Keratinocytic lesions suite
This advanced suite comprises AI models and algorithms dedicated to the in-depth analysis and diagnosis of Basal Cell Carcinoma (BCC) and Squamous Cell Carcinoma (SCC). Utilizing cutting-edge technology, each component is designed to enhance the accuracy and efficiency of skin cancer diagnosis and assessment.
Components and Capabilities
BCC and SCC Classifier:
Utilizes sophisticated AI algorithms to differentiate between Basal Cell Carcinoma and Squamous Cell Carcinoma.
Trained on a comprehensive dataset of 6,353 whole slide images (WSIs), including 1,208 SCC and 5,145 BCC cases, annotated by professional dermatopathologists.
Internal validation dataset of 1,459 WSIs (307 SCC, 1,152 BCC) showed:
Accuracy: 0.96
Weighted Precision: 0.96
Weighted Recall: 0.96
Basal Cell Carcinoma Segmentation Model
Precisely localizes BCC regions on histopathological slides.
Trained on a dataset of 501 WSIs containing BCC tumors, with segmentation masks crafted by professional dermatopathologists.
Performance metrics:
Intersection over Union (IoU): 0.9
Accuracy: 0.95
Sensitivity: 0.92
Specificity: 0.98
Squamous Cell Carcinoma Segmentation Model
Expertly identifies and localizes SCC regions in slide images.
Trained on 442 WSIs specific to SCC, with professional dermatopathologist-created segmentation masks.
Model demonstrates:
IoU: 0.83
Accuracy: 0.92
Sensitivity: 0.88
Specificity: 0.97
Dataset Used and Metrics
For both BCC and SCC models, datasets were meticulously curated and annotated by expert dermatopathologists, ensuring the highest level of detail and accuracy.
The models were validated internally and externally to verify their robustness and reliability in clinical settings.
Technical Stuff
Compatible with most modern web browsers and mobile devices.
Optimal performance observed with a minimum of 2GB RAM and a stable internet connection.