Visual Parathyroid Identification in Thyroidectomy: A Decision Curve Analysis of Gland Preservation.
Pithuksurachai Paveena P, Pongsapich Warut W, Triprateepsilp Noppadon N, Chongkolwatana Cheerasook C et al.
To evaluate the clinical utility and net benefit of intraoperative visual identification for parathyroid preservation during thyroidectomy using decision curve analysis (DCA). Suspected parathyroid tissues (n = 50) were identified intraoperatively by junior staff and residents and photographed for independent consensus review by three expert otolaryngologists. Fine-needle aspiration parathyroid hormone washout (electrochemiluminescence immunoassay; ≥ 100 pg/mL) served as the surrogate biochemical reference standard. Diagnostic performance was assessed using contingency tables, logistic regression, receiver operating characteristic analysis, and DCA. Expert visual identification achieved 84.0% sensitivity, 44.0% specificity, and 64.0% accuracy. Positive predictive values were 60.0% for experts, 51.5% for junior staff, and 47.1% for residents (p = 0.632). Logistic regression yielded odds ratios of 1.69 for experts (p = 0.380) and 1.20 for junior staff (p = 0.765); model AUC was 0.56. DCA showed greater net benefit than "Treat All" and "Treat None" strategies at a clinically relevant threshold probability of 27%. Visual identification demonstrates limited diagnostic accuracy that did not differ significantly across experience levels, although this comparison is constrained by differing assessment conditions. The low specificity underscores the need for objective adjunctive confirmation and supports evaluating biochemical and optical adjuncts in future prospective studies.