Quantitative Analysis of Rectal Cancer Biopsies With the Digital Pathology Segmentation Algorithm QuantCRC Associates With Therapy Response and Recurrence

Abstract

We examined whether QuantCRC, a digital pathology segmentation algorithm, on pretherapy rectal cancer biopsies is associated with pathologic complete response (pCR) to neoadjuvant therapy, recurrence-free survival (RFS), and transcriptomic spatial profiling. QuantCRC was evaluated in an observational cohort of 288 pretherapy biopsies and a separate validation cohort of 37 pretherapy biopsies of rectal adenocarcinoma from patients undergoing neoadjuvant therapy. Associations between QuantCRC features and clinical outcomes, pCR, and RFS were analyzed using multivariable logistic regression and Cox proportional hazards modeling, respectively. QuantCRC variables were also correlated with transcriptomic digital spatial profiling of 37 pretreatment biopsies of cT3N⁺ rectal cancer. QuantCRC-derived lymphocytes per square millimeter of tumor epithelium (tumor-infiltrating lymphocytes [TILs]) was significantly associated with pCR (multivariate odds ratio, 1.05; 95% CI, 1.02-1.10; P = .038). QuantCRC-derived TILs were significantly higher in pretherapy biopsies with pCR (91.3 vs 55.9 lymphocytes/mm²; P = .004). The validation cohort confirmed that only QuantCRC-derived TILs in pretherapy biopsies of rectal cancer were significantly associated with complete response to neoadjuvant therapy. QuantCRC %high tumor grade was independently associated with worse RFS (multivariate hazards ratio, 1.27; 95% CI, 1.09-1.47; P = .002). Patients with ≥10.1% high tumor grade identified by QuantCRC had significantly reduced RFS (5-year RFS 69% vs 83%, log-rank P = .007). Transcriptomic profiling identified high interleukin (IL)-6/JAK/STAT3 signaling within immune cells to be associated with worse RFS (adjusted P = .01). Tumors with low IL-6/JAK/STAT3 expression within immune cells had significantly higher TILs compared with tumors with high expression (median, 152 vs 97 TILs/mm²; P = .039). Biopsy-adapted QuantCRC in pretherapy rectal cancer may be helpful in identifying patients who achieve pCR and are at risk for recurrence.