Histology-specific prognostic models for early-stage cervical cancer based on pathologic intermediate-risk factors: a...

Squamous cell carcinoma and adenocarcinoma in early-stage cervical cancer differ in biology and prognosis, yet current guidelines recommend uniform management. This study aimed to develop and internally validate histology-specific prognostic models for squamous cell carcinoma and adenocarcinoma to predict disease-free survival and overall survival, thereby supporting individualized adjuvant therapy for intermediate-risk cervical cancer.

A multi-center retrospective analysis was conducted on patients with early-stage cervical cancer who underwent radical surgery without pathologic high-risk features. Patients were stratified into squamous cell carcinoma and adenocarcinoma cohorts. Histology-specific nomograms for disease-free survival and overall survival were developed using Cox regression and internally validated via 5-fold cross-validation. Model discrimination was assessed using the concordance index (C-index), and prognostic stratification using Kaplan-Meier analysis.

A total of 1053 patients with stage IB to IIA cervical cancer (911 squamous cell carcinoma, 142 adenocarcinoma) treated at 3 tertiary hospitals from 2005 to 2017 were included. Nomograms incorporating lymphovascular space involvement, stromal invasion, tumor size, and adjuvant therapy showed superior accuracy over Sedlis models (C-indices: 0.81/0.80 for disease-free survival/overall survival in squamous cell carcinoma; 0.91/0.96 for disease-free survival/overall survival in adenocarcinoma). Risk scores stratified patients into distinct prognostic sub-groups. In squamous cell carcinoma, high-risk patients had higher recurrence and mortality rates than low-risk patients (5-year recurrence rate: 11.7% vs 1.6%; mortality rate: 7.7% vs 0.6%). In adenocarcinoma, high-risk patients showed worse outcomes (5-year recurrence rate: 13.3% vs 0.0%; mortality rate: 6.1% vs 0.0%). Kaplan-Meier analysis confirmed significant survival differences across risk groups in both sub-types (all p < .05).

Histology-specific nomograms incorporating intermediate-risk factors and adjuvant therapy effectively stratified prognostic sub-groups in early-stage cervical cancer, and may help inform risk-adapted post-operative management. Prospective external validation is needed for broader application.