2514 Background: RECIST criteria evaluate changes in tumor burden and have been utilized in cancer research for decades. Changes in the sum of maximum diameters of “target” lesions are used to designate patients into response categories, including complete response, partial response, stable disease, and progressive disease. However, not much is known about the implications of mixed response, where lesions within a patient show contrasting responses to systemic treatment. Here we evaluate the data from the Dual Anti-CTLA-4 and Anti-PD-1 Blockade in Rare Tumors (DART) SWOG S1609 to investigate the association between mixed response and survival among patients treated with dual checkpoint inhibitor therapy. Methods: A total of 796 patients were enrolled on the S1609 DART trial, a basket trial evaluating ipilimumab plus nivolumab across solid rare tumor subtypes. All patients had RECIST-measurable disease at the study entry. Patients were excluded from the analyses due to: ineligibility, n=47; not receiving any protocol therapy, n=23; death before day 65 (landmark for analysis based on first scan), n=86; 1 target lesion, n=153. Thus, 487 patients were analyzed. A 5mm cut-off for increase/decrease in lesions was used to define a mixed response. Cox regression models were used to evaluate associations between survival (landmarked at day 65, stratified by basket) and response category (reference of Stable Disease). Results: 6 groups were identified among the 487 patients: no lesions changed more than 5 mm (n = 105), all lesions increased more than 5mm (n = 69), all lesions decreased more than 5mm (n = 39), one+ lesion increased more than 5mm and one+ lesion decreased more than 5 mm (n = 24), one+ lesion increased more than 5mm and one+ lesion did not change more than 5 mm (either direction) (n =155), one+ lesion decreased more than 5mm and one+ lesion did not change more than 5 mm (either direction) (n =95). Hazard ratios and median survival (mOS) are shown (Table). Conclusions: This is the first evaluation of the association between mixed response and survival outcomes among patients with various tumors receiving dual checkpoint therapy. Our results suggest that survival outcomes are driven by the “worst” performing lesion; in other words, having an increase in any lesion is associated with worse outcomes, even if not all lesions increase. [Table: see text]