![](https://assets.isu.pub/document-structure/230223030926-a2db0d32b3d01455154fcb592a7edc50/v1/a41d1fd7bcc2919eff53ea6201f4a8a5.jpeg?width=720&quality=85%2C50)
1 minute read
Artificial Intelligence to Help Further Dermatopathology Research
With the opening of the new Dermatopathology Laboratory at 525 West 57th Street scheduled for the end of 2023, the Department looks forward to expanding its services to additional partners in the tristate area. The expansion of our research in dermatopathology will be facilitated by Randie H. Kim, MD, PhD, a double boarded dermatologist/dermatopathologist recently recruited from NYU Langone, and holds promise for the future.
Dr. Kim’s research focuses on cutaneous oncology and the practices of dermatopathology, with collaboration on projects utilizing artificial intelligence and advanced computational techniques to analyze digitized whole slides of melanoma.
Dr. Kim has vast expertise in digital dermatopathology and applications of artificial intelligence and will create standards for slide image archival and develop a digital dermatopathology database for future research.
di erentiate benign skin lesions as compared to malignant skin cancers. She uses these approaches to help her clinical and dermatopathology diagnoses, and plans to integrate these approaches in future research studies and grants.
Computer algorithms have been integrated into diagnostics and prognostics. The use of these methods increases accuracy when detecting skin cancers. In clinical research, AI helps to perform deep phenotyping of patients by collecting data in electronic health records.
Advanced image analysis in dermatopathology includes deep learning techniques applied to whole slide images for tasks such as classification, mutation prediction, and prognosis.
![](https://assets.isu.pub/document-structure/230223030926-a2db0d32b3d01455154fcb592a7edc50/v1/d1b505ba31f5914528ceab6641ce3291.jpeg?width=720&quality=85%2C50)
AI in dermatology is still in its infancy, and will likely continue to grow in the future, but will first require acquisition of a large volume of high-quality images and other data, as there are many dermatologic conditions with varied clinical presentations.