Abstract Background: Patients with metastatic triple negative breast cancer (mTNBC) have a poor prognosis with median survival of 18 months or less. While the combination of immune checkpoint inhibitors (ICIs) and chemotherapy has shown promise in mTNBC, biomarkers beyond PDL-1 are needed to better predict which individuals will benefit from this treatment approach. In this study, we assessed the ability of spatial characteristics in predicting clinical best response in patients with PDL-1 positive mTNBC treated with an ICI and chemotherapy. Methods: Women with advanced unresectable or mTNBC treated with an ICI plus chemotherapy at Emory University between 2019 and 2021 with available biopsy specimens were retrospectively evaluated. Different cell types (tumor, stroma, immune cells) were identified by morphology on H&E staining. A cellular network was created by connecting each cell centroid to its adjacent centroids within a 30-μm distance. The resulting spatial neighborhood network was used to assess tumor density and quantify immune infiltration. The immune infiltration score was defined as the number of immune to tumor cell neighbors divided by the total number of immune cell neighbors in a region of 75-μm. A final immunoscore was calculated for each patient by averaging the immune infiltration scores in regions with high tumor cell density. Tumor infiltrating lymphocytes (TILs) were manually quantified. Responders were defined as those with a complete response (CR), partial response (PR), or stable disease (SD), while those with progressive disease (PD) were categorized as non-responders. A continuous response score was developed from tumor measurements of a target lesion on serial imaging. Pearson’s correlation coefficients were used to assess the relationship between continuous response scores and tumor characteristics. Responders and non-responders were compared using Mann-Whitney U tests. Results: Fifteen women with PDL-1 positive mTNBC treated with ICI plus chemotherapy and available tissue were included. All patients had relapsed disease, and 10 patients (67%) received an ICI and chemotherapy as first line treatment for mTNBC. Eight patients (53%) received atezolizumab and nab-paclitaxel while the remaining 7 (47%) patients received pembrolizumab with an approved chemotherapeutic agent. Seven patients (47%) experienced a PR, 3 (20%) with a CR, and 1 (7%) had SD. Four patients (27%) had PD with no clinical benefit. Higher immunoscores (-0.17, p=0.6) and TILs (-0.21, p=0.5) were numerically associated with continuous response scores. However, there were no significant differences in immunoscores (0.28 vs 0.26, p=0.6) or TIL counts (2 vs 5, p=0.6) between responders (N=11) and non-responders (N=4). Interestingly, patients with response to treatment had lower tumor densities compared to non-responders (7.5 vs 17.3, p=0.02). Conclusions: Spatial analysis of tumor density and immune infiltration, including immunoscore shows evidence of correlation with response. Tumor density was the only parameter significantly associated with response. The study identified novel tumor characteristics that need to be considered in the prediction of response to ICIs plus chemotherapy in mTNBC. The current findings are hypothesis generating and need validation in additional tissue samples to determine the role of tumor density as a predictor of response to ICI. Citation Format: Jeffrey Aldrich, Thomas Hu, Jeffrey Switchenko, Ahmet Coskun, Yesim Gokmen-Polar, Sunil Badve, Manali Bhave. Predicting the likelihood of response in PDL-1 positive metastatic triple negative breast cancer treated with an immune checkpoint inhibitor [abstract]. In: Proceedings of the 2022 San Antonio Breast Cancer Symposium; 2022 Dec 6-10; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2023;83(5 Suppl):Abstract nr P1-04-14.