What AI algorithms have your already produced and what is the testing to support their validity?
Our initial set of algorithms are licensed from SemanticMD, a deep learning medical image analysis platform. We have algorithms across radiology, pathology, dermatology, and ophthalmology. Specific examples include identifying TB and pneumonia in a chest X-ray, detecting diabetic retinopathy in fundus eye images, and identifying suspected areas for melanoma from cellphone camera images. Our AI4TB initiative includes Dr. Paras Lakhani, a radiologist with numerous publications in applying deep learning at Jefferson Radiology and we have deployed and tested our solutions prospectively on patients in South Africa and China. We are pursuing a similar validation path with our AI4DR initiative.