A-Plus ML Algorithm Promises Early Cancer Detection

City of Hope and TGen scientists developed A-Plus, a revolutionary tool for detecting cancer with a tiny blood sample.

By Sunil Sonkar
2 Min Read
A-Plus ML Algorithm Promises Early Cancer Detection

Scientists at City of Hope in California, teamed up with the Translational Genomics Research Institute (TGen) to create a game-changing machine learning tool called A-Plus. This new tool might change how we find cancer. It only needs a little bit of blood to work its magic.


In a recent study published in Science Translational Medicine, A-Plus wowed by showing a 40.5% accuracy in spotting 11 different cancer types with an impressive 98.5% correctness. When combined with aneuploidy and eight common protein markers, the detection rate shot up to 51%, maintaining a high accuracy of 98.9%. They tested it on 7615 samples from 5178 people including those with solid tumors as well as those without cancer.

Dr. Kamel Lahouel, who is part of the research team, pointed out that A-Plus tool is good at diagnosing cancer with very few mistakes. Scientists used RealSeqS to make certain parts of the DNA stand out and then they let A-Plus work its magic to figure out which cells were normal and which had cancer. By looking at free-floating bits of DNA, the algorithm showed it is good at telling the difference between cancer and normal tissue.

They tested A-Plus in different groups and it was great at finding cancers, especially in the esophagus as well as stomach. But the scientists mentioned that the results might be a bit influenced by things like ethnicity and gender. They will be conducting more tests to ensure the tool works well for everyone.

What is exciting is that they plan to do a big test in the summer of 2024 with real patients to compare A-Plus with the usual tests. This test will see if A-Plus can really diagnose cancer early, which could change how we deal with this sickness.

Share This Article
Leave a comment

Leave a Reply

Your email address will not be published. Required fields are marked *