Breast cancer affects millions of women around the world every year. In 2018, more than 2 million new cases of breast cancer were reported. This makes early breast cancer detection critical. Benefits of AI mammogram are currently being investigated by the industry to enable faster diagnosis.
Despite the increasing awareness and advancement in diagnostic techniques, early detection of breast cancer continues to be a challenge for the medical fraternity. The problem aggravated with the pandemic in 2020. Fewer women got a routine scan due to worries of the virus.
At such a time, AI mammogram can prove helpful in the detection of breast cancer and other cases.
Google’s DeepMind AI mammogram debate
A study done in January 2020 by Nature suggests that Google’s DeepMind AI proved to be more accurate in identifying cases of mammogram than doctors. Google had also conducted a study.
They engaged a panel of 6 radiologists who regularly conducted mammograms. The report suggested that the DeepMind test had fewer false positives and false negatives.
However, 19 researchers and the Massive Analysis Quality Control Society board of directors criticised this study in May 2020. They said Google was not able to explain how the AI was developed. Google also did not release DeepMind’s algorithm code.
The researchers also added it undermines the scientific value of the study that was conducted by Google.
Challenges for early detection in the current systems
Mammography employs x-ray imaging to spot and diagnose breast cancer. However, reading these x-ray images can be a difficult task even for experienced professionals.
False-positive and false-negative mammogram reports continue to be a challenge. In the false-positive, the mammogram looks abnormal although there might not be a cancer in the breast.
Similarly, a false-negative mammogram would show everything normal even though breast cancer is present. One of 5 breast cancers go undetected in this manner.
Impact on the patient
In this debate, we cannot leave out the impact of a false-positive or false-negative on the patient. Such reports can lead to unnecessary stress due to incorrect diagnosis. There is also the increased risk to life in case of delayed or. inaccurate diagnosis.
AI to help radiologists with the screenings
THe medical and technology fraternities are hopeful. They are working to create more reliable mamogram screenings with the help of AI. While Google brought some hope with its AI tool, there is nothing reliable as of now. The company is developing an AI based model to spot breast cancer during the mamogram screenings.
Google’s AI Mammogram
Although the AI model is still in its early phases of development, it seems to have a huge potential in our battle against Breast Cancer. Google’s initial research findings published in Nature revealed that its AI model can spot breast cancer in screening mammograms with greater accuracy than the experts.
Gogole’s AI model used a data set of over 90,000 de-identified mammograms. They belonged to women from the United States and the United Kingdom. The large dataset would help them to train it to spot signs of breast cancer in the scans.
The model was later evaluated on the basis of its results for another dataset of de-identified mammograms. This time, they used mammograms of 3,000 women in the US and more than 25,000 women in the UK.
Researchers have claimed that the AI system produced a 1.2% and 2.7% reduction in the false positives and false negatives in the United Kingdom. Furthermore, it also produced a reduction of 5.7% and 9.4% in false positives and false negatives respectively in the United States.
Moreover, medical professionals usually have access to the medical history of the patient which further assist them in the screening process. However, the AI model was provided with only the most recent mammogram and no extra information for evaluation.
Notably, Cancer Research UK Imperial Centre, Northwestern University, Royal Surrey County Hospital, and Deep Mind have also partnered to help Google with clinical research. But, the AI model would require more research, clinical studies, and regulatory approval practical application.
Experts on AI in mammography
A recent study published in Radiology: Artificial Intelligence suggests reading of breast screenings can be enhanced with the help of AI. Radiologists can identify the readings more accurately with an AI-based mammogram. These AI algorithmns were shown positive and negative mammograms.
Developers work on the AI technology with existing images that show abnormalities associated with cancer. The AI tool is taught to identify the cancer readings from benign findings as well.
The study used AI mamogram tool MammoScreen by Therapixel. Their website says that the tool “provides AI-guided decision-making support for breast cancer screening”. Therapixel clinical research manager Serena Pacilè said, “The results show that MammoScreen may help to improve radiologists’ performance in breast cancer detection.”
What do you think of AI for mammogaraphy? Will it enhance cancer detection technology? Tell us in the comments below!
Follow more updates on health technology here.