Evaluating the CAR-hematotox score in patients with refractory-relapsed multiple myeloma treated with bispecific T-cell engager therapy

Abstract Background: The CAR-HEMATOTOX score (HTx) is a validated tool for predicting the risk of severe cytopenias following CAR-T cell therapy in hematologic malignancies. In patients (px) with refractory-relapsed Multiple Myeloma (RRMM), the HTx score has also shown to predict toxicity and response to standard-of-care ide-cel and cilta-cel. Whether this tool can also predict outcomes in px treated with bispecific T-cell engagers (BiTEs) remains less well defined. Here, we evaluated the utility of the HTx score in predicting clinical response, toxicity, and survival in px with RRMM treated with anti-BCMA or anti-GPRC5D BiTEs. Methods: This retrospective study included all px with RRMM receiving a BiTE at our center up to April 2024 in whom a HTx score could be obtained. The HTx score was calculated according to Rjeski et al. using five laboratory parameters: absolute neutrophil count (ANC), hemoglobin, platelet count, C-reactive protein (CRP) and ferritin. These values were obtained within 5 days before BiTE initiation. Demographic data, baseline disease characteristics, previous lines of treatment (LOT), cytogenetics, staging, best response according to IMWG, progression of disease (POD), and toxicities were collected. Statistical analyses and plots were conducted in R. Progression-Free Survival (PFS) was calculated from the first step-up dose to either POD or last follow-up. Results: A total of 41 px were started on a BiTE regimen in our center, of whom n=23 had complete data to calculate the HTx score and were analyzed. BiTE distribution was as follows: elranatamab (n=5, 22%), talquetamab (n=5, 22%), and teclistamab (n=13, 57%). Two HTx groups were identified: low Htx (score 0-1; n=14, 61%) and high HTx (score ≥2; n=9, 39%). No significant differences in demographic features, baseline disease characteristics, performance status, previous lines of treatment (LOT), cytogenetics or staging were seen between HTx groups. Overall, 9 px (39%) developed G3-4 cytopenia post-BiTE. As expected, the rate of severe cytopenia was higher in the high HTx group than in the low HTx group (67% vs. 21%, p=0.015). A total of 16 px (70%) developed CRS, n=7 (30%) had ICANS, n=6 (26%) had G3-4 infections and n=7 (30%) were admitted to the ICU, with no significant differences observed between high and low HTx groups. Remarkably, therapeutic response rates were higher in the low HTx group compared to high HTx px (≥VGPR: 64% vs. 11%, p=0.02). Mortality rates were higher in the high HTx group (deaths: 78% vs. 21%, p=0.01). No significant differences were found in progression rates or PFS between the HTx groups. When analyzing therapeutic responses post-BiTE, we observed that px achieving ≥VGPR had a lower baseline HTx score (median score 0, [IQR 0-1]) compared to px with partial response or less (median score 2, IQR 1-3.5; p=0.01). Baseline hemoglobin was the HTx component most strongly associated with improved response: median hemoglobin 11 g/dL [IQR 10.7–11.8] in ≥VGPR vs. 9.3 g/dL [IQR 8.8–11] in PR/less; p=0.047. Notably, although not part of the HTx criteria, the baseline absolute lymphocyte count (ALC) was higher in ≥VGPR px (median 1.16×10⁹/L, [IQR 0.8–1.77]) than in px with PR/less (median 0.42×10⁹/L, [IQR 0.34–1.17]; p=0.044). Baseline ALC was also significantly lower in the high HTx group compared to the low HTx group (median 0.29×10⁹/L vs. 1.02×10⁹/L; p = 0.02). Further regression analysis revealed that a low HTx score was associated with higher odds of achieving ≥VGPR (OR=14.4, 95% CI 1.8–310; p=0.026). Conclusion: The HTx score can help predict therapeutic responses to BiTEs in px with RRMM, with hemoglobin being a strong predictor. High HTx scores also associate with higher mortality rates. Although not included in the HTx model, our findings suggest that lymphopenia may represent an additional hallmark of a high-risk phenotype during BiTE therapy in RRMM.