Automatic fluorescence microscopic image analyzer: a novel AI-based tool for early diagnosing superficial fungal infections
Abstract Background Superficial fungal infections (SFIs) are highly prevalent globally, affecting approximately 20–25% of the population. Their diverse and often non-specific clinical manifestations necessitate accurate and timely laboratory diagnosis. Traditional methods such as potassium hydroxide...
Saved in:
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
BMC
2025-07-01
|
| Series: | BMC Infectious Diseases |
| Subjects: | |
| Online Access: | https://doi.org/10.1186/s12879-025-11296-5 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Abstract Background Superficial fungal infections (SFIs) are highly prevalent globally, affecting approximately 20–25% of the population. Their diverse and often non-specific clinical manifestations necessitate accurate and timely laboratory diagnosis. Traditional methods such as potassium hydroxide (KOH) microscopy and fluorescence staining, although widely used, are limited by operator dependency, variability in sensitivity, and time-consuming procedures. To improve diagnostic accuracy and efficiency, we evaluated the performance of a novel Artificial Intelligence (AI) -powered Fluorescence Microscopic Image Analyzer (FMIA) in the early detection of SFIs. Results Among 300 patients with suspected SFIs, 241 were confirmed using a comprehensive clinical and mycological reference standard. FMIA achieved the highest diagnostic sensitivity (96.27%), outperforming both fluorescence staining (92.95%) and KOH microscopy (75.52%). FMIA also demonstrated high specificity (96.61%) and an area under the ROC curve of 0.96. In spore-dominant infections such as Malassezia folliculitis, genital candidiasis, tinea capitis, and seborrheic dermatitis, where KOH microscopy showed lower detection rates ranging from 29 to 59%, FMIA achieved significantly higher sensitivities between 83% and 100%. Across various SFI subtypes, FMIA consistently exhibited excellent diagnostic performance, achieving 100% detection for tinea pedis, tinea manuum, tinea faciei, and pityriasis versicolor. The system provided results within three to five minutes at a low per-test cost, while automated focusing and frame validation effectively minimized false positives caused by artifacts. Conclusions The FMIA system offers a highly accurate, rapid, and cost-effective tool for diagnosing SFIs. By addressing the limitations of traditional microscopy and current AI-based image analysis methods, FMIA improves diagnostic precision and efficiency. Its fully automated workflow delivers significant clinical value for routine application in both high-throughput diagnostic laboratories and resource-limited healthcare settings. |
|---|---|
| ISSN: | 1471-2334 |