Histopathological parameters are essential for deciding on adjuvant treatment following breast cancer surgery. We assessed the impact of inter-observer variability on treatment strategy in patients operated for clinically node negative, non-palpable breast carcinomas. In the context of a multicenter randomised controlled trial, clinical and histological data of 310 patients with clinically node negative non-palpable invasive breast cancer were prospectively collected. Histological assessment of the primary tumour and sentinel nodes was first performed in a routine setting, subsequently central review took place. In case of discordance between local en central assessments, we determined the impact on locoregional and systemic treatment strategy. Discordance between local and central review was observed in 13% of the patients for type (kappa 0.60, 95% CI 0.50-0.71), in 12% for grade (k=0.796, 95% CI 0.73-0.86), in 1% for ER status (k=0.898, 95% CI 0.80-1.0), in 2% for PR status (k=0.940 95% CI 0.89-0.99). Discrepancy in the assessment of the sentinel node(s) was seen in 2% of the patients (k=0.954, 95% CI 0.92-0.98). Applying current Dutch Guidelines, central review would have affected locoregional treatment in 2% (7/310), systemic treatment in 5% (16/310) and both in 1% (2/310) of the patients. For the 9 (3%) patients in whom central review would have led to additional systemic treatment, Adjuvant! predicted 10 years mortality and recurrence rate would have decreased with a median of 4.6% and 15%, respectively. Discordance between routine histological assessment and central review of non-palpable breast carcinoma specimens and sentinel nodes was observed in 24% of patients. This inter-observer variation would have impacted locoregional and/or systemic treatment strategies in 8% of the patients.
Keywords: Breast cancer; Histologic assessment; Interobserver variability; Treatment.
© 2013 Elsevier B.V. All rights reserved.