Purpose Locoregional control for inflammatory breast cancers and chest wall recurrences is suboptimal, which has motivated interest in radiosensitization to intensify therapy. Preclinical studies have suggested a favorable therapeutic index when poly (ADP-ribose) polymerase inhibitors are used as radiosensitizers; clinical investigation is necessary to establish appropriate dosing and confirm safety. Patients and Methods We conducted a multi-institutional phase I study of veliparib and concurrent radiotherapy (RT) to the chest wall and regional lymph nodes in 30 patients with inflammatory or locally recurrent breast cancer after complete surgical resection. RT consisted of 50 Gy to the chest wall and regional lymph nodes plus a 10-Gy boost. A Bayesian time-to-event continual reassessment method escalated dose through four levels, with a 30% targeted rate of dose-limiting toxicity (DLT) measured during the 6 weeks of treatment plus 4 weeks of follow-up. DLTs were defined as confluent moist desquamation > 100 cm, nonhematologic toxicity grade ≥ 3, toxicity that requires an RT dose delay > 1 week, absolute neutrophil count < 1,000/mm, platelet count < 50,000/mm, or hemoglobin < 8.0 g/dL if possibly, probably, or definitely related to study treatment. Results Five DLTs occurred: Four were moist desquamation (two each at 100 and 150 mg twice a day), and one was neutropenia (at 200 mg twice a day). The crude rate of any grade 3 toxicity (regardless of attribution) was 10% at year 1, 16.7% at year 2, and 46.7% at year 3. At year 3, six of 15 surviving patients had severe fibrosis in the treatment field. Conclusion Although severe acute toxicity did not exceed 30% even at the highest tested dose, nearly half of surviving patients demonstrated grade 3 adverse events at 3 years, which underscores the importance of long-term monitoring of toxicity in trials of radiosensitizing agents.