Expander-phase morbidity in RFID-enabled two-stage breast reconstruction: A multicenter risk prediction model and clinical nomogram.

BACKGROUND

Two-stage implant-based breast reconstruction (TSBR) using tissue expanders (TE) continues to be a standard procedure. Although TE incorporating non-metallic radiofrequency identification (RFID) ports have been introduced to address the limitations of metallic systems, clinical data are limited, and no validated models exist to guide individualized surgical planning.

OBJECTIVE

To develop a clinical prediction model and risk score for complications during RFID-TE reconstruction.

METHODS

We conducted a multicenter cohort study involving 356 patients undergoing TSBR with RFID-enabled TE. The primary endpoint was overall postoperative complications during the expander phase. Candidate predictors included demographic and treatment-related variables. Independent predictors were identified using multivariable logistic regression. Model performance was assessed using receiver operating characteristic (ROC) analysis and Hosmer-Lemeshow testing. Regression coefficients were converted into an integer-based score to improve risk stratification.

RESULTS

At a mean follow-up of 29.4±16.6 months, complications, most commonly seroma (7.8%), occurred in 21.6% of the patients. Independent predictors of complications included age >50 years (OR: 2.25 years), smoking (OR: 2.00), and radiation therapy (OR: 2.26), whereas use of the submuscular pocket was protective (OR: 0.44). These variables were incorporated into a 0-9 point risk score, which yielded an area under the ROC curve of 0.698 (95% CI: 0.628-0.768; p<0.001), with acceptable calibration (p=0.051). Risk stratification demonstrated complication rates of 8.7 and 39.0% in the low- and high-risk groups, respectively (p<0.001), enabling individualized estimation of expander-phase complication risk during preoperative planning.

CONCLUSION

This study derives the first prediction model specific to RFID-TE reconstruction. By translating independent predictors into a preliminary risk score, the model provides a framework for individualized risk stratification and basis for future internal and external validation.