Abstract
OBJECTIVE: To analyze outcomes of non-malignant concordant, discordant, and indeterminate results of CT-guided biopsies determined by standardized radiology-pathology concordance evaluation.
METHODS: In this study, consecutive patients undergoing CT-guided omental and mesenteric biopsy between March 2005 and August 2021 were included. A standardized radiology-pathology concordance workflow was implemented in July 2016, with retrospective concordance assessment applied to earlier cases. Concordance between pathology results and imaging findings was assessed by procedural radiologists.
DEFINITIONS: concordant, for malignant biopsy results or benign pathology where imaging findings agree; discordant, if pathology results are not congruent with imaging; and indeterminate, if imaging could be explained by pathology, but could also represent malignancy.
RESULTS: 222 biopsies were included. Pathology showed non-malignant results in 43/222 (19%), further classified by radiology-pathology concordance evaluation as discordant in 24/43 (56%), indeterminate in 8/43 (19%), and concordant in 11/43 (26%). One patient was lost to follow-up in the indeterminate category. The prevalence of malignancy on follow-up was higher in discordant (13/24, 54%) and indeterminate (2/7, 29%) groups vs. concordant cases (0/11, 0%), p < 0.001. There were 15/42 (36%) patients with final diagnosis of malignancy that would have been missed if radiology-pathology concordance evaluation had not been performed. Median time to diagnosis was shorter with repeat biopsy (18 days, IQR 9-34) and surgery (41 days, IQR 17-60) vs. imaging (185 days, IQR 107-239) and clinical follow-up (330 days, IQR 240-374), p < 0.001.
CONCLUSION: Radiology-pathology concordance evaluation in CT-guided omental and mesenteric biopsies showed high malignancy rates in initially non-malignant discordant (54%) and indeterminate (29%) cases. This practice prevented missed cancer diagnoses in 36% of patients with initial non-malignant results.