Search

e19574 Background: Breast implant-associated anaplastic large cell lymphoma (BIA-ALCL) was recognised by the WHO in 2016 as a rare sub-type of peripheral T-cell, non-Hodgkin lymphoma (NHL), characterized by an indolent clinical course and excellent prognosis. Little evidence exists on the role of post-treatment imaging surveillance with variable practices across the world. Recent UK guidelines recommend that routine surveillance imaging should not be offered to BIA-ALCL patients, in line with national/international recommendations for other NHLs. The aim of this study was to quantify the direct economic costs (DEC) of post-treatment BIA-ALCL routine radiological surveillance at our institution compared to the DEC if UK guidelines were followed. Methods: Following IRB approval a retrospective analysis of a prospectively maintained database of BIA-ALCL patients at The Royal Marsden Hospital was performed. DECs were estimated using current (2020) NHS tariffs for radiological investigations. Imaging undertaken for symptomatic problems/non-BIA-ALCL related concerns was excluded. Results: Eleven patients [median age: 49 years (IQR 45-52)] were treated for BIA-ALCL between 2015-2020 following cosmetic augmentation (n = 6) or breast reconstruction (n = 5). Median time from first implant surgery to BIA-ALCL diagnosis was 11 years (IQR 8-12). Patients presented with effusion (n = 7), mass (n = 2) or effusion and mass (n = 2). One patient had neoadjuvant CHOP/brentuximab, all 11 had explantation with en bloc total capsulectomy, 1 had adjuvant CHOP. Median follow-up was 38 months (IQR 12-47) with no local or distant relapses. Two patients did not have any radiological surveillance and 1 had follow-up elsewhere. The remaining 8 patients had a combination of PET/CT (n = 10), CT (n = 2), breast ultrasound (n = 6), mammogram (n = 4) and breast MRI (n = 1) as routine imaging follow-up not guided by clinical concerns. This represents evolving practice at our institution as the UK guidelines were published in 2021. Total cost of routine imaging surveillance was £10,396 ($14,396) with median cost of £1,953 ($2,705) per patient [IQR £526-2029 ($728-2,810)]. This cost could have been saved based on current guidelines recommending no routine surveillance for patients with no symptoms or clinical concerns. Conclusions: This data demonstrates that omission of routine post-treatment imaging surveillance, as per the recent UK guidelines, would result in a median DEC saving of £1,953 ($2,705) per patient at our institution. No local or distant relapses where identified within the follow-up period, in line with the existing literature suggesting BIA-ALCL has a very low risk of relapse and excellent prognosis. These findings provide a value-based analysis to further support the recommendation not to perform routine post-treatment imaging surveillance in patients with BIA-ALCL.

TPS471 Background: Oesophagogastric (OG) cancers represent a significant health burden and leading cause of cancer related death. Prognosis in advanced disease is poor and novel therapies are needed to improve outcomes. Molecular features of advanced OG cancer suggest that assessment of DDR (DNA damage repair) targeted agents is warranted. Specifically, ATM and ARID1A defects and mutational scars indicative of homologous recombination defects are present in a subset of OG cancers and are associated with polyadenosine 5’diphosphoribose polymerase inhibitor (PARPi) sensitivity. Methods: SOlar is a multi-centre, open-label, single arm, phase II study of olaparib, a PARPi, in patients with advanced oesophageal, gastro-oesophageal junction and gastric adenocarcinoma. The trial will use a single-arm Simon two-stage design to evaluate the anti-tumour activity of olaparib in advanced OG cancers. The primary endpoint is disease control rate (DCR) at 8 weeks by RECIST v1.1. To rule out a DCR of ≤15% while aiming for DCR ≥30% (alpha = 0.09, power = 89%), 54 patients must be recruited it total. An interim analysis will take place when 27 patients have been accrued, dosed and followed until the 8-week disease evaluation. If 4 or fewer patients have disease control (DC) the study will be terminated. If 5 or more patients have DC, an additional 27 patients will be enrolled to a total of 54 patients. If ≥12/54 have DC in the final analysis then it will be concluded that the treatment has shown anti-tumour activity compatible with 30% and an investigation of potential biomarkers of response will be carried out. Secondary endpoints are ORR, DoR, OS, PFS, time to radiological progression and safety. This highly translational study incorporating serial tumour biopsies will investigate candidate predictive biomarkers of PARPi sensitivity with the aim of identifying responder/non-responder subpopulations. Further exploratory objectives will investigate the predictive role of early FDG-PET/CT in assessing tumour response and the creation of an organoid biobank. The trial opened to recruitment in July 2019 and will recruit up to 54 patients over 3 years. Clinical trial information: NCT03829345.

Purpose To evaluate the addition of cetuximab to neoadjuvant chemotherapy before chemoradiotherapy in high-risk rectal cancer. Patients and Methods Patients with operable magnetic resonance imaging–defined high-risk rectal cancer received four cycles of capecitabine/oxaliplatin (CAPOX) followed by capecitabine chemoradiotherapy, surgery, and adjuvant CAPOX (four cycles) or the same regimen plus weekly cetuximab (CAPOX+C). The primary end point was complete response (CR; pathologic CR or, in patients not undergoing surgery, radiologic CR) in patients with KRAS/BRAF wild-type tumors. Secondary end points were radiologic response (RR), progression-free survival (PFS), overall survival (OS), and safety in the wild-type and overall populations and a molecular biomarker analysis. Results One hundred sixty-five eligible patients were randomly assigned. Ninety (60%) of 149 assessable tumors were KRAS or BRAF wild type (CAPOX, n = 44; CAPOX+C, n = 46), and in these patients, the addition of cetuximab did not improve the primary end point of CR (9% v 11%, respectively; P = 1.0; odds ratio, 1.22) or PFS (hazard ratio [HR], 0.65; P = .363). Cetuximab significantly improved RR (CAPOX v CAPOX+C: after chemotherapy, 51% v 71%, respectively; P = .038; after chemoradiation, 75% v 93%, respectively; P = .028) and OS (HR, 0.27; P = .034). Skin toxicity and diarrhea were more frequent in the CAPOX+C arm. Conclusion Cetuximab led to a significant increase in RR and OS in patients with KRAS/BRAF wild-type rectal cancer, but the primary end point of improved CR was not met.

TPS4139Background: Peri-operative chemotherapy with 5FU, oxaliplatin and docetaxel (FLOT) is a new standard of care in resectable gastro-oesophageal adenocarcinoma (GOA), however with 3y survival r...

Purpose To evaluate neoadjuvant capecitabine/oxaliplatin before chemoradiotherapy (CRT) and total mesorectal excision (TME) in newly diagnosed patients with magnetic resonance imaging (MRI) –defined poor-risk rectal cancer. Patients and Methods MRI criteria for poor-risk rectal cancer were tumors within 1 mm of mesorectal fascia (ie, circumferential resection margin threatened), T3 tumors at or below levators, tumors extending ≥ 5 mm into perirectal fat, T4 tumors, and T1-4N2 tumors. Patients received 12 weeks of neoadjuvant capecitabine/oxaliplatin followed by concomitant capecitabine and radiotherapy. TME was planned 6 weeks after CRT. Postoperatively, patients received another 12 weeks of capecitabine. Results Between November 2001 and August 2004, 77 eligible patients were recruited. After neoadjuvant capecitabine/oxaliplatin, the radiologic response rate was 88%. In addition, 86% of patients had symptomatic responses in a median of 32 days (ie, just over one cycle of capecitabine/oxaliplatin). After CRT, the tumor response rate was increased to 97%. Three patients remained inoperable. Sixty-seven patients proceeded to TME, and all but one patient had R0 resection. Pathologic complete response was observed in 16 patients (24%; 95% CI, 14% to 36%), and in an additional 32 patients (48%), only microscopic tumor foci were found on surgical specimens. Four deaths occurred during neoadjuvant capecitabine/oxaliplatin therapy as a result of pulmonary embolism, ischemic heart disease, sudden death with history of chest pain, and neutropenic colitis. Conclusion Capecitabine/oxaliplatin before synchronous CRT and TME results in substantial tumor regression, rapid symptomatic response, and achievement of R0 resection.

58 Background: Single nucleotide polymorphisms (SNPs) in mir-608 (rs4919510) are prognostic in colorectal cancer (CRC) and hepatocellular cancer. SNPs of the let-7 complementary site 6 (LCS-6) 3'-untranslated region of KRAS (rs61764370) are associated with prognosis in CRC and platinum sensitivity in head and neck cancer. We investigated the prognostic effects of these SNPs in patients treated with surgery and with perioperative epirubicin, cisplatin and 5-fluorouracil chemotherapy plus surgery in the MRC MAGIC trial. Methods: DNA was extracted from tumour tissue using the QIAamp DNA FFPE Tissue Kit (Qiagen, Hilden, Germany). Samples were genotyped using the Taqman assay (Life Technologies, Carlsbad, CA) for rs4919510 in mir-608 and rs61764370 in LCS-6. SNP status was assessed for association with patient characteristics and survival. Results: 305 patients (of 456 operated) had mir-608 and/or LCS-6 results available. SNPs mir-608 and LCS-6 were in Hardy-Weinberg equilibrium. Pathological tumour regression grading was not different for any SNP genotype. No genotype was statistically significantly associated with overall survival in either surgery arm or chemotherapy plus surgery arm of the trial (Table). Conclusions: Germline polymorphisms in mir-608 and LCS-6 showed no evidence of a prognostic effect in patients with gastric and gastroesophageal cancer treated with surgery or chemotherapy plus surgery in the MAGIC trial. However, the small number of patients with selected genotypes means that these analyses may be underpowered to detect survival differences. [Table: see text]

709 Background: International guidelines suggest that RT dose escalation, intraoperative RT or brachytherapy could be considered for LARC pts with positive resection margins, pT4 or unresectable tumours after standard neoadjuvant CRT. However, data to support these approaches are scarce. The potential of systemic CT as salvage treatment after failure of neoadjuvant CRT for LARC has never been explored. We conducted a single-centre, retrospective analysis to address this question. Methods: Pts with newly diagnosed rectal adenocarcinoma who were deemed inoperable or candidates for extensive (i.e. beyond total mesorectal excision, TME) surgery after completion of long-course RT and received salvage systemic CT were included. The primary objective was to estimate the proportion of pts who became potentially suitable for TME after CT. Secondary objectives included the proportion of pts who ultimately underwent TME and survival outcomes. Results: 45 pts (2001-2015) met the study inclusion criteria (39 candidates for extensive surgery and 6 with unresectable tumours). Previous RT was given concurrently with CT in 43 cases (median dose: 54.0 Gy; range: 34.0-55.8). Salvage oxaliplatin-based and irinotecan-based CT was administered in 40 (88.9%) and 5 (11.1%) cases, respectively. 8 pts (17.8%) became suitable for TME based on the MRI after CT, 10 (22.2%) ultimately underwent TME with clear margins and 2 (4.4%) were managed with a watch & wait approach following radiological clinical complete response. Additionally, 13 pts had a beyond-TME surgery with curative intent. 3-year progression-free survival and 5-year overall survival in the entire population were 30.0% (95% CI: 15.0-46.0) and 44.0% (95% CI: 26.0-61.0), respectively. For the curatively resected and watch & wait pts these figures were 52.0% (95% CI: 27.0-73.0) and 67.0% (95% CI: 40.0-84.0), respectively. Conclusions: Systemic CT may be an effective salvage strategy for LARC pts who fail to respond to long-course CRT and are inoperable or candidates for beyond-TME surgery. According to our study, 1 out of 4 pts may become resectable or being spared from an extensive surgery after systemic CT.

TPS221 Background: The MYC proto-oncogene is among the most frequently dysregulated genes in human cancer and is amplified in 10-30% of oesophagogastric (OG) cancers. Such genes which code for transcription factors are challenging to target directly with small molecule inhibitors or monoclonal antibodies. High-throughput siRNA screening of OG cell lines has found silencing of Bruton’s tyrosine kinase (BTK) to result in selective lethality in the presence of MYC amplification. Sensitivity to the orally available BTK inhibitor ibrutinib has been confirmed in-vitro in MYC, HER2 and co-amplified OG cancer cell lines (Chong, Ann Onc 2014), suggesting it may be a potentially effective therapeutic strategy in both MYC and HER2 amplified OG cancer. Methods: i MYC is an open label, phase II non-randomised study to assess the efficacy of ibrutinib in advanced pre-treated OG cancer. A novel FISH assay for tumour MYC amplification has been developed and pts are pre-screened for MYC and HER2 amplification. Eligibility includes metastatic or locally advanced inoperable OG cancer (SCC or adeno), MYC (ratio >2.5) and/or HER2 (ratio>2) amplification, progression after at least 1 prior line of chemotherapy +/- trastuzumab for advanced disease, PS 0-2 and no history of significant cardiovascular or bleeding disorders. Of the first 9 pts, at least 4 will demonstrate MYC amplification and remaining 5 will show either MYC or HER2 amplification, or co-amplification of both. Pts will be treated with ibrutinib monotherapy until progression or unacceptable toxicity. A Simon 2 stage design will be used for the primary endpoint of response rate, with interim analysis after 9 pts and maximum recruitment to 17 pts. Secondary endpoints include PFS, OS and exploratory translational and imaging biomarker analysis. Mandatory biopsy will be obtained at baseline and optional biopsies will be undertaken at day 10-14, week 8 and progression. The trial represents the first attempt at targeting MYC amplification via synthetic lethal gene interactions in OG cancer through the novel application of an existing anti-cancer drug. As of September 2016, 61 patients have consented for pre-screening, 19 completed MYC analysis and first patient has started treatment. Clinical trial information: 02884453.

4064Background: MMRd and MSI are prognostic for survival in many cancers and predict resistance to fluoropyrimidines in colon cancer. We examined the relationship between MMRd, MSI and survival in ...

e15027Background: Tumour heterogeneity may limit the efficacy of targeted therapies. Pulmonary metastases (PM) occur in 10-20% of patients (pts) with colorectal cancer, but little is currently know...

TPS227 Background: The molecular characteristics of a pt’s tumour can determine their suitability for treatment with targeted drugs. Several molecular alterations are only seen in a small percentage of pts and so it is often necessary to screen many pts to identify those who may benefit from targeted therapies. Testing each pt for multiple biomarkers is also increasingly important. Targeted next generation sequencing (NGS) can rapidly interrogate tumours for actionable genetic aberrations and therefore facilitate a personalised treatment approach. FOrMAT aims to assess the feasibility of delivering validated NGS results (to accredited standards) in a clinically meaningful timeframe and how this could be adopted into routine clinical practice in the United Kingdom’s National Health Service. Methods: FOrMAT (ClinicalTrials.gov identifier NCT02112357) is a single-centre translational study in pts with locally advanced/metastatic GI tumours (including gastroesophageal, pancreatic, biliary tract and colorectal). Targeted capture sequencing is performed on archival or fresh biopsy samples to detect mutations, copy number variations and translocations in 45 genes which have prognostic, predictive or pharmacogenomic significance or are targets in current/upcoming clinical trials. Results are discussed by a Sequencing Tumour Board to establish if a pt is potentially suitable for a targeted therapy (e.g. within another clinical trial, as part of standard of care or via a compassionate access program). Changes to pts’ treatment are not mandated. The primary endpoint is the percentage of pts in whom a currently actionable molecular alteration was detected. Secondary endpoints include evaluation of the time required to obtain sequencing results. Blood samples will be collected at baseline and at the time of each tumour response assessment for analysis of molecular markers, including ctDNA and microRNA. Pts may also consent to optional biopsies for exploratory research into tumour heterogeneity and development of pt-derived organoids. The study aims to recruit 200 pts over 2 years. Recruitment started in February 2014 and to date 56 pts have been recruited. Clinical trial information: NCT02112357.

71 Background: In oesophagogastric cancer TP53 mutation may be prognostic and/or predict for chemoresistance, however available data are conflicting. We hypothesised that TP53 status would impact on survival for patients (pts) randomised to surgery alone or perioperative ECF chemotherapy in the MRC MAGIC trial. Methods: Genomic DNA FFPE tissue sections were extracted with QIAmp DNA FFPE Tissue Kit (Qiagen, Hilden, Germany). Mutations in exons 4-9 were screened for by fluorescent PCR -amplification of genomic DNA, followed by Capillary Electrophoresis-Single Strand Conformational Analysis (CE-SSCA). Mutations were characterised by bi -directional Sanger sequencing analysis performed on an independent PCR -reaction and the sequencing results were compared to the reference sequences in the COSMIC database. TP53 status was correlated with demographics and survival. Results: TP53 results were available on 154 pts (50% of pts with available DNA). The remainder failed CE-SSCA due to degraded DNA. Pts with/without TP53 data had similar demographics and survival. TP53 mutation was detected in 51/154 (33%) samples. Exon 4, 5, 6, 7 and 8 mutations occurred in 4%, 42%, 8%, 36% and 10% of specimens with a TP53 result. Pts with TP53 mutations were more likely to have an oesophageal or junctional tumor (vs. gastric) than TP53 wild type (WT) pts (38% vs. 17% respectively p = 0.016). Survival from surgery was comparable for pts with mutant and WT TP53 tumors in both arms of the trial (See Table). P-value for interaction between treatment group and TP53 status was 0.776, indicating no interaction was present. Conclusions: In the MAGIC trial, TP53 mutation status was not prognostic, and did not predict for lack of benefit from chemotherapy. Further study is required in order to evaluate the effect of differing TP53 mutations on treatment and survival. [Table: see text]

78 Background: DNA copy number aberrations (CNAs) are common in oesophageal and gastric adenocarcinomas (OGCs) and display extensive inter-tumour heterogeneity. CNA patterns define gastric cancer molecular subtypes and ERBB2 amplifications, present in a small fraction of patients with OGC, are predictive for ERBB2-targeted drug sensitivity. Together, this suggests a critical role of CNAs determining OGC tumour biology and clinical outcomes. Despite this, predictive and prognostic CNA biomarkers have not been identified for the majority of OGCs, precluding the development of effective personalized therapy approaches for these aggressive tumours. Intra-tumour heterogeneity, characterized by the presence of multiple subclones with distinct genetic profiles within an individual cancer, can hinder the accurate molecular analysis and classification of tumours. The aim of this pilot study was to assess whether chemotherapy-naïve localized OGCs display intra-tumour macroheterogeneity of CNA profiles. Methods: Tissue specimens from four tumour regions representing the macroscopic spatial extent of each of five OGCs were systematically collected after surgical resection. DNA extracted from these FFPE specimens was analysed by molecular inversion probe SNP arrays for high resolution CNA detection. Results: Comparison of genome wide copy number and B-allele frequency profiles suggested highly concordant CNA profiles across the regions from individual primary tumours. Eight driver CNAs leading to amplification of the MET, KRAS, ERBB2, PIK3CA or FGFR2 oncogenes were identified in 4/5 tumours. Only one out of these eight driver CNA’s, harbouring the KRAS oncogene, was heterogeneous within a tumour. Conclusions: Although chromosomal instability is thought to be common in this tumour type, this pilot study suggests that macroheterogeneity is limited and that CNA profiles assessed from a single tumour biopsy are likely to be representative of the dominant CNA profile of localized OGCs. Thus, clinical correlative CNA analyses may be possible from single biopsies of localized OGCs. Mutational heterogeneity and microheterogeneity in microdissected and single cells are currently being investigated.

581 Background: An association between rs4919510, a SNP of mir-608, and prognosis of patients with colorectal cancer has been reported. However, no studies have analysed the role of this SNP in LARC. We conducted a pharmacogenomic analysis of rs4919510 in EXPERT-C, a randomised phase II trial of neoadjuvant CAPOX followed by chemoradiotherapy, surgery and adjuvant CAPOX ± cetuximab in high-risk LARC. Methods: SNP analysis was performed on DNA extracted from tumour tissue. Kaplan-Meier method and Cox regression analysis were used to calculate survival estimates and compare treatment arms. Results: A total of 155/164 (94.5%) patients had tumours assessable for rs4919510 genotyping. 95 (61.3%) were homozygous for CC, 55 (35.5%) heterozygous (CG) and 5 (3.2%) homozygous for GG with no deviation from the Hardy-Weinberg equilibrium (p=0.379). Genotypes were evenly distributed in the two treatment groups and no association with baseline clinico-pathological characteristics or tumour molecular status (including RAS) was observed. Complete response did not differ according to the tumour genotype in the entire population (14.7% for CC, 13.3% for CG/GG) and was not influenced by treatment (12.0% for CC and 14.3% for CG/GG in CAPOX, 17.8% for CC and 12.5% for CG/GG in CAPOX-C) (all p values>0.05). Median follow-up was 64.9 months. In the CAPOX arm the 5-yr PFS and OS rates were 54.6% and 60.7% for CC and 82.0% and 82.1% for CG/GG, respectively (HR PFS 0.13, 95% CI: 0.12-0.83, p=0.02, HR OS 0.38, 95% CI: 0.14-1.01, p=0.05). In the CAPOX-C arm the same survival figures were 73.2% and 82.2% for CC and 64.6% and 73.1% for CG/GG (HR PFS 1.38, 95% CI: 0.61-3.13, p=0.44, HR OS 1.34, 95% CI: 0.52-3.48, p=0.55). An interaction was found between study treatment and rs4919510 genotype for both PFS (p=0.02) and OS (p=0.07). Conclusions: This is the first study investigating rs4919510 in LARC. We found that the CC genotype was associated with worse prognosis compared to the CG/GG genotype in patients treated with chemotherapy alone but not in those treated with cetuximab. The addition of cetuximab to chemotherapy may mitigate the unfavourable prognostic associated with the CC genotype.

62 Background: MSI is prognostic for survival in diverse cancers and predicts resistance to fluoropyrimidine chemotherapy in colon cancer. We examined the interaction between MSI and patient (pt) characteristics and survival for pts randomised to surgery alone or perioperative ECF chemotherapy in the MRC MAGIC trial. Methods: Tumor and normal control DNA was extracted from resection FFPE. MSI status was determined using the Promega MSI Analysis System (NR-21, BAT-26, BAT-25, NR-24, MONO-27). Tumors were classified as MSS when all markers were stable, MSI-L when only one marker was unstable and MSI-H with minimum of instability in two markers. MSI status was correlated with pt characteristics and survival. Results: MSI results were available for 303 pts (66% resected pts). Twenty (6.6%) and 2 (0.6%) pts were MSI-H and MSI-L. All pts with MSI-H had stomach tumors (vs.GEJ/esophagus). KRAS mutation was more common in MSI-H tumours (30% vs. 4%, p-value

4068 Background: Germline polymorphisms may affect chemotherapy efficacy and survival. We examined the effect of polymorphisms in drug metabolism and DNA repair genes on survival for patients (pts)...

3575 Background: After the use of preoperative radiotherapy and TME, survival of LARC has reached a plateau. More effective therapies, predictive/prognostic factors for treatment selection and vali...

3573 Background: FCγRIIa-H131R and FCγRIIIa-V158F SNPs have been reported to enhance the immune-mediated effects of monoclonal antibodies including cetuximab (C) in metastatic colorectal cancer (CR...

4067 Background: REAL3 evaluated EOC + P in advanced oesophago-gastric adenocarcinoma. We previously reported that addition of P was associated with increased diarrhoea, skin rash and mucositis wit...

TPS8618 Background: Diffuse Large B-cell lymphoma (DLBCL) exhibits high sensitivity to first-line chemo-immunotherapy, however approximately one-third of patients (pts) will relapse. A number of sa...

91 Background: The REAL3 trial evaluated the addition of panitumumab (P) to epirubicin, oxaliplatin and capecitabine (EOC) in advanced OGA. We previously reported that addition of panitumumab was not associated with improvement in survival endpoints. In this analysis we aimed to explore factors associated with OS in the REAL3 population. Methods: Analysis performed in ITT population (n=553). OS was evaluated in relation to the baseline characteristics displayed in the table below. Cox regression was used to obtain HRs and 95% CIs. Factors with p100 U/L. Results: Results of the univariate analysis are shown in the table below. In multi-variate analysis, the factors which remained statistically significant for OS were: PS (p=0.080): 1 vs 0 HR 1.24 (95% CI 0.99-1.57); 2 vs 0 HR 1.57 (95%CI 0.97-2.54) Disease extent (p=0.054): Met disease vs LA HR 1.42 (95% CI 0.99-2.02) Baseline global health score (p=0.005): Low vs high HR 1.52 (95% CI 1.16-2.00); middle vs high HR 1.45 (95% CI 1.11-1.89) Patients with an RMH Prognostic Index score of 0 (good prognosis) had median OS of 13.1 mo. This compared to 9.5 mo in patients with 1-2 adverse features (intermediate prognosis) (HR 1.37, 95% CI 1.12-1.68; p=0.002) and 4.7 mo in patients with 3-4 adverse features (poor prognosis) (HR 4.02, 95% CI 1.96-8.22; p

489 Background: Studies indicate that RAS mutations beyond KRAS exons 2-3 may predict anti-EGFRs benefit. EXPERT-C was a randomized phase II trial of neoadjuvant CAPOX and CRT ± C in MRI-defined, high-risk RC. We have previously shown that adding C in KRAS (exons 2-3)/BRAF wild type (WT) patients did not improve complete response (CR) and was associated with a non-significant improvement in progression-free survival (PFS) (HR 0.62, p=0.23) and overall survival (OS) (HR 0.56, p=0.20). The aim of this study was to analyse the impact of RAS mutations on the outcome of C-treated patients in this trial. Methods: Between October 2005 and July 2008, 164 eligible patients were randomly assigned to 4 cycles of CAPOX followed by CRT, surgery, and 4 cycles of adjuvant CAPOX (n=81) or the same regimen plus C (CAPOX-C, n=83). KRAS (exons 2-3) and NRAS (exon 3) mutations were prospectively analysed. Of 90 KRAS/NRAS WT patients, 84 were retrospectively analysed for additional KRAS (exon 4) and NRAS (exons 2/4) mutations by using bi-directional Sanger sequencing. The effect of C on CR, PFS, and OS in patients with RAS WT tumors was analyzed. PFS and OS were estimated with the Kaplan-Meier method and treatment arms compared using a log-rank analysis. Results: Eleven (13%) of 84 patients initially classified as KRAS/NRAS WT were found to have tumours harbouring a mutation in KRAS exon 4 (11%) or NRAS exons 2/4 (2%). Overall, after this retrospective mutation analysis, 78/149 (52%) assessable patients were RAS WT (CAPOX, n=40; CAPOX-C, n=38). After a median follow-up of 63.8 months, in line with the initial analysis, the addition of C in the group of RAS WT patients, was associated with numerically higher, but not statistically significant, rates of CR (15.8% vs. 7.5%, p=0.31), 5-year PFS (78.4% vs. 67.5%, p=0.17) and 5-year OS (83.8% vs. 70%, p=0.20). Conclusions: Although the results of our analysis are potentially affected by the small numbers, in our locally advanced RC population the status of RAS did not appear to significantly improve the selection of patients who may benefit from the use of an anti-EGFR therapy.