Intratumoral Fibrosis Pattern is associated with Lymph Node Metastasis and Recurrence of Colon Cancer

Background: Fibrosis is a significant histopathological feature of colon cancer (CC). However, the association between intratumoral fibrosis and its clinical features has been understudied. We aimed to investigate the histological features of intratumoral tissue fibrosis that reflect lymph node metastasis and recurrence in CC. Methods: Propensity score matching identified 39 stage T3 CC patients from 499 who underwent resection. Azan-Mallory-stained primary tumor tissues were analysed by digital pathology quantitative single-fiber artificial intelligence(AI) FibroNest^TM image analysis platform. Phenotypic Composite Fibrosis Score (Ph-CFS) was calculated from over 300 features of collagen architecture and fiber morphometry to distinguish groups. Results: Ph-CFS differed between patients with lymph node metastasis (LN+) and those without (LN-) (p<0.01). Ph-CFS differentiated the LN+ and LN-groups with 89.5% sensitivity and 89.4% specificity (AUC 0.95). Kurtosis, indicating histogram data concentration, was the most significantly different parameter between the groups. Further, Ph-FCS distinguished recurrence groups post-chemotherapy with 71.4% sensitivity and 100% specificity (AUC 0.88). Among the individual traits examined, the extent of collagen branching was significantly higher in the recurrence group compared to the non-recurrence group (p <0.05). Conclusions: In CC patients who underwent surgery, the fibrotic histological phenotype was different from that of the LN+ metastasis and LN- and recurrence and non-recurrence cohorts. AI-assisted collagen analysis is an effective method for predicting CC outcomes.