Share
Tweet
Share
Share
Breast cancer is the second-most common type of cancer affecting women in the U.S., accounting for roughly 30% of all new female cancers every year, according to the American Cancer Society. As such, the need for new technologies to improve early detection, risk analysis and treatment development is significant.
Artificial intelligence is one of the newest tools being applied to the problem of breast cancer, attacking it from multiple angles to help achieve earlier diagnoses and improved treatments.
October is Breast Cancer Awareness Month, so this is an excellent time to take stock of where current technologies are advancing early detection and better outcomes for patients.
AI applied to mammography
Some of the big headlines we’ve seen recently are about applying AI to mammography, both for early detection and to potentially reduce unnecessary follow-up screening and testing. Radiologists average one cancer case for every 200 mammograms they evaluate, but that may be a bit high.
From a study conducted with Whiterabbit.ai, Washington University School of Medicine in St. Louis recently reported results indicating that AI use could reduce the number of false positives in mammography screening — without missing any true positives.
Researchers developed an algorithm capable of identifying normal mammograms with extremely high sensitivity. Then they ran a simulation using patient data to see what the results would have been if radiologists did not have to review all the very low-risk mammograms, which freed them up to focus on more questionable scans. The simulation found that fewer people would have received callbacks for additional testing, while the same number of true cancer cases would’ve been detected.
Notably, Whiterabbit.ai is also behind a different algorithm marketed as WRDensity, which was approved by the Food and Drug Administration in 2020. WRDensity is designed to help radiologists determine breast density from mammograms so that they can identify people who might be able to benefit from alternative or additional screening.
Digging into the data to predict risk
In addition to improving the reading of mammograms and reducing unnecessary testing, AI is being applied to risk prediction for breast cancer. Working with the Breast Cancer Research Foundation (BCRF), Drs. Constance Lehman and Regina Barzilay developed and tested a deep-learning model called MIRAI on a dataset of diverse patients.
The model analyzed multiple mammogram images over time while integrating risk factor data into its analysis. Lehman and Barzilay showed that MIRAI could generate individualized and cost-effective improvements in risk prediction for breast cancer versus traditional risk models.
Also working with the BCRF, Dr. Barzilay and Dr. Adam Yala are applying MIRAI to predicting which patients are at high risk for breast cancer and then following their progress through MRI screenings. It’s hoped that this study will validate the results from the first study using MIRAI, demonstrating that it can identify high-risk patients over the next five years.
Applying AI to developing treatments
Beyond applying AI to mammography, companies like ImmunePrecise Antibodies (IPA) are applying their AI model to harmonize diverse data sets to help with developing treatments for various types of cancer. In August, the company announced that it was able to engineer antibodies to elusive tumor proteins in silico using its proprietary LENSai model.
IPA reported that its novel antibodies displayed highly specific binding to a tumor microenvironment protein of previously unknown structure. The study was verified in a laboratory setting.
Even though there was no previously known structural information about the target cancer protein, LENSai was able to model its structure and then accurately engineer antibodies tailored to bind to it. This represents a major hurdle in drug discovery and presents hope for breast cancer patients.
IPA hopes to use its LENSai to make safer therapeutics that are instantly accessible and affordable. They also partner with biopharmaceutical companies to hasten their drug discovery efforts.
Spotlighting Breast Cancer Awareness Month
Amid the advent of AI technology, there is even greater hope for breast cancer patients. For example, being able to determine risk based on more than just age will become a critical piece of the puzzle.
Here is some of the good news we can celebrate during Breast Cancer Awareness Month — all thanks to technology and therapeutics, like those developed for metastatic breast cancer, now being treated as a chronic illness.
- Advancements in technology for breast cancer detection and diagnosis, like AI, 3D mammography, and liquid biopsy conducted via a simple blood draw, have led to a 43% drop in deaths over the last 30 years.
- There are currently over 4 million survivors of breast cancer living in the U.S., including those still being treated and those who have finished treatment.
- When caught early in its localized stage, the five-year survival rate is 99%.
Disclosure: Ari Zoldan is CEO of Quantum Media Group LLC, and IPA Therapeutics is a client of Quantum Media Group.
