Structured MRI reporting in patients with deep pelvic endometriosis: performance and clinical impact of the deep pelvic endometriosis index (dPEI).

PURPOSE: To evaluate the feasibility, interobserver reproducibility, reporting efficiency, and perceived clinical impact of the Deep Pelvic Endometriosis Index (dPEI) when implemented as part of structured MRI reporting in routine clinical practice.

MATERIALS AND

METHODS

This retrospective single-center study included 100 consecutive women with surgically confirmed deep pelvic endometriosis who underwent preoperative 3-Tesla pelvic MRI. Three radiologists with different levels of experience (2, 4, and 20 years) independently reviewed all examinations using a structured MRI report based on the dPEI. Interobserver agreement for the global score was assessed using the intraclass correlation coefficient (ICC), while compartment-specific agreement was evaluated using Cohen's and Fleiss' κ statistics. Reporting time was recorded for each reader. Perceived clinical utility and practical limitations were assessed through a structured questionnaire administered to 10 clinicians and 10 surgeons involved in multidisciplinary endometriosis care.

RESULTS

The dPEI was applicable across all cases and readers. Interobserver agreement for the global score was low-to-moderate (ICC = 0.44) and was influenced by reader experience, with higher concordance between senior and intermediate radiologists than between pairs including the junior reader. Agreement was higher in central compartments and lower in lateral and extrapelvic compartments. Reporting time decreased with increasing reader experience and demonstrated a clear learning effect, remaining below 30 min for all readers. Clinicians rated the dPEI as useful for disease stratification and multidisciplinary communication, whereas surgeons reported limited usefulness for advanced surgical planning, mainly due to the lack of quantitative descriptors.

CONCLUSION

The dPEI is a feasible and intuitive MRI-based staging system whose performance is influenced by reader experience, while remaining easy to apply and associated with a rapid learning curve. These characteristics support its use as a complementary structured reporting tool in routine clinical practice and suggest a potential role as an educational and training instrument in deep pelvic endometriosis imaging.