Tool Can Help Predict Futile Surgery in Pancreatic Cancer

TOPLINE:

An easy-to-use web-based prognostic tool, MetroPancreas, may help predict the likelihood of futile pancreatectomy in patients with resectable pancreatic ductal adenocarcinoma and improve patient selection for upfront surgery.

METHODOLOGY:

• Immediate resection is associated with a high incidence of postoperative complications and disease recurrence within a year of surgery in patients with pancreatic ductal adenocarcinoma. Predicting which patients likely won’t benefit from upfront pancreatectomy is important.
• To identify preoperative risk factors for futile pancreatectomy, researchers evaluated 1426 patients (median age, 69 years; 53.2% men) with anatomically resectable pancreatic ductal adenocarcinoma who underwent pancreatic resection between January 2010 and December 2021.
• The patients were divided into derivation (n = 885) and validation (n = 541) cohorts.
• The primary outcome was the rate of futile upfront pancreatectomy, defined as death or disease recurrence within 6 months of surgery. Patients were divided into three risk categories — low, intermediate, and high risk — each with escalating likelihoods of futile resection, worse pathological features, and worse outcomes.
• The secondary endpoint was to develop criteria for surgical candidacy, setting a futility likelihood threshold of < 20%. This threshold corresponds to the lower bound of the 95% CI for postneoadjuvant resection rates (resection rate, 0.90; 95% CI, 0.80-1.01) from recent meta-analyses.

TAKEAWAY:

• The futility rate for pancreatectomy was 18.9% — 19.2% in the development cohort and 18.6% in the validation cohort. Three independent risk factors for futile resection included American Society of Anesthesiologists (ASA) class (95% CI for coefficients, 0.68-0.87), preoperative cancer antigen 19.9 serum levels (95% CI for coefficients, 0.05-0.75), and radiologic tumor size (95% CI for coefficients, 0.28-0.46).
• Using these independent risk factors, the predictive model demonstrated adequate calibration and discrimination in both the derivation and validation cohorts.
• The researchers then identified three risk groups. In the derivation cohort, the rate of futile pancreatectomy was 9.2% in the low-risk group, 18.0% in the intermediate-risk group, and 28.7% in the high-risk group (P < .001 for trend). In the validation cohort, the futility rate was 10.9% in the low-risk group, 20.2% in the intermediate-risk group, and 29.2% in the high-risk group (P < .001 for trend).
• Researchers identified four conditions associated with a futility likelihood below 20%, where larger tumor size is paired with lower cancer antigen 19.9 levels (defined as cancer antigen 19.9–adjusted-to-size). Patients who met these criteria experienced significantly longer disease-free survival (median 18.4 months vs 11.2 months) and overall survival (38.5 months vs 22.1 months).

IN PRACTICE:

“Although the study provides an easy-to-use calculator for clinical decision-making, there are some methodological limitations,” according to the authors of accompanying commentary. These limitations include failing to accurately describe how ASA class, cancer antigen 19.9 level, and tumor size were chosen for the model. “While we do not think the model is yet ready for standard clinical use, it may prove to be a viable tool if tested in future randomized trials comparing the neoadjuvant approach to upfront surgery in resectable pancreatic cancer,” the editorialists added.

SOURCE:

This study, led by Stefano Crippa, MD, PhD, Division of Pancreatic Surgery, Pancreas Translational and Clinical Research Center, San Raffaele Scientific Institute, Vita-Salute San Raffaele University, Milan, Italy, and the accompanying commentary were published online on July 24, 2024, in JAMA Surgery.

LIMITATIONS:

In addition to the limitations noted by the editorialists, others include the study’s retrospective design, which could introduce bias. Because preoperative imaging was not revised, the assigned resectability classes could show variability. Institutional differences existed in the selection process for upfront pancreatectomy. The model cannot be applied to cancer antigen 19.9 nonsecretors and was not externally validated.

DISCLOSURES:

The Italian Association for Cancer Research Special Program in Metastatic Disease and Italian Ministry of Health/Italian Foundation for the Research of Pancreatic Diseases supported the study in the form of a grant. Two authors reported receiving personal fees outside the submitted work. No other disclosures were reported.