D.A. Student Publishes Research on Using AI to Diagnose Tuberculosis

By Claire Curry

Imagine being able to diagnose tuberculosis by using Artificial Intelligence to scan chest x-rays. That’s exactly what Anusuiya Bhorkar ’26 tested in an independent 16-month research study under the mentorship of an Oxford University scholar. Her research paper reveals how technology opens a new door in diagnosing the deadly disease, and it was recently accepted for publication by The Journal of Emerging Investigators and included in the Oxford University Research Archive.

Engaging in Real-World Research

Opportunities for high school students to conduct advanced research with leading global scientists don’t come along every day. So, when Anusuiya learned about a summer research program called Veritas AI from her biology teacher, Dr. Lauren Serpagli, she jumped on it. 

Serpagli, who is co-chair of D.A.’s STEM department and has taught at the school for two decades, said that Dominican Academy has always prioritized science, technology, engineering and math. “We’re really promoting young women in the STEM fields,” she said, adding that the department is continually updating its courses and lab facilities to keep students on the leading edge. “Whenever I learn about extracurricular opportunities like this one, I share them with every student.”

A Deep Dive into AI and Medicine

Anusuiya signed up for the Veritas AI 10-week summer program that combines lectures and training with hands-on team projects and individual research. Created and run by Harvard scholars, it was a perfect fit for the student who loves technology and aspires to become a medical doctor. 

In addition to lectures and coding sessions, Anusuiya worked on a team project that involved segmenting cardiac MRIs. “We developed a code that would be able to look at MRI scans and break apart the images and extract the features of the heart,” she explained, a solution that provides doctors with fast access to vital information to treat their patients. “For the first time, I saw that technology isn’t just something that I have to learn for the future, but something that is making a difference now.”

Next, she worked on her own research project with Oxford Scholar and Harvard Fellow Ricardo Gonzales, Ph.D., as her mentor and advisor. Anusuiya’s study involved using AI to diagnose one of the world’s most deadly infectious diseases: pulmonary tuberculosis. While millions around the world are diagnosed with the disease annually, many go unaware, particularly in developing countries where access to healthcare is limited. 

Inspired by a Family Friend

When she was considering her research topic, Anusuiya said, “I wanted to do research on medicine and healthcare accessibility.” Her interest in tuberculosis was sparked after learning about a family friend who contracted the disease on a fashion internship in an Indian manufacturing facility where there was an outbreak. 

“This is someone who was in prime health and had access to top medical care, and he was bedridden for six months,” Anusuiya explained. “It made me think of other factory workers who don’t have the same access to healthcare. According to the World Health Organization’s 2022 Global Report, there’s a 50% mortality rate for people who contract tuberculosis who don’t receive treatment.”

The Research Process

“Our study aims to enhance diagnostic opportunities by developing a deep-learning model to categorize pulmonary X-ray scans into ‘normal’ and ‘tuberculosis’ classes,” Anusuiya explained. “My model works by a series of code that looks at images and turns that into an array of numbers it can break down and assign. I’ve taught it to assign that one set of numbers refers to a pulmonary x-ray scan showing tuberculosis. The [other] set relates to a pulmonary x-ray scan showing a normal reading.”

Anusuiya developed three models trained on datasets of pulmonary X-ray scans from Alabama, China, and Southeast Asian countries. Her research culminated in the paper titled “Effects of data amount and variation in deep learning-based tuberculosis diagnosis in chest X-ray scans.” Anusuiya served as the lead author of the peer-reviewed paper, which outlines the process and findings. 

“The models trained on more data perform more accurately due to greater data variation and amount,” Anusuiya said. “I hope to use this project as a way to bring awareness to the fact that there are not a lot of publicly accessible tuberculosis datasets and [in order to scale this work to make an impact], we need a huge tuberculosis database.” With that resource, using deep learning models to diagnose tuberculosis could make testing more accessible and ultimately, improve global health.

Committed to Supporting Young Women in STEM

Over the years, Dr. Serpagli has watched Anusuiya and many other students grow and develop confidence in their capabilities. “We never know what the future holds and we have to make sure [they] are continually encouraged to explore their imaginations. Young thinkers are where it’s at. They’re the ones who come up with the great ideas and we have to find ways to inspire that.” 

STEM facilities at D.A. also have adapted to meet the technological moment. The Barbara Robotti Murray ’64 Science Center opened in 2011 and is state-of-the-art space for hands-on learning and enhancing students’ understanding of science, and the adjacent Natalie Frankowski '14 Collaborative Classroom is a custom-built creative space that lets students learn valuable skills in design, engineering, and problem-solving by using 3D printers to bring their ideas to life.

An Experience that Taught More than Science

Juggling her research project alongside a rigorous honors curriculum and many extracurricular activities, including the Pre-med Society, Mock Trial, serving as a Student Ambassador — on top of commuting to school from New Jersey, which involves three trains each way — was no small feat for Anusuiya.

In addition to honing her time management skills and learning to stay calm and focused under pressure, her research mentor encouraged Anusuiya to aim for her personal best.

About Dr. Gonzales, Anusuiya said, “He taught me grit. He pushed me to go the extra mile to do a great research paper. I’d come back from classes and I’d have a coffee on one side and water on the other and I’d be like, ‘I have to get this done.’ It’s something that has really shaped my high school experience.”

As for her future, Anusuiya plans to pursue a career of service, something Dominican Academy instills in its students. “I want to do more research with AI and medicine on a pre-med track, and I also want to look into public health and administration, healthcare accessibility and how we distribute resources as a society. When I envision [working in] the medical field, I feel you need to have the human connection…the human soul.”