Artificial intelligence, developed by researchers at the Washington University School of Medicine in St. Louis, can accurately predict the risk of lung cancer spreading to the brain.
- Researchers created a machine learning algorithm to accurately predict brain metastases using biopsy samples from patients with early-stage non-small cell lung cancer.
- The algorithm has an accuracy of 87% in predicting the development of brain tumors.
- The results need to be validated in a larger study, but researchers believe AI has great potential to provide better care.
Doctors often face a dilemma when treating patients with early-stage lung cancer. They must decide whether to opt for treatments that put strain on the body, such as chemotherapy, radiotherapy or immunotherapy, to eliminate tumors and reduce the risk of spread to the brain or whether lung surgery is sufficient.
However, a team of researchers at Washington University School of Medicine in St. Louis has developed artificial intelligence that could help them find the best option for their patients.
This AI, presented in detail in the review The Journal of Pathologymanages to predict whether lung cancer will spread to the brain or not.
Lung cancer: accurate prediction of brain metastases using AI
The researchers developed the machine learning algorithm using records and lung biopsy samples from 118 patients with early-stage non-small cell lung cancer. Some patients developed brain cancer during a five-year surveillance period, others did not and were in remission. The AI’s ability to predict whether or not there is a risk of brain metastases was measured with lung biopsy samples from 40 other patients.
The algorithm was able to predict the possible development of brain cancer with 87% accuracy. In comparison, four pathologists who participated in the study had an average score of 57.3%. Furthermore, the AI was very reliable in predicting which patients would not develop brain metastases.
“Our results need to be validated in a larger study, but we believe AI has great potential to make accurate predictions and impact care decisions”explains Dr Ramaswamy Govindan, one of the authors of the research, in a communicated.
The expert then adds: “Identifying patients at risk of relapse in the brain can help us develop strategies to intercept cancer early in the metastasis process. We believe AI-based predictions could one day inform personalized treatments” .
The algorithm could help better understand the development of metastases
The developed AI evaluates the characteristics of tumors and healthy cells, in the same way as the human brain by observing scans. However, what the algorithm sees, and allows it to predict the course of the disease, is not yet well understood, the study scientists acknowledge. “This study began as an attempt to find predictive biomarkers”specifies Professor Changhuei Yang of the California Institute of Technology. “But we didn’t find any. Instead, we discovered that AI has the potential to make predictions about cancer progression using biopsy samples already collected for diagnostic purposes .”
The team thus seeks to identify the molecular and cellular characteristics that the AI uses for its predictions. “This knowledge could lead to the development of new treatments and influence the design of imaging instruments optimized for data collection for AI”she says.
Check out our digital video work,
only an account creation is required to access it.
