Computer algorithms identify cancer from tissue images

Artificially intelligent computer algorithms have been found to equal human pathologists in detecting metastases of breast cancer in lymph nodes, as demonstrated during Radboud University Medical Center’s CAMELYON16 challenge.

The CAMELYON16 challenge, held between November 2015 and November 2016, gave machine learning researchers the opportunity to create a computer algorithm that can independently make a diagnosis based on pathology images — specifically, the detection of metastases of breast cancer in lymph nodes. The challenge was accepted by 23 research groups from around the world, with the results published in the journal JAMA.

Participants were given 270 digital images of tissue preparations, of which it was already known whether, and where, metastases could be found. Using these images, the participants developed algorithms which had to distinguish between images with and without metastases, and then locate the exact position of the metastases. They then received 129 new images, which they used to test the algorithms. The 129 test images were also assessed by 11 experienced pathologists, who analysed them under conditions that were similar to a realistic hospital situation. In addition, one pathologist was allowed to take as much time as she wanted to diagnose the images.

The 23 participating research groups submitted a total of 32 computer algorithms, the most successful of which used ‘deep learning’, in which the computer learns to recognise patterns based on a large number of examples. The top algorithms performed significantly better than the pathologists who assessed the images in the realistic hospital work situation.

The winning algorithm, meanwhile, detected metastases as effectively as the pathologist who worked without time pressure. On average, this algorithm generated a false positive (a metastasis that actually did not exist) only 1.25 times per 100 images.

“For the first time we have seen that a computer can make this diagnosis as effectively as a pathologist,” said Jeroen van der Laak, who coordinated the challenge. “A pathologist with this algorithm is therefore better off than a pathologist without. The patient receives the result of the biopsy sooner and the algorithm helps pathologists make better diagnoses, even under time pressure.”

The researchers expect that the technique will be suitable for use in patient care within a few years, and that the computer algorithm will increasingly be capable of diagnoses.

Image credit: ©stock.adobe.com/au/vectorfusionart