2 97 CLINICAL IMPLICATIONS AND RELEVANCE - Objective 1: Accurate and objective MRI diagnosis for adenomyosis The diagnostic modality focussed on within this thesis has been the MRI diagnosis of adenomyosis. It should be said that we do not hereby mean to suggest that all patients with suspected adenomyosis should be referred for an MRI. The primary first line diagnostic method used in clinical practice still is, and should be, TVUS diagnosis. TVUS is more widely available, quicker, and cheaper than MRI, and should form the basis of initial clinical work-up. Due to the widely used and accepted MUSA criteria, TVUS has been shown to be reliable for adenomyosis diagnosis in many cases, especially with improving ultrasound quality and the availability of 3D transvaginal ultrasound (29). There also exists a recently developed multivariate prediction model for TVUS adenomyosis diagnosis showing a similar performance to our developed model (184). It cannot go unmentioned however, that the literature that has been published regarding TVUS diagnosis generally relates mainly to TVUS carried out by expert observers in specialised clinics. The diagnostic accuracy here likely does not reflect the accuracy of TVUS diagnosis by every general gynaecologist in daily practice. This leaves room for further improvement and investigation into objective and reproducible diagnostic modalities. . We hence chose to concentrate our research on MRI diagnosis as there are cases where ultrasound remains unclear, for example due to combined pathology, a (relatively) inexperienced observer, or need for extensive preoperative work-up (199,292). In cases such as these, MRIs are often performed, but adenomyosis diagnosis then becomes difficult due to the current lack of accepted diagnostic criteria. For this reason, we chose to attempt to close this knowledge gap by summarising the diagnostic criteria available. Moreover, we also investigated which of these criteria (with symptoms) were most promising for facilitating accurate MRI diagnosis with our multivariate prediction model. Future (prospective) studies should focus on the further validation and implementation of this model for adenomyosis diagnosis. It may also be possible to further improve upon the model, with the addition of further developing diagnostic markers in future, like texture analysis or elastography (293,294). Improving the useability of the model could also involve exploration the use of automated MRI measurement using AI tools for example.
RkJQdWJsaXNoZXIy MTk4NDMw