Before he even started high school, Fox Chapel Area freshman Adhitya Thirumala was using his technological prowess to develop potential advancements in the medical field.

Thirumala, 13, was named among the Top 300 middle school innovators from across the country in the 2020 Broadcom MASTERS, a prestigious STEM competition.

The contest seeks to motivate young scientists to solve vital 21st century challenges.

There were 3,476 applicants.

“From a young age, I have always loved interacting with computers, using them as a vehicle for my imagination, creativity and problem-solving skills,” said Thirumala of Indiana Township.

His project, “Developing an Early Screening Application to Detect Lung Abnormalities Using Machine Learning,” was completed as an eighth-grader at Dorseyville Middle School.

Nicki Wood, from the Pittsburgh Regional Science and Engineering Fair, said Thirumala devised a machine learning (ML) classifier to detect cardiomegaly in lung x-rays. His initial model had a 60% accuracy level that has grown to 85%, and he plans to create an app that can be used as a screening tool for a wide array of diagnoses.

The teen, whose other interests include golf, guitar and Pokémon, was one of 15 students in the state to be honored.

Thirumala said his passion for coding and robotics was spurred by participating in the First Robotics Challenge team and local summer camps.

“Last summer, I learned how to make a basic ML classifier to automatically determine which numerals were written in a sample of handwritten digits,” he said. “This experience fueled my interest in computational techniques and opened my eyes to the various ways ML could be utilized in fields like health care, business and education.”

He started to brainstorm ways to apply ML to his own interests in medicine, Thirumala said.

Using a data set called ChestX-ray8 from the National Institute of Health and a format of a convolutional neural network, Thirumala built his model that analyzes x-ray abnormalities.

His Broadcom MASTERS project targets medical misdiagnoses and aims to cut health care errors through ML.

“Many people throughout the world live in places where access to proper radiology facilities is limited or these facilities are overwhelmed,” Thirumala said.

“My project will be helpful in these situations because it will allow many x-rays to be analyzed and underlying lung conditions to be detected more efficiently.”

Copyright ©2025— Trib Total Media, LLC (TribLIVE.com)