Inflammatory reactions mimic residual or recurrent lymphoma on [¹⁸F]FDG-PET/CT after CD19-directed CAR T-cell therapy

TO THE EDITOR:

Chimeric antigen receptor T-cell (CAR-T) therapy, an immunotherapy in which T cells are genetically modified to target tumor surface antigen(s), transformed the treatment landscape in hemato-oncology, with durable responses in ∼50% of the patients with relapsed or refractory large B-cell lymphoma (LBCL) and mantle cell lymphoma (MCL).^1-4 

2-Deoxy-2-[¹⁸F]fluoro-D-glucose–positron emission tomography and computed tomography ([¹⁸F]FDG-PET/CT) imaging is routinely used for staging and response evaluation in patients with aggressive lymphoma. The Lugano classification, including the Deauville score (a 5-point scale to grade the intensity of [¹⁸F]FDG uptake), is used for imaging assessment in these patients.^5-7 

It is known that within the field of immunotherapy, adequate determination of treatment outcome can be challenging because of abnormalities on [¹⁸F]FDG-PET/CT caused by active inflammation, resulting in phenomena such as pseudoprogression and sarcoid-like reactions (SLRs).^8-10 This study reports on 2 types of inflammatory reactions (IRs) mimicking residual/recurrent lymphoma on [¹⁸F]FDG-PET/CT after CD19-directed CAR-T therapy.

Based on clinical, radiological, and pathological findings among patients treated with CD19-directed CAR-Ts, 2 IRs, an SLR and a histiocytic reaction (HR), were identified by experts of the Dutch CAR-T tumorboard, a platform for nationwide referral, eligibility assessment, and data collection for CAR-T therapy for adult patients with LBCL. Based on the literature, SLR was defined as mostly symmetric bilateral hilar and mediastinal lymphadenopathy and/or lymphadenopathy in other regions with increased [¹⁸F]FDG uptake and, in a biopsy specimen, noncaseating epithelioid cell granulomatous inflammation without evidence of lymphoma.^11-13 HR was defined as increased [¹⁸F]FDG uptake at the original tumor localization early (ie, ±1 month) after CAR-T infusion, with sheets of (foamy) histiocytes and necrotic lymphoma cells in the biopsy specimen.

We included adult patients with LBCL and MCL treated with CD19-directed CAR-Ts (axicabtagene ciloleucel [axi-cel], tisagenlecleucel, lisocabtagene maraleucel, and brexucabtagene autoleucel [brexu-cel]) from November 2016 to June 2022 in The Netherlands, either in a clinical trial (NCT02348216, NCT03391466, NCT02601313, NCT02445248, NCT0357089, NCT03484702, and NCT03575351) or as a standard of care, with [¹⁸F]FDG-PET/CT abnormalities suspected for an SLR or HR at routine response evaluation(s) after CAR-T therapy. [¹⁸F]FDG-PET/CT abnormalities were defined as increased [¹⁸F]FDG uptake without evidence of residual/recurrent lymphoma, either biopsy proven or after a “wait-and-see” strategy, when biopsy was unfeasible. Only patients with ≥3 months of follow-up, including ≥1 [¹⁸F]FDG-PET/CT scan(s), after the onset of [¹⁸F]FDG-PET/CT abnormalities suspected of a SLR or HR were included. An incidence could be estimated using the nationwide registry of patients treated with commercially available CAR-Ts and the total number of patients treated in the aforementioned clinical trials in The Netherlands. Ethical approval was obtained. Data were retrospectively collected by reviewing medical records with a cutoff date of 31 January 2023. Care was provided according to study protocols or institutional guidelines. More specifically, [¹⁸F]FDG-PET/CTs were performed at ∼1, 3, 6, and 12 months after CAR-T infusion, and additional scans were performed according to study protocol or institutional guidelines. The study was performed in accordance with the Declaration of Helsinki.

Of the 236 patients treated with CD19-directed CAR-Ts (axi-cel: n = 200, brexu-cel: n = 12, lisocabtagene maraleucel: n = 11, and tisagenlecleucel: n = 13), 18 (7.6%) were identified with abnormalities suspected for an SLR or HR on routine [¹⁸F]FDG-PET/CT (median age, 60.5 years; range, 35-73 years). Most patients were diagnosed with diffuse LBCL, followed by transformed follicular lymphoma and MCL; 50% of these patients (n = 9) received bridging therapy before CAR-T therapy (radiotherapy: n = 5, and systemic therapy: n = 4). The best overall response rate was 100% (complete response rate: 89% [n = 16]). All patients were asymptomatic at onset of the IR. No clinical signs of disease progression, such as B symptoms, clinical lymphadenopathy, or >2× upper limit of normal lactate dehydrogenase levels, were present. Median follow-up from the onset of the IR was 657.5 days (range, 133-1558 days). Two patients experienced both an SLR and HR, resulting in a total of 20 IRs observed. Eleven patients were suspected to have an SLR (axi-cel: n = 10, and brexu-cel: n = 1), and 9 patients were suspected to have an HR (axi-cel: n = 9; Table 1). No additional treatment was given for the IRs. Estimated incidences of SLRs and HRs in patients treated with CAR-Ts were 4.7% and 3.8%, respectively, whereas a total of 116 patients (49.2%; 116 of 236) had progressive disease after CAR-T therapy.

In 11 of 18 patients, an SLR was suspected using the aforementioned definition. A representative case is shown in Figure 1A-F. Of these 11 patients, biopsy was performed in 5 and showed noncaseating epithelioid cell granulomatous inflammation without evidence of lymphoma or pathogens such as Mycobacterium tuberculosis, fungus, or yeast (Figure 1G). The median time to onset was 116 days (range, −259 to 548 days) from CAR-T infusion. In retrospect, SLR was already existent before CAR-T infusion in 2 patients but became more evident after CAR-T therapy. The SLR resolved in 8 of 11 patients, with a median time to resolution of 190 days (range, 13-585 days). In 3 patients, the SLR was still ongoing, with a median duration of 989 days (range, 402-1321 days).

Using the aforementioned definition of HR, 9 of 18 patients were suspected of having this IR. A representative case is shown in Figure 1H-M. Histopathological examination, performed in 3 of 9 patients, showed sheets of (foamy) histiocytic cells with necrotic lymphoma cells and without granulomatous processes (Figure 1N). HRs occurred early after CAR-T infusion (median time to onset: 28 days; range, 21-31 days). In 3 of 9 patients, the HR resolved with a median time to resolution of 141 days (range, 102-427 days). In the other 6 patients, the HR was still ongoing, with a median duration of 285.5 days (range, 85-1035 days). All HRs were located in the irradiated field if patients received radiotherapy as bridging strategy (n = 5).

We describe cases of SLR and HR occurring at different time points after CAR-T therapy, illustrating that after CAR-T therapy also, similar to other immunotherapies, IRs can mimic residual/recurrent disease on [¹⁸F]FDG-PET/CT.¹¹ 

Cases of SLRs are extensively described in literature in non-Hodgkin lymphoma at diagnosis, after systemic treatment or stem cell transplantation, and in solid cancers after treatment with immune checkpoint inhibitors.^12,14-16 However, the exact pathophysiological mechanism is still unknown. The association of an interferon-γ–producing T helper 17 (Th17) T-cell subset with the occurrence of sarcoidosis after immune checkpoint inhibitors led to the hypothesis that (re)activation of CD4 Th1 cells could lead to macrophage activation, aggregation, and, thereby, induction of granuloma formation.^11,17 Recently, this hypothesis was supported by the presence of interferon-γ–producing Th17 T-cell subset cells in a case of SLR after anti–B-cell maturation antigen CAR-T therapy for multiple myeloma.¹⁸ However, the reaction of the host to released tumor antigens after treatment is also proposed as an underlying mechanism.¹¹ 

Concerning HR, it has been described that histiocytic infiltration can cause increased [¹⁸F]FDG uptake.¹⁰ The accumulation of necrotic lymphoma cells probably triggers this infiltration. Furthermore, Kuhnl et al have shown that a subgroup of patients who received radiotherapy as bridging strategy with high [¹⁸F]FDG uptake (Deauville score of 4) in the irradiated field on [¹⁸F]FDG-PET/CT early after CAR-T therapy only had a 10% risk of progressive disease at 6 months compared with 46% in the other patients with Deauville score–4 [¹⁸F]FDG uptake.¹⁹ Similarly, we also found a low relapse rate in patients with increased [¹⁸F]FDG uptake in the irradiated field, after receiving radiotherapy as bridging strategy (n = 5).

Moreover, pseudoprogression, a well-known IR after immunotherapy, has been described early after CAR-T infusion (4-21 days) and is associated with a favorable outcome. In contrast to HR, pseudoprogression is characterized by initial tumor enlargement due to immune cell infiltration before decreasing tumor burden.^20-24 

In 2016, an adjusted version of the Lugano classification to account for pseudoprogression and tumor flare in the context of immunotherapies was published: the Lymphoma Response to Immunomodulatory Therapy Criteria.²⁵ Further adjustment of the Lugano criteria, taking into account SLRs and HRs as well, could enhance the identification of IRs after CAR-T.

A limitation of our study is that cases were identified using expert opinion instead of extensive medical record/imaging review of all patients treated with CAR-T therapy in The Netherlands; therefore, the number of IRs could have been underestimated.

In conclusion, our findings highlight that SLRs and HRs can occur after CAR-T therapy and should be considered as potential causes of [¹⁸F]FDG uptake. This emphasizes the importance of histological proof of suspected residual/recurrent lymphoma, when feasible, to prevent potential misclassification of response and unnecessary treatment. This is particularly important, given the growing indications for CAR-T therapy use, probably resulting in a higher incidence of [¹⁸F]FDG uptake because of IRs.

Acknowledgments: The authors thank all patients and their informal caregivers, the CAR-T treatment teams with special attention to the clinical trial offices, the Dutch CAR-T tumorboard, and all the referring centers.

E.R.A.P. is supported by the Innovative Medicines Initiative 2 Joint Undertaking (grant agreement no. 116026), T2EVOLVE. This joint undertaking receives support from the European Union’s Horizon 2020 Research and Innovation program and the European Federation of Pharmaceutical Industries and Associations. J.W.d.B. is supported by a partnership of UMCG-Siemens for building future of health (PUSH 2020).

Contribution: E.R.A.P. and J.W.d.B. designed and performed research, analyzed data, and wrote the manuscript; A.K., A.M.S., J.A.v.D., P.G.N.J.M., M.J., M.W.M.v.d.P., M.T.K., J.A.A., A.D., L.K., S.v.D., and J.S.P.V. contributed to data collection and edited and approved the final manuscript; A.G.H.N. contributed to data collection and interpretation and edited and approved the final manuscript; and M.J.K. and T.v.M. designed the research, contributed to data collection and interpretation, and edited and approved the final manuscript.

Conflict-of-interest disclosure: P.G.N.J.M. received research funding from AstraZeneca and served on the advisory board of GlaxoSmithKline. M.J. received honoraria from Kite/Gilead and Bristol Myers Squibb/Celgene; serves in an advisory role for Janssen; and received research funding from Novartis. M.W.M.v.d.P. received honoraria from Kite/Gilead and Takeda. A.D. received research funding from Takeda. A.G.H.N. discloses conflicts of interest all outside the submitted work with IBA, Philips, Mirada, RaySearch, Siemens, Elekta, Leonie, and Genentech. M.J.K. received honoraria from Bristol Myers Squibb, Celgene, Kite/Gilead, Miltenyi Biotech, Adicet Bio, Novartis, and Roche; received research funding from Kite/Gilead, Roche, Takeda, and Celgene; has an advisory role in Bristol Myers Squibb, Celgene, Kite/Gilead, Miltenyi Biotech, Adicet Bio, Novartis, and Roche; and received travel support from Kite/Gilead, Miltenyi Biotech, Novartis, and Roche. T.v.M. received research funding from Kite/Gilead, Celgene/Bristol Myers Squibb, Genentech, and Siemens, and has served on advisory boards of Kite/Gilead, Celgene/Bristol Myers Squibb, Jansen, and Lilly. The remaining authors declare no competing financial interests.

Correspondence: Tom van Meerten, Department of Hematology, University Medical Center Groningen, PO Box 30.001, Hanzeplein 1, 9700 RB Groningen, The Netherlands; e-mail: t.van.meerten@umcg.nl.