AUTHORS
Rafaella Litvin, MD, Kristina Nasr, MD2, Ran Zhao3, Agrima Mian, MD4, Mikhaila Rice, PharmD5, Joslyn Rudoni6, Allison Winter, MD7, Deepa Jagadeesh, MD, MPH8, Robert Dean, MD9, Taylor R. Brooks, MD10, Louis Samuel Williams, MD11, Shahzad Raza, MD12, Faiz Anwer, MD13, Jason Valent, MD14, Christy Samaras, DO15, Kelsey Pandrangi, MD16, John Carlos Molina, MD, MEd17, Claudio Brunstein, MD, PhD18, Sharan Jhaveri, MD19, Evandro Bezerra, MD20, Diana Basali, MD21, Sandra Mazzoni, DO22, Jack Khouri, MD23, Brian Hill, MD, PhD24, Paolo Fabrizio Caimi, MD25
AFFILIATIONS
1 Cleveland Clinic - Taussig Cancer Center, Cleveland, Ohio, United States
2 Cleveland Clinic - Taussig Cancer Center, Cleveland, Ohio, United States
3 Cleveland Clinic - Taussig Cancer Center, Cleveland, Ohio, United States
4 Memorial Sloan Kettering Cancer Center, New York, New York, United States
5 Cleveland Clinic - Taussig Cancer Center, Cleveland, Ohio, United States
6 Cleveland Clinic - Taussig Cancer Center, Cleveland, Ohio, United States
7 Cleveland Clinic - Taussig Cancer Center, Cleveland, Ohio, United States
8 Cleveland Clinic - Taussig Cancer Center, Cleveland, Ohio, United States
9 Cleveland Clinic - Taussig Cancer Center, Cleveland, Ohio, United States
10 Cleveland Clinic - Taussig Cancer Center, Cleveland, Ohio, United States
11 Cleveland Clinic - Taussig Cancer Center, Cleveland, Ohio, United States
12 Cleveland Clinic - Taussig Cancer Center, Cleveland, Ohio, United States
13 Cleveland Clinic - Taussig Cancer Center, Cleveland, Ohio, United States
14 Cleveland Clinic - Taussig Cancer Center, Cleveland, Ohio, United States
15 Cleveland Clinic - Taussig Cancer Center, Cleveland, Ohio, United States
16 Cleveland Clinic - Taussig Cancer Center, Cleveland, Ohio, United States
17 Cleveland Clinic - Taussig Cancer Center, Cleveland, Ohio, United States
18 Cleveland Clinic - Taussig Cancer Center, Cleveland, Ohio, United States
19 Cleveland Clinic - Taussig Cancer Center, Cleveland, Ohio, United States
20 Cleveland Clinic - Taussig Cancer Center, Cleveland, Ohio, United States
21 Cleveland Clinic - Taussig Cancer Center, Cleveland, Ohio, United States
22 Cleveland Clinic - Taussig Cancer Center, Cleveland, Ohio, United States
23 Cleveland Clinic - Taussig Cancer Center, Cleveland, Ohio, United States
24 Cleveland Clinic - Taussig Cancer Center, Cleveland, Ohio, United States
25 Cleveland Clinic - Taussig Cancer Center, Cleveland, Ohio, United States
ABSTRACT
Introduction
Hematologic toxicity is common with bispecific T-cell engaging antibodies (BsAbs) and can lead to treatment interruptions, discontinuation and potentially worse outcomes. Currently, there are no established strategies to identify patients at increased risk for hematologic toxicity before starting therapy. Using a cohort of 200 patients with lymphoma and multiple myeloma treated with single-agent BsAbs, we developed a clinical risk score to predict cytopenias and evaluated outcomes across different risk groups.
Methods
Data preprocessing and statistical analyses were performed using R (version 4.5.3; R Core Team). A total of 200 patients (144 with multiple myeloma and 56 with lymphoma) meeting inclusion criteria were included. Continuous variables were compared using the Wilcoxon rank-sum test, and categorical variables using Fisher's exact or Pearson's chi-square test. Overall survival (OS) and progression-free survival (PFS) were estimated using the Kaplan-Meier method and compared with the log-rank test. Competing-risk outcomes were compared using Gray's test. Univariate logistic regression was used to estimate odds ratios for associations between individual predictors and the outcome of interest.
For development of the clinical risk score, the dataset was randomly divided into training (75%) and testing (25%) cohorts. To address class imbalance, the Synthetic Minority Over-sampling Technique (SMOTE) was applied to the training set only. Candidate predictors were selected using stepwise multivariable logistic regression and LASSO regression. The final multivariable logistic regression model was used to develop the clinical score based on model coefficients, optimal predictor cutoffs determined by the Youden index, and predictor quantiles. Model performance was assessed in the test cohort using area under the receiver operating characteristic curve (AUC), sensitivity, specificity, precision, F1-score, and accuracy.
All statistical tests were two-sided, with P < 0.05 considered statistically significant. Cytopenia was defined as grade ≥3 anemia (hemoglobin <8 g/dL), grade ≥3 thrombocytopenia (platelet count <50 × 10?/L), and/or any-grade neutropenia (ANC <1.5 × 10?/L) occurring during BsAb therapy or within 90 days of treatment discontinuation, until disease progression, initiation of subsequent therapy, death, or last follow-up.
Results
In the final score, two points were assigned for baseline hemoglobin between 10.8 and 12.4 g/dL, four points for baseline hemoglobin >12.4 g/dL, two points for baseline ANC >3.4 × 10?/L, and one point for baseline platelet count >177 × 10?/L. For baseline values below those thresholds, zero points were given.
Patients with a sum score <3.5 were placed in the high-risk group (HRG), whereas those with a score ≥3.5 were placed in the low-risk group (LRG). Although only baseline blood counts were used to develop the model, additional baseline laboratory differences were observed between the groups. CRP was higher in the HRG (median 2.75 vs. 0.4 mg/dL, P < 0.001), as were ferritin (median 732 vs. 161.5 ng/mL, P < 0.001) and LDH (median 286 vs. 206 U/L, P = 0.007). There were no differences in baseline creatinine, albumin, or D-dimer between the HRG and LRG in the full cohort.
In the test cohort, the score predicted cytopenia with an AUC of 0.893, sensitivity of 0.87, specificity of 0.75, precision of 0.92, recall 0.87, and F1 0.89. In the full cohort, cytopenias occurred in 90% of patients in the HRG compared with 46% in the LRG (P < 0.001). Anemia occurred in 56% versus 12% (P < 0.001), respectively. Severe neutropenia (ANC <0.5 × 10?/L) occurred in 33% of patients in the HRG compared with 10% in the LRG (P < 0.001).
Patients in the HRG had shorter OS (median OS 13.5 mo vs. 27.4 mo, p=0.02) and PFS (median PFS 7.9 mo vs. 16.1 mo, p=0.03). The incidence of CRS and ICANS was numerically higher in the HRG (CRS 63% vs. 49%, P = 0.07; ICANS 16% vs. 6.8%, P = 0.07), although this was not statistically significant. There were no differences in incidence of infection or infection-related mortality between the groups.
Conclusions
The risk score, derived from baseline CBC values, accurately predicted cytopenia and identified patients with inferior survival following BsAb therapy. If validated in independent cohorts, this tool could facilitate pretreatment risk stratification and guide supportive care interventions in patients receiving bispecific antibodies.
DISCLOSURES
Rafaella Litvin, MD: No relevant conflicts of interest to declare.