To explore the characteristics of automatic breast volume scanner (ABVS) imaging that affect the difference of molecular subtypes between core needle biopsy and surgical specimens.

A retrospective analysis was performed on 63 breast cancer patients (63 lesions) who underwent core needle biopsy (CNB) and were examined by ABVS before surgery at the Jinling Hospital between November 2014 and October 2020. The differences in the ABVS imaging characteristics between the discrepancy group (molecular subtypes changed before and after surgery) and the consistency group (the pathological results were consistent before and after surgery) were compared. Logistic regression was used to analyze the independent factors affecting the difference in molecular subtypes between before and after surgery.

According to the surgical pathology, there were 15 (23.8%) luminal A, 39 (61.9%) luminal B, 6 (9.5%) HER-2, and 3 (4.8%) triple-negative subtypes. Among the 63 cases, there were 14 lesions that had different pathological results between CNB and surgical specimens. Significant differences were found between the discrepancy group(n=14) and the consistency group (n=49) in the largest diameter, boundary, morphology, margin, and retraction phenomenon (P＜0.05). The results of Logistic regression analysis showed that retraction phenomenon (odds ratio [OR]=0.113, 95% confidence interval [CI]: 0.015-0.86), indistinct boundary (OR=0.117, 95%CI: 0.015-0.912), and irregular morphology (OR=0.171, 95%CI: 0.029-0.992) were independent protective factors for the difference of molecular subtypes between before and after surgery, while spiculated margin (OR=12.576, 95%CI: 1.783-88.717) was the independent risk factor (P＜0.05).

ABVS can help doctors to perform CNB better, reduce the difference with postoperative molecular typing, improve the accuracy of CNB, and provide reliable diagnostic basis and reference information for clinicians.