More than 5 million diagnoses of skin cancer are made each year in the United States, about 106,000 of which are melanoma of the skin*. Although melanoma occurs more rarely than several other skin cancers, patients’ long term survival rate is extremely low if the diagnosis is missed. Diagnosis, however, is complicated by a high discordance rate among pathologists when distinguishing between melanoma and benign melanocytic lesions.
To address these challenges, Proscia has developed a tool (pathology deep learning system, or PDLS) that allows pathology labs to sort and prioritize melanoma cases in their workflow and potentially improve turnaround time by prioritizing challenging cases and routing them directly to the appropriate subspecialist.
This deep learning system achieved:
- Area Underneath the ROC Curve (AUC) values of 0.93 classifying Melanocytic Suspect specimens on the reference lab, 0.95 on the first validation lab, and 0.82 on the second validation lab.
- Automatic sorting and triaging of skin specimens with high sensitivity to Melanocytic Suspect cases, indicating that a pathologist would only need between 30% and 60% of the caseload to address all melanoma specimens.
Authors:
- Sivaramakrishnan Sankarapandian | Proscia
- Saul Kohn | Proscia
- Vaughn Spurrier | Proscia
- Sean Grullon | Proscia
- Rajath E. Soans | Proscia
- Kameswari D. Ayyagari | Proscia
- Ramachandra V. Chamarthi | Proscia
- Kiran Motaparthi | Department of Dermatology, University of Florida College of Medicine
- Jason B. Lee | Department of Dermatology, Sidney Kimmel Medical College at Thomas Jefferson University
- Wonwoo Shon | Department of Pathology and Laboratory Medicine , Cedars-Sinai
- Michael Bonham | Proscia
- Julianna D. Ianni | Proscia
*American Cancer Society. American Cancer Society. Cancer Facts and Figures 2021.
