Diagnostic Performance of Magnetic Resonance Imaging for Pediatric Ovarian Neoplasms: A Multi-Institutional Review

Study Objective: To assess the diagnostic performance of MRI to predict ovarian malignancy alone and compared with other diagnostic studies. Methods: A retrospective analysis was conducted of patients aged 2-21 years who underwent ovarian mass resection between 2009 and 2021 at 11 pediatric hospitals. Sociodemographic information, clinical and imaging findings, tumor markers, and operative and pathology details were collected. Diagnostic performance for detecting malignancy was assessed by calculating sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) for MRI with other diagnostic modalities. Results: One thousand and fifty-three patients, with a median age of 14.6 years, underwent resection of an ovarian mass; 10% (110/1053) had malignant disease on pathology, and 13% (136/1053) underwent preoperative MRI. MRI sensitivity, specificity, PPV, and NPV were 60%, 94%, 60%, and 94%. Ultrasound sensitivity, specificity, PPV, and NPV were 31%, 99%, 73%, and 95%. Tumor marker sensitivity, specificity, PPV, and NPV were 90%, 46%, 22%, and 96%. MRI and ultrasound concordance was 88%, with sensitivity, specificity, PPV, and NPV of 33%, 99%, 75%, and 94%. MRI sensitivity in ultrasound-discordant cases was 100%. MRI and tumor marker concordance was 88% with sensitivity, specificity, PPV, and NPV of 100%, 86%, 64%, and 100%. MRI specificity in tumor marker-discordant cases was 100%. Conclusion: Diagnostic modalities used to assess ovarian neoplasms in pediatric patients typically agree. In cases of disagreement, MRI is more sensitive for malignancy than ultrasound and more specific than tumor markers. Selective use of MRI with preoperative ultrasound and tumor markers may be beneficial when the risk of malignancy is uncertain. Concise Abstract: This retrospective review of 1053 patients aged 2-21 years who underwent ovarian mass resection between 2009 and 2021 at 11 pediatric hospitals found that ultrasound, tumor markers, and MRI tend to agree on benign vs malignant, but in cases of disagreement, MRI is more sensitive for malignancy than ultrasound.