COMPARATIVE DIAGNOSTIC PERFORMANCE OF TRANSVAGINAL ULTRASONOGRAPHY AND MAGNETIC RESONANCE IMAGING IN ADENOMYOSIS: AN EVIDENCE-INTEGRATED REVIEW OF ACCURACY, REAL-WORLD EVIDENCE, AND CLINICAL IMPLEMENTATION

Abstract

 Adenomyosis  is  a  heterogeneous  uterine  disorder  characterized  by  endometrial-like tissue within the myometrium. Diagnostic uncertainty has consequences  for pain management, heavy menstrual bleeding, infertility counselling, and selection of  uterus-sparing  therapy.  Objective:  To  critically  compare  transvaginal  ultrasonography  (TVUS) and magnetic resonance imaging (MRI), and to establish a rigorous framework  for evaluating diagnostic performance in routine care. Methods: This integrative evidence  review  prioritizes  comparative  diagnostic  meta-analyses,  the  revised  Morphological  Uterus  Sonographic  Assessment  (MUSA)  consensus,  clinical  guidance,  and  relevant  methodological and regulatory publications. We distinguish published empirical findings from  proposed  analytical  protocols;  no  original  patient  cohort,  new  meta -analysis,  or  causal treatment-effect analysis was performed. Results: In a 2023 head-to-head metaanalysis encompassing six studies and 595 women undergoing both modalities before  hysterectomy histopathology, pooled sensitivity and specificity were 75% and 81% for  TVUS  versus  69%  and  80%  for  MRI,  without  statistically  significant  modality  differences (P = 0.7509); between-study heterogeneity and selection bias limit universal 
extrapolation.  Revised  MUSA  terminology  separates  direct  signs,  including  intramyometrial cysts, echogenic islands, and subendometrial lines or buds, from indirect  myometrial  responses.  MRI  improves  anatomical  mapping  when  sonography  is  technically  constrained  or  complex  coexisting  pathology  requires  clarification,  but  junctional-zone thickness thresholds alone are insufficiently robust across physiological  contexts. Routine-care datasets may assess diagnostic pathways, interval to treatment, and  downstream  outcomes,  provided  verification  bias,  indication  bias,  missingness,  and  treatment  changes  are  addressed.