September 1, 2022 – It’s hard to know what the road ahead for a cancer patient will look like. Lots of evidence is considered, such as patient health and family historyTumor grade, tumor stage, and tumor cell characteristics. But in the end, the expectations are up to the health professionals who analyze the facts.
This can lead to “widespread diversity,” says Faisal Mahmoud, PhD, assistant professor in the department of computational pathology at Brigham and Women’s Hospital. He says patients with similar cancers can end up with very different prognosis, with some being more (or less) accurate than others.
That’s why he and his team have developed an artificial intelligence (AI) program that can form a more objective – and possibly more accurate – assessment. The aim of the research was to find out if artificial intelligence was a practical idea, and the team’s findings were published in cancer cells.
And because diagnosis is key in determining treatments, more accuracy could mean more treatment success, Mahmoud says.
“[This technology] It has the potential to generate more objective risk assessments and, in turn, more objective treatment decisions, he says.
Building artificial intelligence
The researchers developed the AI using data from the Cancer Genome Atlas, a general index of profiles of different types of cancer.
Their algorithm predicts cancer outcomes based on Histology Genomics (a description of the tumor and how fast the cancer cells grow) and genomics (using DNA sequencing to assess Tumor at the molecular level). Mahmoud points out that histology has been the diagnostic standard for over 100 years, while genomics is being used more and more.
“Both are now commonly used for diagnosis in major cancer centers,” he says.
To test the algorithm, the researchers selected the 14 types of cancer with the most available data. When histology and genomics were combined, the algorithm gave more accurate predictions than it did with either of the two sources of information alone.
Not only that, but the AI used other markers — such as a patient’s immune response to treatment — without being asked to do so, the researchers found. Mahmoud says this could mean that AI can detect new signs that we don’t know about yet.
While more research is needed – including large-scale testing and Clinical trials Mahmoud is confident that this technology will be used in real patients’ lives one day, probably in the next 10 years.
“From now on, we will see large-scale AI models capable of ingesting data from multiple modalities,” he says, such as radiology, pathology, genomics, medical records, and family history.
Mahmoud says that the more information an AI can take into account, the more accurate its assessment will be.
“Then we can continually assess a patient’s risk in a mathematical and objective way.”