The role of infiltrating immune cells within the tumor microenvironment has received considerable attention, but their function in cervical cancer remains to be elucidated; thus, comprehensive evaluation of their predictive value is needed. Using cervical cancer samples from 406 patients, immune cell infiltration was evaluated via immunohistochemistry. CD3+, CD4+, CD8+, CD20+, CD57+, CD68+, and CD163+ cell infiltration was compared in samples from adjacent tissues and the tumor center. The associations between immune cell distributions in the tumor center, clinicopathological features, and prognosis were correlated among immune cell types. Using three immune features, an immune model was constructed based on a Cox regression analysis with the least absolute shrinkage and selection operator (lasso) penalty to derive immune risk scores. Immune cells that infiltrated the tumor center correlated with clinicopathological characteristics and prognosis. The immune risk scores were an independent prognostic indicator and were found to predict cervical cancer prognosis as well as the effects of chemoradiotherapy. We classified patients into either high- or low-risk subgroups (namely CD4+highCD163+highCD57+low and CD4+lowCD163+lowCD57+high, respectively) based on their immune scores. Significant differences were found in the 3-year overall survival of patients with high- and low-risk scores (83.0% vs. 96.6%; P < 0.001). High immune risk scores resulted in decreased overall survival for patients in stages IB1+IIA1, IB2+IIA2, and IIB-IV (P = 0.001, P = 0.008, and P = 0.044, respectively). Overall survival was significantly worse following chemoradiotherapy in high-scoring patients in stages IB1+IIA1 and IB2+IIA2 (P = 0.001 and P=0.008, respectively). Moreover, overall survival was significantly worse after radiotherapy or chemotherapy in high-scoring patients in stage IB1+IIA1 (P = 0.03). Our work reveals that the distribution of infiltrating immune cells affects their function in cervical cancer. Our tumor center-centric immune model effectively predicted survival, suggesting its potential use in identifying suitable candidates for chemoradiotherapy.
