Dynamic ¹⁸F-FDG PET response to preoperative neoadjuvant chemotherapy in potentially resectable pancreatic ductal adenocarcinoma may predict overall survival

Objectives: Dynamic PET, combined with kinetic parameter modeling, enables valuable quantification of tumor blood flow (early uptake) and metabolism (late uptake). In this work, we evaluated kinetic response to preoperative neoadjuvant chemotherapy (NAC) via dynamic FDG PET imaging, to predict overall survival following pancreaticoduodenectomy in patients with potentially resectable pancreatic ductal adenocarcinoma (PDAC).

Methods: A total 18 patients with PDAC (mean age = 66; M/F = 11/7) underwent preoperative 60-min dynamic FDG PET/CT (injected dose =555±37 MBq) at both pre- and post-NAC [Gemcitabine and Paclitaxel with or without Hydroxychloroquine (n=8 and n=6, respectively) as well as Gemcitabine and Hydroxychloroquine (n=4)], followed by pancreaticoduodenectomy (lymph node-negative vs -positive = 9, in each). Dynamic PET kinetics were analyzed using the PMOD software (version 3.7; PMOD Technologies Ltd., Adliswil, Switzerland). The Kaplan-Meier survival product-limit estimate and Spearman’s coefficient of rank correlation (rho) were utilized.

Results: As shown in the figure, postoperative overall survival was successfully predicted by pathologic regional nodal involvement (pN) (p = 0.03) as well as preoperative FDG metabolic response to NAC (defined as 25% decrease of SUV) (p = 0.06); however, preoperative CA 19-9 response to NAC (defined as 75% decrease of CA 19-9) failed to reach significance (p = 0.2). Perfusion status of K₁ value and its changes during NAC, as quantified from 60-min dynamic PET using two-tissue compartment model, demonstrated excellent correlation with those from only 2-min first pass perfusion using one-tissue compartment model (rho = 0.93 and 0.95; p < 0.0001, respectivley), which could enable to shorten as initial only 15 min dynamic PET with conventional 60-min static PET. SUV showed inverse correlation with pathologic tumor diameter in regional node-positive PDAC (rho = -0.9; p < 0.001) and in overall (rho = -0.5; p = 0.02), but not in node-negative PDAC (rho = 0.0; p = ns), which may represent changing relationship between perfusion and metabolism (switch) over tumor growth.

Conclusion: Preoperative FDG PET metabolic response to neoadjuvant chemotherapy as well as operative pathologic nodal involvement successfully predicted postoperative overall survival in potentially resectable PDAC. 60-min dynamic PET protocol could be shortened as initial only 15 min dynamic for perfusion in addition to conventional 60-min static PET for metabolism. Further evaluation is considered for perfusion and metabolic switch over tumor growth. Research Support: UPCI In Vivo Imaging Facility supported in part by award NIH P30CA047904. Department of Energy Grant DE SC0008833, and NIH R01 CA181450 $$graphic_B4224EB5-DFA5-4E58-BF41-0F454604F319$$