Outcome Prediction in Patients With Large B-cell Lymphoma Undergoing Chimeric Antigen Receptor T-cell Therapy : HemaSphere

Journal Logo


Outcome Prediction in Patients With Large B-cell Lymphoma Undergoing Chimeric Antigen Receptor T-cell Therapy

Voltin, Conrad-Amadeus1; Gödel, Philipp2,3; Beckmann, Laura2,3; Heger, Jan-Michel2,3; Kobe, Carsten1; Kutsch, Nadine2,3; Borchmann, Peter2,3; Dietlein, Markus1; Herrmann, Ken4; Stelljes, Matthias5; Rahbar, Kambiz6; Lenz, Georg5; Reinhardt, H. Christian3,7; Teichert, Marcel7; Noppeney, Richard7; Albring, Jörn C.5; Seifert, Robert4,6; von Tresckow, Bastian3,7; Flossdorf, Sarah8; Hanoun, Christine7

Author Information
HemaSphere 7(1):p e817, January 2023. | DOI: 10.1097/HS9.0000000000000817
  • Open



Chimeric antigen receptor (CAR) T-cell therapy can induce long-term remission in a considerable proportion of patients with relapsed or refractory large B-cell lymphoma.1–3 However, potential biomarkers for identifying individuals most likely to benefit from this novel treatment are still under investigation. In the phase II JULIET trial, elevated lactate dehydrogenase (LDH) before infusion of tisagenlecleucel was associated with significantly shorter progression-free (PFS) and overall survival (OS).4 Moreover, a study by Locke et al5 showed that pretreatment tumor burden estimated on computed tomography (CT) as defined in the Cheson criteria, LDH, and interleukin-6 correlate with the probability of durable response to axicabtagene ciloleucel.

Positron emission tomography (PET) using 18F-fluorode oxyglucose (FDG) has become an essential diagnostic modality for the staging and response assessment of large B-cell lymphoma.6–9 It is highly accurate in detecting both nodal and extra-nodal (EN) sites of disease.10,11 Some recently introduced PET-derived biomarkers might further improve risk stratification and have been shown to provide significant prognostic value for individuals undergoing chemoimmunotherapy.12–15 There is only limited data available on the role of novel parameters such as metabolic tumor volume (MTV) in identifying patients best suited for CAR T-cell treatment.16–18

Therapy decisions should ideally be based on both tumor characteristics and individual patient factors. Yet, there is no established predictive model for the specific context of CAR T-cell treatment. We therefore set out to examine potential risk factors taking PET metrics into account.


Study cohort

After approval by the institutional ethics committee, patients treated through January 31, 2021, were enrolled in our study based on the following criteria:

  • (1) relapsed or refractory, biopsy-proven large B-cell lymphoma;
  • (2) PET scan performed within 30 days of tisagenlecleucel or axicabtagene ciloleucel infusion;
  • (3) no cytoreductive treatment between imaging and final product administration besides fludarabine and cyclophosphamide or local radiotherapy.

The three participating German university centers identified a total of 47 individuals suitable for analysis. All of them provided written informed consent before PET examination and CAR T-cell therapy.

Quantitative PET evaluation

MTV measurements were performed semi-automatically without lower lesion volume limit by two expert readers using the ACCURATE tool (PETRA consortium, Amsterdam, The Netherlands) and syngo.via (Siemens Healthcare GmbH, Erlangen, Germany) based on a standardized uptake value (SUV) threshold of 4.0. In lymphoma tissue surrounded by areas of high physiological FDG uptake, particularly intracerebral lesions, manual correction was needed to avoid overestimating the tumor volume. Moreover, we documented the maximum SUV (SUVmax) as an additional metabolic parameter.

Definition of predictive factors and statistical analysis

Patient outcomes following CAR T-cell treatment were measured in our study by PFS as well as OS, using the Kaplan–Meier method for assessment of one-year survival and log-rank test to compare risk groups. Median follow-up duration was calculated with a reverse Kaplan–Meier estimator.

We identified the most predictive model for PFS in multivariable analysis based on forward selection through Akaike’s information criterion (AIC) using Cox and logistic regression. Here, tumor stage, the presence of more than one EN disease site, LDH elevation, patient age, Eastern Cooperative Oncology Group score, treatment lines, C-reactive protein (CRP) values, and response after bridging therapy were considered as factors that might affect the outcome. The optimal threshold of CRP, MTV, and SUVmax for PFS was determined by Cox as well as logistic regression based on AIC and receiver operating characteristic (ROC) analysis with Youden’s index, respectively. Our study also examined potential correlations between factors and whether higher metabolic tumor burden represents a risk for the development of cytokine release (CRS) or immune effector cell-associated neurotoxicity syndrome (ICANS) using Pearson’s coefficients.

Baseline data were collected from the electronic medical records and evaluated through descriptive methods. Moreover, we assessed the inter-observer variability in PET volume measurements based on a test set of 24 individuals, using linear regression to ensure that MTVs calculated by the two readers could be pooled. All statistical analysis was conducted with SAS software version 9.4 (SAS Institute, Cary, NC).


Patient characteristics

Median age at time of treatment was 61 years and 18 individuals were female (38%). The majority had diffuse large B-cell lymphoma (n = 33, 70%), followed by transformed follicular lymphoma (n = 10, 21%), primary mediastinal B-cell lymphoma (n = 2, 4%), and high-grade B-cell lymphoma (n = 2, 4%). Bridging therapy was administered in all enrolled patients and 26 were refractory thereafter (55%). While 43 individuals received tisagenlecleucel (91%), four underwent treatment with axicabtagene ciloleucel (9%). Further baseline characteristics and disease-related information are provided in Table 1.

Table 1 - Patient and Disease Characteristics Before CAR T-cell Administration
Age (y) Median 61
Range 19–82
Sex Female 18 (38)
Male 29 (62)
Lymphoma subtype Diffuse large B-cell lymphoma 33 (70)
Transformed follicular lymphoma 10 (21)
Primary mediastinal B-cell lymphoma 2 (4)
High-grade B-cell lymphoma 2 (4)
IPI factors Age >60 y 25 (53)
Ann Arbor stage III or IV 30 (64)
More than one EN lesion 25 (53)
ECOG performance status ≥2 10 (21)
Elevated LDH 32 (68)
IPI score 1–2 20 (43)
3–5 27 (57)
Bulky diseasea Yes 6 (13)
No 41 (87)
Treatment lines Median 3
Range 2–12
Prior stem-cell transplantation Autologous 15 (32)
Allogeneic 1 (2)
Response to bridging therapy Complete remission 6 (13)
Partial response 15 (32)
Stable disease 5 (11)
Progressive disease 21 (45)
Data are n (%) unless specified otherwise.
aNodal mass with a diameter of .7.5 cm in at least one axis.
CAR = chimeric antigen receptor; ECOG = Eastern Cooperativ

Outcome prediction using patient and tumor-related factors

Forward selection identified EN disease as the most predictive factor with a hazard ratio (HR) of 1.71 (95% confidence interval [CI] 0.92-3.17). No other patient or tumor characteristic considered was of added value in our analysis set.

After a median follow-up time of 17.03 months, one-year PFS and OS probabilities were 8.0% (95% CI 2.1-30.2%) and 38.6% (95% CI 23.3-63.9%) for the subgroup of individuals who had more than one EN lesion (Figure 1A and B). Patients free of this risk factor achieved a one-year PFS rate of 18.2% (95% CI 7.5-44.1%, P = 0.083) and OS probability was significantly superior (P = 0.012) with 67.0% (95% CI 49.6-90.5%).

Figure 1.:
Kaplan–Meier curves for PFS as well as OS according to (A, B) EN disease status, (C–F) PET parameters, and (G, H) combined risk stratification. EN = extra-nodal; MTV = metabolic tumor volume; OS = overall survival; PET = positron emission tomography; PFS = progression-free survival; SUVmax = maximum standardized uptake value.

PET measurements and association with survival

The patients included underwent PET scanning a median of 7 days (range 0–29 days) before CAR T-cell administration. Linear regression analysis yielded a slope coefficient and R-squared of 0.91 and 0.98, respectively, for metabolic lymphoma burden indicating comparable values between the two expert readers. Median MTV before CAR T-cell treatment was 44 mL in our cohort (range 0–1831 mL). Based on the AIC score, we selected a tumor volume of 11 mL as optimal threshold regarding PFS with a HR of 2.57 (95% CI 1.16-5.67). ROC analysis confirmed this cut-off and revealed a sensitivity and specificity of 75.0% and 79.1%, respectively (Figure 2).

Figure 2.:
ROC curves for predicting PFS based on PET parameters. MTV = metabolic tumor volume; PET = positron emission tomography; PFS = progression-free survival; ROC = receiver operating characteristic; SUVmax = maximum standardized uptake value.

One-year PFS and OS probability was 5.7% (95% CI 1.5-21.9%) and 46.6% (95% CI 32.3-67.3%), respectively, for individuals with elevated MTV (Figure 1C and D). In contrast, patients who had lower tumor load achieved a significantly superior PFS (P = 0.016) with 33.3% (95% CI 15.0-74.2%) free of disease progression and death after 12 months, while the one-year OS probability was 66.7% (95% CI 44.7-99.5%, P = 0.22). However, we did not observe any association between higher metabolic tumor burden and the development of CRS (r = 0.015) or ICANS (r = 0.044).

High MTV and EN disease were weakly correlated (r = 0.27) and can thus be considered as independent predictors of survival. A particularly favorable outcome was noted for patients with no risk factors who had one-year PFS and OS rates of 50.0% (95% CI 25.0-100%) and 87.5% (95% CI 67.3-100%), respectively (Figure 1G and H). Moreover, SUVmax predicted one-year PFS (P = 0.002), but not OS (P = 0.301), after CAR T-cell treatment when using the optimal threshold of 16.7 (Figure 1E and F).


CAR T-cell therapy heralds a fundamental shift in the management of relapsed and refractory large B-cell lymphomas. However, this novel cellular treatment requires specialized infrastructure as well as complex patient preparation and carries a risk of relevant toxicities.19 Thus, candidates should be selected carefully according to individual factors that are currently under investigation. Tumor load before CAR T-cell therapy, either measured through the surrogate parameter LDH or using CT based on the Lugano criteria, has emerged as an independent predictor of outcome.5,20,21 However, besides these indirect and rather imprecise methods to quantify the extent of disease, metabolically active lymphoma burden can be accurately assessed by PET.

Our study shows that higher MTV is associated with inferior one-year PFS following CAR T-cell treatment. A similar observation has been made by Dean et al16 who additionally report significantly shorter OS for individuals with elevated metabolic lymphoma burden. Unlike these authors, we did not categorize patients by median of MTV but identified a markedly lower tumor volume as most predictive. Another small retrospective analysis suggested a similarly low MTV threshold for the specific context of CAR T-cell treatment.22 The association of higher metabolic tumor load with CRS development, reported by a group from China, could not however be observed in our study.23 Recently, tumor and systemic immune dysregulation have been described as forming a potential mechanistic link between elevated lymphoma burden and lower response rates.24 These findings indicate that effective reduction of tumor load before final product administration is particularly important in patients with higher MTV. Hence, the value of debulking should be investigated more thoroughly. While the recent study by Lutfi et al25 found no association between bridging therapy and survival benefits but instead revealed a link to prolonged cytopenia, others report favorable outcome, especially for individuals achieving deep remission before their planned treatment.26,27 We performed PET prior to lymphodepleting chemotherapy in all except one patient who was scanned on the day of product infusion. This is a time point at which cytoreductive treatment can still be considered if the individual has high MTV. Moreover, systemic and radiotherapy shortly after CAR T-cell administration may be feasible alternatives in heavily pretreated cases.

Multivariable analysis identified EN disease as the most predictive of conventional tumor and patient characteristics, with significantly shorter OS after therapy. Similar results have been published by Vercellino et al17 who examined the larger French cohort and proposed a model based on EN involvement as well as high metabolic lymphoma burden. Lesions in the skeleton, lung, liver, or skin may be more difficult to penetrate for CAR T-cells which reduces their efficacy at these tumor sites independent of MTV. Some studies on solid cancers have already shown that the migration of T-cells into extra-lymphatic tissue is limited.28 Alternative strategies are therefore required to enhance the penetration of CAR T-cells. Interestingly, our results also confirm findings reported by Cohen et al29 regarding the predictive value of SUVmax before product administration. However, the association between elevated CRP and shorter PFS observed within another French cohort could not be validated in our patients.30

A further important conclusion which emerged from the present analysis was that MTV can be measured reliably with an SUV threshold of 4.0 even by two observers using different software tools for tumor delineation. In a recently published study, the highest concordance has been achieved through cut-offs such as SUV 2.5 and 4.0, thereby also supporting the argument for fixed thresholds.31 Barrington et al32 showed a significantly higher success and lower failure rate for the cut-off we applied.

Several limitations of this work should be considered when interpreting the results: Our analysis is retrospective in nature and covers a limited follow-up time. Moreover, construct-specific subgroup analyses were not possible, since we included a rather low number of patients treated for various diagnoses, including transformed follicular lymphoma. It is already known that certain co-stimulatory domains promote rapid CAR T-cell expansion.33 Correlation between metabolic tumor burden and kinetic parameters is likely to improve our understanding of response heterogeneity. Future research should particularly examine whether the optimal MTV thresholds differ between constructs, as this PET-derived biomarker might then be a valuable parameter for product selection.

In conclusion, MTV could guide therapy by identification of those individuals likely to benefit from upfront debulking. It might also become useful for defining the optimal CAR T-cell infusion time point. Moreover, our study found a risk profile, including higher tumor load and SUVmax as well as EN disease, that should be considered in the management of candidates, since these factors are associated with significantly inferior patient outcome.


C-AV, PG, and CH conceptualized the study. SF performed statistical analysis. C-AV, PG, SF, and CH drafted the manuscript. All authors provided patients or collected data, participated in revising the manuscript, and approved its final version.


PG: Gilead Sciences, Novartis, travel support. J-MH: Incyte, Novartis, research funding; Gilead Pharmaceuticals, Novartis, travel support. NK: AstraZeneca, paid honoraria; Gilead Sciences, Inc., research funding; Celgene, Gilead Sciences, Inc., Janssen, travel support. PB: Bristol Myers Squibb, Celgene, Gilead Sciences, Janssen, Miltenyi Biotec, Novartis, paid honoraria. KR: ABX advanced biochemical compounds GmbH, ABX-CRO, Advanced Accelerator Applications, Bayer HealthCare, Janssen-Cilag, SIRTeX Medical Europe GmbH, consultation and lectureship honoraria. HCR: AbbVie, AstraZeneca, Merck, Roche, Vertex Pharmaceuticals, consultation and lectureship honoraria; AstraZeneca, Gilead Pharmaceuticals, research funding; CDL Therapeutics GmbH, co-founder. All the other authors have no conflicts of interest to disclose.


This study was supported by Novartis Pharma GmbH, Nuremberg, Germany.


1. Schuster SJ, Bishop MR, Tam CS, et al. Tisagenlecleucel in adult relapsed or refractory diffuse large B-cell lymphoma. N Engl J Med. 2019;380:45–56.
2. Abramson JS, Palomba ML, Gordon LI, et al. Lisocabtagene maraleucel for patients with relapsed or refractory large B-cell lymphomas (TRANSCEND NHL 001): a multicentre seamless design study. Lancet. 2020;396:839–852.
3. Locke FL, Miklos DB, Jacobson CA, et al. Axicabtagene ciloleucel as second-line therapy for large B-cell lymphoma. N Engl J Med. 2022;386:640–654.
4. Westin JR, Tam CS, Borchmann P, et al. Correlative analyses of patient and clinical characteristics associated with efficacy in tisagenlecleucel-treated relapsed/refractory diffuse large B-cell lymphoma patients in the Juliet trial. Blood. 2019;134:4103.
5. Locke FL, Rossi JM, Neelapu SS, et al. Tumor burden, inflammation, and product attributes determine outcomes of axicabtagene ciloleucel in large B-cell lymphoma. Blood Adv. 2020;4:4898–4911.
6. Raanani P, Shasha Y, Perry C, et al. Is CT scan still necessary for staging in Hodgkin and non-Hodgkin lymphoma patients in the PET/CT era? Ann Oncol. 2006;17:117–122.
7. Cheson BD, Fisher RI, Barrington SF, et al. Recommendations for initial evaluation, staging, and response assessment of Hodgkin and non-Hodgkin lymphoma: the Lugano classification. J Clin Oncol. 2014;32:3059–3068.
8. Dührsen U, Müller S, Hertenstein B, et al. Positron emission tomography-guided therapy of aggressive non-Hodgkin lymphomas (PETAL): a multicenter, randomized phase III trial. J Clin Oncol. 2018;36:2024–2034.
9. Kostakoglu L, Martelli M, Sehn LH, et al. End-of-treatment PET/CT predicts PFS and OS in DLBCL after first-line treatment: results from GOYA. Blood Adv. 2021;5:1283–1290.
10. Moog F, Bangerter M, Diederichs CG, et al. Lymphoma: role of whole-body 2-deoxy-2-[F-18]fluoro-D-glucose (FDG) PET in nodal staging. Radiology. 1997;203:795–800.
11. Kaddu-Mulindwa D, Altmann B, Held G, et al. FDG PET/CT to detect bone marrow involvement in the initial staging of patients with aggressive non-Hodgkin lymphoma: results from the prospective, multicenter PETAL and OPTIMAL>60 trials. Eur J Nucl Med Mol Imaging. 2021;48:3550–3559.
12. Mikhaeel NG, Smith D, Dunn JT, et al. Combination of baseline metabolic tumour volume and early response on PET/CT improves progression-free survival prediction in DLBCL. Eur J Nucl Med Mol Imaging. 2016;43:1209–1219.
13. Schmitz C, Hüttmann A, Müller SP, et al. Dynamic risk assessment based on positron emission tomography scanning in diffuse large B-cell lymphoma: post-hoc analysis from the PETAL trial. Eur J Cancer. 2020;124:25–36.
14. Cottereau AS, Meignan M, Nioche C, et al. Risk stratification in diffuse large B-cell lymphoma using lesion dissemination and metabolic tumor burden calculated from baseline PET/CT. Ann Oncol. 2021;32:404–411.
15. Kostakoglu L, Mattiello F, Martelli M, et al. Total metabolic tumor volume as a survival predictor for patients with diffuse large B-cell lymphoma in the GOYA study. Haematologica. 2022;107:1633–1642.
16. Dean EA, Mhaskar RS, Lu H, et al. High metabolic tumor volume is associated with decreased efficacy of axicabtagene ciloleucel in large B-cell lymphoma. Blood Adv. 2020;4:3268–3276.
17. Vercellino L, Di Blasi R, Kanoun S, et al. Predictive factors of early progression after CAR T-cell therapy in relapsed/refractory diffuse large B-cell lymphoma. Blood Adv. 2020;4:5607–5615.
18. Sesques P, Tordo J, Ferrant E, et al. Prognostic impact of 18F-FDG PET/CT in patients with aggressive B-cell lymphoma treated with anti-CD19 chimeric antigen receptor T cells. Clin Nucl Med. 2021;46:627–634.
19. Shimabukuro-Vornhagen A, Gödel P, Subklewe M, et al. Cytokine release syndrome. J Immunother Cancer. 2018;6:56.
20. Nastoupil LJ, Jain MD, Feng L, et al. Standard-of-care axicabtagene ciloleucel for relapsed or refractory large B-cell lymphoma: results from the US lymphoma CAR T consortium. J Clin Oncol. 2020;38:3119–3128.
21. Rabinovich E, Pradhan K, Sica RA, et al. Elevated LDH greater than 400 U/L portends poorer overall survival in diffuse large B-cell lymphoma patients treated with CD19 CAR-T cell therapy in a real world multi-ethnic cohort. Exp Hematol Oncol. 2021;10:55.
22. Iacoboni G, Simó M, Villacampa G, et al. Prognostic impact of total metabolic tumor volume in large B-cell lymphoma patients receiving CAR T-cell therapy. Ann Hematol. 2021;100:2303–2310.
23. Hong R, Tan Su Yin E, Wang L, et al. Tumor burden measured by 18F-FDG PET/CT in predicting efficacy and adverse effects of chimeric antigen receptor T-cell therapy in non-Hodgkin lymphoma. Front Oncol. 2021;11:713577.
24. Jain MD, Zhao H, Wang X, et al. Tumor interferon signaling and suppressive myeloid cells are associated with CAR T-cell failure in large B-cell lymphoma. Blood. 2021;137:2621–2633.
25. Lutfi F, Holtzman NG, Kansagra AJ, et al. The impact of bridging therapy prior to CD19-directed chimeric antigen receptor T-cell therapy in patients with large B-cell lymphoma. Br J Haematol. 2021;195:405–412.
26. Nader A, Lee H, Sellmyer M, et al. Association of PET/CT response assessment prior to CAR T-cell infusion with outcomes after CAR T-cell therapy in aggressive B-cell lymphomas. J Clin Oncol. 2021;39:e19568.
27. Khurana A, Al Saleh AS, Gandhi S, et al. Response to bridging therapy (BT) before CAR-T cell infusion predicts outcomes for relapsed/refractory (R/R) aggressive B-cell non-Hodgkin lymphoma (NHL). Blood. 2020;136:30.
28. Salmon H, Franciszkiewicz K, Damotte D, et al. Matrix architecture defines the preferential localization and migration of T cells into the stroma of human lung tumors. J Clin Invest. 2012;122:899–910.
29. Cohen D, Luttwak E, Beyar-Katz O, et al. [18F]FDG PET-CT in patients with DLBCL treated with CAR-T cell therapy: a practical approach of reporting pre- and post-treatment studies. Eur J Nucl Med Mol Imaging. 2022;49:953–962.
30. Sesques P, Ferrant E, Safar V, et al. Commercial anti-CD19 CAR T cell therapy for patients with relapsed/refractory aggressive B cell lymphoma in a European center. Am J Hematol. 2020;95:1324–1333.
31. Tutino F, Puccini G, Linguanti F, et al. Baseline metabolic tumor volume calculation using different SUV thresholding methods in Hodgkin lymphoma patients: interobserver agreement and reproducibility across software platforms. Nucl Med Commun. 2021;42:284–291.
32. Barrington SF, Zwezerijnen BGJC, de Vet HCW, et al. Automated segmentation of baseline metabolic total tumor burden in diffuse large B-cell lymphoma: which method is most successful? A study on behalf of the PETRA consortium. J Nucl Med. 2021;62:332–337.
33. Shuford WW, Klussman K, Tritchler DD, et al. 4-1BB costimulatory signals preferentially induce CD8+ T cell proliferation and lead to the amplification in vivo of cytotoxic T cell responses. J Exp Med. 1997;186:47–55.
Copyright © 2023 the Author(s). Published by Wolters Kluwer Health, Inc. on behalf of the European Hematology Association.