Wearable Devices Detect Cytokine Release Syndrome 7 Hours Earlier in CAR-T Recipients, Pilot Study Shows
Key Insights
Wearable sensors detected cytokine release syndrome (search) a median of 7 hours before standard nursing assessments in multiple myeloma (search) patients receiving CAR T-cell therapy.
The best-performing model identified 90% of CRS episodes with a sensitivity of 0.72 and specificity of 0.80 in a 25-patient evaluable cohort.
Interferon gamma emerged as a consistent biomarker, with doubling values reliably predicting CRS onset and offering potential therapeutic targeting opportunities.
A pilot study from Mount Sinai (search) researchers has demonstrated that wearable biosensors can detect cytokine release syndrome (search) (CRS) — a common and potentially life-threatening complication of CAR T-cell therapy — a median of 7 hours before standard hospital nursing assessments. The findings, published by Rajeeve et al in JCI Insight, could reshape how patients are monitored following CAR-T infusion and accelerate the shift toward outpatient treatment models.
“This result was mind-blowing,” said Samir Parekh, MD, director of the Center of Excellence for Multiple Myeloma (search) at Icahn School of Medicine at Mount Sinai (search) and senior corresponding author of the study. “I expected the wearables would be able to detect CRS, but I did not anticipate they would give us such a lead time. This makes all the difference and really changes the game.”
The Clinical Challenge
CAR T-cell therapy, approved in the United States since 2017, has improved outcomes for many patients with heavily pretreated relapsed or refractory blood cancers. Yet only about 20% of eligible U.S. patients actually undergo the procedure. Administration remains concentrated at tertiary centers, constrained by geographical limitations, provider confidence, and significant logistical and financial burdens.
Most institutions administer CAR-T through an inpatient model requiring 10 to 14 days of hospitalization for close monitoring of acute adverse events. CRS — a systemic inflammatory response triggered when the highly activated immune system releases excessive cytokines into the bloodstream — represents one of the most serious risks. Fever is the earliest and most universal symptom, with other flu-like manifestations including vomiting, nausea, chills, and fatigue. If left untreated, cardiovascular, respiratory, or neurological complications can ensue, with ICU admission required for severe cases.
The extended hospitalization “poses a huge infrastructure challenge” for treatment centers facing increasing resource constraints, Parekh noted. It also pushes the total cost of CAR-T — with base drug acquisition typically between $375,000 and $475,000 — to as high as $1 million.
Study Design and Results
Parekh and colleagues conducted a prospective, single-center pilot study enrolling 30 patients (median age, 59.5 years; range, 40–88; 53.3% women; 66.6% white) with relapsed or refractory multiple myeloma (search) who underwent CAR-T at The Mount Sinai (search) Hospital. All participants had ECOG performance status of 1 and had received a median of 5.5 prior lines of treatment. Median hospital stay reached 13 days (interquartile range, 12–14).
Wearable devices continuously monitored underarm and skin temperatures, heart rate, breathing rate, oxygen levels, and motion. Investigators also collected blood samples before and after CAR-T infusion to evaluate peripheral blood cytokines.
Among 25 evaluable patients with complete data, the best-performing wearable model identified 18 of 20 CRS episodes (90%), demonstrating a sensitivity of 0.72 (mean, 0.75; 95% CI, 0.60–0.91) and specificity of 0.80 (mean, 0.76; 95% CI, 0.68–0.84). The model detected signs of CRS a median of 7 hours before standard nursing recognition.
Median adherence to wearable monitors was 67% overall and 71% (range, 55%–84%) during high-risk periods.
Biomarker Insights
Analyses of 92 analytes at multiple time points revealed that cytokine changes occurred in parallel with temperature elevations and CRS development. Interferon gamma (IFN-γ) emerged as a particularly consistent biomarker, with a doubling of the interferon gamma value reliably predicting CRS.
“Continuous monitoring gave us a clearer picture of how CRS develops in real time,” said co-corresponding author Adriana Rossi, MD, Associate Professor of Medicine at Icahn School of Medicine at Mount Sinai (search). “Earlier detection could allow clinicians to intervene sooner, potentially reducing complications and improving the patient experience.”
Beyond predictive utility, the interferon gamma finding carries therapeutic implications. Parekh noted the availability of interferon gamma blockers that could potentially interrupt the downstream cytokine cascade.
Toward Outpatient CAR-T
If validated in larger cohorts, the technology could enable patients to recover in hotels or apartments within a 30-minute to 1-hour radius of treatment centers, with hospital transfer triggered only when CRS is detected. This model could substantially reduce healthcare utilization while maintaining safety.
“The cost of holding a hospital bed each day, plus the nursing resources required for surveillance, is far greater than the cost of a simple wearable and monitoring system,” Parekh said. “As we develop more intelligent application systems to read the signals, we will be able to depend less on having a huge human component to monitor these patients.”
The researchers acknowledged study limitations, including low statistical power due to the small sample size and the single-center design. They are now working to validate the findings in larger cohorts and to evaluate generalizability across different CAR-T products and indications beyond multiple myeloma (search). Investigators have also applied to the NIH for funding to evaluate wearables in patients treated with T-cell engaging bispecific antibodies, which behave similarly to CAR-T and can also result in CRS.
“This work highlights the potential of combining wearable technology with biologic markers to improve cancer care,” said co-corresponding author Alessandro Laganà, PhD, Assistant Professor of Genetics and Genomic Sciences and Oncological Sciences at Icahn School of Medicine at Mount Sinai (search). “Our long-term goal is to develop smarter monitoring systems that help clinicians predict toxicity earlier, personalize care, and improve outcomes for patients receiving advanced cancer therapies.”