A simple risk score for detecting radiological occult metastasis in patients with resectable or borderline resectable pancreatic ductal adenocarcinoma.

Background: We advocated carbohydrate antigen (CA) 19-9 ≥ 150 U/mL and tumor size ≥30 mm as "high-risk markers" for predicting unresectability among patients with radiologically resectable (R) or borderline resectable (BR) pancreatic ductal adenocarcinoma (PDAC). The main aim is to establish a risk scoring system for occult abdominal metastasis (OAM) in R/BR PDAC.

Methods: Predictors of OAM were investigated retrospectively in an experiment cohort from 2006 to 2018. The proposed risk scoring system was validated in another cohort from 2019 to 2020.

Results: Five hundred and thirteen eligible patients were divided into the experimental (405 patients; OAM, 22%) and validation cohorts (108 patients). Multivariate analysis identified tumor location of body/tail (odds ratio [OR] 4.45, P < .0001) and "high-risk markers" (OR 2.07, P = .011) as independent predictors of OAM. A scoring system consisting of body/tail (yes: 1, no: 0) and "high-risk markers" (yes: 1, no: 0) was constructed. In the validation cohort, when staging laparoscopy (SL) was performed for patients with scores 1/2, the eligibility for SL, sensitivity, and negative predictive value of OAM were 55%, 91%, and 96%, respectively.

Conclusions: Tumor location of body/tail and "high-risk markers" were independent predictors of OAM, composing our simple and reproducible risk scoring system.