When it comes to the future of delivering cancer care, augmented intelligence (AI) is on the horizon for improved diagnosis and treatment of cancer.
When the words “artificial intelligence” or “augmented intelligence” come to mind, you might think of large metal machines or robots. However, while artificial intelligence is designed to create intelligent machines that work better than humans, augmented intelligence is meant to enhance human intelligence rather than replace it.
In oncology’s case, AI-assisted technology can detect high-risk lesions and tumors earlier than other traditional approaches, especially in breast and lung imaging. “AI is also being used in pathology to augment detection of abnormal cells and to classify tumors better. AI is being used to assist with medication adherence and with ongoing support such as lifestyle modification and mental health support,” explains Dr. Ajeet Gajra, M.D., vice president and chief medical officer of Cardinal Health Specialty Solutions.
However, using AI for oncology decision support is also something that is becoming more common.
“Decision-making in cancer care can be complex and relies on a variety of data points such as the patient’s specific cancer that includes stage, mutations and sites of spread. Since the clinical status of patients fighting can change rather quickly, it would be especially helpful to predict and detect those who are at the highest risk for deterioration so we can prevent potentially bad outcomes and modify trajectories, like we see in our research for support in palliative care,” Gajra continues.
Many other factors also contribute to each individual patient’s cancer care, such as underlying health conditions; performance status; age and physiologic function; environmental factors such as housing distance, social support and health care literacy; and cultural factors.
AI in cancer care can be especially helpful to predict and detect patients whose status may cause them to be at the highest risk for deterioration. The more patterns the AI technology can recognize, the more powerful it is in helping both warriors and doctors.
When it comes to end-of-life (EOL) care and palliative care for patients, AI technology can also help.
EOL decision-making is difficult for patients and medical teams because no one wants to give up hope, so discussions about this type of care often happen too late. There is also research demonstrating that physicians often tend to overestimate patients’ life expectancy. A major risk factor that AI technology can help with is identifying the appropriate patients for palliative care and instituting it.
Gajra explains that, “Using AI, we may be able to identify covert factors that are responsible for patient deterioration e.g., an occult infection – some of these can be effectively managed. However, if the underlying clinical condition cannot be remedied ( e.g., advanced cancer, heart failure), then it is critical that the patient be offered and guided away from the acute care path and on to a palliative care path.”
“AI can help to augment human intelligence in these situations, evaluating health and socioeconomic high risks for patients, while also serving as a reminder or nudge to providers to evaluate and judge if the patients are good candidates for palliative care. Ultimately, this can improve quality of life for patients and caregivers,” says Gajra.
Yolaine Jeune-Smith, Ph.D., the director of scientific writing and strategic research at Cardinal Health Specialty Solutions, explains how AI will also help doctors and hospitals with patient data to ensure better assessment of care.
“In our current health care system, we have the ability to collect much more real-world data, even having patients submit this data directly to providers. However, while the data can be collected, it’s very hard for clinical teams to collect and process the data. AI can help identify meaningful and actionable data much faster,” Jeune-Smith says. Real-world data may include electronic health records; disease or medication/device registries; and insurance billing and claims.
Gajra notes that future uses of AI may also include identifying high-risk patients, assisting with pain management for cancer patients, matching the right dose or therapy with decision tree tools and clinical trial matching.
“The latter is particularly important in this era of precision oncology whereby using technology such as AI can help match patients to clinical trials of newer therapies—something that they may otherwise not have access to,” expresses Gajra.
Many patients may also struggle with depression as a result of all the changes and health issues they are going through. AI can assist in identifying these patients and help them get the care they need.
Jeune-Smith explains that a way for identifying such patients is through a questionnaire: “AI could be reading the submitted questionnaires, looking for patterns and responses. AI is monitoring changes in patient health in EMRs; there could be small changes that won’t be big enough for the human eye, but for machine learning, it may be showing trends.”
The innovative work Cardinal Health is undertaking with augmented intelligence bodes well for the future development of cancer care, especially when it comes to palliative care and management of mental health.