Data Science and AI are revolutionizing healthcare and breast cancer is no exception. Irene Acerbi talks about how these new innovations enable more personal and precise breast care.
In healthcare the ability to recognize patterns in vast amounts of data can be the difference between life and death. Technological advances in data science and artificial intelligence promise to make medicine more personal and precise. However, change is slow in the highly regulated healthcare industry and many disruptive ideas are far from affecting actual patient care. The future of medicine is, well, still in the future.
The Athena Network combines healthcare and research
The University of California’s Athena Breast Health Network is an exception to this rule. Combining state-of-the-art breast health screening and care with scientific research, the network uses data science to improve patient care today instead of in the distant future. Athena is about “learning from practice and patient data to improve care,” says Irene Acerbi, Program Manager at Athena Breast Health Network.
Irene came to UCSF from Italy. She graduated high school the year Dolly the Sheep was born, and the historic event ignited her passion for bioengineering. She completed her Ph.D. in biomedical engineering back in Italy before moving to San Francisco on a grant from the UCSF Center for Bioengineering and Tissue Regeneration to study breast cancer.
She enjoyed working with an interdisciplinary team so much that she finally decided to ditch the lab coat and go into project management for the Athena Network.
Athena is a collaboration of all medical centers of the University of California. The idea is to provide personalized care for each patient based on the assessment of individual breast cancer risk factors. “We have integrated electronic medical records. When scheduling an appointment at our medical center, women receive an email to complete the intake form for their appointment online,” Irene explains.
Building a database for risk assessment
Instead of having to fill out paper forms at the doctor’s office, patients can fill out the forms at home. All information that Athena has on a patient is stored in these medical records. For example, the records contain “personal information and medical history at different points in time, such as annual screening mammograms.”
The network has created a large database from the patient data. “We started six or seven years ago and today the database contains over 150,000 patient records,” Irene says — and you can tell she’s proud of this number.
The concept of Athena is to learn from this data and to enable scientific research using it. “The enhancement of care in Athena’s mission means that we want to enable research to learn from our experience,” Irene says.
The Athena Network uses data science to analyze the patient data in their database and identify women who are at elevated risk for breast cancer. “If anybody’s at high-risk, we send a letter home to these women to inform them,” Irene explains. “This way, we empower patients [to make] their own decisions.”
Introducing the WISDOM Study
But Irene and the Athena Network don’t want to stop there. “Our results are good, but it’s still not enough to identify [everybody] who’s at elevated risk,” Irene admits.
They still have patients that receive annual screening but one day “come to the clinic with a sudden aggressive breast cancer,” she says. These women could receive better and earlier help if their risk would have been assessed more accurately and if they had received more frequent screening.
To address this issue and further improve risk assessment and preventive measures, the UCSF and other partnering branches of the University of California have developed the WISDOM Study. The study is open to women from age 40 to 74 who have never had breast cancer. Participants receive an at-home kit to donate a saliva sample. Just like with Athena, all data collection is done online or through the mail to simplify the process for the participants.
Irene explains what the sample is used for: “We take the genetic information from nine different genes which can contribute to breast cancer risk, as well as a number of single nucleotide polymorphisms (SNPs).” SNPs are small variations in the genetic code of a person that can influence the risk for developing certain diseases, including breast cancer. “[Checking for SNPs] is an innovative aspect [of WISDOM]. This is something commercial genetic tests for cancer risk don’t check today,” Irene says.
Catching the most aggressive cancers earlier could save lives
The genetic information is added to the Athena database to improve the risk assessment. This improved information is used for the second innovative aspect of WISDOM: depending on their risk level, women receive personalized screening schedules. The higher the risk, the more often the screenings, and the lower the risk, the less often the screenings.
This way, Irene hopes, aggressive and fast-growing cancers can be identified earlier. If successful, treatment could start at an earlier stage, significantly increasing the chance of survival and reducing the need for invasive treatments. It could also lead to reduced costs in healthcare, as expensive later-stage treatments would be less frequent.
Irene and her colleagues hope to recruit 100,000 women to participate in WISDOM. The study consists of the personalized screening, as well as a control group, where women receive the normal annual mammogram.
With WISDOM, women have a choice
While in most other studies participants get randomly selected to one group or the other, WISDOM offers women to choose which group they want to be part of. “Randomization is the best way to get scientific evidence,” Irene explains. “But, there are analyses we can do even if [a woman] chooses [which group she wants to be in].”
Offering this choice, the researchers hope to include more women in the study. Irene emphasizes that WISDOM is open to any eligible woman in California, regardless of their insurance or which clinic they go to.
Both the Athena Network and WISDOM make it a priority to create rich datasets from their patient data. They use this data themselves to conduct risk analyses, but another focus is to make the data available to research groups that use data science and artificial intelligence to improve cancer care.
AI researcher use Athena data to find new solutions
“We are providing the database to mine,” Irene says. “For example, a [research] group at UCSF that works on AI recently submitted a proposal to us because they want to analyze the mammography images [in the database].” If a proposal is granted, the AI researchers get access to the anonymized data and collaborate with the Athena and WISDOM staff.
The groups share the results of the research, so both sides can benefit from them. This way, the Athena Network can incorporate new findings into their risk assessment models and refine it. “We are the platform and AI kind of happens to us,” Irene says and laughs.
Screening will become more personal in the future
In the future, Irene believes biomarkers and DNA analyses will be key factors to improve cancer care further. She emphasizes that the most important goal has to be finding people that are at risk before they develop cancer and monitor them.
Biomarkers could help to build even more fine-grained risk assessment models. They could also help to develop new drugs to target tumors in a more focused way. “I see that at the intersection of technology and care, data mining and AI can help identify how these new biomarkers and drugs can be used,” Irene says.
The Athena Network and WISDOM, with their unique combination of applied care and scientific research, are a role model for advancing healthcare. While many other medical innovations that use data science are still years away from reaching patients, Athena is already here. Their approach may be less disruptive and revolutionary than other concepts, but the network is an impressive example of where the future of healthcare might lead us.