A Clinical Prediction Model for Prognosticating Salvage of the Infected Implant in Alloplastic Breast Reconstruction.

Abstract

Periprosthetic infection (PPI) is a complication of alloplastic breast reconstruction that can result in reconstructive failure, delay of adjuvant therapies, and reoperation. Despite multiple observational cohort studies, optimal PPI management remains unclear. The aim of this study was to develop a clinical prediction tool to guide clinical management. A multicenter retrospective cohort study was conducted. Consecutive patients with breast cancer who underwent immediate alloplastic breast reconstruction between 2010 and 2020 were included. Collected data included patient, oncologic, and reconstructive factors for patients whose postoperative course was complicated by cellulitic infection or PPI. Two models were created for prediction of progression from cellulitis to PPI and from breast infection to reconstructive failure. A total of 1438 patients (2165 breasts) were included. The incidence of infection was 7.1% ( n = 145). Implant reconstruction was salvaged in 67.1% ( n = 104) of cases. The first model, predicting progression from cellulitis to PPI, had good predictive accuracy, with an area under the receiver operating characteristic curve of 0.61 (95% CI, 0.53 to 0.70; P < 0.001). The second model, predicting progression to reconstructive failure, also had good predictive accuracy, with an area under the receiver operating characteristic curve of 0.79 (95% CI, 0.71 to 0.87; P < 0.001). This study presents novel clinical prediction tools with good predictive accuracy. Application of these prediction tools can assist the clinician in making evidence-based management decisions for treatment of PPI after alloplastic breast reconstruction. Future research will aim to prospectively validate treatment algorithms and improve reconstructive success.