Artificial intelligence diagnoses breast cancer 30 times faster than a doctor

Researchers have developed a machine learning software that can correctly diagnose a patient's risk of breast cancer 30 times faster than doctors. Diagnosis is based on results of mammograms and personal medical records.

The machine learning system (data analysis method that will automate the construction of analytical models) helps doctors make better diagnoses at the first time, avoiding errors and false positive cases with cancer.

Stephen Wong, a member of the research team from Methodist Research Institute in Houston, USA, said: "This software accurately evaluates millions of records in a short period of time. This allows us to identify "The risk of breast cancer is more effective than using a patient's mammogram results. This breakthrough finding offers the potential to reduce other unnecessary biopsy tests."

Mammography is a method of using low doses of X-rays, with high contrast combined with a computer to obtain detailed images of the breast. Mammograms play an important role in early detection of breast cancer cells.

Patients with breast cancer may hope to reduce unnecessary biopsies through artificial intelligence.(Photo: GagliardiImages / Shutterstock.com).

In many countries, women over the age of 50 are advised to screen for whole body cancer every 2 years. This system is good, but up to 50% of tests in the US give false positive results. This means that one of the two women is misdiagnosed when they conclude they are likely to have cancer.

According to the researchers, about 20% of biopsies in the US are done unnecessarily every year. And artificial intelligence is designed to significantly reduce those biopsies by making more accurate diagnoses from the first time.

Computer software has been playing an active role in helping the team of doctors analyze mammogram images to identify cancer cells. However, the system can take a step further, considering doctors' reports on scanned images as well as the patient's entire medical record to better determine breast cancer risk.

In the most recent proof, the researchers tested artificial intelligence on the results of cancer patients' 500 mammograms and pathological reports. Within a few hours, the software has made a diagnosis, especially, it recognizes every type of breast cancer that every patient has.

Researchers then had to re-examine the diagnosis that artificial intelligence made in clinical results. And surprisingly, the diagnostic software is 99% accurate. In fact, the same analysis is taking away 500 hours of hard working doctors.

Artificial intelligence can help doctors make more accurate diagnoses for patients.

"It's hard to believe, but it is possible that accurate evaluations of cancer charts may not be possible without artificial intelligence," Wong said .

Clearly, until now, artificial intelligence has not been tested in a real context - the results of this study come from patients with breast cancer. However, the researchers claim they have provided sufficient evidence to ensure follow-up tests.

Artificial intelligence cannot completely prevent a false positive diagnosis or suspicious mammogram results because sometimes there is not enough information to properly diagnose the system. But it can certainly help doctors make more accurate diagnoses for patients.

"We are looking forward to seeing how artificial intelligence can diagnose breast cancer as well as other types of cancer faster and more accurately. Most patients suspected of having cancer are always tense. straightforward waiting and having to undergo unnecessary biopsies, therefore, artificial intelligence is expected to resolve this situation soon, " Stephen Wong said.