AUTHORS
Ben Zion Katz, PhD, Chen Glait Santar, PhD2, Tanya Badelbayev3, Olga Pozdnyakova, MD, PhD4, Michael Feldman, MD PhD5, Claudio Mosse, MD PhD6, Adam Seegmiller, MD PhD7, Irit Avivi Mazza, MD8
AFFILIATIONS
1 Gray Faculty of Medical and Health Sciences, Tel Aviv University, Tel Aviv, Israel, Tel Aviv Sourasky University Medical Center, Tel Aviv, Israel
2 Tel Aviv Sourasky University Medical Center, Tel Aviv, Israel
3 Tel Aviv Sourasky University Medical Center, Tel Aviv, Israel
4 Hospital of the University of Pennsylvania, Philadelphia, United States
5 University of Indiana School of Medicine, Indiana, United States
6 Vanderbilt University Medical Center,, Nashville, United States
7 Vanderbilt University Medical Center,, Nashville, United States
8 Gray Faculty of Medical and Health Sciences, Tel Aviv University, Tel Aviv, Israel, Tel Aviv Sourasky University Medical Center, Tel Aviv, Israel
ABSTRACT
Background
Lymphocytes are reported by routine hematology analyzers as a single undifferentiated population in all complete blood count (CBC) testing. Quantification of T-, B-, and NK-cell subsets is currently performed by flow cytometry, which requires reagents, sample preparation, specialized instrumentation, and trained operators, limiting its use to specialized testing rather than routine surveillance. Automated reporting of T-, B-, and NK-cell counts within the routine complete blood count would make immune-cell profiling available at every blood draw, enabling longitudinal surveillance, which is increasingly relevant in the era of immune-directed therapy. Raman scattering is an intrinsic, label-free phenomenon that captures intracellular biochemical information, including signals from DNA, RNA, proteins, lipids, and metabolites. Raman scatter weak signal has historically prevented routine measurement in clinical practice. We report the first performance evaluation of the PhenoTA Analyzer, a system designed to operate in line with routine CBC testing. By enabling automated, high-signal to noise ratio acquisition of Raman fingerprints, the platform quantifies mature T-, B-, and NK-cell subsets directly from the CBC tube, paving the way for broader routine cellular subset screening and profiling.
Study Design and Methods
De-identified residual whole blood samples were analyzed in parallel, using the PhenoTA Analyzer and conventional flow cytometry under IRB approval. Samples represented a broad clinical range, including normal controls, hematologic malignancies, and other clinical conditions. Raman signals were acquired using a system optimized for Raman cellular measurements, together with a unique calcium fluoride flow chamber that serves as an optically transparent substrate for the cells and enables approximately 15-fold higher signal-to-noise ratio than a conventional Raman measurement configuration. More than 200 lymphocytes were automatically measured per sample and classified as mature T, B, or NK cells within the lymphocytes population. The results were accompanied by an intracellular scatter plot that visualizes differences in intracellular biochemical signatures.
Results
A total of 929 samples were included in the analysis (56.5% males, mean age of 61.9 ± 17.3 years (range, 0–99)). The cohort was comprised of 380 normal controls (40.9%), 450 samples from patients with hematologic malignancies (48.4%, at diagnosis, during therapy and in remission), and 99 samples from patients with other medical conditions (10.7%). Within the hematologic malignancies cohort, cases included lymphoma (n=165), plasma cell dyscrasias (n=79), leukemia (n=187), and others (n=19). The leukemia subset was further classified as CLL/SLL (n=140), AML (n=21), ALL (n=12), and other variants (n=14; e.g. HCL and CMML). Performance analysis demonstrated strong agreement between the PhenoTA Analyzer and reference flow cytometry. For T-cell percentages, the correlation was r = 0.988, with a regression slope of 1.006 (95% CI, 0.992–1.020) and an intercept of 0.48 (95% CI, −0.45 to 1.42). For B-cell percentages, the correlation was r = 0.994, with a regression slope of 1.006 (95% CI, 0.994–1.017) and an intercept of 0.05 (95% CI, −0.33 to 0.42). The mean absolute differences were 3.79, 2.49, and 2.79 percentage points for T-, B-, and NK-cell percentages, respectively. Patient age, sex, baseline diagnosis and therapy type did not affect performance.
Conclusions
Our study introduces intracellular Raman spectroscopy as a novel analytical modality for clinical hematology, expanding routine cell counting to intracellular biochemical phenotyping. It enables automated quantification of T-, B-, and NK-cell subsets directly from the CBC tube with performance comparable to reference flow cytometry. The ability to generate an automated 8-part differential from whole blood without antibodies, specialized reagents, or dedicated immunophenotyping workflows represents a fundamental advancement in hematology testing. Strong agreement with flow cytometry across nearly 1,000 clinically diverse samples demonstrates the feasibility of scalable immune profiling within the routine hematology laboratory. These findings establish intracellular biochemical phenotyping as a new dimension of routine hematology testing, with potential applications extending beyond immune profiling to disease screening, detection, classification, and therapeutic monitoring.
DISCLOSURES
Ben Zion Katz, PhD: PhenoTA - Current holder of stock options in a privately-held company; PhenoTA - Consultancy (Includes expert testimony)