Radiomics modeling to predict the tumor immune-microenvironment of mucinous adenocarcinoma of gastric-type cervical cancer

Gastric-type adenocarcinoma (GAS) of the cervix is usually diagnosed at advanced stages and has a poor prognosis due to resistance to standard therapies. Tumor-infiltrating lymphocytes (TILs) are an important component of the tumor immune microenvironment (TME) and have been associated with prognosis across multiple cancer types. Histopathological assessment of GAS is challenging because it often arises in the upper cervix. This study aimed to explore the association between MRI-based radiomics features and the TME in GAS. We enrolled 16 patients with GAS treated at our institution. TILs were evaluated by immunohistochemistry, quantifying them with the modified Immunoscore (mIS). Fourteen patients with usual endocervical adenocarcinoma (UEA) served as controls. A total of 1,302 radiomic features were extracted from each primary tumor and peritumoral region on each pre-treatment MRI image. After feature selection within leave-one-out cross-validation (LOOCV), regression models were developed to predict the TME in GAS, followed by exploratory clustering of the mIS groups. GAS exhibited significantly lower T-cell infiltration than UEA, particularly in early-stage tumors. The regression model showed a correlation with observed TIL densities in the full cohort ( r = 0.87, P < 0.001); however, LOOCV showed limited out-of-sample performance ( r = 0.36). Exploratory full-cohort clustering showed 81.3% concordance with the mIS groups (low vs. intermediate-to-high). These findings may provide a basis for the future development of MRI-based radiomic models for assessing the TME in GAS.