非心臓手術後の周術期主要臨床イベント予測のためのルーチンPET/CTにおける視覚的順序冠動脈石灰化スコアリング

Background: PET/CT is commonly performed during oncologic workup and provides an opportunity for coronary artery calcification (CAC) evaluation. However, technical constraints typically preclude standard Agatston scoring using the attenuation-correction CT images obtained during PET/CT examinations. Objective: To assess the prognostic utility of a simple visual ordinal CAC scoring system applied to PET/CT for perioperative risk stratification after noncardiac surgery. Methods: This retrospective study included 972 patients (559 male, 413 female; median age, 58 years) who underwent [¹⁸F]FDG PET/CT with a nongated low-dose CT acquisition followed by intermediate- or high-risk noncardiac surgery within 12 months (median interval, 4 days [IQR 2-7 days]) from April 2013 to June 2024. Three radiologists independently reviewed CT images to assign a visual ordinal CAC score (scale, 0-3) to each of four major coronary arteries; these scores were used to derive a visual CAC grade (none, mild, moderate, severe) for each patient. Multivariable logistic regression analyses were performed to identify the role of visual CAC grades (based on consensus assessments) in predicting 30-day perioperative major clinical events (MCEs, defined as all-cause mortality or in-hospital troponin I elevation), adjusting for age, sex, and revised cardiac risk index (RCRI; a traditional perioperative risk stratification tool). Results: Interobserver agreement for the visual CAC grade was high (kappa=0.875). At least mild CAC and moderate or severe (hereafter, moderate/severe) CAC were present in 46.3% and 16.0% of patients, respectively. Perioperative MCEs occurred in 3.2% of patients. Perioperative MCEs occurred in 1.9%, 3.1%, 5.8%, and 11.3% of patients with no, mild, moderate, and severe CAC, respectively. Moderate/severe CAC was an independent predictor of perioperative MCEs (adjusted OR=2.44; 95% CI, 1.11-5.38). Among 718 patients with RCRI of 1, frequency of perioperative MCEs was higher among those with moderate/severe versus no or mild CAC (8.3% vs 2.3%, respectively; p=.005). The AUC for predicting perioperative MCEs was 0.608 for RCRI and 0.652 for the visual CAC grade (p=.36). Conclusions: Simple visual CAC scoring on routine PET/CT was associated with perioperative events after noncardiac surgeries, independent of the RCRI. Clinical Impact: Standardized CAC reporting on preoperative PET/CT may help refine risk stratification and support clinical decision-making regarding the intensity of perioperative care.