant and show a shorter PFS. Future studies are needed to deeply investigate the clinical and biological meaning of HER2 gene status deregulation in mCRC.

3514 Background: Individual variation in colorectal cancer (CRC) survival may in part be explained by germline determinants. To date, the literature investigating such effects is sparse, and the fe...

3554 Background: Response to the epidermal growth factor receptor (EGFR) inhibitor, cetuximab appears to correlate with the intensity of the associated skin reaction. This randomized multicenter study investigated the effect of cetuximab dose escalation on EGFR activation and downstream signaling in skin and tumor biopsies, on activity and skin reactions in patients (pts) with EGFR-expressing mCRC, failing prior irinotecan (I). Methods: Pts with ≥1 measurable lesion suitable for biopsy were eligible and were randomized 22 days after starting treatment with cetuximab (initial dose 400mg/m2 then 250mg/m2/week(w)) plus I (2-weekly regimen 180 mg/m2) if they had not experienced >grade 1 skin reaction, any other >grade 2 cetuximab-related toxicity and were not intolerant to I. Pts were randomized to standard cetuximab dose (arm A; 250 mg/m2/w) or dose-escalation (arm B; cetuximab dose increased by 50 mg/m2 q 2 w up to 500 mg/m2, until >grade 2 toxicity or tumor response). Pts not randomized (arm C) continued with the standard regimen after day 22. Skin and tumor biopsies were taken pre- and during treatment. Results: Of 273 screenedpts, 221 were EGFR-expressing and 166 were recruited. 45 pts were randomized to arm A, 44 to arm B and 76 were in arm C. By August 2005, 18 pts had reached the highest dose (500mg/m2/w), with no concerns over safety. All baseline and second biopsies are available. Safety results are presented for the first 36 pts in arm B. M/F 23/13, median age 60 y (34–79), median Karnofsky performance score 90 (80–100). All pts reported here had received cetuximab for at least 6 weeks. Maximum dose reached at w 6: 350mg/m2/w (300mg/m2 in 36 pts at w 4; 350mg/m2 in 31 pts at w 6). Relevant grade 3/4 toxicities were diarrhea (2.8%) and fatigue (5.6%). No grade 3/4 skin reactions occurred. Conclusions: Cetuximab at a dose of 350 mg/m2 is well tolerated in pts who have had no or mild skin reactions in response to standard dose treatment. Further dose escalation is possible. No early increase of skin toxicity is observed. Analysis of efficacy and pharmacokinetics/dynamics is ongoing. [Table: see text]