Related Experiment Video
Updated: Sep 3, 2026
Establishing a Competing Risk Regression Nomogram Model for Survival Data
Published on: October 23, 2020
A novel bivariate nomogram for predicting sequela vaginal lesions after surgery in patients with HPV-associated
Yueheng Liu1, Yue Shi1, Yunqiang Zhang1
1Department of Gynecology Oncology, The Obstetrics and Gynecology Hospital of Fudan University, Shanghai, China.
Objective:
To study and predict the risk of further vaginal lesions after surgery in patients with HPV-associated cervical cancer.
Methods:
Medical records of women who underwent surgery for cervical cancer between January 2018 and December 2022 at the Obstetrics and Gynecology Hospital of Fudan University were analyzed. Incidence and genotype of persistent HPV infection were recorded and patients with further vaginal lesions were analyzed for clinicopathological risk factors. Of the patients, 70% were randomly grouped into a training cohort, and predictive prognostic models for vaginal lesions were constructed through machine learning. The model with the highest area under the receiver operator curve (AUC) was screened out in the testing cohort. The nomogram and its calibration curve presented the risk of sequela vaginal lesions. R 4.2.0 software was used for all data processing.
Results:
Within five years after surgery, 29.94% of patients remained persistently infected with HPV, with annual rates fluctuating around 22%. In addition, 10.2% of patients were diagnosed with vaginal intraepithelial neoplasia (VaIN), and 320 cases (78.35%) were low-grade squamous intraepithelial lesions (LSIL). The annual incidence of vaginal lesions decreased gradually from 6.97% in the first year (Y1) to 2.96% at year 5 (Y5). Ovarian preservation (OP) during hysterectomy and adenocarcinoma histology were found to be protective from further vaginal lesions, while elder age, FIGO stage II, and positive vaginal incision margin were significant risk factors. As for persistent HPV infection, both single and multiple genotype remarkably increased the risk of vaginal lesions, and a α-9 HPV infection (OR = 18.20) brought higher risk than non-α-9 HPV (OR = 11.76). Then we built three predictive models; multiple logistic regression was optional, with its AUC at 0.7955 in the ROC curve.
Conclusion:
The predictive model constructed in our study could identify populations at high risk of vaginal lesions and precisely guide clinical interventions.
More Related Videos
06:46Competing-Risk Nomogram for Predicting Cancer-Specific Survival in Multiple Primary Colorectal Cancer Patients after Surgery
Published on: September 27, 2024
07:13Comparison of Predictive Performance of Three Lymph Node Staging Systems in Colorectal Signet Ring Cell Carcinoma Based on Machine Learning Model
Published on: April 18, 2025
