Background: Collectively, sarcomas have demonstrated less benefit from immunotherapy than many other cancers. Pts with AS have limited options, particularly after taxane chemotherapy. Although AS is a vascular malignancy, anti-VEGF therapies have minimal monotherapy clinical activity. Some cutaneous AS have a high UV damage signature, but other AS often do not. Cabozantinib (C) is a multikinase inhibitor that may alter PD-1 expression in regulatory T cells, promoting an immune permissive environment when combined with nivolumab (N), a fully human anti PD-1 MAB. We hypothesize that the combination of C+N will be an effective therapy in AS, across AS subtypes. Methods: This open-label multi-arm study enrolled pts with locally advanced/metastatic AS. Arm 3, reported here, enrolled pts who had received a taxane for AS (including adjuvant) prior to study enrollment. Pts were not restricted on number of prior lines, but prior anti-VEGF and immunotherapies were not allowed. Eligible pts received C (40 mg po daily) with N (480 mg IV every 4 wks). Treatment was permitted beyond progressive disease (PD) in the 1^st 12 wks (4 wk confirmatory scan), but PD response was censored at 12 wks. The Simon 2-stage design (null and alternative overall response rate (ORR) were 10% and 35%, respectively) required ≥1 confirmed response in 9 pts in the 1^st stage, and >4 of 18 pts (91.4% power, alpha 0.095) for the primary endpoint. Secondary endpoints were adverse events (AEs), progression free, overall survival (PFS, OS), and pt-reported outcomes (PRO). Results: 21 eligible pts [median age 66 yrs (32-92), 10 female] received ≥1 dose of C+N. Primary disease sites: 12 cutaneous (scalp/face), 1 liver, 2 breast, 6 other. All pts received prior taxanes (11/21 = 52% as adjuvant therapy) and 5/21 (24%) had also received prior anthracycline (all relapsed AS). Primary endpoint: 13 of the first 18 evaluable patients (72%) experienced objective response (OR, 95%CI: 47%-90%). After a median follow-up of 11.2 mo, 13 of 21 pts achieved OR (11 partial response, 2 complete), for ORR 62% (95%CI: 38-82%). Responses were seen in pts with primary cutaneous disease 7/12 (58%) and noncutaneous disease 6/9 (67%). Median PFS was 9.6 mo (5.3-NR), and OS 20.5 mo (14.4-NR). Off treatment reasons include: progressive disease (14), comorbid condition unrelated to AS or study (1); 6 pts remain on study. Grade (G) 3 hypertension was the only possibly treatment related AE (TRAE) occurring in ³10%. No G 4/5 TRAE were reported. Conclusions: C + N demonstrated significant antitumor activity in taxane-pretreated AS and was well tolerated without new safety concerns. Activity of the combination was seen across AS subtypes. PRO and exploratory analyses are ongoing. Support: U10CA180821, U10CA180882, U24 CA196171; U10CA180820; https://acknowledgments.alliancefound.org. U10CA180868; U10CA180888. NCT04339738. Clinical trial information: NCT04339738.