Clear Search

The Future of Breast Cancer Treatment is Personalized

By BCRF | April 22, 2024

How AI and predictive analytics will shepherd in a new era of highly precise and nimble diagnosis and treatment

Without a doubt, artificial Intelligence (AI) has already made positive contributions to cancer care. So far, our “StAI Tuned” series has explored the basics of AI’s potential in healthcare and how it’s being used in screening and diagnostic procedures specifically for breast cancer. But so many more applications are on the horizon.

In this third installment of our series, we delve into how AI can be leveraged for predictive analytics and to develop personalized treatment plans—revolutionizing breast cancer care.

What is predictive analytics and what does AI have to do with it?

Put simply, predictive analytics refers to the processing of large amounts of data to determine the likelihood of future outcomes. It falls under the AI umbrella and leverages several complex techniques:  

  • Data mining that pulls together disparate data to identify patterns and trends
  • Statistics or mathematical assessments of probabilities
  • Data modeling that conceptualizes relationships between different datasets
  • Machine learning, which trains machines to perform certain tasks and develop a generalized learning pattern that can look at new situations and problem solve

Predictive analytics uses these techniques to parse data and predict possible outcomes, so its potential in healthcare is substantial. Imagine how these techniques could analyze patient data and formulate a precise diagnosis or assess a patient’s risk of developing certain conditions including breast cancer. The power to process multiple variables and patterns quickly and repeatedly means that AI-driven analytics may potentially predict disease progression, identify specific patients who are more likely to respond positively or negatively to treatment, or perhaps even predict outcomes.

The result? Patients receive the right interventions and targeted screenings at the right time—leading to more accurate diagnoses and improved prognoses.

Real-world applications of predictive analytics

Breast cancer is a complex disease, and multiple factors, both known and unknown, influence how it starts, progresses, responds to treatment, and potentially develops resistance. We’ve made incredible progress investigating the intricacies of the disease. But, thanks to its ability to process huge amounts of data, AI is going to help researchers level up.

BCRF is currently supporting several projects that are leveraging AI-driven predictive analytics to answer some pressing questions.

In part 2 of our series, we described how AI is already being used to enhance screening and digital pathology methods as well as coupling screening and risk prediction. But can AI play other roles in breast cancer diagnosis?

Drs. Sohail Tavazoie and Elizabeth Comen are conducting studies to answer this question. Recently technology has revealed that circulating RNAs in the blood can be important biomarkers of breast cancer progression. Based on these findings, the team has used machine learning to analyze samples and pinpoint small RNAs that can predict which patients have benign versus malignant disease, as well as those who are at risk for (or already have) metastatic disease. Since more than one million women in the U.S. undergo breast biopsies to confirm or rule out breast cancer every year, their results could significantly impact clinicians’ ability to diagnose the disease with a minimally invasive test.

Dr. Britta Weigelt and her colleagues are applying AI technology to develop a robust and reproducible approach to diagnosing metaplastic breast cancers, which are rare but aggressive. Because they’re notorious for their complex genetic makeup and resistance to current treatments, it’s particularly important to diagnose these cancers early. Her team has already described the complex genetic architecture specific for this cancer cell type. Now they hope to develop an AI algorithm that can recognize these characteristics, thereby enhancing diagnostic precision for patients with metaplastic breast cancer.

BCRF investigators Dr. Constance Lehman and Regina Barzilay have also harnessed predictive analytics for early detection. Read more about their work in part two of our series here.

Why do some women develop breast cancer at a very young age?

Dr. Gad Rennert is compiling data from younger patients with breast cancer including DNA mutations in cancer-related genes, genetic variants in tumor tissue, participants’ immunological status, demographics, and behavioral risk factors. AI will then analyze this vast data to develop a predictive analytic profile that can help clinicians determine the likelihood a young woman will develop breast cancer. With this knowledge, doctors could potentially intervene earlier with preventive measures or tailor treatment plans based on individual risk profiles.

Can AI predict how a patient will respond to chemotherapy?

Drs. Nadine Tung and Stuart Schnitt assembled a large number of breast cancer samples, conducted extensive molecular analyses, and analyzed clinical outcome data. Through AI’s ability to process the extensive resulting data, the investigators identified distinct tumor characteristics that correlated with tumor sensitivity to chemotherapy. They plan to refine this prediction tool in ongoing studies.

Can AI predict survival outcomes?

Researchers are examining this question through different strategies. A group led by BCRF investigator Dr. Charles Perou is developing computational models using multiple data types and machine learning methods to identify new biomarkers that can predict survival outcomes for patients with triple-negative breast cancer.

Others are examining the gene expression patterns of epigenetic factors in tumors— factors that influence how genes are turned on or off. In fact, a team at the UCLA Health Jonsson Comprehensive Cancer Center has developed an AI model (not BCRF-supported) that can categorize epigenetic factors into distinct groups. Further, they correlated each group with outcome data across multiple cancer types including breast cancer. The team demonstrated that stratification of epigenetic factors can predict patient outcomes successfully and more effectively than traditional measures like cancer grade and stage. Without AI’s ability to simultaneously process data, it would have been a daunting task to decipher epigenetic changes that impact the more than 20,000 genes in the human body.

How AI can advance precision medicine

Researchers are constantly seeking ways to develop personalized treatments and identify those most likely to benefit from targeted therapies. Matching the right patient to the right treatment at the right time improves outcomes. Apart from predictive analytics detailed above, in the future, AI algorithms that help doctors in the clinic stand to substantially improve patient care and treatment outcomes—and ultimately advance precision medicine by leaps and bounds.

AI’s analytic prowess provides a way for researchers to assemble a wide range of patient-specific data and analyze relationships between specific factors (such as genetics, biomarkers, comorbidities, and molecular signatures) associated with breast cancer. These can be correlated with clinical information and outcomes data to identify those most likely to benefit from targeted therapies—thereby informing breast cancer care. Further, AI-driven analysis could lead to personalized treatment plans that are tailored to each patient, offering potentially more effective treatments while incurring fewer side effects.  

AI can also provide valuable insights as doctors seek ways to monitor how a patient’s treatment is going in real time—allowing them to adjust strategies more nimbly and efficiently when necessary. AI-powered wearable devices or algorithms that can analyze electronic health records and patient-reported outcomes could prove valuable tools to assess whether a treatment plan is effective and support timely adjustments or interventions.

AI-powered systems can help healthcare providers make evidence-based decisions by analyzing vast amounts of clinical data, research literature, and treatment guidelines. These systems can be programmed to make recommendations or appropriate interventions and alert healthcare professionals to potential risks or adverse events.

AI algorithms continuously learn and adapt. In breast cancer care, that means AI-driven treatment planning can improve over time, incorporating new evidence, clinical guidelines, and best practices. This iterative process means treatment approaches are constantly refined and optimized—ultimately leading to better patient outcomes and improved healthcare practices.

Looking forward

We are in an age of rapid progress in breast cancer that will only be accelerated by AI and the imaginations of researchers who leverage it. AI’s real-world, tangible future depends on the partnership between technology and people—a partnership that can deliver enhanced diagnostic strategies, tailored treatments, and improved treatment response prediction.

Even better: It can empower collaboration between doctors and patients who can more easily participate in their own breast cancer care decisions. The possibilities are truly endless.