Abstract
Background: There has been enormous effort to develop a prognostic genomic classification of colorectal cancer (CRC). As a result, mismatch repair (MMR) status was established as a prognostic marker in addition to TNM-staging. Phenotypic subtypes were recently proposed that stratified patient survival based on assessing the immune infiltrate (KM grade), tumour growth (Ki67) and stromal infiltrate (TSP). However their clinical utility had not been compared to other proposed classification systems including consensus molecular subtypes (CMS) or current prognostic markers including MMR status.
Methods: Three patient cohorts, a pilot cohort of 237 stage I-III CRC patients, a validation cohort of 879 stage I-III CRC patients, and the AMC-AJCCII-90 cohort with 81 stage II colon cancer patients were utilised to investigate associations between phenotypic subtypes, MMR status, CMS and patient survival.
Results: In the pilot cohort, phenotypic subtype stratified cancer-specific survival (P<0.001). In the validation cohort, phenotypic subtype stratified overall (p=0.003) and cancer-specific survival (CSS, p<0.001). This stratification of CSS was irrespective of adjuvant chemotherapy treatment (no chemotherapy - p<0.001 and chemotherapy received - p=0.006) and tumor location (right colon -p<0.001, left colon - p=0.007 and rectal p<0.001). However, MMR status only stratified CSS in the right colon (p=0.023) and was significantly associated with right-sided colon cancer (p<0.001) upon chi-squared analysis. Therefore further analysis was restricted to this subset of patient (n=380). The immune subtype had the highest MMR deficiency and the stromal subtype was mainly MMR competent (p=0.001). Furthermore, phenotypic subtype and not MMR status was independently prognostic in the full cohort (p<0.001) and right-sided colon cancer (p<0.001). In the AMC-AJCCII-90 cohort phenotypic subtypes aligned with CMS (p<0.001) and stratified overall survival better than CMS (p=0.125 v p=0.487 respectively). However, discordance was seen between CMS1 and the immune subtype, potentially due to differences in the site of inflammation assessed. To address this the KM grade was replaced with the immunoscore within the phenotypic subtypes and the pilot cohort assessed for survival. This method of subtyping stratified CSS (p<0.001) in a similar manner to the original (p<0.001), suggesting that the method of assessing the immune infiltrate is not influencial; therefore as KM grade is more clinically translatable this was chosen as the best method.
Conclusions: Phenotypic subtype is a more effective prognostic classification than CMS and MMR status. Phenotypic subtype is clinically relevant to all patients irrespective of site and adjuvant treatment. Phenotypic subtypes should be incorporated alongside MMR status as a clinical aid for the prognosis of patients with CRC.
Citation Format: Antonia Kathryn Roseweir, James H. Park, Sanne ten Hoorn, Arfon G. Powell, Campbell S. Roxburgh, Donald C. McMillan, Paul G. Horgan, Louis Vermeulen, Joanne Edwards. Phenotypic subtypes successfully stratify prognosis of patient with colorectal cancer: A step towards precision medicine [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2018; 2018 Apr 14-18; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2018;78(13 Suppl):Abstract nr 4615.