Prospective evaluation of 3T MRI findings for primary adnexal lesions and comparison with the final histological diagnosis

• Published in: Archives of gynecology and obstetrics 2014-02, Vol.289 (2), p.357-364
• Main Authors: Zhang, He, Zhang, Guo-Fu, He, Zhi-Yan, Li, Zheng-Yu, Zhang, Gui-Xiang
• Format: Article
• Language: English
• Subjects: Accuracy; Adnexal Diseases - diagnosis; Adnexal Diseases - pathology; Adult; Contrast Media; Cysts; Cysts - diagnosis; Cysts - pathology; Diagnosis, Differential; Diffusion Magnetic Resonance Imaging; Endocrinology; Female; Gadolinium DTPA; Gynecologic Oncology; Gynecology; Human Genetics; Humans; Magnetic Resonance Imaging - methods; Medicine; Medicine & Public Health; Middle Aged; NMR; Nuclear magnetic resonance; Obstetrics/Perinatology/Midwifery; Ovarian Neoplasms - diagnosis; Ovarian Neoplasms - pathology; Predictive Value of Tests; Prospective Studies; Sensitivity and Specificity; Teratoma - diagnosis; Teratoma - pathology; Tumors
• ISSN: 0932-0067; 1432-0711
• DOI: 10.1007/s00404-013-2990-x

Purpose We prospectively investigated the diagnostic accuracy of magnetic resonance imaging (MRI) at 3.0 Tesla (3T) for the detection of suspected primary adnexal masses in a large cohort of patients. Methods This prospective clinical study included 223 patients with suspected gynaecological disease who were referred for 3T MRI assessments before laparoscopy or laparotomy. Fifty-nine patients were excluded. All detected adnexal pathologies on MRI were categorized into the four groups (endometric cysts, teratomas, benign tumours and malignant tumours). Histological findings were used as the comparative reference standard. As measures to detect or rule out primary adnexal masses, accuracy, sensitivity, specificity, positive predictive values (PPV) and negative predictive values (NPV) were determined by lesion-based evaluations. Results The reference standard method detected 141 primary adnexal lesions in 125 patients. The areas under the receiver operating characteristic curve of the lesion-based evaluations for endometric cysts, teratomas, benign lesions and malignant lesions were 92.8, 93.6, 95.1 and 94.4 %. Lesion-based evaluation yielded an accuracy of 90.3 %, sensitivity of 92.7 %, specificity of 89.3 %, PPV of 77.6 % and NPV of 96.8 % in differentiating malignancies from non-malignant lesions. The diagnostic value of 3T MRI for detecting malignancies was superior to that for benign tumours. Conclusions 3T MRI well categorize the characteristics of primary adnexal lesions and may be a reliable modality for distinguishing malignancies from benign tumours.