Abstract INTRODUCTION: Pancreatic ductal adenocarcinoma (PDAC) remains a deadly disease with poor survival and limited therapy options. While traditional cytotoxic agents target cancer cells, there has been growing interest in tumor microenvironment (TME) modulation given that desmoplastic stroma and immune suppression are hallmarks of PDAC. Recently, vitamin D agonism has been identified as a potential treatment to normalize the PDAC TME. We have previously described tumor-intrinsic and cancer-associated fibroblast (CAF) subtypes of PDAC that independently predict survival, but whether these subtypes are associated with outcomes after TME modulation has not been studied. METHODS: We evaluated the association of treatment outcomes in a Phase 2 trial of neoadjuvant nab-paclitaxel (A), gemcitabine (G), cisplatin (C), and paricalcitol (P) (NCT03138720) and a Phase 2 trial of A+G+C+hydroxychloroquine (HCQ) (NCT04669197) in patients with non-metastatic PDAC. RNA-sequencing was performed on 28 pretreatment biopsy samples. PurIST and DeCAF were used to determine tumor-intrinsic and CAF subtypes, respectively. Clinical variables such as CA19-9 levels, resectability, and overall survival (OS) were examined. RESULTS: For tumor-intrinsic subtypes, 5 tumors (18%) were basal-like and 23 (82%) were classical. For CAF subtypes, 11 (39%) were tumor-permissive (perm)CAF and 17 (61%) were tumor-restraining (rest)CAF. Neither tumor-intrinsic or CAF subtypes were associated with CA19-9 normalization or resectability at time of diagnosis in either trial. OS was available for NCT03138720; patients with basal-like tumors receiving A+G+C+P had worse median survival of 12.3 months compared to 20.9 months for those with classical tumors, although this difference did not reach statistical significance (log-rank P = 0.065). In two independent cohorts of 925 and 129 patients, those with permCAF tumors had an increased hazard ratio of at least 1.5 (P < 0.001) compared to patients with restCAF tumors for OS. Compared to this historical data, we did not find a difference in OS between patients with permCAF tumors (19.5 months) compared to restCAF tumors (19.7 months, log-rank P = 0.56). CONCLUSION: In the study of A+G+C+P, there was a disparity in OS between patients with basal-like vs. classical tumors, in contrast to the interim results of the COMPASS and PASS-01 trials, where patients with basal-like tumors receiving A+G had similar OS compared to patients with classical tumors. Compared to prior analysis, there was no difference in OS for patients with permCAF tumors vs. those with restCAF tumors, which may suggest that either paricalcitol (or the cytotoxic combination) may be of benefit for patients with permCAF tumors and normalize the CAF landscape. Single-agent studies with paired pre- and post-treatment samples are needed to better understand how to best tailor therapy regiments for patients based on tumor subtypes. Citation Format: Jaewon J Lee, Priscilla S Chan, Changfei Luan, Elena V Kharitonova, Naim U Rashid, Xianlu L Peng, Erkut H Borazanci, Jen Jen Yeh. Paricalcitol improves survival in tumor-permissive cancer-associated fibroblast subtype of pancreatic cancers [abstract]. In: Proceedings of the AACR Special Conference in Cancer Research: Advances in Pancreatic Cancer Research; 2024 Sep 15-18; Boston, MA. Philadelphia (PA): AACR; Cancer Res 2024;84(17 Suppl_2):Abstract nr A048.