If we can train machine models for vision systems to the same parity as a pigeon brain (no offence to pigeons, but surely this is possible) then it makes sense that we can use artificial intelligence to automate or even more accurately conduct this task. IBM Watson has already shown how it's vision system has been trained to spot melanoma: https://www.ibm.com/cognitive/au-en/melanoma/
Diagnosing cancer has a lot to do with vision. Pathologists study biopsy samples looking for cancerous cells. Radiologists scan x-rays and MRIs for possible malignancies. The authors of the new study were curious how these specialists acquire the skill to identify features and qualities of an image that signify cancer. To find out they turned to pigeons. Though primates and pigeons haven't shared a common ancestor for over 300 million years (it looked something like this) our visual processing biology is surprisingly similar to that of birds.