Google’s AI Can Read Mammograms to Detect Breast Cancer

According to statistics, breast cancer is among the most common and second leading cause of deaths among females. Amid this when Google's research came forward which said that artificial intelligence can read mammograms better than radiologists, it was a sigh of relief for most. Recently, DeepMind and Google Health have developed a new AI system that can detect breast cancer at an early stage. The researchers trained an algorithm on mammogram images from female patients in the US and UK. They trained it to perform better than human radiologists and hence the result has arrived.

According to the MIT Technology Review, "In tests, the AI system decreased both types of error. For US patients, it reduced false negatives and positives by 9.4% and 5.7%, respectively; for UK patients it reduced them by 2.7% and 1.2%. In a separate experiment, the researchers tested the system's ability to generalize: they trained the model using only mammograms from UK patients, and then evaluated its performance on US patients. The system still outperformed human radiologists, reducing false negatives and positives by 8.1% and 3.5%."

Significance

In the United Kingdom, there is a shortage of skilled radiologists and this AI system can assist doctors and specialists when reading mammograms. This move will subsequently improve the overall accuracy of mammography reading and reduce the radiologists' workload.

The real value-added potential from using AI for mammography is "immediate feedback in the screening setting", claims study author. Providing the result immediately will enhance the standard of care. Also with AI, all mammograms could become "diagnostic" while releasing the results of mammograms on patients' schedules.

As the AI system is efficient enough in the reading film as an accompaniment to radiologists, it could give more time to experts to actually discuss to patients in person about their results. AI systems can be used to streamline the process of a breast cancer diagnosis. It can also help reduce the wait time for a biopsy.

Similarity with NYU's Research

Moreover, last October, NYU researchers published a similar study which demonstrates that the AI system can screen breast cancer similar to the skills of human radiologists. However, the difference between both the studies is that NYU only used mammograms from US patients, and it compared the system's performance with human expert diagnoses conducted in an artificial lab environment whereas Google and DeepMind compared performance with real-world diagnoses, notes MIT Technology Review.

Eventually, both the studies present the same conclusion that AI breast cancer screenings should be used in tandem with human radiologists and the combination of humans and AI can achieve the most accurate diagnostic results and reduce the workload on human radiologists. This could help free up their time to focus more on patient care.

Expert's Opinion

Dr. Mozziyar Etemadi, a research assistant professor of anesthesiology and biomedical engineering at Northwestern University and one of the paper's co-authors says, "health care is being squeezed with the number of patients increased and the amount of time that doctors have to see patients decreasing. So tools like these are what every physician is hoping for. We just have to better understand when tools like AI help and when it doesn't and ultimately come up with the combination of technology and human contributions what will ultimately improve care and make it more efficient."