Objective: This study aimed at comparing the reliability of diagnostic fluid hysteroscopy, transvaginal sonography (TVS), and magnetic resonance imaging (MRI) to assess pre-operatively the presence of cervical involvement by endometrial carcinoma.
Methods: Cervical involvement was assessed by diagnostic fluid mini-hysteroscopy, TVS and MRI before surgery in 100 patients with histological diagnosis of endometrial carcinoma. Results were compared with pathological examination on surgical specimen. The sensitivity, the specificity, the positive and negative predictive values, the accuracy, the positive and negative likelihood ratios (LR) of the three techniques for recognizing the cervical involvement by the carcinoma were calculated.
Results: At histology cervical involvement was found in 15 cases. Compared to TVS and MRI, hysteroscopy showed the highest sensitivity (0.53, 0.67 and 0.93, respectively). The specificity of MRI was significantly higher than both hysteroscopy and TVS (0.95, 0.88 and 0.82, respectively). The diagnostic accuracy of hysteroscopy (0.89) and MRI (0.91) was similar and significantly higher than TVS (0.78). The LR for a positive result of MRI was 14.16, that was 2.08 and 4.68 times higher than that of hysteroscopy and TVS, respectively.
Conclusions: In conclusion, this study demonstrates that in women with endometrial carcinoma the exclusion of cervical canal involvement at hysteroscopy is more reliable than at MRI and TVS while MRI is the most reliable technique for predicting cervical involvement. In the pre-surgical work-up of patients affected by endometrial carcinoma hysteroscopy and MRI are both useful for staging and planning the correct surgical strategy.