Radiation and Reconstructive Failure: Long-Term Outcomes in Two-Stage Flap Versus Implant Breast Reconstruction

Authors: Shammas RL, Wang J, Boe LA, Levy J, Graziano FD, Allen RJ Jr, Cordeiro PG, Disa JJ, Mehrara BJ, Nelson JA

Affiliation: Memorial Sloan Kettering Cancer Center, New York, NY

Journal: Annals of Surgical Oncology, March 2026

PMID: 41191309

Key takeaways

• In radiated patients, those who underwent implant reconstruction had substantially greater predicted chance of reconstructive failure than flap reconstruction: 12% vs 1.4% at 72 months.

• Implant reconstruction was associated with a 7.57-fold higher adjusted hazard of failure compared with flap reconstruction.

• Three-year BREAST-Q satisfaction with breasts favored flap reconstruction by 8 points, exceeding the 4-point minimal clinically important difference used in the study..

• All patients began with the same initial tissue-expander stage, during which 5.4% of the overall cohort experienced reconstructive failure before definitive implant or flap reconstruction.

Background

Radiation increases complications after breast reconstruction, particularly with prosthetic reconstruction. Long-term comparative data on reconstructive failure and patient-reported outcomes after a tissue-expander-first strategy followed by either implant or flap reconstruction remain limited.

Objective

Compare long-term reconstructive failure and patient-reported outcomes between two-stage implant and flap reconstruction in patients receiving radiation therapy.

Methods

• Design: Single-institution retrospective cohort study at Memorial Sloan Kettering Cancer Center, 2017 to 2023.

• Population: 1,210 patients receiving immediate tissue-expander (TE) placement and radiation.

  • Median age: 48 years

  • Median BMI: 23.7 kg/m²

  • Bilateral reconstruction: 57%

  • Radiation after TE placement: 90%

  • Median follow-up: 49.8 months

• Definitive reconstruction: Of 1,210 patients undergoing TE placement, 930 proceeded to implant reconstruction and 215 to flap reconstruction; 65 experienced TE-stage failure.

• Exclusions: Immediate flap, direct-to-implant reconstruction, delayed TE placement, mixed/hybrid bilateral reconstruction, and pedicled flap reconstruction.

• Primary endpoint: Time to reconstructive failure, defined as TE or implant loss without replacement, total flap loss, or implant-to-flap conversion because of significant pain, capsular contracture, or aesthetic dissatisfaction.

• Secondary endpoints: Surgical complications and BREAST-Q patient-reported outcomes.

• Statistics: Wilcoxon rank sum test for continuous variables; Chi-squared or Fisher’s exact test for categorical variables; Kaplan-Meier curves to estimate cumulative incidence of long-term reconstructive failure; log-rank test to compare groups; multivariable Cox proportional hazards regression to evaluate association between reconstruction type and reconstructive failure; multivariable-adjusted linear mixed effects model for longitudinal BREAST-Q analysis. Statistical significance was P < 0.05.

Results

• Overall reconstructive failure: 157/1,210 patients (13%).

  • TE stage: 65/1,210 (5.4%)

  • Flap stage: 3/215 (1.4%)

  • Implant stage: 89/930 (9.6%)

• Among implant failures, 51 (5.5%) patients had the implant removed without replacement and 38 (4.1%) converted to flap reconstruction.

• Most common complications across the entire reconstructive course were infection (15%), seroma (9.7%), hematoma (3.8%), and prosthetic exposure (3.8%).

• Most common contributing factors among patients who experienced reconstructive failure were infection (34.4%), prosthetic exposure (31.8%), capsular contracture (11.5%), aesthetic dissatisfaction (9.6%), and chronic implant-related pain (9.6%).

• Estimated 72-month failure: 12% for implant reconstruction versus 1.4% for flap reconstruction (log-rank p= 0.003).

• After multivariable adjustment, implant reconstruction was associated with markedly increased failure versus flap reconstruction: HR 7.57 (95% CI 2.36-24.2; p< 0.001).

• Infection independently increased failure risk: HR 2.61 (95% CI 1.58-4.31; p < 0.001).

• BREAST-Q satisfaction with breasts at 3 years: flap 65 vs implant 57 (P = 0.014), an 8-point difference exceeding the 4-point minimal clinically important difference; after adjusting for covariates, still favored flap reconstruction for satisfaction with breasts (β = -6.5 for implant vs flap; 95% CI -9.4 to -3.6; p < 0.001).

• BREAST-Q physical well-being of chest at 3 years: flap 68 vs implant 76 (P=0.035); however, after adjusting for covariates, no significant differences noted (β = 0.48 for implant vs flap; 95% CI -2.4 to 3.3; p = 0.7).

Conclusion

Among patients undergoing radiation and an initial tissue-expander reconstruction, definitive implant reconstruction was associated with higher predicted long-term reconstructive failure and lower breast satisfaction than definitive flap reconstruction. The authors emphasize counseling radiated patients about these long-term risks when choosing definitive reconstruction.

Strengths

• Includes a large cohort of more than 1,200 radiated patients undergoing immediate two-stage reconstruction.

• Captures reconstructive failures and outcomes over an extended period (median follow-up 49.8 months), with cumulative failure estimates reported through 72 months.

• Directly compares outcomes among patients who all began with the same tissue-expander (TE) stage before diverging to definitive implant or flap reconstruction.

• Includes validated BREAST-Q outcomes and clinically meaningful longitudinal failure data.

Limitations

• Single-center, retrospective, nonrandomized study, creating important potential for selection bias and residual confounding.

• Definitive reconstruction was selected after radiation based partly on evolving clinical factors and patient preferences, so the implant and flap groups were not exchangeable at baseline.

• Cohorts differed significantly in BMI, race, chemotherapy timing, reconstructive laterality, radiation timing, and TE pocket plane despite multivariable adjustment. 

• Although analyses adjusted for confounding variables, unmeasured factors such as variations in radiation technique could have influenced outcomes.

• BREAST-Q questions about the natural appearance of a breast may be less generalizable to implant than autologous reconstruction. 

• Factors that may influence satisfaction, including fat grafting, nipple reconstruction, and psychological diagnoses, were not uniformly available and therefore were not included in the multivariable models.

• The study was underpowered to definitively state differences in long-term BREAST-Q scores.

• Follow-up did not extend beyond 72 months.

• Lack of nuanced examination of radiation timing relative to implant exchange. 

Clinical relevance

For patients who require radiation and are candidates for either approach, these findings strengthen counseling in favor of autologous reconstruction when minimizing long-term reconstructive failure and maximizing breast satisfaction are priorities. The increased risk of failure with implant reconstruction extends beyond the perioperative period and persists throughout the long-term reconstructive course. Importantly, not every patient is a candidate for or desires autologous reconstruction; therefore, these findings are most useful for individualized reconstructive planning and counseling regarding the long-term risks and expected patient-reported outcomes of each approach.

Editorial notes

This study quantifies the long-term impact of radiation on implant reconstruction: failure continued to rise to approximately 12% at 6 years, compared with 1.4% for flap reconstruction.

However, the adjusted HR of 7.57 should not be interpreted causally. Reconstruction type was selected after radiation based on clinical factors and patient preference, creating potential confounding by indication. Patients with more substantial radiation-related tissue changes may have been considered poor implant candidates and preferentially directed toward flap reconstruction, potentially underestimating the full adverse effect of radiation on implant-based reconstruction. The failure definition also favored detection of problems in the implant group because elective implant-to-flap conversion counted as failure, with no directly comparable endpoint for flaps.

The key practical finding is the absolute difference in failure and satisfaction: approximately 12% versus 1.4% failure at 6 years and an 8-point BREAST-Q advantage for flaps. These results support discussing autologous reconstruction with suitable radiated patients but do not establish that flaps are preferable for everyone or that the findings will generalize beyond high-volume microsurgical centers.