ACCURACY OF IMAGE-BASED DEEP LEARNING IN ORAL SQUAMOUS CELL CARCINOMA DIAGNOSIS: A SYSTEMATIC REVIEW AND META-ANALYSIS.
J Evid Based Dent PractJun 13, 2026 (epub)
Hai-Tao Zhu, Pei-Chi Chen, Rui-Min Wang et al.
✦ AI-curated · Sources linked
In 30 seconds
This systematic review and meta-analysis evaluated the effectiveness of image-based deep learning (DL) in diagnosing oral squamous cell carcinoma across 52 studies. The findings indicate that DL shows strong diagnostic performance, with a pooled sensitivity of 0.99 and specificity of 0.98 for Raman spectral images, and similar high performance for oral photographs and histopathological images, all with an area under the curve (AUC) of 0.99.
Key findings
- For Raman spectral images, DL achieved a pooled sensitivity of 0.99 and specificity of 0.98.
- DL based on oral photographs yielded a pooled sensitivity of 0.97 and specificity of 0.93.
- Histopathological images showed a pooled sensitivity of 0.98 and specificity of 0.97 with DL.
Why it matters
The promising diagnostic accuracy of image-based deep learning for oral squamous cell carcinoma highlights its potential as an adjunct tool in clinical practice, potentially improving early detection and patient outcomes.