A new machine learning model that can classify lung cancer subtypes could help pathologists achieve quicker, more accurate diagnoses, according to a study published in Scientific Reports.
For the study, researchers from Dartmouth University's Norris Cotton Cancer Center in Lebanon, N.H., created an algorithm to help identify the subtypes of lung adenocarcinoma — the most common form of lung cancer.
Researchers used 422 whole-slide images of cancerous lung tissue from Dartmouth-Hitchcock Medical Center in Lebanon to train, develop and test the machine learning algorithm. Of 143 slides used for testing, the algorithm identified the same diagnosis as human pathologists about 67 percent of the time. By comparison, human pathologists agreed with each others' diagnoses about 62.7 percent of the time.
"Our study demonstrates that machine learning can achieve high performance on a challenging image classification task and has the potential to be an asset to lung cancer management," lead study author Saeed Hassanpour, PhD, an assistant professor of biomedical data science at Dartmouth, said in a press release. "Clinical implementation of our system would be able to assist pathologists for accurate classification of lung cancer subtypes, which is critical for prognosis and treatment."