PARIS • A computer program can identify breast cancer from routine scans with greater accuracy than human experts, researchers said in what they hope could prove a breakthrough in the fight against the global killer.
Breast cancer is one of the most common cancers in women and regular screening is vital in detecting the earliest signs of the disease in patients who do not show any obvious symptoms.
In Britain, women over the age of 50 are advised to get a mammogram every three years, the results of which are analysed by two independent experts.
However, interpreting the scans leaves room for error and a small percentage of all mammograms either return a false positive - misdiagnosing a healthy patient as having cancer, or false negative - missing the disease as it spreads.
Now, researchers at Google Health have trained an artificial intelligence (AI) model to detect cancer in breast scans from thousands of women in Britain and the United States.
The images had already been reviewed by doctors in real life, but unlike in a clinical setting, the machine had no patient history to inform its diagnoses.
The team found that their AI model could predict breast cancer from the scans with a similar accuracy level to the expert radiographers.
Further, the AI showed a reduction in the proportion of cases where cancer was incorrectly identified - 5.7 per cent in the US and 1.2 per cent in Britain, respectively.
It also reduced the percentage of missed diagnoses by 9.4 per cent among American patients and by 2.7 per cent in Britain.
"The earlier you identify a breast cancer, the better it is for the patient," Mr Dominic King, British lead at Google Health, said.
"We think about this technology in a way that supports and enables an expert, or a patient, ultimately, to get the best outcome from whatever diagnostics they have had."
In Britain, all mammograms are reviewed by two radiologists, a necessary but labour-intensive process.
The team at Google Health also conducted experiments comparing the computer's decision with that of the first human scan reader.
If the two diagnoses were the same, the case was marked as resolved.
Only with discordant outcomes was the machine then asked to compare with the second reader's decision.
The study by Mr King and his team, published in Nature, showed that using AI to verify the first human expert reviewer's diagnosis could save up to 88 per cent of the workload for the second clinician.
"Find me a country where you can find a nurse or a doctor who isn't busy," said Mr King.
"There's the opportunity for this technology to support the existing excellent service of the (human) reviewers."
The team noted that further research is needed, but they hope that the technology could one day act as a "second opinion" for cancer diagnoses.