A new study published in Cell Systems shows the potential of artificial intelligence-based tools to revolutionize the pathology of cancer and a variety of other diseases. Specifically, these new artificial intelligence algorithms have been developed by researchers at UT Southwestern to assess the metastatic potential in skin cancers .
Reverse engineering
Methods based on deep learning (deep learning) able to distinguish small differences in images that are essentially invisible to the human eye, researchers have proposed using this latent information to look for differences in the characteristics of the disease that could provide information on prognoses or guide treatments.
The differences that distinguish AI are generally not interpretable in terms of specific cellular characteristics, a drawback that has made AI difficult to sell for clinical use, but here AI was used to look for differences between images of melanoma cells with high and low metastatic potential, a feature that can mean life or death for skin cancer patients, and they then reverse engineered their findings to find out which features in these images were responsible for the differences .
Using tumor samples from seven patients and available information on their disease progression, including metastasis, the researchers took videos of approximately 12,000 random cells living in Petri dishes , generating about 1,700,000 raw images. The researchers then used an artificial intelligence algorithm to extract 56 different abstract numerical features from these images.
They then found a characteristic that could precisely discriminate between cells with high and low metastatic potential. By manipulating this abstract numerical feature, they produced artificial images that exaggerated the visible features inherent in metastasis that human eyes cannot detect . Specifically, highly metastatic cells produced more extension in the form of pseudopods (a type of finger-like projection) and had greater scattering of light, an effect that may be due to subtle rearrangements of cell organelles.