Accuracy and clinical impact of MRI in early-stage cervical cancer after cervical conization, a retrospective study.

Abstract

To assess the accuracy of MRI in detecting residual disease after conization in patients with early-stage cervical cancer (CC) and to evaluate the impact of MRI on FIGO staging and surgical management. Between January 2018 and December 2023, a consecutive series of patients with early-stage invasive CC undergoing conization, pre-operative pelvic MRI, and surgery were included in this retrospective study. Two experienced radiologists reviewed the MRI scans for the presence of residual tumor, assessed on T2-weighted, diffusion-weighted, and contrast-enhanced sequences, if available. MRI findings were compared with surgical pathology to evaluate diagnostic performance and clinical impact. A total of 108 patients were included in the study. MRI detected residual disease with an accuracy of 78.7% (95% CI 71.0-86.4), sensitivity of 66%, specificity of 90.9%, positive predictive value of 87.5%, and negative predictive value of 73.5%. The use of contrast agent showed no significant difference in overall accuracy, with an accuracy of 74.5% for contrast-enhanced MRI (CE-MRI) and 83% for non-CE-MRI (p = 0.28), respectively. MRI correctly staged 73.8% of cases, suggested the correct extent of radicality in 79.7% of cases following the 2018 ESGO guidelines, and the appropriate type of hysterectomy (simple vs. radical) in 90.9% of cases following the SHAPE study. Non-CE MRI is a reliable and sufficient tool for detecting residual tumor after conization in early-stage CC, as contrast administration offers no additional diagnostic value. Accurate pre-operative identification of residual disease may refine FIGO staging and assist in tailoring surgical strategies. Question MRI evaluation of residual tumor after cervical conization in early-stage cervical cancer is critical to optimize surgical planning noninvasively and avoid overtreatment. Findings MRI correctly identified residual disease with high specificity (90.9%) and moderate sensitivity (66%); contrast enhancement did not significantly improve diagnostic accuracy. Clinical relevance MRI, particularly non-contrast protocols, reliably excludes post-conization residual disease and aids in selecting appropriate surgical treatment, reducing overtreatment in early-stage cervical cancer.