AJR Am J Roentgenol. 2016 Jun;206(6):1217-21. doi: 10.2214/AJR.15.14540. Epub 2016 Apr 8.
Chalian M1, Del Grande F2, Thakkar RS2, Jalali SF2, Chhabra A3, Carrino JA4.
OBJECTIVE:
The purpose of this article is to evaluate the significance of subspecialty second-opinion consultations for CT and MRI examinations in musculoskeletal (MSK) radiology.
MATERIALS AND METHODS:
All 3165 MSK CT and MRI examinations referred to one academic institution for second-opinion consultation during a 24-month period were reviewed by three MSK-trained radiologists. Outside and inside reports were compared by two independent MSK radiology fellows using a previously published 5-point scale. Clinically important differences (categories 4 and 5) were defined as those likely to change patient management. Statistical comparisons of rates were performed using a chi-square test with Bonferroni corrections. Interobserver reliability was reported using linear weighted kappa statistics and the percentage of agreement.
RESULTS:
Of all second-opinion examinations, 73.5% (2326/3165) had an outside report available for comparison and inclusion in this study. There were 610 of 2326 (26.2%) examinations with clinically important differences. The rate of clinically important discrepant readings was even higher in oncologic cases (36.3%; 331/911). When the final diagnosis was determined from pathology reports performed after internal interpretation, the second-opinion consultation was noted to be correct in 82.0% (334/407) of examinations with category 4 or 5 discrepancies. There was very good agreement (κ = 0.93) in scoring the discrepancies between second-opinion consultants.
CONCLUSION:
The subspecialty second-opinion consultation was more accurate than outside reports in 82.0% of examinations when pathologic confirmation was made. A moderate rate (26.2%) of discrepant interpretations was noted between outside and inside MSK imaging examinations, especially in tumor cases (36.3%). Most discrepancies were in interpreting rather than detecting abnormalities.