Comprehensive clinical, pathological and molecular characterization of a cohort of locally advanced rectal cancer patients: Towards an integrative classification for rectal cancer management.

A. LLERA; M. INK; M. ELETA; V. MIKOLAITIS; U. GUALDRINI; J.M. SENDOYA; E. ROCA; E. FERNANDEZ; O. PODHAJCER; S. IRMAS; M. RIZZOLO; C. ROTONDARO; A. CABANNE; R. SALANOVA; M KUJARUK; M. GOLUBICKI; M. CORAGLIO; S. ISEAS

Congreso; ESMO 19th World Congress on Gastrointestinal Cancer; 2017

Introduction: The response of Locally Advanced Rectal Cancer (LARC) topreoperativeCRT is heterogeneous. Prognosis of patients depends on several knownfactors: cT,nodal stage, EMVI, CRM, tumor regression grade and surgical margins.But these factorsdon?t have a large impact on individualizing the therapy. Some patientsmay requirean intensified regimen to increase tumor response, whereas standard5-FU-basedchemoradiotherapy may be sufficient for others. Molecular markers topredict treatmentresponse in advance, could contribute to a better understanding of theunderlyingmechanisms of rectal cancer and its response to preoperative treatment.Gene expressionis linked to cellular phenotype and tumour behaviour. The fourconsensus molecularsubtype (CMS) groups represent the current best description of CRCheterogeneity at the gene-expression level. Our aim is to integrate CMSpredictionswith clinical and response assessment available data from LARC.Methods: We analyzed gene expression data from the first 12/42consecutive patientswith rectal adenocarcinoma from an ongoing recruiting prospectivetranslational researchtrial in rectal cancer patients treated in a single institution sinceNov 2015 inArgentina. All patients were radiologically staged usinghigh-resolution abdominal-pelvicMRI, chest CT scans and CEA levels. LARC patients were stratified toarm A: preoperativechemoradiotherapy (CRT: 50Gyþ capecitabine 825mg/m2/bid) or arm B:induction chemotherapy (ICT: CAPOX 3 cycles) followed by CRT then TMEsurgery.MRI asessment was performed after ICT and 6-8w postCRT considering asGoodresponse:mrTRG1-2, moderate response:mrTRG3, poor response: mrTRG4-5.Mutations in RAS exons 2, 3 and 4/BRAF V600E were measured by PCREntrogen mutationpanel. IHQ Protein Expression Deficit (MLH1,MSH2, MSH6,PMS2) was measuredby Cell Marque monoclonal primary antibodies. High quality RNAextractedfrom baseline primary tumor biopsies was labeled and hybridized intoAgilentAgi4x44K Whole Human Genome microarrays. We then analyzed thenormalized datasets with the R package ?CMSclassifier? -developed by theColorectal CancerSubtyping Consortium-, using the similarity-to-centroid approach.Results: We analyzed 12 patients with rectal adenocarcinomahistologically confirmed,with a median age of 48 yo, male 67%, PS 0-1, and a median distance toanal verge:80mm. 10/12 patients were LARC who completed preoperative treatment.Metastaticdisease at diagnoses was observed in 2/12p. CMS 2-Canonical subtype:10p (82%).Stage II/III:8p (T3-T4N0: 3p, N1/2: 5p, EMVIþ: 4p, CRMþ:7p). StageIV:2p . MSH2Defficit protein expresi[1]on: 1/10, K-RAS mut:1/10. Arm A: 3p(CRT): Good response:1p, Moderate response:1p and poor response 1p.) ARM B:5p (ICTþCRT): 5p(Goodresponse: 2p, moderate response:1p, poor response 2p).CMS1- MSI immunesubtype:1/12p (8%), Stage III (mrT3N2,EMVI-,CRMþ), Lynch?s syndrome confirmedbyMSH2 germinal mutation. Disease progression was observed afterinduction chemotherapy.CMS3-Metabolic subtype: 1/12p (8%). Stage III: mrT3N1, EMVIþ,CRMþ.MSS, KRAS mut. Poor response after ICTþCRT was observed.Conclusion: Our results confirmed CMS2 is the more prevalent group inrectal Cancer.CMS4 was not identified in our cohort probably due to the small sample.Within CMS2patients, a heterogeneous grade of response was seen. We expect that anincrease of analyzedcases can help us establishing molecular features to correlate withneoadjuvantresponse.