Sophomore Opportunity Leads to Published Research on A.I. in Cancer Studies

April 1, 2022

Yvonne Gay

A young college students works inside of a vent.
Ella Halbert '23 at work in professor Mary Garvin's biology lab.
Photo credit: Yvonne Gay

A published manuscript coauthored by Ella Halbert ’23 offers software tools that pathologists could use to assist with a tedious, yet necessary process in the study of disease.

“The study of disease often involves analyzing tissue samples to diagnose specific diseases,” explains Halbert, a biology and Hispanic studies major. “Typically, a pathologist will analyze slides of tissue samples to determine any abnormalities, but this process can be made more efficient through the use of digital pathology, which relies on images of slides, image processing, and machine learning. Essentially, machine learning programs can be trained to recognize and identify abnormalities in tissue samples, saving time and energy, and supporting pathologists’ workloads.”

Halbert’s contribution to this effort began last February, when she connected with Jacob Rosenthal ’18 through SOAR. The college’s Sophomore Opportunities and Academic Resources program provides opportunities for students to connect their interests inside and outside the classroom as well as offers instruction on how to use Oberlin resources, and provides individualized help with resumés. 

A portrait of a college student in a lab
Ella Halbert '23. Photo credit: Yvonne Gay

After submitting her resumé and interest materials, Rosenthal, an imaging data scientist and data engineer, invited Halbert to intern at the Dana-Farber Cancer Institute. She was joined by nine other members of the group who included scientists, public health professionals, and professors of research in pathology from the Dana-Farber Cancer Institute, Massachusetts Institute of Technology, Weill Cornell Medicine, and Harvard T.H. Chan School of Public Health. 

As the only undergraduate student member of such an experienced team, the two-month internship would afford Halbert a broader understanding of pathology and introduce her to a facet of medicine that was new to her: data science. 

“I knew a little about pathology going into the internship, but I didn’t know anything about the challenges of combining data science, imaging techniques, and pathology,” says Halbert. “This experience has made me think more deeply about how the process of treating patients is changing as technology advances.”

To prepare for her internship, Halbert independently studied the basics of Python—a high-level, general-purpose programming language. The skill set made her a welcome addition to the group’s artificial intelligence operations team, where she worked closely with Rosenthal and Renato Umeton, associate director of Artificial Intelligence Operations and Data Science Services at the Dana-Farber Cancer Institute.

The manuscript written by the group is based on PathML, a software toolkit developed by Rosenthal that processes and analyzes pathology slides. “Most importantly, this toolkit is meant to lower the barrier to entry for digital pathology so that pathologists with limited programming experience can utilize this powerful tool for their own research or clinical practice,” says Halbert. 

Although the team was unable to work in person, one-on-one virtual meetings held several times a week and weekly group sessions kept the lines of communication flowing.

The completed abstract highlights three themes to guide development of computational tools: scalability, standardization, and ease of use. The group then applied these principles to develop PathML, describe the design of the software’s framework, and demonstrate applications in diverse use cases.

In December 2021, the group’s completed manuscript—“Building Tools for Machine Learning and Artificial Intelligence in Cancer Research: Best Practices and a Case Study with the PathML Toolkit for Computational Pathology” was published in Molecular Cancer Research, a monthly journal produced by the American Association for Cancer Research. Halbert received author credit for her work with the PathML software.

“I think global awareness and cultural competencies are really important for any field of study, particularly the sciences,” says Halbert. “I’m planning to pursue a medical degree, and being able to relate with patients across cultural differences is a vital skill.”

Halbert currently studies the ecology of disease in professor Mary Garvin’s biology lab. She has accepted a summer research position at the Mountain Lake Biological Station (MLBS) in Pembroke, Virginia, where she will be working with Chloé Lahondère on her research of Mosquito Thermal Biology and Interactions with Plants and Herpetofauna.

You may also like…

Friday Afternoon with Lab Crawl

October 30, 2023

This year’s Lab Crawl drew what may be the largest crowd in the event’s illustrious history, with some 500 students and dozens of faculty taking part. Missed out on the excitement? It looked something like this.
A person demonstrating on a whiteboard to students in a laboratory.