50 Characteristics of Included Studies: The study characteristics and outcomes for the included studies are summarized in Tables 2.S5 -2.S7 of the supplementary file. The MRI sequences most often implemented in order to assess adenomyosis on MRI were T1- and T2- weighted MRI. Eight studies reported using 3.0 Tesla coil MRI, the remaining studies used 1.0 or 1.5 Tesla coil MRI. Methodological quality of included studies: Only the studies investigating diagnostic accuracy were assessed for methodological quality. A graphical summary of the quality assessment is shown in Figure 2.2, as well as the assessment per included study in Figure 2.3. In the domain of patient selection, two studies were deemed to have a high risk of bias due to their retrospective design and/or unclear exclusion criteria (116,117). For the index test domain (MRI), two studies had a high risk of bias, as no definition of adenomyosis prior to MRI evaluation was reported (118,119). As for the reference standard domain, many studies did not clearly report if pathologists were blinded. Two studies were deemed to have a high risk of bias as in one (120) the assessment of the reference test was not blinded, and in the other (121) the reference diagnosis was only made based on myometrial biopsy instead of hysterectomy. For patient flow and timing, most studies did not provide enough information to assess this domain properly. Hamimi et al. (117) had a high risk of bias on this domain as not all patients received the same reference standard diagnosis. Two studies (117,121) were arguably less applicable with regards to the analysis of diagnostic accuracy. Hamimi et al. did not compare the index test to histopathological diagnosis in many cases, and Phillips et al. only investigated the diagnostic performance of MRI in relation to adenomyomas, and not adenomyosis generally. Because of their low applicability and quality, the results of these two studies were not included in the meta-analysis. Complete details per study can be found in the supplementary file (Tables 2.S8-2.S11).
RkJQdWJsaXNoZXIy MTk4NDMw