The Siemens Foundation named Manan Shah (12) as a national finalist in the Siemens Competition on Nov. 7.

Manan earned top individual honors and a $3,000 scholarship for developing a computational model that would help pathologists more efficiently assess breast cancer severity in patients.

“[My project] predicts the scale both on a categorical level, so one, two or three, where three is the most severe, and on a molecular level, so investigating how likely this tumor is to proliferate based on biological factors from samples [and] biopsy tissues,” he said.

After his grandfather was diagnosed with prostate cancer following a long period of examinations, Manan was inspired to create a way to expedite the process.

“In India and other developing countries, it’s difficult for a lot of patients to get diagnosed for these types of cancers because tissue examinations take a long time and are inefficient for pathologists to grade,” he said. “To obviate the subjectivity and the underlying time [three or four hours] involved in this kind of process, I hoped to develop a method that would fix those concerns.”

He worked with his mentor, Dayong Wang, and other researchers to further develop his approaches and methods at the Beck Lab, a part of the Beth Israel Deaconess Medical Center of the Harvard Medical School.

“The P.I. [Principal Investigator] at the lab was Dr. Andrew Beck, and he was also pretty helpful in helping me write the paper and create the presentation,” he said. “The other pathologists helped verify my methods, so it was like a collaborative effort from fields of biology and computer science.”

Though most of the problems he encountered while developing his project were minor computational errors, an issue that required deeper research and experimentation involved accounting for the various methods used internationally to stain tissues extracted from biopsies. 

“In different parts of the U.S. and also internationally, these stains may vary with different colors and intensities, and so my methods that I developed wouldn’t generalize to other intensities,” he said. “I corrected the staining normalization by employing Dr. Bejnordi’s algorithm to normalize all of the stains, resulting in an improved accuracy.”

Upper school physics and science research teacher Chris Spenner and science and biology teacher Anita Chetty also helped Manan throughout his project by editing his drafts and providing feedback on his presentations.

“[Mr. Spenner and Ms. Chetty] have provided me with both a really solid biological and computational base,” he said. “[They’ve also] provided me feedback on my papers and revisions, so the end product was something to be proud of.”

Manan will move on to present his work alongside five other individual finalists and six team finalists at George Washington University in Washington, D.C. on Dec. 2 to 6 for a chance to become a national winner of the Siemens Competition.