A premier academic medical center in New York City, renowned for its excellence in clinical care, research, and medical education, consistently ranking among the top hospitals in the nation.

Researchers faced difficulties in detecting melanoma metastasis due to the need for each user group to develop their own models for identifying cancer in test animals. This manual process slowed down research efforts and led to inconsistencies in results, hampering progress in developing treatments and cures.

The client, with Turing’s support, implemented an AI-powered solution to automate the detection of melanoma metastasis using only microscopic images:

• Image scanning and annotation: Human xenograft models were developed, and metastasis in specific organs (brain, liver) was scanned and annotated by pathologists, marking tumor and non-tumor regions.
• Deep learning model development: A deep learning network was trained on these manually annotated images to detect metastasis accurately.
• Web interface for image detection: A user-friendly web interface was developed, allowing users to upload large images for automatic detection and masking, enabling broader use across various research groups.
• TensorFlow for deployment: The solution was deployed leveraging TensorFlow for model training, image pre-processing, and prediction, ensuring scalability and efficiency.

• Enhanced diagnostic capabilities: Improved accuracy in detecting melanoma metastasis, resulting in better patient outcomes and expanding use in diagnosing other types of cancer.
• Accelerated research and development: Joint research initiatives led to advancements in melanoma treatment and the development of new diagnostic tools and therapies.
• Broader adoption and impact: Increased global adoption of the technology contributed significantly to the fight against melanoma metastasis and improved cancer care.