Mapping patterns of local recurrence after pancreaticoduodenectomy for pancreatic adenocarcinoma: a new approach to adjuvant radiation field design

Abstract

Purpose: To generate a map of local recurrences after pancreaticoduodenectomy (PD) for patients with resectable pancreatic ductal adenocarcinoma (PDA) and to model an adjuvant radiation therapy planning treatment volume (PTV) that encompasses a majority of local recurrences.

Methods and materials: Consecutive patients with resectable PDA undergoing PD and 1 or more computed tomography (CT) scans more than 60 days after PD at our institution were reviewed. Patients were divided into 3 groups: no adjuvant treatment (NA), chemotherapy alone (CTA), or chemoradiation (CRT). Cross-sectional scans were centrally reviewed, and local recurrences were plotted to scale with respect to the celiac axis (CA), superior mesenteric artery (SMA), and renal veins on 1 CT scan of a template post-PD patient. An adjuvant clinical treatment volume comprising 90% of local failures based on standard expansions of the CA and SMA was created and simulated on 3 post-PD CT scans to assess the feasibility of this planning approach.

Results: Of the 202 patients in the study, 40 (20%), 34 (17%), and 128 (63%) received NA, CTA, and CRT adjuvant therapy, respectively. The rate of margin-positive resections was greater in CRT patients than in CTA patients (28% vs 9%, P=.023). Local recurrence occurred in 90 of the 202 patients overall (45%) and in 19 (48%), 22 (65%), and 49 (38%) in the NA, CTA, and CRT groups, respectively. Ninety percent of recurrences were within a 3.0-cm right-lateral, 2.0-cm left-lateral, 1.5-cm anterior, 1.0-cm posterior, 1.0-cm superior, and 2.0-cm inferior expansion of the combined CA and SMA contours. Three simulated radiation treatment plans using these expansions with adjustments to avoid nearby structures were created to demonstrate the use of this treatment volume.

Conclusions: Modified PTVs targeting high-risk areas may improve local control while minimizing toxicities, allowing dose escalation with intensity-modulated or stereotactic body radiation therapy.

Fig. 1

Local recurrence map. (A) Anterior-posterior and (B) lateral views of local recurrence plots in relation to the celiac artery (yellow) and superior mesenteric artery (blue) after pancreaticoduodenectomy for patients receiving no adjuvant therapy (red), chemotherapy alone (orange), and chemoradiation (green).

Fig. 2

Stepwise planning process. *ROI region of interest, ^†organ at risk, ^‡clinical target volume, ^§planning target volume.

Fig. 3

Demonstration of proposed adjuvant plan for 3 simulated patients. Radiation treatment plans were created using the stepwise process. Axial, sagittal, and coronal views are shown for patient 1 (A, B, and C), patient 2 (D, E, and F), and patient 3 (G, H, and I). PTV90-final (green), PTV80-final (dark blue), 33 Gy isodose line (red), 25 Gy isodose line (orange), 15 Gy isodose line (khaki), proximal stomach (light blue), proximal bowel (purple), and proximal jejunum (pink) are indicated on each plan.

Fig. 4

A standard Radiation Therapy Oncology Group 0848 clinical target volume (orange) and planning target volume (red) are shown simultaneously with the proposed PTV80-final (blue) and PTV90-final (green) of this study on an anterior-posterior digitally reconstructed radiograph (A) and on axial (B), sagittal (C), and coronal (D) computed tomographic sections of 1 simulated patient as an example of where areas could potentially be reduced to minimize the toxicity of adjuvant treatment. PTV80-final = planning target volume containing 80% of mapped recurrences with avoidance of proximal organs at risk; PTV90-final = planning target volume containing 90% of mapped recurrences with avoidance of proximal organs at risk.