부산대학교 도서관

This study aims to analyze the sensitivity of vaginosonography (VGS) and magnetic resonance imaging (MRI) in the preoperative local evaluation of early-stage cervical cancers and to assess their accuracy in the detection of tumors, size of the lesions and stromal invasion by comparing them with the final histopathology report. This single-center study included 56 consecutive patients with cervical cancer who underwent VGS and MRI from November 2012 to January 2021. VGS significantly overestimated the lesion size by 2.7 mm (p = 0.002), and MRI underestimated it by 1.9 mm (p = 0.11). Both MRI and VGS had a good concordance with the pathology report (Cohen's kappa of 0.73 and 0.81, respectively). However, MRI had a false-negative rate (38.1%) that was greater than VGS (0%) in cases of cervical tumor size <2 cm. We found a good concordance between histology and VGS in the stromal infiltration assessment, with 89% sensitivity (95% CI 0.44–0.83) and 89% specificity (95% CI 0.52–0.86). VGS is a simple, inexpensive, widely available, and fast execution method that can complement ultrasound in particular cases and show a good correlation with MRI in the assessment of tumor dimensions, with a better performance in detecting small tumors (<2 cm). [ABSTRACT FROM AUTHOR]