Sensitive Detection and Quantification of CAR-T Cells in Treated Patients

Flow cytometry bottoms out near 1% and PCR cannot resolve individual cells; this AACR 2019 poster from Crown Bioscience reports a rare-cell imaging workflow that counted CAR-T cells in a treated patient's blood at densities where flow cytometry could not separate signal from background.

Presented by Crown Bioscience.

Read or download to see:

• A spike-in model set the detection floor before any patient sample was run. Human MCF-7 cells were spiked into one million mouse white blood cells and recovered linearly from 5 to 245 cells at y = 0.9374x - 0.3412, R² = 0.9961; the sensitivity limb found 2 of 2 spiked cells, and the specificity denominator came from two blank samples that returned 0.
• Where CAR-T cells were abundant, imaging and flow cytometry agreed. In purified CD19 CAR-T cells mixed into CD3+ T cells, flow cytometry gated 57.98% CAR-positive while the RareCyte platform scored 56.20% across 10 imaged fields, counting an average 137.80 CAR-T cells against 245.20 CD3+ cells per field.
• In a treated patient's blood, the two methods parted company at low density. Whole blood from a DLBCL patient given CD19 CAR-T was serially diluted into healthy donor blood: the RareCyte platform counted 141 CAR-T cells undiluted and 0 in the healthy control, while flow cytometry read 10.26% undiluted but still returned 0.21% on that same control — a background floor its dilutions straddled rather than cleared, at 0.98% down to 0.18%.