Background: Although breast cancer survivors are living longer, they have an increased risk of treatment-related second cancers including rare but deadly thoracic soft tissue sarcomas (tSTS), such as angiosarcomas, compared with the general population. Limited studies have examined the relationship between surgery type and radiotherapy on tSTS risk, due to the rarity of this event. We undertook this large-scale study to analyze the joint effects of surgery type and radiotherapy with tSTS risk among older breast cancer patients followed in SEER-Medicare. Methods: The SEER-Medicare analytic cohort included 162,507 patients diagnosed with a primary breast cancer between 2000-2017, aged 66-84 years, who survived 1 year, and were followed until 2018. We calculated the cumulative incidence of tSTS and angiosarcoma accounting for competing risks. We then examined the joint effects of breast surgery and radiotherapy receipt (mastectomy + no/unknown radiotherapy, mastectomy + radiotherapy, BCS + no/unknown radiotherapy, and BCS + radiotherapy), with risk of developing any tSTS or angiosarcoma specifically. We also explored the influence of extent of mastectomy (removal of whole breast only, removal of whole breast and chest wall). We calculated hazard ratios (HRs) from multivariable Cox proportional hazard models, with attained age as the time scale, and adjusted for year of diagnosis (2000-2004, 2005-2009, 2010-2017), chemotherapy (yes, no/unknown), and stage (in situ, localized, regional/distant, unknown). Results: Mean age of breast cancer diagnosis was 73.68 (SD: 5.18) years. Patients were diagnosed with in situ (17.74%), localized (59.89%), and regional/distant (21.94%) stage. Median follow-up time was 6.75 years (IQR: 2.00-3.92), and 139 patients developed a tSTS (95 angiosarcomas). By 5 years after diagnosis, cumulative incidence of tSTS was 0.036% (95% CI: 0.027-0.48) and angiosarcomas was 0.019% (95% CI: 0.012-0.029). Compared with patients who received mastectomy + no/unknown radiotherapy, patients receiving mastectomy + radiotherapy were not at elevated risk (HR: 2.18, 95% CI: 0.71-6.70), whereas patients who received BCS without radiotherapy (HR: 3.04, 95% CI: 1.19-7.76) or with radiotherapy (HR: 7.01, 95% CI: 3.55-13.84) were at elevated risk of developing tSTS. Patients who underwent removal of whole breast and chest wall had a lower risk of developing tSTS (HR: 0.32, 95% CI: 0.11-0.94), compared to whole breast only. Nearly all angiosarcomas occurred among patients treated with BCS + radiotherapy. Conclusions: Clinically, these results, while rare, indicate the greatest risk of tSTS observed after BCS + radiotherapy and the need for closer monitoring and heightened awareness of tSTS. Additional studies are needed to investigate biological underpinnings of tSTS and the potential role of other cancer treatments and comorbidities on development of tSTS.