Background: Chimeric antigen receptor T-cell (CAR-T) and bispecific antibody (BsAb) therapies represent burgeoning treatment options in the hematologic malignancies and, more recently, solid tumor treatment landscapes. Unique toxicities associated with these therapies, namely cytokine release syndrome (CRS) and immune effector cell-associated neurotoxicity syndrome (ICANS), necessitate frequent in-person monitoring at specialized medical centers with expertise in management and limit widespread deployment and access to these potentially life-saving therapies.

Methods: We designed the CAR-SEET (Chimeric Antigen Receptor Side Effect Evaluation Tool) Pilot Study to evaluate the feasibility and clinical utility of remote monitoring using a commercially available wearable device (Apple Watch) for the prediction and diagnosis of CRS. Primary aims included: 1) Feasibility- defined by the ability to obtain data on >80% of patients wearing the device for >12 hours per day every day during the study period, and 2) Sensitivity of vital sign derangements detected by Apple Watch in supporting the diagnosis of CRS using one or more of the following criteria: fever (temperature of ≥38.0°C), heart rate of >100 beats per minute (bpm) sustained for >1 hour, and saturation of peripheral oxygen (SpO2) <90%. Other exploratory biometrics included the evaluation of the predictive value of decline in heart rate variability (HRV) as detected by Apple watch via Standard Deviation of Normal-to-Normal (NN) Intervals (SDNN) in the diagnosis of CRS. CRS was defined clinically based on daily in-person evaluations using standard diagnostic criteria. All adult patients treated with CAR-T or BsAb therapies at Sidney Kimmel Comprehensive Cancer Center at Johns Hopkins after November 2024 were eligible. Consented participants were provided with the wearable device with instructions to wear for 7 days prior to, and 30 days following, day 0 of CAR T-cell infusion or cycle 1 day 1 of BsAb therapy. At the end of the 30-day period, devices were retrieved and data was exported for analysis.

Results: Fifteen patients were enrolled between November 2024 and May 2025. Patients ranged in age from 24-84 years (median, 63), and 40% were female. Diagnoses included acute leukemia (2), lymphoma (3), multiple myeloma (8), and small cell lung cancer (2). Therapy consisted of CAR-T in 11 patients and BsAb therapy in 4. A total of 13 patients have completed therapy with sufficient follow up for study evaluation. Devices were successfully retrieved for data analysis in 9 patients (69%). Among these 9 patients who completed treatment with available data from the wearable device, the average daily use of the wearable device was 16.21 hours every day over 30 days. In these patients, sustained HRV depression (calculated automatically by Apple Watch and defined as a decline in HRV > 1 standard deviation below the patient's pre-infusion HRV mean, sustained for > 2 consecutive values) occurred in 100% (4/4) of patients who developed CRS at an average of 13.75 hrs (range, -31 to +1) prior to clinical diagnosis of CRS by the gold standard of daily in-person assessment. Among the 5 evaluable patients who did not develop CRS, sustained HRV depression was seen in only 1; this patient had a diagnosis of persistent atrial fibrillation, a condition which is known to influence HRV. Additional vital sign analysis, including temperature, heart rate, and oxygen saturation will be presented at a later date.

Conclusions: This small pilot study of patients undergoing CAR-T or BsAb therapy supports that remote patient monitoring is feasible among those patients where the device could be retrieved. Study challenges included loss of data for patients where device retrieval was not possible and technical challenges related to data retrieval in a usable format for certain vital sign measurements. Future investigation will further assess the association between sustained HRV depression and CRS; evaluate options for triggering direct communication between wearable device alerts and providers; and explore gyrometric-based assessments for the evaluation of ICANS. We will also utilize a HIPAA-compliant backend server for data storage, obviating the reliance on device retrieval and minimizing data loss.

This content is only available as a PDF.
2025
Sign in via your Institution