Four factors came together to make digital pathology a must-have instead of a nice-to-have. First, COVID-19 sent pathologists home and challenged them to figure out new remote workflows. Second, cloud storage got cheaper and more robust, allowing for the sharing of massive images. Third, the development of advanced artificial intelligence has enabled pathologists to gain more insights out of a digital pathology system. Finally, cancer cases are on the rise and pathologists are on the decline, creating a critical gap that’s only going to grow in coming years.

AI use in pathology so far has led to interesting observations. Reading a pathology image is a combination of speed and accuracy. Pathologists that take their time and carefully examine the image seem to have better accuracy and in studies have shown to be equivalent to the algorithms. But, algorithms are producing that level of accuracy much faster. However, pathologist plus algorithms have shown improved accuracy and significantly more speed in interpreting digital path.  Much like radiology where AI may be able to help in the actual acquisition of the images, the use of algorithms to improve the quality of out of focus images in pathology may be useful. Also, as molecular diagnostics become part of the diagnosing and classification of path results, machines can integrate info about ‘omics (DNA sequence, RNA sequence, proteomics, methylation patterns, etc) into the final diagnosis. AI can ingest and analyze all of this information to yield better diagnosis. AI has shown to be effective in relating specific mutations to treatments.