ESTRO 2021 Abstract Book

S436

ESTRO 2021

post-operative reRT+HT. Mean age was 64±11 years, 85% had received systemic treatment, 63% had salvage mastectomy, and 37% had a local resection after previous mastectomy. ReRT was given to a total dose of 46Gy in 23 (n=78), or 32Gy in 8 fractions (n= 34), combined with 4-5 weekly HT sessions guided by invasive thermometry. Patients were divided in ‘low’ (n=56) and ‘high’ thermal dose (TD) groups (n=56) by the session with the highest invasive dose CEM43T50 (=median cumulative equivalent minutes at 43°C) <7.2 and ≥7.2minutes, respectively. LR control, OS and late toxicity incidence according to Common Terminology Criteria for Adverse Events version 5.0, were evaluated. Backward multivariable Cox regression was performed to identify outcome-associated patient and treatment characteristics. Results Median follow-up period was 43 months (range 1-107). Actuarial 3-year rate LR control and OS were 83.2% and 85.4%, respectively. Three-year LR control was 74.0% vs. 92.3% in the low and high TD group, respectively ( p =0.008) (Figures 1+2). Three-year LR control was also significantly different for the low and high TD groups, 81.6% and 97.3% ( p =0.013) for 46 Gy in 23 fractions, and 55.6% and 81.2% ( p =0.033) for 32 Gy in 8 fractions, respectively. After 3 years, 25% and 0.9% of the patients had late toxicity grade 3 and 4, respectively. TD was associated with LR control ( p =0.0013), but not with OS ( p =0.29) or late toxicity ( p =0.74). Multivariable analysis showed that presence of distant metastases (HR 17.6; 95% CI 5.2 - 60.2), lymph node involvement (HR 2.9; 95% CI 1.2-7.2), site of recurrence (chest wall vs. breast; HR 4.6; 95% CI 1.8-11.6) and TD (low vs. high; HR 4.1; 95% CI 1.4-11.5) were associated with worse LR control.