内科: オープンアクセス
オープンアクセス

概要

Soft-Tissue Tumor Reporting and Data System (ST-RADS): MRI Reporting Guideline with Multi-Institutional Validation Study of Musculoskeletal Extremity Tumors

Avneesh Chhabra, Oganes Ashikyan, Raghu Ratakonda, Gitanjali Bajaj, Uma Thakur, Parham Pezeshk, Yin Xi, Marwa Zaid, Alexandra Callan, William Murphy, Rajendra Kumar, Behrang Amini

Background: To our knowledge, no current guideline exists in outlining follow-up or interventional strategies in musculoskeletal soft tissue tumor management. To develop and validate Soft-tissue Tumor Reporting and Data System (ST-RADS) with the hypothesis that the proposed guideline reliably and accurately assists in separating benign from malignant musculoskeletal soft tissue tumors.

Methods: This is a multi-institutional cross-sectional study of soft-tissue masses. An expert consensus agreement was reached for ST-RADS categories using the terminology from WHO classification. Adipocytic tumors, T2-hyperintense and T2-hypointense masses of extremities with a wide spectrum of histologies were assessed. MRI categories were: STRADS 0-incomplete imaging, I-no lesion identified, II-definitely benign, III-probably benign, IV-indeterminate or suspicious for malignancy, V-highly suggestive of malignancy, and VI-known biopsy-proven malignancy or recurrence. Eight readers’ evaluated cases and ICC (Intra-Class Correlation) and AUC (Area Under the Curve) were calculated.

Results: 200 soft tissue masses were tested. There was good inter-reader agreement with ICC (Intra-Class Correlation)=0.72 [95% CI=0.64-0.79] and 0.69 [95% CI=0.59-0.70] for adipocytic and T2-hyperintense, and fair, 0.48 [95% CI=0.35-0.62] for T2-hypointense masses. The sensitivity and specificity for detection of malignancy were 96% and 63%, 93% and 71%, 64% and 84% for adipocytic, T2-hyperintense, and T2-hypointense masses, respectively. The AUC (Area Under the Curve) was 0.79-0.89.

Conclusion: ST-RADS guideline using standardized terminology stratifies musculoskeletal tumors into benign and malignant categories and provides management strategy. This MRI-based guideline is meant to be a "dynamic" document that can be further refined and updated in response to future user feedback and as new scientific data becomes available.