Meet Vatsal Sodha

Knee radiograph with a diagram of a deep-learning network.
Vatsal Sodha's interest in machine learning for medicine was sparked by a college project and has taken him from Gujarat to Arizona to a bicoastal telecommute.

Our Center welcomes a new assistant research scientist, Vatsal Sodha.

Vatsal holds a master's in computer science from Arizona State University, where he co-developed machine learning methods for disease detection in CT, MRI, and X-ray images. He hopes that computer-science tools will help meet the world’s demand for clinical expertise.

Sodha became interested in machine learning applications to medical imaging during undergraduate studies in computer science in Gujarat, India. A friend pulled him into a research project on detection of lung nodules, and Sodha found himself excited about the technology. The friend's grandfather had received a late-stage lung cancer diagnosis, underscoring the stakes and promise of automated detection.

"I come from a country that does not have many radiologists available per person," said Sodha in a recent video call. “My degree allows me to somehow reduce that gap.”

Motivated by his experience in Gujarat, Sodha went on to graduate studies at Arizona State. While there, he performed research at the Mayo Clinic in Scottsdale and co-created self-supervised networks trained on unlabeled data.

"What we wanted to do is reduce the amount of time radiologists need to spend labeling stuff," Sodha said, referring to manual annotation—a time-consuming, laborious endeavor required by many machine learning models.

Self-supervised networks that need little or no manual labeling may benefit patients by allowing medical imaging research to advance without pulling human radiologists away from clinical cases. The project earned Sodha and co-authors a 2019 Young Scientist Award from the Medical Image Computing and Computer Assisted Intervention (MICCAI) Society and won one of MICCAI's 2020 best paper awards.

At our Center, Sodha will delve into development of semi-supervised and unsupervised deep learning classification models for knee osteoarthritis in order to more accurately predict whether and how far the disease is likely to progress. The project is led at our Center by Cem Deniz, PhD, and supported by the National Institute of Arthritis and Musculoskeletal and Skin Diseases.

Sodha is also curious about the PhD in Data Science—Medical Track program, offered by NYU Center for Data Science in partnership with NYU Grossman School of Medicine. The program's imaging component is taught by our Center's faculty, making CAI2R an excellent place to contemplate potential commitment to PhD training.

Initially, Sodha is telecommuting from San Jose, California. We are looking forward to welcoming him in person at our New York headquarters later in 2021.


Philanthropic Support

We gratefully acknowledge generous support for radiology research at NYU Langone Health from:
• The Big George Foundation
• Bernard and Irene Schwartz

Go to top