[en] Purpose: In previously reported studies using radiation therapy (XRT) and hyperthermia (HT) for treatment of superficial metastases from adenocarcinoma of the breast, we have identified several pretreatment and treatment parameters that correlated with rate of initial complete response (ICR) recorded at 3 weeks and duration of local control (DLC). These parameters include minimal intratumoral temperature, T_min, and the temperature exceeded by 90% of the measured intratumoral temperatures, T₉₀. Recently, others have shown that thermal dose defined as the cumulative time of isoeffective treatments with T₉₀ = 43 deg. C (CUM EQ MIN T₉₀ 43) was predictive of complete response in superficial tumors. We have assessed the prognostic value of several formulations of this parameter for both ICR and DLC in a relatively uniform patient population treated with XRT-HT. Methods and Materials: The corresponding EQ MIN T₉₀ 43 were calculated for 332 HT treatments in 111 HT fields in 83 patients who started treatment between October 1982 and May 1992. Each field contained only one measurable superficially located nodular tumor recurrence or metastasis from adenocarcinoma of the breast that was treated with XRT-HT, had mapped or multiple point temperatures recorded, and had at least one posttreatment follow-up evaluation. The thermal doses from all treatments delivered to a field were added to obtain the total thermal dose, SUM EQ MIN T₉₀ 43. Logistic and life-table multivariate analyses were performed to determine which pretreatment parameters (including initial T-stage, prior XRT, and tumor volume at the time of HT) and treatment parameters (including XRT dose, T_min, T₉₀, thermal dose, and hormonal therapy) best correlated with ICR and DLC. Results: Of the treatment parameters tested, SUM EQ MIN T₉₀ 43 had the strongest correlation with both ICR (p = 0.0002) and DLC (p = 0.0014). Also, SUM EQ MIN T₉₀ 43 contributed to the best multivariate models predictive of ICR and DLC. Conclusion: For this relatively uniform patient population, we have confirmed that SUM EQ MIN T₉₀ 43 is the treatment parameter most strongly correlated with not only response following XRT-HT, but also duration of local control. This formulation of thermal dose should permit prescriptions to be written for HT treatments. Prospective trials designed to confirm this thermal dose relationship are to be encouraged