The objective is to prospectively determine the factors responsible for reconstruction failure and capsular contracture in mastectomized breast cancer patients who underwent immediate two-stage breast reconstruction with a tissue expander and implant, followed by radiotherapy. This is a multicenter, prospective, non-randomized study. Between February 1998 and September 2006, we prospectively examined 141 consecutive patients, each of which received an implant after mastectomy, followed by chest wall radiotherapy at 46-50 Gy in 23-25 fractions. Radiotherapy was delivered during immediate post-mastectomy reconstruction. Patients were evaluated by both a radiation oncologist and a surgeon 24-36 months after treatment. The median follow-up duration was 37 months. According to Baker's classification, capsular contracture was grade 0, 1, or 2 in 67.5% of cases; it was grade 3 or 4 in 32.5% of cases. In total, 32 breast reconstruction failures required surgery. In univariate analysis, the following factors were associated with Baker grade 3 and 4 capsular contraction: adjuvant hormone therapy (P = 0.02), the surgeon (P = 0.04), and smoking (P = 0.05). Only one factor was significant in multivariate analysis: the surgeon (P = 0.009). Three factors were associated with immediate post-mastectomy breast reconstruction failure in multiple logistic regression analysis: T3 or T4 tumors (P = 0.0005), smoking (P = 0.001), and pN+ axillary status (P = 0.004). Patients with none, 1, 2, or all 3 factors have a probability of failure equal to 7, 15.7, 48.3, and 100%, respectively (P = 3.6 x 10(-6)). The model accurately predicts 80% of failures. Mastectomy, immediate reconstruction (expander followed by implant), and radiotherapy should be considered when conservative surgery is contraindicated. Three factors may be used to select patients likely to benefit from this technique with a low failure rate.