Sonographic prediction of intraductal papillary carcinoma with partially cystic breast lesions

Abstract

Background Intraductal papillary carcinoma (IDPC) is a rare but fatal disease. Preoperative ultrasound diagnosis of IDPC remains challenging and meaningful. The aim of the study was to determine an effective ultrasound model to predict intraductal papillary carcinoma (IDPC) in patients with partially cystic breast lesions on ultrasound. Methods We reviewed female patients with breast nodules who underwent biopsy or surgery between 2004 and 2019, and pathological results were used as the reference standard. We finally included 21 IDPC patients with partially cystic lesions on preoperative ultrasound matched to 40 patients with intraductal papilloma. The association of ultrasound features with IDPC was analysed. Results Posterior echo enhancement (P < 0.001), tumour size (P = 0.002), irregular shape (P = 0.003), wide base (P = 0.003), solid-mainly component (P = 0.013), rich Doppler flow (P < 0.001) and multiple lesions (P = 0.044) were associated with IDPC by univariate analysis. Based on univariate analysis, variables were included in the regression analysis to obtain independent factors. The regression analysis showed that microcalcification, multiple lesions, posterior echo enhancement, wide base of solid components and rich colour Doppler flow were predictors for IDPC (P < 0.001). The collective model of the independent factors (microcalcification, multiple lesions, posterior echo enhancement, wide base of solid components and rich colour Doppler flow) could predict IDPC with an area under the curve (AUC) of 0.99 (95% CI 0.95–1.00). The collective model had a better net benefit demonstrated by the decision curve. Conclusion Ultrasonic features may be an applicable model for predicting IDPC with partially cystic breast lesions on ultrasound and has a better potential to facilitate decision-making preoperatively.