Navigating Second-Line CAR-T Candidacy and Sequencing in DLBCL

Chimeric antigen receptor (CAR) T-cell therapy has moved from a later-line rescue option to a guideline-preferred second-line strategy for relapsed/refractory diffuse large B-cell lymphoma (DLBCL). Yet candidacy remains anything but straightforward: some patients with substantial disease burden are still appropriate candidates, while others carry comorbidities or face social barriers that make CAR T-cell therapy a poor fit. Translating that nuance into consistent referral practice and choosing between 2 effective but seldom head-to-head-compared products depends heavily on how individual physicians and multidisciplinary teams work through the calculus together.

In a recent Satellite Session focused on the University of Pennsylvania Health System in Philadelphia, Pennsylvania, a panel of community and academic hematologists/oncologists and advanced practice providers worked through second-line decision-making for DLBCL. The session was moderated by Emily Tomasulo, DO, FACP, assistant professor of medicine, hematology and oncology, at Penn Medicine, University of Pennsylvania Health System, and program lead for the Cellular and Transplant Program at Pennsylvania Hospital.

Panelists included Moshe Chasky, MD, FACP, hematologist/medical oncologist at Alliance Cancer Specialists in Bensalem, Pennsylvania; Douglas Eric Guggenheim, MD, hematologist/medical oncologist at Abramson Cancer Center in Cherry Hill, New Jersey; Daniel Helbig, MD, hematologist/medical oncologist at Abramson Cancer Center; Sparsha Kukunoor, MD, hematologist/medical oncologist at Penn Hematology/Oncology Bucks County in Yardley, Pennsylvania; Emilie Potter, CRNP, nurse practitioner at Pennsylvania Hospital in Philadelphia; Michael Wysota, MD, assistant professor of malignant hematology–lymphoid at Thomas Jefferson University in Philadelphia; John Youssef, MD, hematologist/medical oncologist at Penn Hematology/Oncology Bucks County; Howard Zipin, MD, hematologist/medical oncologist at Alliance Cancer Specialists in Sellersville, Pennsylvania; and Stephen E. Zrada, MD, hematologist/medical oncologist at Penn Hematology/Oncology Sewell in Sewell, New Jersey.

The panel worked through candidacy criteria, reviewed the pivotal second-line CAR-T trials, discussed holding and bridging therapy, and applied all of their insights to a real-world case.

CAR-T Candidacy: Weighing Function Over Age

Tomasulo opened by asking what factors are considered when referring a patient with lymphoma relapsed within 12 months of frontline therapy for CAR-T evaluation. Youssef described weighing early interventions that facilitate immediate relief against longer-term treatments offering durable benefit; he leans toward early referral so patients understand their options. Zrada emphasized social factors such as caregiver availability and a patient’s willingness to travel into the city as important considerations.

Wysota noted that certain comorbidities, including significant chronic infection and pericardial disease, may be too significant to safely tolerate CAR-T–associated toxicity. He added that, at Jefferson, patients with preexisting neurocognitive impairment have shown inferior outcomes following T-cell–directed therapies such as bispecifics and CAR-T. Zrada added that it’s important to distinguish whether impaired performance status stems from the disease itself, which may improve with effective treatment, or from underlying organ dysfunction. Tomasulo referred to this conundrum as one of the hardest parts of oncology practice: determining whether treatment will reverse disease-related functional decline or accelerate morbidity.

Zipin noted that disease trajectory also matters: an asymptomatic patient with incidental progression on a routine scan is a different case than one presenting with rapid, symptomatic disease, where early establishment with a cellular therapy program is especially valuable.

Kukunoor explained that performance status is a more useful consideration than chronological age, with Helbig adding that he has referred patients as old as 88 years for CAR-T evaluation with successful outcomes.

“There’s definitely a difference between physiologic age and your chronologic age,” Tomasulo said, noting the program is currently evaluating patients in their 90s.

Identifying Patients Earlier in the Disease Course

The panel agreed on early referral as a near-universal default. Kukunoor and Zipin both said they refer as soon as a patient relapses, and Helbig estimated his group refers 75% to 100% of relapsed patients, attributable in part to the region’s geographic access to multiple cellular therapy programs.

Discussion also turned to whether an incomplete response to frontline chemotherapy should prompt earlier referral. Wysota described a patient with double-hit lymphoma whose 17-cm mass shrank clinically over several cycles of chemotherapy, but whose interim PET scan still showed a standardized uptake value (SUV) in the 20s. “This woman is not going to get to a [complete response],” he thought, and referred her rather than continuing the toxic chemotherapy she was unlikely to benefit from.

“Not all [partial responses] are the same,” Tomasulo summarized. A lesion trending from an SUV of 20 toward 3 is likely headed toward complete response, while high-risk molecular features such as double-hit status may shorten response duration and support earlier CAR-T referral as consolidation.

The group also discussed minimal residual disease (MRD) testing platforms, including clonoSEQ and Signatera, as potential tools for risk-stratifying patients. Tomasulo described a patient found to be MRD-positive at end of treatment despite no formal end-of-treatment scan, which prompted earlier CAR-T evaluation. Wysota said he would argue for referring “anyone with MRD-positive [disease] at the end of treatment,” while Chasky and Zrada cautioned that roughly a third of patients who don’t clear MRD early still go on to achieve clearance, and that standardized use of MRD in DLBCL remains an evolving practice.

Reviewing the ZUMA-7 and TRANSFORM Data

Tomasulo walked through the phase 3 ZUMA-7 trial (NCT03391466), which randomly assigned patients with primary refractory or early-relapsing (within 12 months) LBCL to axicabtagene ciloleucel (axi-cel; Yescarta) or investigator’s-choice platinum-based chemoimmunotherapy.¹ Notably, 74% of enrolled patients had primary refractory disease.

The median event-free survival (EFS) was 10.8 months with axi-cel vs 2.3 months with standard of care, and the median overall survival (OS) was not reached with axi-cel vs 31.1 months with standard of care regardless of transplant eligibility.

Tomasulo noted any-grade cytokine release syndrome (CRS) occurred in 92% of axi-cel-treated patients, but grade 3 or higher CRS was uncommon. Any-grade immune effector cell-associated neurotoxicity syndrome (ICANS) occurred in 61%, with grade 3 or higher instances in 21%. She pointed out that neurologic events were also reported in the standard-of-care arm, suggesting ICANS may be over-attributed to CAR-T in this population. Zipin expected grade 3 or higher CRS rates to be higher than observed, and Helbig suggested trial patients were somewhat selected, since reaching the trial itself implied some baseline disease control, and expected higher CRS rates in an older, higher-disease-burden real-world population.

Tomasulo then reviewed the phase 3 TRANSFORM trial (NCT03575351)of lisocabtagene maraleucel (liso-cel; Breyanzi), which enrolled a similar population but permitted secondary CNS lymphoma and saw 60% of standard-of-care patients cross over to CAR-T after treatment failure.² With a median follow-up of 17.5 months, the median EFS was not reached with liso-cel vs 2.4 months with standard of care. At the 3-year follow-up, the EFS was 29.5 months vs 2.4 months (HR, 0.375; 95% CI, 0.259-0.542).³ Neutropenia rates were slightly higher with liso-cel than typically seen with axi-cel, which Tomasulo characterized as a reasonable expectation for second-line CAR-T generally. CRS rates trended lower with liso-cel.

When asked whether the biology differed between patients who were primary refractory or early-relapsed, Tomasulo said roughly 75% of patients with primary refractory disease across both trials are largely driving the separation between the CAR-T and standard-of-care curves. “They’re the hardest to treat,” she said. “They’re the ones that give me goosebumps even when it’s 100 degrees.”

Choosing a Bridging Strategy Ahead of CAR-T

Tomasulo distinguished holding therapy, given before apheresis to control disease without compromising T-cell collection, from bridging therapy, given after apheresis while CAR-T cells are manufactured, to avoid clinical deterioration. Her practical rule: avoid bendamustine, since data suggest patients who receive it within 12 months of CAR-T respond more poorly, likely due to effects on T-cell integrity and function.

Helbig noted that bridging in this region is less often the community doctor’s independent call, since proximity to Penn allows closer real-time collaboration with the cellular therapy team on regimen choice. Guggenheim described retrospective data suggesting bendamustine as bridging therapy, as opposed to holding therapy, does not appear to affect overall or complete response rates, though he noted comparable data in DLBCL for its use as holding therapy remain limited.

Wysota raised the question of bispecific antibodies in the interim period, saying his group at Jefferson has debated this extensively without reaching consensus. Tomasulo’s group uses bispecifics for both holding and bridging, noting the approach requires re-biopsy to confirm CD20 expression, since bispecific antibodies won’t work if CD20 expression on immunohistochemistry falls below 10%. She cited a median time to response with bispecific monotherapy of about 1.4 months, which she called too long for a patient experiencing rapidly progressing disease without pairing it with cytoreductive chemotherapy such as gemcitabine-oxaliplatin (GEMOX).

Assessing CAR-T Candidacy for an Early Relapse

Tomasulo presented a case of a 52-year-old man with stage IV, de novo germinal center B-cell (GCB) subtype DLBCL, diagnosed 9 months earlier with a bulky mediastinal mass (8.2 cm) and bone marrow involvement (International Prognostic Index score of 2). He was CD20-positive, expressed BCL2, BCL6, and MYC, and had TP53-mutated disease, without a MYC rearrangement.

Previously treated with 6 cycles of polatuzumab vedotin (Polivy), rituximab, cyclophosphamide, doxorubicin, and prednisone (Pola-R-CHP), the patient achieved a complete response with resolved grade 2 peripheral neuropathy. He then relapsed with dyspnea, bilateral axillary and retroperitoneal adenopathy, a 4.2-cm mediastinal mass (Deauville score 4), suspected marrow involvement, no CNS involvement, and good performance status.

“It’s very CAR-eligible,” Helbig said. He cited the relapse within a year, the patient’s age, and good performance status. Potter outlined the additional baseline workup CAR-T requires beyond standard pre-chemotherapy evaluation: echocardiogram, pulmonary function testing, infectious disease testing (including hepatitis B and CMV), and dental clearance, the last of which, Helbig noted, can hold up treatment entirely if not obtained. Tomasulo estimated roughly 6 weeks as a feasible timeframe, encompassing insurance authorization, workup, and the 2-to-3-week manufacturing window, though Zipin observed that timeline can realistically stretch to 8 weeks, suggesting this patient likely needs bridging therapy.

Kukunoor and Zipin both raised whether a repeat biopsy is needed. Tomasulo said it’s ideal when feasible, since it can both confirm CD20 status to inform whether a bispecific bridge would work and reveal whether the disease phenotype has changed.

On product selection between axi-cel and liso-cel, Chasky defers entirely to the treating academic center given the absence of head-to-head data: “I really lean on the academic center to make that formal decision.” Tomasulo acknowledged tertiary centers face the same uncertainty, though she has favored axi-cel in young, fit patients with aggressive disease because of its roughly 1-week shorter manufacturing time.

Additionally, Zrada stated that he would still favor CAR-T for a patient with TP53-mutated disease, though Youssef noted the calculus differs meaningfully between a 13-month and an 8-year relapse. Zrada cited retrospective data suggesting transplant may outperform CAR-T long-term in chemosensitive, late-relapsing patients. Tomasulo said this nuance reflects disease biology more than anything captured cleanly by National Comprehensive Cancer Network (NCCN) category.

Barriers to Referral and the Choice of Holding or Bridging Therapy for This Patient

When asked what would make referring this patient for CAR-T challenging in the second line, Kukunoor pointed to logistics such as knowing whom to call and the patients’ willingness to travel into the city. Guggenheim described a small cohort of patients who refuse to come into Philadelphia under any circumstances, sometimes requiring workarounds, and noted that New Jersey Medicaid coverage can restrict where patients are treated, directing some to MD Anderson Cancer Center at Cooper rather than Penn. Tomasulo noted the recent reduction of the required post-infusion monitoring period, during which a patient must remain within an hour of the treatment center, from 30 days to 14 days as an improvement for patients reluctant to relocate temporarily.

Zrada described the workflow for holding or bridging therapy administration as collaborative, proposing a regimen = to the tertiary center for input before proceeding locally. Tomasulo confirmed that radiation is frequently used this way, with local radiation oncologists coordinating with Penn’s lymphoma-focused radiation team so patients can receive bridging radiotherapy close to home. Zipin reiterated that his group also routes first-line treatment for double- and triple-hit patients to Penn specifically because of their elevated relapse risk.

References

1. Westin JR, Oluwole OO, Kersten MJ, et al. Survival with axicabtagene ciloleucel in large B-cell lymphoma. N Engl J Med. 2023;389(2):148-157. doi:10.1056/NEJMoa2301665
2. Abramson JS, Solomon SR, Arnason J, et al. Lisocabtagene maraleucel as second-line therapy for large B-cell lymphoma: primary analysis of the phase 3 TRANSFORM study. Blood. 2023;141(14):1675-1687. doi:10.1182/blood.2022018730
3. Kamdar M, Solomon SR, Arnason J, et al. Lisocabtagene maraleucel versus standard of care for second-line relapsed/refractory large B-cell lymphoma: 3-year follow-up from the randomized, phase III TRANSFORM study. J Clin Oncol. 2025;43(24):2671-2678. doi:10.1200/JCO-25-00399