There are a multitude of existing applications for artificial intelligence (AI) in oncology, often for administrative tasks or screening support, but we are increasingly seeing its potential to help cancer providers make important decisions about patient care. As oncology stakeholders fight to address a variety of pressing challenges (such as rising costs, rapid clinical innovation, and unnecessary utilization), these AI innovations have the potential to help.

While these applications are still in the earlier days of experimentation, here are three ways that providers may someday be able to utilize AI to streamline clinical decision-making and push precision medicine in oncology forward.

Though oncology treatment innovation is advancing rapidly, the resulting positive impact on patient outcomes is delayed by the difficulty of getting these new treatments to the correct patients. Providers often have trouble keeping up with changing standards of care, making it challenging to select the best possible treatments for their patients.

Knowing which patients are likely to benefit from new treatments can promote the adoption of promising treatments into the clinical setting. That's where AI can help.

In January of this year, Massive Bio announced the launch of a new tool that uses AI to match patients to drugs or clinical trials based on real-time patient health information. The technology works by extracting patient data from the EHR, sorting through all active clinical trial protocols and approved drugs, and matching the patient to the optimal clinical trial or drug for their specific needs.

The tedious task of manually reviewing patient records and clinical trial protocols has historically been a barrier to clinical trial enrollment and appropriate treatment selection. The process is prone to error and too time consuming in a field where there are already staffing shortages; therefore, Massive Bio's innovation could help free up workforce time while also ensuring that each patient receives the most appropriate treatment.

This type of tool will likely become more common as the industry continues to learn how best to utilize the vast amounts of available clinical data.

A variety of more effective and targeted cancer treatments have been developed over the past decade, and some have become standard of care for certain cancer types. Despite their popularity, these treatments aren't always effective. This is especially problematic when the treatment in question happens to be extremely expensive.

Advanced non-small cell lung cancer (NSCLC) is a great example of this conundrum, because its first-line treatment often includes immunotherapy, which can be very effective — but only for a certain subset of patients. Because immunotherapy is incredibly expensive, identifying which patients might benefit from immunotherapy before they receive the treatment is a huge opportunity to avoid financial waste.

Stakeholders are seizing this opportunity to lower cancer spending and improve patient outcomes by developing AI tools that can accurately predict whether a standard cancer treatment is likely to be effective for a specific patient. 

For example, researchers at Emory University performed a retrospective AI-based analysis of NSCLC patients' imaging from before and after immunotherapy treatment. The study found a new imaging biomarker that can identify which patients are likely to benefit from immunotherapy. While this biomarker is newly discovered, it demonstrates how AI can help physicians further differentiate patients and personalize their treatment plans.

AI is being used to discover more biomarkers for other treatments and tumor sites, which will optimize treatments for patients while reducing spending on ineffective treatments.

AI will likely help clinicians predict not only which treatments will work, but also when to administer those treatments.

This guidance on treatment timing could be especially impactful when applied to rapidly metastasizing cancers like glioblastoma multiforme (GBM), which is one of the deadliest cancers. Because GBM is so aggressive, oncologists often start treatment immediately instead of waiting to understand how the disease might progress. This makes it difficult for clinicians to optimize treatment timing and for researchers to develop precision therapeutics.

Researchers at the University of Waterloo have recently attempted to address this issue by combining existing mathematical models with novel AI to create a tool that can predict the tumor progression in patients with GBM. Using AI to examine and compare disease progression in patients who chose not to undergo treatment, the researchers created an AI-informed predictive pathway for other patients with GBM, which can help inform clinicians about the ideal timing of treatment.

While this is still a very new application of AI, it highlights the other promising areas that someday may help clinicians treat patients even more effectively. 

Recent innovations have shown that AI will increasingly be incorporated into almost all aspects of cancer care. These examples are just the beginning.

As cancer care becomes increasingly personalized, cancer programs and other oncology stakeholders should consider how they can utilize existing AI to help them stay competitive.

There are always challenges and concerns with new innovations, but we believe AI will become a must-have for clinical decision-making. There is a long way to go before clinical decision-making technology can be easily integrated into provider workflows, so organizations should thoughtfully weigh the pros and cons of introducing AI-based innovations before jumping in. However, this is an area to keep an eye on in the coming years.