Can computers identify lung cancer type, severity, survival rate?

Over the year’s computers have become faster and smarter, now being able to predict lung cancer in patients thanks to researchers from Stanford University School of Medicine.

Researchers used 2,186 images from the Cancer Genome Atlas to “teach” a computer to identify features pertaining to lung cancer by identifying 10,000 different traits, many more than the few hundred beings assessed by pathologists. The computer is able to examine cell's size and shape, the texture of the nucleus, and its spatial relationship to neighboring tumor cells. These identifying factors allow the computer to predict the type of lung cancer, its severity, patient survival times and differences between cancerous and healthy tissue.

"Pathology as it is practiced now is very subjective," said Michael Snyder, PhD, professor and chair of genetics. "Two highly skilled pathologists assessing the same slide will agree only about 60 percent of the time. This approach replaces this subjectivity with sophisticated, quantitative measurements that we feel are likely to improve patient outcomes."

The study only focused on the detection of lung cancer, though researchers are hopeful such programs may also be used to identify other types of cancer.

"Ultimately this technique will give us insight into the molecular mechanisms of cancer by connecting important pathological features with outcome data," said Snyder. "We launched this study because we wanted to begin marrying imaging to our 'omics' studies to better understand cancer processes at a molecular level. This brings cancer pathology into the 21st century and has the potential to be an awesome thing for patients and their clinicians."