OBJECTIVE: The purpose of this article is to evaluate the concordance between tumor grade found on ultrasound-guided core biopsies of invasive ductal carcinomas of the breast and subsequent excision specimens. MATERIALS AND METHODS: We retrospectively studied 300 consecutive invasive ductal carcinomas (274 women) that were biopsied under sonographic guidance, using 14-gauge core needles exclusively, and that were subsequently excised surgically. A minimum of four cores were taken per lesion. Core biopsy grades were compared with final surgical grades (reference standard). Tumor grade was assigned using the standard modified Scarff-Bloom-Richardson system. The agreement rate was expressed in percentages and in kappa statistics; the rates of overestimation and underestimation were also assessed. The correlation between tumor size (small, ≤ 0.5 cm; medium, 0.6-2.4 cm; and large, ≥ 2.5 cm) and agreement rate was also evaluated. RESULTS: The overall agreement between core biopsy and surgical pathology grade was 69% (simple κ = 0.46; 95% CI, 0.36-0.54). Agreement by biopsy grade was 86% (55/64) for grade 3, 66% (118/180) for grade 2, and 55% (23/42) for grade 1. Core biopsy underestimated 24% (70/286) and overestimated 7% (20/286) of the lesions. When discordant, core biopsy differed from excision by no more than one grade. Large tumors were more likely to show underestimation rather than overestimation when discordant (rate of underestimation, 92% for large, 81% for medium, and 33% for small tumors; p < 0.0031). CONCLUSION: Ultrasound-guided core biopsy accurately predicts high-grade breast tumors but is moderately accurate for lower-grade lesions. Large tumor size negatively impacts the accuracy of tumor grade found on biopsy and is associated with underestimation.
OBJECTIVE: The purpose of this article is to evaluate the concordance between tumor grade found on ultrasound-guided core biopsies of invasive ductal carcinomas of the breast and subsequent excision specimens. MATERIALS AND METHODS: We retrospectively studied 300 consecutive invasive ductal carcinomas (274 women) that were biopsied under sonographic guidance, using 14-gauge core needles exclusively, and that were subsequently excised surgically. A minimum of four cores were taken per lesion. Core biopsy grades were compared with final surgical grades (reference standard). Tumor grade was assigned using the standard modified Scarff-Bloom-Richardson system. The agreement rate was expressed in percentages and in kappa statistics; the rates of overestimation and underestimation were also assessed. The correlation between tumor size (small, ≤ 0.5 cm; medium, 0.6-2.4 cm; and large, ≥ 2.5 cm) and agreement rate was also evaluated. RESULTS: The overall agreement between core biopsy and surgical pathology grade was 69% (simple κ = 0.46; 95% CI, 0.36-0.54). Agreement by biopsy grade was 86% (55/64) for grade 3, 66% (118/180) for grade 2, and 55% (23/42) for grade 1. Core biopsy underestimated 24% (70/286) and overestimated 7% (20/286) of the lesions. When discordant, core biopsy differed from excision by no more than one grade. Large tumors were more likely to show underestimation rather than overestimation when discordant (rate of underestimation, 92% for large, 81% for medium, and 33% for small tumors; p < 0.0031). CONCLUSION: Ultrasound-guided core biopsy accurately predicts high-grade breast tumors but is moderately accurate for lower-grade lesions. Large tumor size negatively impacts the accuracy of tumor grade found on biopsy and is associated with underestimation.