Front Oncol. 2021 ;11 774459
Background: Perineural invasion (PNI) is associated with a poor prognosis for cervical cancer and influences surgical strategies. However, a preoperative evaluation that can determine PNI in cervical cancer patients is lacking.
Methods: After 1:1 propensity score matching, 162 cervical cancer patients with PNI and 162 cervical cancer patients without PNI were included in the training set. Forty-nine eligible patients were enrolled in the validation set. The PNI-positive and PNI-negative groups were compared. Multivariate logistic regression was performed to build the PNI prediction nomogram.
Results: Age [odds ratio (OR), 1.028; 95% confidence interval (CI), 0.999-1.058], adenocarcinoma (OR, 1.169; 95% CI, 0.675-2.028), tumor size (OR, 1.216; 95% CI, 0.927-1.607), neoadjuvant chemotherapy (OR, 0.544; 95% CI, 0.269-1.083), lymph node enlargement (OR, 1.953; 95% CI, 1.086-3.550), deep stromal invasion (OR, 1.639; 95% CI, 0.977-2.742), and full-layer invasion (OR, 5.119; 95% CI, 2.788-9.799) were integrated in the PNI prediction nomogram based on multivariate logistic regression. The PNI prediction nomogram exhibited satisfactory performance, with areas under the curve of 0.763 (95% CI, 0.712-0.815) for the training set and 0.860 (95% CI, 0.758-0.961) for the validation set. Moreover, after reviewing the pathological slides of patients in the validation set, four patients initially diagnosed as PNI-negative were recognized as PNI-positive. All these four patients with false-negative PNI were correctly predicted to be PNI-positive (predicted p > 0.5) by the nomogram, which improved the PNI detection rate.
Conclusion: The nomogram has potential to assist clinicians when evaluating the PNI status, reduce misdiagnosis, and optimize surgical strategies for patients with cervical cancer.
Keywords: biomarker; cervical cancer; nomogram; perineural invasion; predictive model