Diagnostic Performance of Common Biochemical Screening Tests for Mild Autonomous Cortisol Secretion in Adrenal Incidentalomas: A Systematic Review and Hierarchical SROC Meta-analysis.

Abstract

Objective: Mild autonomous cortisol secretion (MACS) is a biochemical phenotype in adrenal incidentalomas associated with cardiometabolic risk. Serum and salivary dexamethasone suppression testing (DST), late-night salivary cortisol (LNSC), and urinary free cortisol (UFC) are used, but interpretation is limited by nonuniform disease definitions and reference standards.

Methods: PubMed, Embase, Web of Science, and Scopus were searched through November 19, 2025, in accordance with PRISMA-DTA. QUADAS-2 assessed risk of bias. Bivariate random-effects models generated HSROC estimates and likelihood ratios, and Fagan analyses modeled post-test probabilities. Sensitivity assessments examined reference-standard dependence, diagnostic era, and threshold selection.

Results: Twelve studies comprising 1,420 patients were included. Serum DST had sensitivity of 97.9% (95% CI, 88.6-99.6), specificity of 76.8% (64.5-85.7), and a negative likelihood ratio of 0.03 (0.00-0.16). Salivary DST had sensitivity of 74.9% and specificity of 80.5%; LNSC sensitivity was 62.7%, and UFC estimates were highly imprecise. All three serum DST studies used reference definitions that incorporated DST or a related dexamethasone-suppression criterion, leaving no fully independent-reference serum DST pool. Threshold-specific data showed substantial sensitivity-specificity tradeoffs.

Conclusion: Serum DST produced the lowest pooled negative likelihood ratio against study-defined biochemical reference standards, suggesting potential rule-out value. Incorporation bias, threshold heterogeneity, indirect comparisons, and sparse study counts preclude claims of diagnostic superiority. Positive results require clinical context and confirmatory evaluation when appropriate.