The Sequential Radiographic Effects of Preoperative Chemotherapy and (Chemo)Radiation on Tumor Anatomy in Patients with Localized Pancreatic Cancer

Prior to initiation and following the completion of preoperative chemotherapy and (chemo)radiation, anatomic disease staging for all patients was accomplished using multidetector CT or magnetic resonance imaging (MRI) and standard protocols optimized for imaging pancreatic tumors. Multiplanar reconstructions were used as necessary to visualize the vascular anatomy of each tumor.

All treatment decisions were made by the multidisciplinary teams at both centers. Systemic chemotherapy was routinely recommended as primary therapy to patients with a BR or LA tumor, and was also generally recommended to all patients with a resectable tumor at MD Anderson Cancer Center. At Verona University Hospital, it was administered more selectively to such patients, primarily to those with a radiographic interface between their tumor and superior mesenteric vein or portal vein. However, at both institutions, systemic chemotherapy was routinely administered to patients with a resectable tumor and one or more of the following: (1) imaging studies demonstrating findings suspicious but not diagnostic for extrapancreatic disease; (2) a depressed performance status or significant comorbidity profile; and (3) a carbohydrate antigen (CA) 19-9 level (in the absence of jaundice) suggestive of disseminated cancer.

Systemic chemotherapy consisting of FOLFIRINOX and/or GA was administered to all evaluated patients. (Chemo)radiation therapy was administered selectively at both centers and was delivered more frequently to patients with tumors who had any degree of mesenteric vascular involvement. Radiation therapy consisted of external-beam radiation therapy (total dose, 50.4 Gy delivered over 6 weeks, or 30 Gy delivered over 2 weeks) with concurrent administration of 5-fluorouracil, capecitabine, or gemcitabine or stereotactic body radiation therapy (SBRT) delivered over 5 days without a radiosensitizer. Within 8 weeks after completing preoperative therapy, the patients’ disease was clinically and radiographically restaged. Patients without evidence of disease progression and with adequate performance statuses were considered for surgical resection. Pancreatoduodenectomy, distal pancreatectomy, or total pancreatectomy was performed using standardized techniques at both centers.9,10

Gastrointestinal pathologists used standardized protocols to evaluate all surgical specimens.11 R1 margin status was defined as evidence of cancer cells at the inked bile duct or pancreatic parenchymal margin or within 1 mm of the superior mesenteric artery margin.

CT images of all patients were reviewed for this study by two research associates (GP and AC) blinded to treatments and outcomes.

To evaluate the cumulative response to preoperative therapy, the images of each patient obtained before preoperative therapy and before surgery were compared. In addition, the radiographic changes associated with chemotherapy and (chemo)radiation were assessed independently. To evaluate the changes associated with chemotherapy alone, the pretreatment images were compared with the post-chemotherapy images (obtained prior to surgery or prior to radiation therapy). To evaluate the changes associated with radiation therapy alone, the post-chemotherapy images were compared with the preoperative images (Fig. 1).

Each tumor was staged radiographically as potentially resectable, BR, or LA according to two different classification systems—the MD Anderson Classification and the National Comprehensive Cancer Network (NCCN) guidelines.2,12 Downstaging was defined as a change from BR to resectable disease, or from LA to BR or resectable disease. Tumor size was measured using the longest (L) and shortest (W) axial diameters and the craniocaudal diameter (H), and the volume of each tumor was calculated using the formula for a typical ellipsoid: volume = π/6 × L × W × H.13 The radiographic interface between the tumor and each mesenteric vascular structure was characterized as either no contact, abutment (≤ 180° of the circumference), encasement (> 180° of the circumference), or occlusion.14 To measure the average attenuation in Hounsfield units, a circular region of interest encompassing one-half to two-thirds of the tumor’s area was drawn in the center of the tumor on the section with the largest surface area on each portal venous phase CT images. To characterize radiographic changes associated with preoperative therapy, the volume of the primary tumor, interface between the tumor and each mesenteric vascular structure, and tumor attenuation on pretreatment images were compared with those on post-treatment images for each patient. The change in tumor volume after preoperative treatment was calculated as the percentage of the baseline volume. Changes were also described using the modified Response Evaluation Criteria in Solid Tumors (RECIST; version 1.1).15 Progressive disease (PD) was defined as an increase of at least 20% in the primary tumor’s largest dimension (or absolute increase ≥ 5 mm); a partial response (PR) was defined as a decrease of at least 30% in the primary tumor’s largest dimension; stable disease (SD) was defined as an increase or decrease in tumor size insufficient to qualify as PD or a PR, respectively; and complete response (CR) was defined as total disappearance of the primary tumor.

Serum CA 19-9 levels (normal range, 0–37 U/mL) were measured prior to and following treatment. Patients in whom the CA 19-9 level was lower than 1 U/mL both prior to and following treatment were defined as nonproducers.

Continuous data were expressed as medians and ranges, whereas categorical data were expressed as frequencies and percentages. Continuous variables were compared using a t test if normally distributed and a nonparametric Mann–Whitney U test if not. Categorical variables were compared using a Pearson Chi square test (or Fisher’s exact test when appropriate). Statistical analyses were performed using the SPSS software program version 24.0 (IBM Corporation, Armonk, NY, USA), and p values < 0.05 were considered significant. All p values were two-sided.

The MD Anderson radiographic stage of 197 tumors (87%) did not change following the administration of preoperative therapy, whereas five tumors (2%) were upstaged due to local progression. The tumors of 24/82 patients (29%) who had BR or LA cancer were downstaged after preoperative therapy. We observed radiographic downstaging of BR and LA tumors after chemotherapy in 20 patients (24%) and after subsequent (chemo)radiation in 4 patients (6%) [Fig. 2]. These rates were similar when we staged the tumors using the NCCN criteria.

The volume of 180 primary tumors (80%) decreased after preoperative therapy. Patients experienced a median baseline tumor volume loss of 36% over the entire preoperative treatment course. Loss of volume was equally common following chemotherapy and (chemo)radiation (73% vs. 81%; p = 0.7); however, the median tumor volume loss after chemotherapy was significantly greater than that after (chemo)radiation (28% vs. 17%; p < 0.01). Notably, volume loss following chemotherapy was similar between patients who received only chemotherapy and those who received chemotherapy followed by (chemo)radiation (28% vs. 28%; p = 0.2). In addition, there was no difference in terms of median volume loss between 50.4 Gy and 30 Gy (chemo)radiation (20% vs. 16%; p = 0.2), and between ‘conventional’ (chemo)radiation and SBRT (19% vs. 11%; p = 0.2).

After receiving chemotherapy with or without (chemo)radiation, 193 patients (85%) underwent resection of their primary tumors and regional lymph nodes, whereas 33 (15%) did not due to intraoperative identification of metastases (n = 25, 11%) or a local tumor anatomy that precluded resection (n = 8, 4%). Patients who did not undergo resection were similar to those who did in terms of treatment response (according to RECIST 1.1), decrease in tumor volume, increase in tumor attenuation, changes in tumor-vessel interfaces, and other potential indirect radiographic signs of tumor response to preoperative therapy (all p > 0.05) [Table 3]. The median post-treatment CA 19-9 level in patients whose surgery was aborted was higher than that in those in whom resection was completed (37 U/mL vs. 25 U/mL; p = 0.04).