Abstract
Cytokine release syndrome (CRS), neurotoxicity (eg, ICANS), and prolonged neutropenia constitute major complications of CAR-T therapy. Lymphocytes with aberrant morphology, including pre-activated lymphocytes, are variably present in the CAR-T cell product due to patient-specific factors at the time of lymphocyte harvesting as well as cellular processes occurring during CAR-T preparation and storage.
The aim of this study was to evaluate the impact of aberrant lymphocytes in CAR-T products on inflammatory complications following CAR-T infusion.
We prepared slides from the washed CAR-T infusion bag for all consecutive patients who received CAR-T therapy at our center from March 2025 to July 2025.
Residual cells were collected by washing the CAR-T bag post-infusion with normal saline, centrifuging at 500 bpm, fixing onto a slide, and staining using May-Grunwald-Giemsa. Slides were scanned by the Siemens Scopio X100 digital morphology platform using the full-field setting to obtain an automated differential count of 200 intact nucleated white blood cells. The differential count included the following lymphocyte subsets: normal lymphocytes, atypical lymphocytes, large granular lymphocytes, and aberrant lymphocytes.
Clinical data was obtained from electronic medical records and included patient diagnosis, CAR-T product, max grade CRS, max grade ICANS, doses of Tocilizumab, doses of Dexamethasone, use of growth factor support (G-CSF) or Anakinra, and white blood cell count on the day of infusion and on day +10.
Slides were successfully prepared and scanned for a total of 16 patients, while 2 patients were excluded as a result of insufficient sample preparation yielding a differential count of less than 200 cells.
Indications for CAR-T therapy included diffuse large B cell lymphoma (n=8), follicular lymphoma (n=3), mantle cell lymphoma (n=3), or rheumatologic disease (n=2). CAR-T products included axi-cel (Yescarta) (n=10), brexu-cel (Tecartus) (n=2), liso-cel (Breyanzi) (n=1), tisa-cel (Kymriah) (n=1), or CD19-Car_Lenti (produced by the Officina Farmaceutica of Bambino Gesù Children's Hospital) (n=2). One patient received Anakinra in prophylaxis.
Proportions of aberrant lymphocytes in the differential counts were compared between groups using the t-test following arcsine square root transformation of the percentage values.
The median count of aberrant lymphocytes in the overall cohort was 7% (range 0-17%) (10% for axi-cel, range 3-17%).
Patients who experienced CRS of grade ≥2 had increased aberrant lymphocytes compared with those with no CRS or CRS grade 1 (14% vs. 6%, p=0.04). CRS of grade ≥2 remained associated with increased aberrant lymphocytes in the subgroup of patients who received axi-cel (17% vs. 7%, p=0.02).
Increased aberrant lymphocytes were also associated with prolonged neutropenia requiring G-CSF support in the overall cohort (16% vs. 6%, p=0.004) and in the subgroup of patients treated with axi-cel (16% vs. 7%, p=0.006).
Associations were not found between number of aberrant lymphocytes and development of neurotoxicity (eg, ICANS) nor with doses of Tocilizumab, Dexamethasone, or Anakinra or with WBC.
While the number of patients is small, presence of increased aberrant lymphocytes, as pre-classified by the Scopio full-field digital morphology platform, appears to be associated with increased inflammatory complications following CAR-T infusion, including CRS of grade ≥2 and prolonged neutropenia requiring G-CSF support.
Digital morphology combined with artificial intelligence may represent a novel platform for predicting clinical outcomes in CAR-T therapy.
