Abstract
Despite the well-known prognostic value of the tumor-immune microenvironment (TIME) in colorectal cancers, objective and readily applicable methods for quantifying tumor-infiltrating lymphocytes (TIL) and the tumor-stroma ratio (TSR) are not yet available. We established an open-source software-based analytic pipeline for quantifying TILs and the TSR from whole-slide images obtained after CD3 and CD8 IHC staining. Using a random forest classifier, the method separately quantified intraepithelial TILs (iTIL) and stromal TILs (sTIL). We applied this method to discovery and validation cohorts of 578 and 283 stage III or high-risk stage II colorectal cancers patients, respectively, who were subjected to curative surgical resection and oxlaliplatin-based adjuvant chemotherapy. Automatic quantification of iTILs and sTILs showed a moderate concordance with that obtained after visual inspection by a pathologist. The K-means-based consensus clustering of 197 TIME parameters that showed robustness against interobserver variations caused colorectal cancers to be grouped into five distinctive subgroups, reminiscent of those for consensus molecular subtypes (CMS1-4 and mixed/intermediate group). In accordance with the original CMS report, the CMS4-like subgroup (cluster 4) was significantly associated with a worse 5-year relapse-free survival and proved to be an independent prognostic factor. The clinicopathologic and prognostic features of the TIME subgroups have been validated in an independent validation cohort. Machine-learning-based image analysis can be useful for extracting quantitative information about the TIME, using whole-slide histopathologic images. This information can classify colorectal cancers into clinicopathologically relevant subgroups without performing a molecular analysis of the tumors.
Highlights
Despite the remarkable improvement in the diagnosis and treatment of colorectal cancer, it remains a major health burden worldwide
Automatic quantification of intraepithelial TILs (iTIL) and stromal TILs (sTIL) showed a moderate concordance with that obtained after visual inspection by a pathologist
The K-means–based consensus clustering of 197 tumor–immune microenvironment (TIME) parameters that showed robustness against interobserver variations caused colorectal cancers to be grouped into five distinctive subgroups, reminiscent of those for consensus molecular subtypes (CMS1-4 and mixed/intermediate group)
Summary
Despite the remarkable improvement in the diagnosis and treatment of colorectal cancer, it remains a major health burden worldwide. In the case of colorectal cancer patients, it has often been observed that the prognostic stratification of stage II and III patients was poorly performed, even with the use of the latest TNM system [3]. This unmet need for patient stratification probably explains the fact that a diagnosis of colorectal cancer alone is the leading cause of cancer-related deaths in colorectal cancer patients, even though the majority of the patients are diagnosed at operable stages, with an estimated 5-year survival rate exceeding 60% [4,5,6]
Sign-in/Register to view highlights & summary Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
