Atypical Ductal Hyperplasia of the Breast on Core Needle Biopsy - Risk of Malignant Upgrade on Surgical Excision.

Abstract Purpose: Atypical ductal hyperplasia (ADH) is a high risk lesion with an increased risk of developing breast cancer. This study aims to identify factors predictive of malignant upgrade for ADH diagnosed on core needle biopsy (CNB) and to develop a nomogram to facilitate evidence-based decision making.Methods: Retrospective analysis of women with CNB diagnosed ADH at the National Cancer Centre Singapore between 2010 and 2015 was performed. Cox proportional hazards regression was used to identify independent clinical, radiological and histological factors associated with malignant upgrade. A nomogram was constructed and multivariable logistic regression coefficients were used to estimate the predicted probability of upgrade for each factor combination. Combinations with the lowest predicted probabilities (≤5%) were identified as low risk. Model sensitivity, specificity, positive and negative predictive values were assessed.Results: From 2010-2015, 238,122 women underwent screening under the national breast cancer screening programme. 29,564 women were recalled and 5742 CNBs were performed, of which 2686 were performed at NCCS. 88 patients (90 lesions) were diagnosed with ADH. 26 lesions were upgraded to a breast malignancy on excision biopsy. On univariate analysis, presence of a mass on either ultrasound (p= 0.018) or mammogram (p=0.026), presence of mammographic microcalcifications (p=0.047), diffuse microcalcification distribution (p=0.034), mammographic parenchymal density (p=0.008), presence of microcalcifications on biopsy (p=0.037) and three or more separate foci of ADH found on biopsy (p=0.024) were associated with malignant upgrade. Mammographic parenchymal density (Hazard ratio= 0.04, 95% CI 0.005-0.35, p=0.014), presence of a mass on ultrasound (Hazard ratio= 10.50, 95% CI 9.21-25.2, p=0.010) and number of foci of ADH (Hazard ratio = 1.877, 95% CI 1.831-1.920, p=0.002) remained significant on multivariate analysis and were included in the normogram which demonstrated good discrimination with C-statistic of 0.81 [95% CI, 0.74 to 0.88].Conclusion: Our model provides good discrimination of breast cancer risk prediction in patients with ADH on CNB. A subset of women at low risk (<5%) of upgrade to cancer may avoid surgical excision following a core-needle biopsy diagnosis of ADH.