How AI Can Expand Early Breast Cancer Detection And Reduce Maternal Mortality

While many people worry that Artificial Intelligence might take away crucial jobs, it’s worth considering how the technology could revamp healthcare—particularly around detection of breast cancer, which affects millions of women around the world.

I recently spoke with Dr. Kristina Lång, an Associate Professor in Diagnostic Radiology Diagnostics at Lund University, Sweden, who shared with me exciting developments in the field of breast cancer detection. Dr. Lång’s team has developed an algorithm based on AI that utilizes a low-cost, handheld ultrasound probe connected to a mobile phone to detect breast cancer tumors.

Ultrasound technology is already a standard breast diagnostic method in high-income countries, but it requires advanced equipment and well-trained personnel. The use of handheld ultrasound devices is up to 25 times cheaper, making it a viable option for smaller clinics, which brings healthcare closer to the patient. However, the cost of the device is just one factor; another issue is the lack of trained personnel who can operate the devices and interpret the results. For example, Tanzania has a population of about 58 million people, but only 60 radiologists in the entire country. This is approximately 50 times fewer radiologists per capita than in the United States.

The great potential of AI is that it can assist less trained health care personnel in guiding the examination to find the right location for imaging, and even analyze the images to get a preliminary diagnosis. The algorithm that Dr. Lång and her team has trained is able to differentiate between breast cancer and other lumps such as benign lesions or normal breast tissue, and thereby removing the need to involve an expert in the initial phase.

Breast cancer is the world’s most prevalent cancer, and according to the World Health Organization, there were 7.8 million women in 2020 who were diagnosed with breast cancer within the past 5 years. Breast cancer is however not just a women’s health issue, it is also a major contributor of more orphans worldwide. Since the 1980’s, there has been great progress in diagnosis and treatment for breast cancer, but there are huge disparities in survival rates between regions. In high-income countries, the survival rate is above 90%, while in other countries, such as India and South Africa, it drops significantly to 66% and 40%, respectively.

One of the main reasons for these differences in survival rate is the timing of diagnosis. Early detection of cancer leads to more efficient treatment which significantly raises the chances of a positive outcome. Many high-income countries have screening programs in place, and simple measures such as breast self-exams can also contribute to early detection. Therefore, raising awareness and educating women globally is crucial.

Once a woman detects the first signs of a tumor, she should receive a diagnosis and start treatment quickly. However, in many remote areas with poorly equipped hospitals, the initial diagnosis can often not be made in smaller clinics, and women may need to travel long distances to receive a diagnosis and get treated. This can be a significant obstacle for many women who may need to take time off work or leave their families for extended periods. As a result, this delay in diagnosis can significantly affect treatment outcomes. Therefore, being able to do a first assessment in a smaller clinic closer to the patient could have a huge impact on the survival rates.

However, the method developed by Dr. Lång still has some way to go before it hits the market. As with all AI tools for medicine, it needs to be thoroughly validated and approved as a medical device before it can be safely used in the clinics. This means that not only the algorithm, but also the final application together with the hand-held device, needs to be tested on a large number of women directly in the clinics to verify that the results are accurate and correctly interpreted by the practitioners.

Several AI systems for assisting doctors in analyzing mammograms for finding breast cancer are already on the market, but for low-cost, hand-held ultrasound devices, this would be a first. While higher-income countries may worry about AI taking away jobs, this is a great example on how it can play a crucial role in democratizing global healthcare and supporting the WHO’s goal of reducing the number of maternal deaths due to cancer, particularly breast cancer.

SOURCE: Forbes