Rapid PCR-based diagnosis of superficial mycoses with ROC-optimized cutoffs.

Fluorescence microscopy and fungal culture are the primary diagnostic methods for superficial fungal infections, but they are limited by low sensitivity and long turnaround time, respectively. We determined the optimal cycle threshold (Ct) cutoff for a novel multiplex PCR assay (iFIND) using receiver operating characteristic analysis and evaluated whether a "microscopy + iFIND" combined strategy improves diagnostic performance. This prospective study enrolled 113 patients with suspected superficial mycoses. Specimens underwent fluorescence microscopy, fungal culture, and automated iFIND. Using a composite reference standard, we calculated positive percent agreement (PPA), negative percent agreement, and overall percent agreement. Cohen's kappa and McNemar's test assessed inter-method agreement. To mitigate verification bias, we performed a sensitivity analysis excluding cases diagnosed solely by clinical response. Internal bootstrap validation (1,000 resamples) was used to assess overfitting of the Ct cutoff. The optimal Ct cutoff for any fungal infection was 32.66 (area under the curve [AUC] = 0.746). Bootstrap validation confirmed robustness (optimism-corrected AUC = 0.7431, optimism = 0.00275, bias-corrected and accelerated 95% confidence interval for cutoff: 32.5-33.7). iFIND showed excellent discrimination for IMPORTANCE: Fungal skin infections affect one in four people worldwide, but current diagnostic methods are either insensitive (microscopy) or painfully slow (culture, taking 2-4 weeks). This study validates a rapid PCR test that delivers results in under 3 h and, when combined with microscopy, matches the accuracy of culture. By establishing a data-driven cutoff value for test positivity and using rigorous statistical methods to confirm its reliability, we provide a practical framework that clinics can adopt immediately. The combined strategy reduces diagnostic delay, enabling same-day treatment decisions and reducing unnecessary antibiotic use. This is especially critical, given the global spread of antifungal-resistant strains, where rapid genus-level identification can guide appropriate therapy and curb resistance. Our approach offers an actionable, culture-comparable solution for routine clinical practice, directly addressing the urgent need for faster, more accurate fungal diagnostics worldwide.