Using computers to predict lung cancer type and severity from sample analysis could be a more effective method than relying on the judgement of human pathologists.
This is according to a new study from Stanford University Medical Center, which evaluated a machine-learning approach to identifying critical disease-related features in lung cancer samples.
“A new study has offered evidence that a machine-learning approach can predict lung cancer type and severity more accurately than human pathologists.“
Researchers used 2,186 images from patients with either adenocarcinoma or squamous cell carcinoma to train a computer software program to identify cancer-specific characteristics.
Nearly 10,000 individual traits were identified - many times more than the several hundred usually assessed by pathologists, due to the fact that many of these characteristics cannot be detected by the human eye.
As such, the computers were able to accurately differentiate between the two types of lung cancers and predict patient survival times much better than the standard approach of pathologists classifying tumours according to grade and stage.
Dr Michael Snyder, professor and chair of genetics at the university, said: "This brings cancer pathology into the 21st century and has the potential to be an awesome thing for patients and their clinicians."
With over 20 years of experience within the science market, we at Zenopa have the knowledge skill and expertise to help find the right job for you. To find out more about the current science roles we have available, you can search for the latest job roles, register your details, or contact the team today.See all the latest jobs in Science