Intraoperative Cytologic Sampling for Resected Pancreatic and Periampullary Adenocarcinoma with Implications for Locoregional Recurrence-Free Survival.

Abstract

We hypothesized that pancreatic and periampullary adenocarcinoma recurrence after surgical resection may be affected by the shedding of malignant epithelial cells during surgical dissection and that this may have implications for disease recurrence and survival. In this ongoing, investigator-initiated prospective randomized controlled trial, patients with pancreatic and periampullary adenocarcinoma were randomized intraoperatively, postresection into 3 study arms: peritoneal lavage using 10 L normal saline or distilled water, or control group with no lavage. Peritoneal fluid was sampled for cytologic analysis (cytospin, cellblock, immunohistochemistry-Ber-EP4 antibody) at 4 stages: (1) abdominal entry pre-dissection, (2) resection bed after tumor extirpation, (3) ex vivo resected specimen, and (4) resection bed postlavage. Between April 2016 and May 2018, 193 patients who underwent randomization for the study also underwent the described cytologic sampling. Of these, 167 patients (86.5%) were ultimately found to have pancreatic or periampullary adenocarcinoma. Before dissection (1) on cytospin analysis, 4.9% were positive, which rose to 10.2% intraoperatively (2), 16.7% ex vivo (3), and decreased to 4.3% (4) after lavage. Lymph node metastasis, margin involvement, and perineural invasion did not correlate with locoregional recurrence (LR). Tumor cells in the ex vivo cytospin (3) correlated with LR (odds ratio 3.8 [95% CI 1.6 to 9.2], p = 0.005) and LR disease-free survival (p = 0.007). Cox regression analysis revealed ex vivo cytospin positivity to be strongly associated with poorer LR disease-free survival (hazard ratio 2.26 [95% CI 1.16 to 4.42], p = 0.017). Cytologic sampling from ex vivo specimen irrigation after surgical resection of pancreatic and periampullary adenocarcinoma may have implications for LR, survival, and treatment, suggesting a possible cancer cell shedding phenotype.