Background: Chemotherapy is an integral part of multimodality treatment for resectable pancreatic ductal adenocarcinoma (PDAC). Assessing the impact of chemotherapy dose reductions or missed cycles on survival is difficult as few trials prospectively address the prognostic value of this important metric. Methods: SWOG S1505 (NCT02562716) was a randomized, phase II study of perioperative (12 weeks neoadjuvant and 12 weeks adjuvant) chemotherapy with mFOLFIRINOX or gemcitabine/nab-paclitaxel for resectable PDAC. Dose density (DD) was defined as the estimated percent of chemotherapy dose received of the total planned per protocol. Thresholds were based on published analyses and balance in sample size across treatment arms. Cox regression analysis of overall survival (OS) according to DD was performed in patients alive at two landmark time points: after surgery (among those who completed surgery) and at 40 weeks (when protocol therapy was to be completed). Results: Of the 103 eligible patients enrolled, median age was 64 years, 81% underwent pancreaticoduodenectomy, 15% distal pancreatectomy, and 3% a total pancreatectomy. In the 73 patients who underwent surgery, median neoadjuvant chemotherapy DD was 89%. Patients with ≥ 85% DD had higher median OS from surgery (38.1 v. 17.2 months, p = 0.039) compared to patients with < 85% DD. Of the 82 patients who survived to 40 weeks post randomization, 67 underwent surgery and 58 started adjuvant chemotherapy. The median DD for all perioperative chemotherapy was 67%; DD was significantly higher for neoadjuvant compared to adjuvant therapy (83% vs. 58%, p < 0.001). In this cohort, DD ≥ 70% was associated with better median OS from 40 weeks post randomization (32.2 v. 14.0 months, p = 0.017). DD was not significantly associated with pathologic response, margin status, or lymph node negativity. There were no significant differences in DD between the mFOLFIRINOX and gemcitabine/nab-paclitaxel arms. Conclusions: This is the first study to identify a prognostic role for chemotherapy DD in patients undergoing perioperative chemotherapy for resectable PDAC. Patients who received ≥ 85% DD preoperatively and/or ≥70% DD perioperatively had significantly higher mOS than those receiving a smaller proportion of protocol therapy. With greater survival benefit in patients who received higher chemotherapy DD, and higher DD being likelier in preoperative vs. postoperative treatment, these data support use of neoadjuvant chemotherapy for resectable PDAC. Moreover, these findings suggest that DD should be considered in prospective studies for localized PDAC to optimize therapy to be received.