Can AI Help Monitor Chemotherapy Side Effects?

By Marie DeFreitas

With the potential to reduce workloads, streamline services, and aid in patient care quality, AI is rapidly becoming a popular tool in healthcare.

The technology has now entered the cancer care unit. The Mercy healthcare system is using AI to help cancer patients – by texting them.

Often, chemotherapy patients find themselves struggling with side effects such as general pain, fever, diarrhea, fatigue, and vomiting; these are red flags that frequently lead to hospitalizations. According to a study by the National Institutes of Health, “of 18,486 patients who received chemotherapy for metastatic cancer, 92% were hospitalized at least once for any reason, including 51% hospitalized for a likely toxicity.”

Care units need a way to track these symptoms before they lead to patient hospitalization. Enter the Chen Chemotherapy model.

Named after lead data scientist Jiajing Chen, who lost their own battle with cancer in 2023, the model notifies doctors before these symptoms become severe, keeping patients out of the hospital.

The program works by creating a risk score for non-leukemia chemotherapy patients over 18 years old. As the program learns, it’s able to predict how likely it is that a patient who is experiencing symptoms will be hospitalized within 30 days after their chemotherapy treatment.

Prior to this model, providers were oblivious as to which patients were experiencing problems until patients called or showed up at the emergency room.

This tool allows providers to be more involved in the process of chemotherapy recovery.

“The Chen Chemotherapy Model and smart texting allows us to proactively manage these patients and identify when they are having problems earlier in the journey,” Jay Carlson, DO, medical director of Mercy oncology service line, said in a press release. “This means they may be able to be treated in the office, recover faster and feel better overall.”

The success of AI in cancer care has led to the development of several other tools by different creators. Last year a neurobiology and human genetics professor at the University of Utah, along with a 20 person team, created an AI algorithm to help identify more than 200 ‘micro-symptoms’ for cancer patients, such as behavior, speech, and vocal patterns. These range from abnormal neurological phenotypes and eye movements to sadness in the vocal tone. A clinical trial of the tool is scheduled to start in January 2024 at the Moffitt Cancer Center. Identifying these small changes can help assess how patients are handling the treatment and can even predict changes in future symptoms.

Cancer is the second leading cause of death in the U.S., compelling the need for new innovations and screening technology. In recent years the U.S. Food and Drug Administration has approved more than 500 AI and machine learning-enabled medical devices, ranging from imaging software to remote cardiac monitoring devices.

Marie DeFreitas is an associate content specialist at HealthLeaders.