Remote monitoring for cancer therapies: Faster early detection of complications using AI-supported wearables

The treatment of cancer with chemotherapy and modern immunotherapies harbours a risk of life-threatening complications that are often only recognised late, particularly in the outpatient sector, due to inadequate monitoring. In a study, a research team led by Dr Malte Jacobsen (doctor at the Clinic for Cardiology, Angiology and Internal Intensive Care Medicine at RWTH Aachen University Hospital and PhD student at Witten/Herdecke University) and Prof. Dr Guido Kobbe (Head of the Cell Therapy Department at RWTH Aachen University) were able to identify the risk of life-threatening complications. Guido Kobbe (Head of Cell Therapy at the Department of Haematology, Oncology and Clinical Immunology at Düsseldorf University Hospital) have successfully demonstrated that remote monitoring of vital parameters in cancer patients using artificial intelligence and smart technology can detect and predict potential complications at an early stage. The key to success is wearables - devices that are worn by the patient and allow real-time monitoring. The results of the research work have now been published in the journal npj digital medicine .

Hospitalised for cancer treatment? That is not absolutely necessary. Modern cancer therapies such as CAR-T cell therapy may be able to be carried out on an outpatient basis in the future. Until now, inpatient monitoring has been necessary due to frequent and sometimes severe side effects. After discharge, cancer patients visit an oncology day clinic or a specialised outpatient clinic on a weekly basis. Between treatment appointments, they can go about their daily lives at home as usual. Dr Jacobsen, first author of the study, explains the problem:

"In the outpatient therapy setting in particular, monitoring of patients is patchy and based primarily on their self-assessment. Neither the course of treatment nor the patient's condition can be monitored by hospital staff during this period. Treatments for haematological malignancies using chemotherapy, radiotherapy or modern immunotherapies always carry a high risk of serious complications such as infections, cardiac events and immunological dysregulation, which are therefore only detected at a late stage. This is associated with complex medical measures and a more severe course of the disease." The conventional solutions to this problem to date have been based on frequent, routine presentations of patients in a specialised outpatient clinic. However, this approach means a considerable amount of work for patients and healthcare professionals. The need for innovative concepts is correspondingly great.

AI-controlled wearables

To what extent can artificial intelligence and smart sensor technology contribute to the early detection and therapeutic treatment of life-threatening complications in patients with haematological malignancies, i.e. cancers that affect the haematopoietic system? Scientists from RWTH Aachen University Hospital and Düsseldorf University Hospital, with the participation of the Chair of Cardiology in Medicine at UW/H led by Prof. Dr Melchior Seyfarth (also Head Physician of the Heart Centre at Helios University Hospital Wuppertal), have investigated this central question in their research work "Wearable based monitoring and self-supervised contrastive learning detect clinical complications during treatment of hematologic malignancies".

The use of artificial intelligence (AI) and so-called wearables, which record medical data in real time through non-invasive, continuous monitoring, makes this process possible. Medical wearables are electronic devices worn directly on the body that are able to collect biophysical data. With the help of deep learning - a form of machine learning based on artificial intelligence - artificial neural networks are trained to identify patterns and recognise anomalies that indicate complications.

Recording vital signs and real-time monitoring

A total of 79 adult cancer patients, 54 from the inpatient and 25 from the outpatient sector, were included in the study at Düsseldorf University Hospital and equipped with wearables during their intensive treatment for haematological malignancies. "With the help of AI-controlled sensor technology, vital parameters such as heart rate, temperature and respiratory rate as well as physical activity were continuously monitored, recorded and analysed," explains Prof. Kobbe from Düsseldorf University Hospital and last author of the study. "With our work, we were able to show that complications that occur during intensive cancer therapies lead to characteristic changes in the wearable data at an early stage. We were able to reliably detect these changes using an AI-based analysis and predict them 48 hours before clinical diagnosis," says the physician. The AI algorithms were largely developed in collaboration with Prof. Dr Markus Kollmann at the Institute for Mathematical Modelling of Biological Systems at Heinrich Heine University Düsseldorf.

Improving quality of life and making cancer treatment safer

"Remote patient monitoring with medical wearables represents a novel option for non-invasive remote monitoring of vital data and physical activity. Wearables provide high-resolution health data that expand monitoring options and enable the detection of complications in real time using classification models," summarises Dr Jacobsen.

In the future, automated early detection and permanent remote monitoring of patients in oncology using a digital monitoring system of this kind could lead to preventative complication management. "Ideally, the recorded data would be analysed in real time to provide usable information for early and effective treatment. This could improve clinical pathways, for example by introducing on-demand clinic visits, which could reduce the workload of doctors and nurses. In addition, minimising the frequency of blood sampling during the treatment of patients with haematological malignancies is conceivable, as recent research indicates a good correlation between the vital signs recorded by wearable devices and the results of laboratory measurements," he says optimistically.

Prospective confirmatory studies are needed to prove the benefits of this approach in clinical practice, which can not only have a positive impact on the quality of life of those affected, but also make outpatient cancer treatment safer.

You can find the original publication here: https://www.nature.com/articles/s41746-023-00847-2

Note: The original press release comes from the University Hospital of RWTH Aachen. The UW/H has adapted it.