Deep Learning for Melanoma Detection Using Optical Coherence Tomography
Author Information
Author(s): Lai Pei‐Yu, Shih Tai‐Yu, Chang Yu‐Huan, Chang Chung‐Hsing, Kuo Wen‐Chuan
Primary Institution: National Yang Ming Chiao Tung University
Hypothesis
Can a convolutional neural network (CNN) effectively identify melanoma and predict risk using optical coherence tomography (OCT) imaging?
Conclusion
The CNN achieved high sensitivity and specificity in classifying melanoma and healthy tissues, indicating its potential for early diagnosis.
Supporting Evidence
- The CNN model classified melanoma and healthy tissues with a sensitivity of 0.99 and specificity of 0.98.
- Longitudinal OCT imaging allowed for monitoring melanoma progression over time.
- The study is the first to apply deep learning to OCT images for melanoma detection.
Takeaway
This study shows that a computer program can help doctors find skin cancer early by looking at special pictures of skin.
Methodology
The study used a convolutional neural network to analyze OCT images from melanoma and control mice, performing longitudinal tests.
Potential Biases
Potential biases due to the limited sample size and the specific animal model used.
Limitations
The study used a small number of mouse models, which may not fully represent human melanoma diversity.
Participant Demographics
Mice models including melanoma mice, dysplastic nevus mice, and their respective controls.
Statistical Information
P-Value
0.001
Statistical Significance
p<0.05
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website