A CAR-T response prediction model for r/r B-NHL patients based on a T-cell subset nomogram

doi:10.21203/rs.3.rs-2728417/v1

Download PDF

Research Article

A CAR-T response prediction model for r/r B-NHL patients based on a T-cell subset nomogram

https://doi.org/10.21203/rs.3.rs-2728417/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

Background

Chimeric antigen receptor-modified T-cells (CAR-T) have shown good clinical benefit in the treatment of patients with refractory or relapsed B-cell non-Hodgkin’s lymphoma (r/r B-NHL). However, the factors that influence the clinical response to CAR-T therapy are not fully understood. The aim of this study is to investigate the independent influencing factors of the efficacy of CAR-T cell infusion in the treatment of r/r B-NHL and to establish an early prediction model.

Methods

A total of 43 r/r B-NHL patients were enrolled in this retrospective study to develop and verify the accuracy of the predictive model for CAR-T treatment response. The patients’ pre-treatment baseline data, hematological indicators, T-cell collection baseline indicators, and other general data were recorded, and the patients’ treatment response was mainly observed. The independent factors of complete remission (CR) and partial remission (PR) were investigated by univariate and binary logistic regression analysis, and the prediction model of the probability of CR was constructed according to the determined independent factors. Receiver operating characteristic (ROC) curve and calibration plot were used to assess the discrimination and calibration of the established model.

Results

Univariate analysis and binary logistic regression analysis of 43 patients showed that the ratio of Tcm and Tn in Tc cells was an independent risk factor for response to CAR-T cell therapy in r/r B-NHL. The AUC of Tcm in Tc and Tn in Tc nomogram model was 0.914 (95%CI 0.832-0.996), the sensitivity was 83% and the specificity was 74.2%, which had excellent predictive value.

Conclusions

The ratio of central memory T cells (Tcm) and naïve T cells (Tn) in cytotoxic T cells (Tc) was found to be able to predict the treatment response of CAR-T cells in r/r B-NHL, we have established a nomogram model for the assessment of CD19 CAR-T therapy response which presented a high specificity and sensitivity.

refractory/relapsed B-cell non-Hodgkin's lymphoma

chimeric antigen receptor T cells

predictive models

treatment response

Chimeric antigen receptor-modified T (CAR-T) cells have shown significant advantages in the treatment of hematological malignancies(1). The US Food and Drug Administration (FDA) had approved two CD19 CAR-T products, axicabtagene ciloleucel and tisagenlecleucel to treat refractory/relapsed B-cell non-Hodgkin's lymphoma (r/r B-NHL), demonstrating durable remissions with overall response rates (ORR) of 52–82% and complete response rates (CR) of 40–58%(2, 3). However, some patients do not respond to treatment and up to 60% of patients eventually relapse. This is partly due to the personalized CAR-T cell preparation(4) and the heterogeneous influence of functional T cells, tumor microenvironment(5) and patient baseline status. Therefore, early prediction of response and selection of the most beneficial ones is crucial. In addition, due to the high cost of CAR-T cell therapy, early clinical response evaluation is also important to improve the cost-effectiveness due to the high cost of CAR-T therapy.

The factors that influence the clinical response to CAR-T therapy are not fully understood. Several studies have attempted to determine the impact of these clinical factors(5–11), but no systematic study has evaluated these factors as potential predictive or prognostic indicators of this novel therapy. Here, we analyzed patients with B-NHL who received CD19 CAR-T therapy at our center and screened out independent factors of CAR-T treatment response, found that the proportion of baseline T-cell subsets collected before infusion was positively correlated with patients’ early treatment response, and developed a simple and easy-to-use model to predict the clinical outcome of these patients for the first time.

Patients and study design

A total of 102 patients with refractory and relapsed B-NHL who received CD19 CAR-T therapy at Tianjin First Central Hospital from September 2019 to September 2022 were retrospectively collected. The main inclusion criteria were as follows: (1) B-NHL was diagnosed by histology and immunohistochemistry according to the 2008 WHO classification of hematological malignancies; (2) Refractory or relapsed disease according to the National Comprehensive Cancer Network guidelines; (3) Autologous T cells were selected to prepare CAR-T products; (4) Medical records within 1 month of CAR-T treatment were complete and accessible. Of these, 59 patients were excluded due to incomplete data or death or donor T cell selection within 3 months of CAR-T infusion. A total of 43 patients with r/r NHL were included in the cohort. Clinical data included clinical characteristics and baseline hematological parameters of each subject, and information on the CAR-T cells used at each infusion was collected. ALL data were accurately extracted from the patient’s electronic medical record system.

Study design

The overall devise of the study is shown in Figure 1. Baseline characteristics and hematologic parameters were collected for all 43 patients, and peripheral-blood T cells were tested for T cell subsets, as shown in Table 1. Univariate Mann-Whitney test and binary logistic regression analysis were then used to screen and determine the factors associated with CR and PR in lymphoma patients after CAR-T therapy, including age, gender, baseline tumor burden, bone marrow involvement, previous lines of treatment, ECOG score, baseline T-cell subset proportions, and CAR-T cell dose. In order to determine the independent risk factors affecting CAR-T efficacy, binary logistic regression was used to perform multivariate analysis and calculate the odds ratio and 95% confidence interval. Subsequently, graphpad software was used to draw forest maps with risk factors derived from multivariate analysis. At the same time, the prediction model was generated by odds ratio, and then the ROC curve and the corresponding area under the curve (AUC) were used to evaluate the discriminative ability of the model. Calibration curves were used to estimate the agreement between actual observations and model predictions. The Hosmer-Lemeshow chi-squared test (c2) was also performed to assess accuracy and goodness of fit. The R software was used to construst a nomogram to obtain a predictive model influencing the efficacy of CAR-T treatment, and the calibration curve was drawn to assess the authenticity and accuracy of the model. The ROC curve was then drawn to calculate the specificity, sensitivity, Youden index and kappa value to evaluate the nomogram prediction model.

CAR-T products and lymphatic failure therapy

Peripheral blood mononuclear cells (PBMC) were collected from patients and CD3-positive T cells were sorted by CD3-coated magnetic beads. CD3, CD28-coated magnetic beads and interleukin-2 (IL-2) were added to stimulate T-cell expansion. CAR expression was detected in the laboratory during culture, and a defined number of CAR-T were infused into patients after 7-10 days of culture. In addition, lymphodepleting chemotherapy with fludarabine and cyclophosphamide was administered for an average of 3 days prior to CAR-T infusion. The specific doses were 25mg/m2/day and 250mg/m2/day.

Assessment of response

43 patients were evaluated 1-3 months after CAR-T cell therapy. Evaluation criteria included: (1) CR: Patients had no fever and clinical manifestations showed complete disappearance of lymphoma markers; according to the results of imaging examination, no radioactive enhancement of lymph nodes was found. (2) PR: Patients had no fever and clinical manifestations showed significant reduction in tumor markers; imaging results showed that the radioactive enhancement was reduced by more than 50% compared to the last time, which had a durable therapeutic effect. (3) No response (NR): There was no change or a slight reduction in lymphoma markers, imaging results did not show a significant reduction in radioactivity enhancement, and there was no durable response.

Statistical analysis

Statistical software SPSS 22.0 (SPSS Inc, Chicago, IL, USA) was used for univariate and binary logistic regression analysis, and graphing software GraphPad Prism 8 was used for statistical analysis and mapping of data. The Mann-Whitney test was used for comparison between samples. Binary logistic regression analysis was performed using the Hosmer-Lemeshow test with SPSS software (SPSS 22.0). For each infusion, patients can obtain the probability score of CR and PR generated by the constructed model, and then the clinical outcome of each infusion is evaluated according to the cut-off value of the ROC curve when the Youden index reaches the maximum value. The R software used was R (4.2.2), RStudio (version 1.4.1717), and the following R packages: rms, readxl and nomogram Formula to perform the nomogram in order to obtain a predictive model that influenced the treatment effect of lymphoma patients after CAR-T treatment, and the calibration curve was drawn to evaluate the authenticity and accuracy of the model. The ROC curve was then drawn, and the specificity, sensitivity, Youden index and other evaluation indicators were calculated. The consistency and reliability of the model prediction was assessed by calculating the Lappa value. All reported P values are two-tailed, and a P value of less than 0.05 was considered significant.

Clinical characteristics of the study population

The Table 1 shows the 43 patients’ baseline characteristics and the CAR-T cells they initially received. There were 36 (83.7%) cases of diffuse large B-cell lymphoma, 2 (4.7%) cases of mantle cell lymphoma, 2 (4.7%) cases of follicular lymphoma, 2 (4.7%) cases of B- cell lymphoblastic lymphoma, and 1 (2.2%) case of Burkitt’s lymphoma. There were approximately three times as many men as women and the median age was 56 years (range 26-85 years). Ten patients (23.3%) had bone marrow involvement prior to CAR-T cell infusion. 23 (62.7%) patients had high tumor burden. The median number of lines of therapy prior to CAR-T cell infusion was 3 (range,1-7).

Table1 patient covariates(N=43)
Covariates	n^*
Age，median(range)	56（26-85）
Gender， n(%)
Male	31（72.1）^*
Female	12（27.9）^*
Disease
DLBCL	36（83.7）^*
MCL	2（4.7）^*
FL	2（4.7）^*
B-LBL	2（4.7）^*
Burkitt	1（2.2）^*
Tumor burden prior CAR-T
Low	16（37.2）^*
High	27（62.7）^*
Ann Arbor stage
I-II	8（15.4）^*
III-IV	35（82.0）^*
ECOG score
＜2	19（44.2）^*
≥2	24（55.8）^*
Lines of prior therapies, median(range)	3(1-7)
Bone marrow infiltration
Yes	10（23.3）^*
No	33（76.7）^*
Baseline blood count, median(range)
LDH	225.3（116.4-7477.8）
CRP	4.52（0.54-341.19）
Ferritin	379（34.3-2222）
WBC（×10^9/L）	3.97(1.48-13.85)
Hemoglobin（g/dL）	109(59-145)
Platelet（×10^9/L）	154(23-670)
Baseline T cell subset proportions，median（range）
CD4/CD8	1.1（0.15-13.43）
Treg（%）	5.71（0.07-31.47）
Tcm in Th（%）	7.68（0.38-62.95）
Tcm in Tc（%）	31.88（3.17-64.77）
Tn in Th（%）	13.55（1.15-46.77）
Tn in Tc（%）	8.64（0-48.68）
Teff in Th（%）	2.88（0.16-50.38）
Teff in Tc（%）	33.43（0-79.28）
Tem in Th（%）	34.11（5.72-92）
Tem in Tc（%）	35.12（0-84.97）
CAR-T cell dose（×10^6/kg）	3.82(0.85-12.79)
Outcome variable
CR/PR	31（72.1）^*
No response	12（27.9）^*
（）^*, percentage; DLBCL, diffuse large B cell lymphoma; MCL, mantle cell lymphoma; FL, follicular lymphoma; B-LBL, B cell lympho-blastic lymphoma; ECOG, eastern cooperative oncology group; LDH, lactate dehydrogenase; CRP, C-reaction protein; WBC, white blood cells

The median dose of infused CAR-T cells was 3.82×10^6/kg (range, 0.85-12.79). The median WBC, hemoglobin, platelet, LDH, CRP and ferritin counts before lymphodepletion (baseline blood count) were 3.97×10^9/L, 109g/L, 154×10^9/L, 225.3U/L, 4.52mg/L and 379ng/ml. Before treatment, the median CD4/CD8 ratio of peripheral blood T cells were 1.1 (0.15-13.43), the median proportion of Treg cells was 5.71% (0.07-31.47), the proportion of Tcm in Th and Tc cells was 7.68% (0.38-62.95) and 31.88% (3.17-64.77), respectively. The proportion of Tn in Th and Tc cells was 13.55% (1.15-46.77) and 8.64% (0-48.68), respectively. The proportion of Teff in Th and Tc was 2.88% (0.16-50.38) and 33.43% (0-79.28), respectively. The proportion of Tem in Th and Tc cells was 34.11% (5.72-92) and 35.12% (0-84.97), respectively. 31 (72.1%) patients achieved CR or PR one to three months after CAR-T cell infusion.

Factors associated with CR and PR in r/r NHL patients after CAR-T therapy

To search for possible factors for CAR-T treatment in r/r NHL patients, we first performed univariate logistic regression analysis on patients’ clinical characteristics (Table 2), baseline hematological parameters (Table 3), and first infusion of CAR-T cells (Table 4). The results of univariate analysis showed that the patient’s baseline tumor burden, ECOG score, the proportion of Treg cells in peripheral blood T cell subsets, and the proportion of Tcm and Tn in Tc cells at diagnosis were significantly associated with remission after CAR-T cell infusion (p<0.05). As demonstrated in Figure 2B, 2C and 2D, the proportion of Treg cells, as well as Tcm and Tn in Tc, in the response group was significantly lower than in the non-response group. Although age and the proportion of Tn in Th cells showed in figure 2E present correlation with remission, the difference was not statistically significant. However, in Figure 2a, the CD4/CD8 ratio was not different between the two groups, which may be related to the small sample size.

Table2 Univariate logistic regression analyses of the clinical characteristics of r/r NHL patients associated with CR or PR
Variables	Category	Remission	No-remission	Z	P
Age(years)		60（46，68）^#	46.5（41，58.5）^#	-1.818	0.071
Gender	Male	22（70.97）	9（75）	-0.261	0.841
Gender	Female	9（29.03）	3（25）	-0.261	0.841
Tumor burden	High	16（51.61）	1（8.33）	-2.409	0.043
Tumor burden	Low	15（48.39）	11（91.7）	-2.409	0.043
ECOG	＜2	18（58.06）	1（8.33）	-3.396	0.002
ECOG	≥2	13（41.94）	11（91.7）	-3.396	0.002
Bone marrow infiltration	Yes	6（19.35）	4（33.33）	-0.962	0.495
Bone marrow infiltration	No	25（80.65）	8（66.67）	-0.962	0.495
Lines of prior therapies	＜4	19（61.29）	3（25）	-1.422	0.174
Lines of prior therapies	≥4	12（38.71）	9（75）	-1.422	0.174
（）^#inter-quartile range，percentage for other covariates

Table3 Univariate logistic regression analyses of the blood index of r/r NHL patients associated with CR or PR
Variables	Remission	No-remission	Z	P
WBC	3.53（2.57，5.93）	5（3.625，6.39）	-1.476	0.142
Hemoglobin	111（95，126）	108（90.5，118）	-0.555	0.584
Platelet	153（118，212）	184.5（141.25，284.75）	-1.489	0.142
CRP	4.52（1.12，26.04）	6.57（1.135，40.7025）	-0.108	0.926
Ferritin	321（196，535）	551（289，692.3）	-1.543	0.127
LDH	217.2（191，334.4）	334.05（231.575，484.925）	-1.462	0.149
（）percentage

Table4 Univariate logistic regression analyses of the baseline information of CAR-T therapy and CAR-T cell associated with CR or PR
Variable	Remission	No-remission	Z	P
CD4/CD8	1.1（0.55，1.78）	1.14（0.4225，1.4675）	-0.081	0.947
Treg	4.99（1.05，6.84）	9.225（5.9875，18.76）	-3.222	0.001
Tcm in Th	7.68（4.68，16.67）	8.155（2.535，15.1525）	-0.596	0.565
Tcm in Tc	36.18（29.16，50.05）	16.67（7.1575，28.4775）	-3.574	0.000
Tn in Th	15.28（8.59，28）	8.07（5.8775，18.4125）	-1.814	0.071
Tn in Tc	12.77（7.01，21.34）	3.275（0.9625，6.3225）	-3.385	0.000
Tem in Th	34.11（24.49，49.9）	29.985（8.6175，51.13）	-1.164	0.254
Tem in Tc	40.19（21.6，59.47）	33.765（22.6，47.69）	-0.839	0.414
Teff in Th	2.84（1.4，7.09）	3.175（1.59，6.5975）	-0.271	0.8
Teff in Tc	33.43（18.47，52.72）	34.245（17.12，46.315）	-0.460	0.659
CAR-T cell dose（×10^6/kg）	4（2.05，6.65）	3.65（1.58，6.12）	-1.090	0.287
（）inter-quartile range

Figure 2 The proportions of T cell subsets in NHL patients. A. There was no significant difference in the proportion of CD4+ and CD8+T cells between response and non-response group (p=0.6268). B. The proportion of Treg cells in the response group was significantly lower than in the non-response group (p<0.05). C-D. The proportion of Tcm and Tn in Tc cells was significantly higher than that in the non-response group (p<0.05). E. The proportion of Tn in the response group was higher than that in the non-response group, but the difference was not statistically significant (p>0.05).

Identification of independent factors influencing response to CAR-T therapy

To further identify the independent factors for response after CAR-T therapy, the statistically significant factors in the above univariate analysis (univariate logistic p<0.05) were included in the binary logistic regression analysis (Table 5). The results of the binary analysis showed that the changes in the proportions of T cell subsets collected from the patients’ peripheral blood of patients before CAR-T infusion were significant independent factors for the therapeutic effect of CAR-T cells, among which the proportions of Tcm and Tn in Tc cells significantly influenced remission (p<0.05) (Figure 3A). ROC curve results showed that the AUC of Tcm in Tc and Tn in Tc for predicting the efficacy of CAR-T cell therapy was >60%, the AUC of Tcm in Tc was 0.855(95%CI 0.737-0.9730), and the AUC of Tn in Tc was 0.836(95%CI 0.712-0.960) (Table 6) (Figure 3B).

Table5 Multiple logistic regression analysis of the variables of clinical characteristics of r/r NHL patients and CAR-T cell associated with CR or PR
Variable	B	Wald	P	OR	95CI%
Tumor burden (Low=1；High=0)	-1.155	0.331	0.565	0.315	0.006-16.139
ECOG (＜2=1；≥2=0)	-3.116	3.667	0.056	0.044	0.002-1.076
Treg (＜5.71=1；≥5.71=0)	-1.560	1.583	0.208	0.210	0.018-2.388
Tcm in Tc (＜31.88=1；≥31.88=0)	2.934	3.942	0.047	18.794	1.038-340.137
Tn in Tc (＜8.64=1；≥8.64=0)	2.777	4.252	0.039	16.075	1.147-225.226

Table6 Area under the ROC curve of Tcm and Tn in Tc for predicting the outcome of lymphoma patient
Variable	AUC	Standard error	P	95%CI
Tcm in Tc	0.855	0.060	0.000	0.737-0.973
Tn in Tc	0.836	0.063	0.001	0.712-0.960

Figure 3 Discriminative ability of Tcm and Tn in Tc between CR and NR. A. Multivariate logistic regression analysis showed that Tcm and Tn in Tc were closely related to CAR-T treatment response (p<0.05). B. The ROC curve showed the ability of the T subsets to discriminate between the response and non-response groups. Tcm and Tn in Tc represented good discriminative power with AUC>0.6, respectively.

Establishment and validation of an early prediction model for CAR-T treatment response

Potential biomarkers with concurrent AUC>0.6 were identified based on the selection criteria of p<0.05 between the remission and non-remission groups. We therefore selected Tcm in Tc cells and Tn in Tc cells as predictive biomarkers of early response after CAR-T therapy. Based on the baseline percentage of Tcm and Tn in Tc cells, we developed an early prediction model of treatment effect to analyze the likelihood of remission. Then, to assess remission response in clinical practice, we developed a nomogram that could help predict the likelihood of CR and PR for each individual in clinical practice. Finally, we internally validated the original data and drew a calibration curve to further evaluate the authenticity and accuracy of the prediction model in the actual situation. The results showed good agreement between the likelihood of predicting remission and the observed early response (Figure 4A and 4B). The total scores of Tcm and Tn were calculated in the nomogram, and the ROC curve was drawn to predict the treatment effect. The AUC result was 0.914 (95%CI 0.832-0.996), showing good predictive ability (Figure 4C). The results showed that the prediction model had high sensitivity and specificity in predicting the effect of CAR-T treatment (sensitivity 83%, specificity 74.2%), and the maximum Youden index was 0.742, and the prediction results were in good agreement with the actual situation (kappa, 0.616), indicating that the model has good predictive ability and has some clinical application value.

Chimeric antigen receptor-modified T-cells (CAR-T) have become increasingly effective in the treatment of malignant tumors, particularly hematological tumors. At present, CAR-T cell therapy is expensive. If the therapeutic effect of CAR-T cells can be predicted before infusion and the appropriate treatment plan can be formulated for patients, CAR-T cell therapy can better serve the clinic, further move towards precision treatment, achieve disease management and avoid unnecessary medical burden. Therefore, finding new markers of CAR-T cell efficacy in lymphoma patients is a current medical need. These include the tumor burden of the tumor itself and the antigen density on the surface of the tumor cells(12, 13), as well as factors related to the CAR structure, such as the length of the hinge region and the choice of co-stimulatory factors(14–16). At the same time, the differentiation status of the patient’s own T cells also plays an important role. On the one hand, the enrichment of CAR-T cell products in T cells with more immature phenotypes may be related to the improvement of the long-term response. On the other hand, the increase of T cells with effector or depleted phenotypes may lead to decreased persistence of CAR-T cells and reduced clinical response(6, 17–19). Therefore, we conducted this study.

In this study, we constructed a risk model to predict treatment response in r/r B-NHL patients treated with CAR-T. We analyzed the patients’ clinical characteristics, baseline hematological parameters and the proportion of T-cell subsets to determine the independent risk factors affecting the efficacy of CAR-T therapy. In combination with univariate and multivariate analyses, changes in the proportion of T-cell subsets were considered independent clinical factors associated with early clinical response after CAR-T therapy. We analyzed the subpopulation proportions of T cells isolated from peripheral blood before treatment and identified predictive biomarkers of Tcm in Tc and Tn in Tc associated with treatment response. Finally, we developed and validated an early prediction model to analyze the likelihood of CR based on the proportion of Tcm in Tc and Tn in Tc in isolated T cells. This model helps to predict the early treatment response in r/r B-NHL patients. Quantifiable phenotypes associated with infusion products can be used to take early action by enriching or depleting undesirable cell populations or functional states during the manufacturing process to personalize treatment for corresponding patients, thereby reducing the failure rate of CAR-T therapy for r/r B-NHL.

In this retrospective study, we found no significant differences in age, sex, disease type, number of prior lines of therapy, bone marrow involvement, baseline blood count, baseline biochemical measures, and CD4/CD8 cell ratio between the remission and non-remission groups (Tables 2, 3, 4). It has been reported that age and gender are not risk factors for response to CAR-T cell therapy, which is consistent with our results. However, there was no significant difference in LDH, ferritin and CRP levels between remission and non-remission patients. This is not consistent with previous reports describing lactate dehydrogenase LDH before leukopenia as a strong predictor of CR after CD19 CAR-T cell therapy(20). Extramedullary infiltration is an independent risk factor for death and relapse after CAR-T cell therapy, especially in patients with central nervous system involvement(21), but in this study there was no statistically significant difference between the two groups, possibly due to the short observation period, some patients achieved transient remission in the early stage after treatment, and their long-term efficacy was not observed.

The cellular and molecular diversity of the infused CAR-T cell products is an important factor contributing to the efficacy of LBCL treated with CD19 CAR-T cells. Previous studies have reported that infused product cells from patients with durable CR at 3 months had an enriched CD8 memory T cell phenotype compared to patients with PR/PD(22). At the same time, the baseline clinical characteristics of patients such as tumor burden(23) and ECOG score, also have a certain impact on the treatment effect. Among them, patients with high tumor burden have higher exhaustion markers of autologous T cells. At the same time, patients with a high tumor burden are less likely to achieve a CR after CAR-T therapy. In conclusion, patients with low tumor burden have a better response to CAR-T cell therapy and a set of perfect criteria is needed to identify patients with low tumor burden. ctDNA, ECOG score and EASIX score (24) can reflect the tumor burden of patients at different levels and predict the efficacy of CAR-T cell therapy before infusion. The current study confirmed these findings (Table 2). At present, there is no clear result on the effect of CD4/CD8 ratio on the efficacy of CAR-T therapy, but some scholars believe that CD8 + T-cell-mediated immune response against CAR transgenic products after CAR-T cell infusion limits the persistence of CAR-T cells and increases the risk of relapse in some patients(25). In this study, no significant difference in the ratio of CD4 to CD8 cells was observed between the remission and non-remission groups, but the proportion of CD4-positive T cells was slightly higher in the remission group than in the CD8-positive T cells, confirming the above study (Fig. 2A), and the levels of Tcm and Tn in Tc cells were significantly higher in the remission group than in the non-remission group. In addition, the level of Treg cells was lower than that in the non-remission group, and the difference was statistically, significant (Table 4) (Fig. 2B, C, D). Combined with the results of the univariate analysis, binary logistic regression analysis was performed to determine Tcm in Tc and Tn in Tc as biomarkers. The early response prediction model was constructed and verified. The results showed that the constructed model could predict early CAR-T response well. The AUC was 0.914 (95%CI 0.832–0.996) (Fig. 4C). However, this study has certain limitations. First, the number of patients was relatively small, and the prediction model we constructed needs to be further validated in a larger prospective study. In addition, most of the patients were discharged within 1–3 months after CAR-T treatment, and we could only observe the short-term treatment response after CAR-T treatment, and there was no long-term outcome of the patients. At the same time, we will pay attention to the long-term prognosis after CAR-T treatment and evaluate the long-term impact of changes in the proportion of T-cell subsets on the response to CAR-T treatment.

CAR-T: Chimeric antigen receptor T cell

r/r B-NHL: refractory or relapsed B-cell non-Hodgkin’s lymphoma

CR: complete remission

PR: partial remission

ROC: receiver operating characteristic

Tcm: central memory T cell

Tn: naïve T cell

Tc: cytotoxic T

FDA: Food and Drug Administration

ORR: overall response rates

AUC: area under the curve

PBMC: Peripheral blood mononuclear cells

NR: No response

Availability of data and materials

The data that support the findings of this study are available on request from the corresponding author.

Acknowledgements

Not applicable

Funding

This work was supported by grants from the General Project of the National Natural Science Foundation of China (81970180 to MZ), the Science and Technology Project of Tianjin Municipal Health Committee (TJWJ2022QN030 to MZ), the Key projects of Tianjin Applied Basic Research and Multi-Investment Fund (21JCZDJC01240), the Science and Technology Project of Tianjin Municipal Health Committee (TJWJ2022XK018 to MZ), and the Key Science and Technology Support Project of Tianjin Science and Technology Bureau (20YFZCSY00800 to MZ), as well as Tianjin Key Medical Discipline (Specialty) Construction Project (TJYXZDXK-056B).

Author information

Xiaomei Zhang and Rui Sun contributed equally

Authors and Affiliation

School of medicine, Nankai University, Tianjin, China

Xiaomei Zhang, Rui Sun & Mingfeng Zhao

First Central Clinical College, Tianjin Medical University, Tianjin, China

Yifan Zhao, Xinping Cao, Ruiting Guo & Yi Zhang

Department of Hematology, Tianjin First Central Hospital, Tianjin, China.

Mingfeng Zhao

Contributions

XMZ and RS wrote the manuscript. YFZ, XPC, RTG and YZ have substantively revised the manuscript. MFZ conceived this review and is accountable for all aspects of the work. The final manuscript read and approved by all authors.

Corresponding author

Correspondence to Mingfeng Zhao.

Ethics declarations

The study was approved by the Ethics Committee of Tianjin First Central Hospital. And our study complies with the Declaration of Helsinki. The need for informed consent was waived due to the retrospective nature of the study and because the data were anonymously analyzed. Written informed consent was obtained from the individual for the publication of any potentially identifiable images or data included in this article.

Consent for publication

Not applicable

Competing interests

All authors declare that there are no conflicts of interest

Avanzi MP, Brentjens RJ. Emerging Role of CAR T Cells in Non-Hodgkin's Lymphoma. Journal of the National Comprehensive Cancer Network : JNCCN. 2017;15(11):1429-37.
Schuster SJ, Bishop MR, Tam CS, Waller EK, Borchmann P, McGuirk JP, et al. Tisagenlecleucel in Adult Relapsed or Refractory Diffuse Large B-Cell Lymphoma. The New England journal of medicine. 2019;380(1):45-56.
Locke FL, Ghobadi A, Jacobson CA, Miklos DB, Lekakis LJ, Oluwole OO, et al. Long-term safety and activity of axicabtagene ciloleucel in refractory large B-cell lymphoma (ZUMA-1): a single-arm, multicentre, phase 1-2 trial. The Lancet Oncology. 2019;20(1):31-42.
Kaartinen T, Luostarinen A, Maliniemi P, Keto J, Arvas M, Belt H, et al. Low interleukin-2 concentration favors generation of early memory T cells over effector phenotypes during chimeric antigen receptor T-cell expansion. Cytotherapy. 2017;19(6):689-702.
Yan ZX, Li L, Wang W, OuYang BS, Cheng S, Wang L, et al. Clinical Efficacy and Tumor Microenvironment Influence in a Dose-Escalation Study of Anti-CD19 Chimeric Antigen Receptor T Cells in Refractory B-Cell Non-Hodgkin's Lymphoma. Clinical cancer research : an official journal of the American Association for Cancer Research. 2019;25(23):6995-7003.
Rossi J, Paczkowski P, Shen YW, Morse K, Flynn B, Kaiser A, et al. Preinfusion polyfunctional anti-CD19 chimeric antigen receptor T cells are associated with clinical outcomes in NHL. Blood. 2018;132(8):804-14.
Hirayama AV, Gauthier J, Hay KA, Voutsinas JM, Wu Q, Gooley T, et al. The response to lymphodepletion impacts PFS in patients with aggressive non-Hodgkin lymphoma treated with CD19 CAR T cells. Blood. 2019;133(17):1876-87.
Jia Q, Qin D, He F, Xie Q, Ying Z, Zhang Y, et al. Peripheral eosinophil counts predict efficacy of anti-CD19 CAR-T cell therapy against B-lineage non-Hodgkin lymphoma. Theranostics. 2021;11(10):4699-709.
Breen WG, Hathcock MA, Young JR, Kowalchuk RO, Bansal R, Khurana A, et al. Metabolic characteristics and prognostic differentiation of aggressive lymphoma using one-month post-CAR-T FDG PET/CT. Journal of hematology & oncology. 2022;15(1):36.
Jackson Z, Hong C, Schauner R, Dropulic B, Caimi PF, de Lima M, et al. Sequential single cell transcriptional and protein marker profiling reveals TIGIT as a marker of CD19 CAR-T cell dysfunction in patients with non-Hodgkin's lymphoma. Cancer discovery. 2022.
Vercellino L, Di Blasi R, Kanoun S, Tessoulin B, Rossi C, D'Aveni-Piney M, et al. Predictive factors of early progression after CAR T-cell therapy in relapsed/refractory diffuse large B-cell lymphoma. Blood advances. 2020;4(22):5607-15.
Majzner RG, Rietberg SP, Sotillo E, Dong R, Vachharajani VT, Labanieh L, et al. Tuning the Antigen Density Requirement for CAR T-cell Activity. Cancer discovery. 2020;10(5):702-23.
Locke FL, Rossi JM, Neelapu SS, Jacobson CA, Miklos DB, Ghobadi A, et al. Tumor burden, inflammation, and product attributes determine outcomes of axicabtagene ciloleucel in large B-cell lymphoma. Blood advances. 2020;4(19):4898-911.
Kawalekar OU, O'Connor RS, Fraietta JA, Guo L, McGettigan SE, Posey AD, Jr., et al. Distinct Signaling of Coreceptors Regulates Specific Metabolism Pathways and Impacts Memory Development in CAR T Cells. Immunity. 2016;44(2):380-90.
Guedan S, Posey AD, Jr., Shaw C, Wing A, Da T, Patel PR, et al. Enhancing CAR T cell persistence through ICOS and 4-1BB costimulation. JCI insight. 2018;3(1).
Alabanza L, Pegues M, Geldres C, Shi V, Wiltzius JJW, Sievers SA, et al. Function of Novel Anti-CD19 Chimeric Antigen Receptors with Human Variable Regions Is Affected by Hinge and Transmembrane Domains. Molecular therapy : the journal of the American Society of Gene Therapy. 2017;25(11):2452-65.
Garfall AL, Dancy EK, Cohen AD, Hwang WT, Fraietta JA, Davis MM, et al. T-cell phenotypes associated with effective CAR T-cell therapy in postinduction vs relapsed multiple myeloma. Blood advances. 2019;3(19):2812-5.
Turtle CJ, Hanafi LA, Berger C, Hudecek M, Pender B, Robinson E, et al. Immunotherapy of non-Hodgkin's lymphoma with a defined ratio of CD8+ and CD4+ CD19-specific chimeric antigen receptor-modified T cells. Science translational medicine. 2016;8(355):355ra116.
Sommermeyer D, Hudecek M, Kosasih PL, Gogishvili T, Maloney DG, Turtle CJ, et al. Chimeric antigen receptor-modified T cells derived from defined CD8+ and CD4+ subsets confer superior antitumor reactivity in vivo. Leukemia. 2016;30(2):492-500.
Nastoupil LJ, Jain MD, Feng L, Spiegel JY, Ghobadi A, Lin Y, et al. Standard-of-Care Axicabtagene Ciloleucel for Relapsed or Refractory Large B-Cell Lymphoma: Results From the US Lymphoma CAR T Consortium. Journal of clinical oncology : official journal of the American Society of Clinical Oncology. 2020;38(27):3119-28.
An F, Wang H, Liu Z, Wu F, Zhang J, Tao Q, et al. Influence of patient characteristics on chimeric antigen receptor T cell therapy in B-cell acute lymphoblastic leukemia. Nature communications. 2020;11(1):5928.
Deng Q, Han G, Puebla-Osorio N, Ma MCJ, Strati P, Chasen B, et al. Characteristics of anti-CD19 CAR T cell infusion products associated with efficacy and toxicity in patients with large B cell lymphomas. Nature medicine. 2020;26(12):1878-87.
Thieblemont C, Chartier L, Dührsen U, Vitolo U, Barrington SF, Zaucha JM, et al. A tumor volume and performance status model to predict outcome before treatment in diffuse large B-cell lymphoma. Blood advances. 2022;6(23):5995-6004.
Zhang W, Wang W, Han X, Gan Y, Qian L, Zhang Y, et al. Circulating tumor DNA by high-throughput sequencing of T cell receptor monitored treatment response and predicted treatment failure in T cell lymphomas. International journal of laboratory hematology. 2021;43(5):1041-9.
Turtle CJ, Hanafi LA, Berger C, Gooley TA, Cherian S, Hudecek M, et al. CD19 CAR-T cells of defined CD4+:CD8+ composition in adult B cell ALL patients. The Journal of clinical investigation. 2016;126(6):2123-38.

Download PDF

Version 1

posted

You are reading this latest preprint version

A CAR-T response prediction model for r/r B-NHL patients based on a T-cell subset nomogram

Status:

Version 1

Abstract

Figures

Background

Methods

Results

Discussion

Abbreviations

Declarations

References

Status:

Version 1