Related Experiment Videos
Characterization of hemorrhagic adnexal lesions with MR imaging: blinded reader study
E Outwater1, M L Schiebler, R S Owen
1Department of Radiology, Hospital of the University of Pennsylvania, Philadelphia 19104.
Abstract:
To assess the performance of published criteria for differentiation of endometriomas from other hemorrhagic adnexal lesions with magnetic resonance (MR) imaging, 74 lesions with pathologic proof in 46 patients were evaluated by means of a 1.5-T system with use of conventional T1-weighted and either T2-weighted spin-echo (SE) or fast SE sequences. The MR images were independently presented to each of three blinded readers, and receiver operating characteristic (ROC) curve analysis was performed. The criterion of a cyst with high signal intensity on short repetition time (TR)/echo time (TE) images and shading on long TR/TE images had a mean sensitivity of 68%, mean specificity of 83%, and mean accuracy of 76% for diagnosis of endometrioma. Ancillary findings previously reported to be suggestive of endometrioma (eg, low-signal-intensity rims, adhesions, bowel tethering, and implants) were proved inaccurate and showed low interobserver agreement. ROC analysis showed similar area measurements among the three radiologists for diagnosis of endometrioma. MR imaging had only moderate accuracy in distinction of endometriomas from other hemorrhagic adnexal lesions.
Insights
Magnetic resonance (MR) imaging shows moderate accuracy in differentiating endometriomas from other hemorrhagic adnexal lesions. Key criteria achieved 76% accuracy, but ancillary findings were unreliable.
Area of Science:
- Radiology
- Medical Imaging
- Gynecologic Imaging
Background:
- Endometriomas are common gynecologic conditions.
- Accurate differentiation from other hemorrhagic adnexal lesions is crucial for appropriate management.
- Magnetic resonance (MR) imaging is a key modality for evaluating adnexal lesions.
Purpose of the Study:
- To evaluate the diagnostic performance of established MR imaging criteria for endometriomas.
- To assess the reliability of ancillary MR imaging findings in endometrioma diagnosis.
- To compare the accuracy of MR imaging in distinguishing endometriomas from other hemorrhagic adnexal lesions.
Main Methods:
- Retrospective evaluation of 74 pathologically proven lesions in 46 patients using a 1.5-T MR system.
- Utilized conventional T1-weighted and T2-weighted spin-echo (SE) or fast SE sequences.
- Independent assessment by three blinded readers with receiver operating characteristic (ROC) curve analysis.
Main Results:
- The primary criterion (high signal intensity on short TR/TE images and shading on long TR/TE images) yielded a mean sensitivity of 68%, specificity of 83%, and accuracy of 76%.
- Ancillary findings (e.g., low-signal-intensity rims, adhesions) demonstrated low interobserver agreement and were inaccurate.
- ROC analysis indicated moderate overall accuracy for MR imaging in differentiating endometriomas from other hemorrhagic adnexal lesions.
Conclusions:
- Published MR imaging criteria offer moderate accuracy for endometrioma diagnosis.
- Ancillary findings are unreliable and should not be solely relied upon for diagnosis.
- MR imaging has limitations in definitively distinguishing endometriomas from other hemorrhagic adnexal lesions.