Abstract
PURPOSE: Biomarkers estrogen receptor (ER), progesterone receptor (PR), and human epidermal growth factor receptor 2 (HER2) may undergo alteration on reassessment, with significant impact on management. There is a paucity of large-scale paired data on biomarker changes from low- and middle-income countries (LMICs). MATERIALS AND METHODS: We performed a retrospective audit on 1,107 paired samples, wherein biomarkers were performed at least twice between: core needle biopsy (CNB) and upfront resection (category 1, n = 277); CNB and postchemotherapy resection (category 2, n = 104); primary (CNB/resection) and recurrence (local/metastatic; category 3, n = 702); and initial and subsequent distant metastasis (category 4, n = 24). Concordance was noted for individual receptors and surrogate molecular classification (hormone receptor+HER2–, hormonal receptor+HER2+, hormonal receptor–HER2+, triple-negative). RESULTS: Overall concordance for ER, PR, and HER2 was 85.4% ( k value = 0.693), 77.1% ( k value = 0.541), and 93.8% ( k value = 0.827), respectively. For HRs, higher concordance was in category 1 > 3 > 2 > 4 (ER k value = 0.764 > 0.684 > 0.591 > 0.515, respectively; PR k value = 0.68 > 0.495 > 0.482 > 0.329, respectively), while HER2 was relatively constant across categories ( k value range, 0.808-0.882). Molecular classification showed overall 79.5% concordance ( k value = 0.688). Discordance was 27.1% in triple-positive (highest) and 17.7% in HER2+/hormonal receptor– (lowest). Univariate and multivariate analyses showed poorer concordance for therapy ( v no therapy; odds ratio [OR], 0.437 [95% CI, 0.255 to 0.749]; P = .003) and specific/targeted therapy ( v CT alone; OR, 0.126 [95% CI, 0.073 to 0.217]; P = .001). Shorter time interval (6 months' time interval, or poorly fixed material, which is of particular relevance in LMICs.

This publication has 41 references indexed in Scilit: