Can MRI Be Used as a Sole Diagnostic Modality in Determining Clinical Stage in Cervical Cancer?

Abstract

Objective The objective of this study was to compare staging by MRI to clinical staging in patients with cervical cancer and to determine the histological accuracy of staging by MRI and examination under anesthesia (EUA) in early stage disease. Methods This was a retrospective cohort study of patients diagnosed with cervical cancer between 2010 and 2020 at the Radboud University Medical Centre, the Netherlands. Pretreatment stage (FIGO 2009) by MRI was compared with staging by EUA. Diagnostic accuracy in terms of sensitivity, specificity, positive, and negative predictive value was calculated for MRI and EUA in patients undergoing surgery (early stage disease) with histological results as a reference standard. Results A total of 358 patients were included in the study and MRI-based stage differed from EUA stage in 30.7%. In 12.3% this meant a discrepancy in treatment assignment between MRI and EUA. Diagnostic accuracy of MRI in terms of sensitivity and specificity for detecting early stage disease was comparable to EUA in surgical patients. Further analyses showed that premenopausal status, early stage disease and a tumor diameter of <2 cm were associated with improved comparability of MRI and EUA (98%). Conclusion There is still a large discrepancy between MRI and EUA. In patients with suspected early stage disease, diagnostic accuracy of MRI is similar to EUA, especially for premenopausal women with early stage disease and a tumor diameter of <2 cm.

Systematic Review—Role of MRI in Cervical Cancer Staging

“…All the radiologists were accredited core members of the gynae-oncology multi-disciplinary team and/or PET/CT experts. Smits et al [ 21 ] reported MRI assessed by experienced radiologists, with the type not specified. Anfinan [ 24 ] and Zhu et al [ 26 ] did not specify the experience of their radiologist.…”

“…Three articles reported on stromal invasion; however, their primary outcomes were heterogenous, which reduced the comparability and we were unable to pool ( Table 1 ). Smit et al investigated the performance of MRI in early disease staging (FIGO 2009 stage IA1, IA2, IB1 and IIA1), reporting 94.8% sensitivity, 38.5% specificity, 94.8% PPV and 38.5% NPV [ 21 ]. Ran et al investigated the ability of MRI in detecting FIGO 2018 stage IB and IIA, which showed an accuracy, sensitivity and specificity of 88.6%, 98.8% and 97.0% for stage IB and 94.8%, 100% and 95.5% for stage IIA [ 22 ].…”

“…Five articles reported on parametrial invasion ( Table 1 ) [ 6 , 21 , 22 , 23 , 24 ]. Four of these articles used parametrial invasion as a measurable outcome, and for the reference standard, three of these articles used histopathology and one used clinical examination.…”

“…Smits et al reported a sensitivity, specificity, PPV and NPV of 33.3%, 96.3%, 25% and 97.5%, respectively, for MRI in detecting parametrial involvement when compared to histopathology [ 21 ]. This was the lowest reported sensitivity but the highest specificity for detecting parametrial invasion.…”

“…Our review initially aimed at looking at the diagnostic accuracy of MRI in determining the tumor depth and size, which can be used preoperatively to predict a high GOG score that would direct treatment toward chemoradiation therapy as this has been shown to have improved outcomes compared to upfront surgery with subsequent adjuvant chemoradiation therapy [ 27 ]. Of the publications from the last 5 years, only 2 articles directly investigated the diagnostic accuracy of MRI in detecting early-stage disease [ 21 ] and deep stromal invasion ( Table 2 ) [ 6 ]. A third article reported on the diagnostic ability of MRI in detecting FIGO 2018 Stage IB and IIA cervical cancer rather than measuring stromal invasion itself [ 22 ].…”