Technology has always augmented humans’ capabilities to achieve the milestones that were once considered impossible. In medical science, AI has been matching shoulders with humans to redefine the way we diagnose, analyze and cure diseases. And now the bar has been raised much higher by AI as it has achieved near 100% accuracy in diagnosing signs of cancer in humans.

The news comes from an international team of scientists, including from Charles Darwin University in Australia, which has developed a new AI model called ECgMPL that can assess microscopic images of cells and tissues and identify the signs of endometrial cancer, which manifests as a reproductive tumour. The team of researchers achieved an accuracy of 99.26% and claimed that the model can also replicate its success for oral cancer and colorectal cancer. A major accomplishment for this model is that exceeds the accuracy of other existing models and is computationally efficient. Using ablation studies, efficient training and self-attention mechanisms, the model generalizes quite well across several histopathology datasets which makes it robust model to deal with cancer diagnosis.

In other words, the model enhances the quality of the microscopic scans enough for spotting the problematic growth areas that signal towards possibility of cancer. With this novel AI emerging in the healthcare landscape, doctors will find an additional weapon in their arsenal against cancer and will be able to make more informed data-driven decisions to help people battle cancer effectively and swiftly.