Association of decreased peak SUV on PET after neoadjvuant therapy in patients with borderline resectable pancreatic carcinoma with pathologic response.

Cancers of the Pancreas, Small Bowel, and Hepatobiliary Tract
Session Type and Session Title: 
Poster Session B: Cancers of the Pancreas, Small Bowel, and Hepatobiliary Tract
Abstract Number: 


Poster Board Number: 
Poster Session B Board #L17
J Clin Oncol 34, 2016 (suppl 4S; abstr 423)
Evan Scott Glazer, Omar Maen Rashid, Pamela Joy Hodul, Jose Mario Pimiento, Mokenge Peter Malafa; Moffitt Cancer Center, Tampa, FL

Abstract Disclosures


Background: Despite the use of PET scan in metastatic staging, its utility as a biomarker for borderline resectable pancreatic carcinoma (BRPC) is unclear. The peak standardized uptake value (SUVpeak) is the average SUV around the most avid portion of the tumor. We hypothesized that tumor pathologic response (PR) to neoadjuvant therapy will correlate with SUVpeak changes after therapy. Methods: We reviewed our database of patients with BRPC who completed neoadjuvant chemotherapy (GTX) followed by radiation and had pre- and post- neoadjuvant PET scans. Changes in the SUVpeak were calculated by subtracting the post-neoadjuvant/pre-resection PET scan from the pre-neoadjuvant PET scan. PET responders were defined as those patients whose SUVpeak decreased by 4 units (the overall average decrease in SUVpeak) or more. Results: We identified 74 patients- 32 patients developed progressive disease and 42 patients underwent surgery. The average age was 66 ± 9 years; 61% of patients were male. On the final pathologic specimens, 5 patients had a complete PR to therapy while 23 patients had a moderate PR (1-49% viable tumor) and 14 patients had a minimal PR (≥ 50% viable tumor). PR is associated with a decrease in SUVpeak (p = 0.02; Table). The area under the receiver operator characteristic curve for changes in SUVpeakand complete PR was 0.83 ± 0.08. Of the PET responders who underwent surgery, 29% had a complete PR while none of the PET non-responders had a complete PR (p = 0.004). Of the unresected patients, PET responders had a longer overall survival compared to non-responders (18 ± 7 vs. 12 ± 7 months, p < 0.05). Nodal disease was not associated with PR (p > 0.3). Conclusions: PR was significantly associated with a decrease in SUVpeak post-neoadjuvant therapy in patients with BRPC. Even in patients with unresectable disease, SUVpeak provides useful prognostic data. We believe that pre- and post-neoadjuvant therapy PET scans may be used as a prognostic biomarker in patients with BRPC.

SUVpeak after neoadjuvant therapy by pathologic response.

Pathologic responsesChange in SUVpeak
Minimal (>50% viable)-2.5 ± 2.0
Major (1% to 49% viable)-5.0 ± 4.3
Complete (0% viable)-7.5 ± 3.8 [p = 0.02]