Clinical Correlates and Nonlinear Effects Associated With Short-Term Surgical Outcomes After Breast-Conserving Surgery for Breast Cancer: A Single-Center Retrospective Study.

Abstract

Aim: To explore clinical correlates of short-term surgical outcomes after breast-conserving surgery for breast cancer, and to examine whether the association between preoperative ultrasound tumor size and these outcomes is nonlinear.

Methods: This single-center retrospective cohort study included 150 patients with breast cancer who underwent first breast-conserving surgery with curative intent between January 2021 and December 2025. Preoperative clinical, imaging, core needle biopsy, and laboratory data were collected. The primary outcome was a short-term adverse surgical outcome, defined as reoperation for margin-related reasons within 90 days, a strict positive margin, or a clinically meaningful close/indeterminate margin or suspected residual disease identified on postoperative pathology or imaging-pathology review when available, according to documented postoperative margin-related assessment. All association analyses were performed using Firth penalized logistic regression. A prespecified multivariable model was constructed with preoperative ultrasound tumor size as the main exposure, and restricted cubic spline (RCS) analysis was used as an exploratory analysis to examine potential nonlinearity.

Results: Among 150 patients, 15 (10.0%) experienced a short-term adverse surgical outcome. Compared with the no-adverse-outcome group, patients with the adverse outcome had larger preoperative ultrasound tumor size and a higher frequency of multifocal/multicentric disease on imaging. Univariable Firth regression showed that each 5 mm increase in preoperative ultrasound tumor size was associated with the adverse outcome (odds ratio (OR) = 2.32, p < 0.001). Multifocal/multicentric disease on imaging was also associated with the outcome (OR = 4.68, p = 0.026), whereas a ductal carcinoma in situ (DCIS) component on core biopsy showed borderline statistical significance (OR = 2.87, p = 0.050). In the prespecified primary model adjusted for mammographic microcalcifications and a DCIS component on core biopsy, preoperative ultrasound tumor size remained associated with the outcome (OR = 2.16, p = 0.001). RCS analysis showed an overall association and suggested potential nonlinearity, but the spline findings were exploratory. Exploratory subgroup analyses suggested a stronger association among patients without mammographic microcalcifications and a similar exploratory pattern among those without a DCIS component on core biopsy.

Conclusions: Preoperative ultrasound tumor size may be associated with poorer short-term surgical outcomes, and this association may not be fully explained by a linear relationship. Multifocal/multicentric disease on imaging and a DCIS component on core biopsy may also be relevant risk-related features. For lesions characterized by microcalcifications or accompanied by a DCIS component, preoperative evaluation based on ultrasound tumor size alone may be insufficient. Therefore, imaging findings and pathological information should still be interpreted in an integrated manner. These exploratory findings require validation in larger cohorts.