Sonoelastography of Musculoskeletal Soft Tissue Masses: A Pilot Study of Quantitative Evaluation.

Abstract

To evaluate quantitative sonoelastography of benign and malignant musculoskeletal soft tissue masses. We conducted a prospective study of 50 patients from a specialist sarcoma center who had extremity soft tissue masses referred for biopsy. After consent, the quantitative shear wave velocity (meters per second) was measured in longitudinal and transverse planes (3 readings in each plane and mean calculated). All masses subsequently underwent biopsy, excision, or both, with the histologic diagnosis taken as the reference standard. At a subsequent sitting, all anonymized B-mode sonograms were scored independently by 2 radiologists as benign or malignant with agreement by consensus if necessary. Of the 50 masses, 15 were malignant and 35 benign. Nine masses had incomplete velocity readings. Intraclass correlation coefficients for intra-reader reliability of velocity measurements were highly repeatable. There was preliminary evidence that the longitudinal shear wave velocity of malignant masses was on average 30% slower than that of benign masses (P< .10). Longitudinal and transverse shear wave measurements were moderately associated with each other (P = .003). There was no evidence that shear wave velocity varied with patient age, sex, or mass volume. For B-mode assessment of malignancy, sensitivity (Wilson 90% confidence interval) was 73.3% (52.1%, 87.4%), and specificity was 77.1% (63.8%, 86.6%). Interobserver agreement was substantial (κ= 0.86). Four of 15 malignant masses (26.6%) were incorrectly classified as benign on B-mode assessment (all grade 1 liposarcomas). These data suggest that shear wave velocity measurement is reproducible and that malignant masses may have slower longitudinal shear wave velocities than benign masses. The sample size of this pilot study precludes adjusted analysis but should form the basis for larger study designs.