INTRODUCTION — Most patients with diffuse large B cell lymphoma (DLBCL) are cured with initial treatment using rituximab plus cyclophosphamide, doxorubicin, vincristine, and prednisone (R-CHOP) [1]. However, 10 percent of patients do not respond adequately to initial treatment (primary refractory disease) and nearly one-third of patients will later relapse after achieving a complete response.
This topic will discuss evaluation and management of medically-fit patients with suspected first relapse of DLBCL or primary refractory DLBCL.
Management of patients who have second or later relapse or who are not medically-fit is discussed separately. (See "Diffuse large B cell lymphoma (DLBCL): Second or later relapse or patients who are medically-unfit".)
EVALUATION AND DIAGNOSIS — Refractory DLBCL should be suspected in a patient who does not achieve a complete response (CR) to initial treatment. Relapsed DLBCL should be suspected in a patient with progressive lymphadenopathy, organomegaly, organ dysfunction, unexplained cytopenias, or systemic symptoms after achieving a CR or partial response (PR).
Initial evaluation and diagnosis of DLBCL are described separately. (See "Epidemiology, clinical manifestations, pathologic features, and diagnosis of diffuse large B cell lymphoma", section on 'Diagnosis'.)
Refractory DLBCL — Refractory DLBCL refers to disease that did not achieve a CR with initial therapy, based on positron emission tomography (PET)/computed tomography (CT) (table 1) and according to Lugano criteria (table 2).
●Setting and diagnosis – Refractory DLBCL may be suspected by history (eg, persistent B symptoms) or physical examination (eg, an incomplete nodal response or a new site of disease), but the diagnosis requires PET/CT confirmation, according to Lugano criteria (table 2). (See "Pretreatment evaluation and staging of non-Hodgkin lymphomas", section on 'Positron emission tomography (PET)'.)
●Confirmatory biopsy – A repeat biopsy is generally performed to confirm the diagnosis and to exclude a reactive, post-treatment inflammatory response.
However, it may not be necessary to repeat a biopsy if PET/CT reveals clear progression of size and metabolic activity of a disease site. If a biopsy is not performed, the morphology, immunophenotype, and cytogenetic/molecular features of the initial biopsy specimen should be reviewed to ensure that DLBCL was correctly diagnosed. (See "Epidemiology, clinical manifestations, pathologic features, and diagnosis of diffuse large B cell lymphoma", section on 'Diagnosis'.)
Relapsed DLBCL — Relapsed DLBCL refers to disease that recurs after CR was documented by PET/CT.
●Clinical settings – Relapse is usually suspected by history or physical examination; only rarely is relapse identified solely on the basis of routine follow-up imaging [2,3].
Most relapses occur in the first two years after completing treatment, but up to one-fifth occur more than five years after treatment [4,5]. Most relapses are symptomatic and are associated with clinical findings, such as systemic B symptoms (ie, fever, night sweats, weight loss); enlargement of lymph nodes, liver, or spleen; development of an extranodal mass; and/or unexplained cytopenias. Clinical manifestations of DLBCL are discussed separately. (See "Epidemiology, clinical manifestations, pathologic features, and diagnosis of diffuse large B cell lymphoma", section on 'Clinical presentation'.)
●Biopsy is required to diagnose relapse – A biopsy is required to confirm relapsed DLBCL, to exclude other conditions (eg, other types of lymphoma, carcinoma, sarcoidosis, tuberculosis, fungal infection) and acquisition of new mutations.
Selection of a biopsy site, the importance of an adequate specimen, and analysis of the specimen are described separately. (See "Clinical presentation and initial evaluation of non-Hodgkin lymphoma", section on 'Lymph node and tissue biopsy' and "Epidemiology, clinical manifestations, pathologic features, and diagnosis of diffuse large B cell lymphoma", section on 'Diagnosis'.)
PRETREATMENT EVALUATION — Pretreatment evaluation includes assessment of medical fitness, restaging, and prognosis.
Clinical/laboratory/pathology
●History and physical examination – The presence of B symptoms (ie, fever, sweats, weight loss) and lymph node and/or organ involvement should be documented by history and physical examination. (See "Initial treatment of advanced stage diffuse large B cell lymphoma", section on 'Pretreatment evaluation'.)
●Neurologic abnormalities – Patients with neurologic symptoms or abnormalities on neurologic examination should undergo neuroimaging and lumbar puncture (LP), as described separately. (See "Secondary central nervous system lymphoma: Clinical features and diagnosis".)
Lymphomatous involvement of the central nervous system (CNS) influences the choice of salvage chemotherapy, as described below. (See 'CNS involvement' below.)
●Laboratory
•Complete blood count (CBC) with differential
•Serum electrolytes, glucose, blood urea nitrogen (BUN) and creatinine, calcium, uric acid, and liver function tests, including lactate dehydrogenase (LDH)
•HIV and hepatitis B testing
●Clinical tests
•Cardiac – Echocardiogram or radionuclide ventriculogram (RVG) should be performed, if clinically indicated. If the patient will undergo hematopoietic cell transplantation (HCT), assessment of cardiac function is required prior to transplantation.
•Pregnancy testing, if appropriate.
Medical fitness — We assess and classify fitness for treatment based on functional status and comorbid conditions. Age, per se, is not a measure of fitness.
Age is not a barrier to treatment of relapsed or refractory (r/r) DLBCL, but some institutions limit autologous HCT to patients ≤75 years.
Assessment — Assessment of fitness should include:
●Performance status – Eastern Cooperative Oncology Group (ECOG) performance status (PS) (table 3).
●Comorbidity score – Charlson comorbidity index (CCI) or the HCT-specific comorbidity index (HCT-CI) scoring systems (table 4) are helpful for assessing fitness in some patients.
Geriatric assessment may be useful for judging medical fitness of some patients in this setting [6]. (See "Acute myeloid leukemia: Management of medically-unfit adults", section on 'Pretreatment evaluation'.)
Fitness categories — There are no clear distinctions between categories of medical fitness, and some measures of PS or comorbid illnesses can apply to different fitness categories. In assigning a fitness category, we seek to protect frail patients from treatment that they are unlikely to tolerate, while not depriving others from the opportunity to achieve a meaningful response and prolonged survival.
We classify fitness as follows:
●Medically-fit patients are considered able to tolerate intensive treatments (eg, chimeric antigen receptor [CAR]-T cell therapy, intensive salvage chemotherapy, and autologous HCT) based on both of the following:
•ECOG: 0 to 2 (table 3)
•CCI: 0 to 2 (table 4)
●Medically-unfit, but not frail patients include a broad range of physical and functional status. Some patients have only modest, recent, or transient impairment of functional status, while others have substantial comorbid illnesses, cognitive impairment, or other conditions that can affect their tolerance for therapy. Patients who are medically-unfit, but not frail may be able to tolerate CAR-T cell therapy but are generally not candidates for autologous HCT.
We judge patients to be medically-unfit, but not frail if either of the following applies:
•ECOG: 3 (table 3)
•CCI: 3 (table 4)
●Frail patients are those whose debility or comorbid conditions would not permit treatment aimed at modifying the disease course, as reflected by both the following:
•ECOG: ≥3
•CCI: ≥3
Management of r/r DLBCL in patients who are medically-unfit or frail is discussed separately. (See "Diffuse large B cell lymphoma (DLBCL): Second or later relapse or patients who are medically-unfit".)
Suitability for transplantation — All patients, except those who are frail, should be evaluated for eligibility by transplant specialists soon after the diagnosis of r/r DLBCL.
No specific age excludes autologous HCT, but many institutions limit transplantation to patients ≤75 years who have adequate heart, kidney, and liver function and performance status, as discussed separately. (See "Determining eligibility for autologous hematopoietic cell transplantation".)
Restaging — Restaging is based on clinical evaluation, laboratory studies, bone marrow examination, and positron emission tomography/computed tomography (PET/CT), according to the Lugano criteria (table 5). Disease stage at relapse should be designated by subscript R (R).
●Imaging – PET/CT (table 5) should be scored according to the Deauville five point scale (table 1).
●Bone marrow examination is not required for pretreatment staging of r/r DLBCL, because PET is a good predictor for marrow involvement. If a bone marrow examination is performed during pretreatment staging, it should be examined by microscopy, cytogenetics (using fluorescence in situ hybridization [FISH] or Giemsa-stained chromosomes), and molecular studies, as described separately. (See "Epidemiology, clinical manifestations, pathologic features, and diagnosis of diffuse large B cell lymphoma", section on 'Diagnosis'.)
Note that a bone marrow examination may again be required prior to HCT to assess for the presence of myelodysplasia, as discussed below. (See 'Autologous HCT' below.)
Prognosis — We apply the International Prognostic Index (IPI) (table 6) at the time of diagnosis of r/r DLBCL. The IPI has been validated as a prognostic tool in this setting [7,8].
In the CORAL trial, event-free survival (EFS) at three years was lower in patients with age-adjusted IPI score 2 to 3 compared to patients with IPI score 0 to 1 (18 versus 40 percent, respectively) [9]. The prognostic value of the IPI score for relapsed disease was also validated in the Parma trial [10].
Other factors that have been associated with outcomes in r/r DLBCL include:
●Duration of remission – Longer duration of initial remission (ie, longer time to relapse) is generally associated with a more favorable response to salvage chemotherapy.
In the Parma study, which was conducted prior to the rituximab era, overall response rates to salvage chemotherapy were 40 and 69 percent, respectively, for patients who relapsed <12 versus >12 months after initial diagnosis; it is uncertain if such differences would be present in the rituximab era [11]. In a retrospective analysis of 162 patients with relapsed DLBCL, overall survival (OS) did not differ between those whose initial duration of remission was ≥5 versus <5 years [12].
●Pathologic features – In the CORAL trial, compared with patients whose tumors did not have MYC rearrangement, those with a MYC rearrangement had inferior rates of four-year progression-free survival (18 versus 42 percent, respectively) and four-year OS (29 versus 62 percent) [13].
CHOICE OF TREATMENT — The preferred treatment varies with the timing of the relapse or refractory (r/r) DLBCL (algorithm 1):
●Early relapse (eg, <12 months after completing initial therapy) or primary refractory DLBCL. (See 'Relapse <12 months or primary refractory DLBCL' below.)
Versus
●Later relapse (eg, ≥12 months). (See 'Relapse ≥12 months after treatment' below.)
For patients with documented central nervous system (CNS) involvement, the CNS disease should be controlled before initiating the treatments described below. Management of CNS involvement by DLBCL is described separately. (See "Secondary central nervous system lymphoma: Treatment and prognosis", section on 'Diffuse large B cell lymphoma'.)
Management of second or later relapse and treatment of medically-unfit patients are discussed separately. (See "Diffuse large B cell lymphoma (DLBCL): Second or later relapse or patients who are medically-unfit".)
Relapse <12 months or primary refractory DLBCL — For patients with early first relapse of DLBCL or primary refractory DLBCL, we recommend CD19-directed chimeric antigen receptor (CAR)-T cell therapy using lisocabtagene maraleucel (liso-cel) or axicabtagene ciloleucel (axi-cel), if available, rather than autologous hematopoietic cell transplantation (HCT) (algorithm 1). This recommendation is based on superior disease control with liso-cel or axi-cel compared with autologous HCT in phase 3 trials [14,15]. Notably, the CD19-directed CAR-T cell product, tisagenlecleucel (tisa-cel), was not better than autologous HCT in another phase 3 trial [16].
We recognize that these therapies do not yet have regulatory approval for this indication and, as such, autologous HCT is acceptable for patients who do not have access to these treatments. (See 'Autologous transplantation' below.)
CAR-T cell therapy is a form of immunotherapy in which the patient's own T lymphocytes are transfected with a gene that encodes a CAR to direct the patient's T cells against the lymphoma. The manufacturing process is complex and expensive, administration is limited to qualified institutions, and the preferred product differs between institutions. CAR-T cell therapy can cause potentially fatal cytokine release syndrome (CRS) and neurologic toxicity. Administration, toxicity, and outcomes with CAR-T cell therapy for r/r DLBCL are provided below. (See 'Anti-CD19 CAR-T cell therapy' below.)
Based on the following reports, we consider either axi-cel or liso-cel acceptable for treatment of early relapse or primary refractory DLBCL; we suggest not using tisa-cel in this setting. No trial has directly compared efficacy and toxicity of the various CAR-T cell products. We generally treat with axi-cel in younger, more-fit patients and with liso-cel for older patients or those with comorbidities.
●Axi-cel – A phase 3 trial compared axi-cel versus standard care (two or three cycles of investigator-selected, protocol-defined salvage chemotherapy, followed by autologous HCT) in 359 patients with primary refractory DLBCL or relapse ≤12 months after initial therapy [14]. With follow-up >2 years, median event-free survival (EFS) was 8.3 months for axi-cel and 2.0 months for standard care; rates of 24-month EFS were 41 and 16 percent, respectively (hazard ratio [HR] for event or death 0.40 [95% CI 0.31-0.51]). Estimated two-year overall survival (OS) was 61 percent for the axi-cel group versus 52 percent for standard care (HR 0.73 [95% CI 0.53-1.01]). For the axi-cel group, grade ≥3 CRS was reported in 6 percent and grade ≥3 neurologic events in 21 percent; there were no deaths related to CRS or neurologic events.
●Liso-cel – The TRANSFORM trial randomly assigned 184 patients with primary refractory DLBCL or early relapse (<12 months) to liso-cel versus standard care (salvage chemotherapy followed by autologous HCT); cross-over to liso-cel was permitted for patients who did not achieve complete response (CR) or partial response (PR) after three cycles of salvage chemotherapy or did not achieve CR after autologous HCT [15]. Compared with the standard care arm, patients who received liso-cel had superior median EFS (10.1 versus 2.3 months), overall response rate (ORR; 79 versus 44 percent), CR (61 versus 36 percent), median progression-free survival (PFS; 14.8 versus 5.7 months), and OS (not reached versus 16.4 months after six-month follow-up; HR, 0.51 [95% CI 0.26–1.00]). Grade ≥3 treatment-related adverse events (AE) were reported in 79 percent of patients with standard care and 85 percent of patients with liso-cel.
●Tisa-cel – A phase 3 trial that included 322 patients with primary refractory DLBCL or early relapse reported that outcomes were similar for tisa-cel versus standard care (salvage chemotherapy followed by autologous HCT) [16]. Median EFS for both groups was 3.0 months and response rates were 46 percent for tisa-cel versus 43 percent for standard care. Deaths from treatment-related AEs were reported in 10 patients with tisa-cel and 13 patients with standard care.
The US Food and Drug Administration (FDA) has approved liso-cel, axi-cel, and tisa-cel for treatment of r/r DLBCL after ≥2 treatments.
Relapse ≥12 months after treatment — For patients with a late first relapse of DLBCL (eg, ≥12 months after completing therapy), we suggest salvage chemotherapy followed by autologous HCT, rather than CAR-T cell therapy or salvage chemotherapy alone (algorithm 1). Salvage chemotherapy followed by autologous HCT has long been the standard care in this setting, is associated with long-term survival in approximately half of patients with first relapse of DLBCL, and achieves superior OS compared with salvage therapy alone. Autologous HCT has not been directly compared with CAR-T cell therapy for late relapse of DLBCL (although it has been compared for early relapse and refractory disease, as described above).
In the phase 3 Parma trial, which was performed prior to routine use of rituximab, autologous HCT achieved superior OS compared with salvage therapy alone [17]. Among 109 patients with chemotherapy-sensitive lymphoma (most of whom had DLBCL), patients randomly assigned to transplantation had superior OS (53 versus 32 percent), ORR (84 versus 44 percent), and EFS (46 versus 12 percent), compared with four additional cycles of R-DHAP chemotherapy (rituximab, dexamethasone, high dose cytarabine, cisplatin). However, transplantation was more toxic than chemotherapy alone; 4 of 49 patients died with transplant-related effects (three infections, one cardiac toxicity), while none of the chemotherapy-only patients died from treatment [17]. However, it is important to note that this trial was performed prior to the use of mobilized peripheral blood progenitor/stem cells (PBPCs) as a stem cell source; contemporary approaches, which also feature improved supportive care, are associated with lower mortality risk (eg, 1 to 2 percent) [18].
Autologous HCT for late relapse of DLBCL involves the following:
●Intensive salvage chemotherapy – Two or three cycles of salvage therapy are given to reduce the burden of disease and determine if the relapsed DLBCL is sensitive to chemotherapy, according to positron emission tomography (PET)/computed tomography (CT). Choices of salvage chemotherapy are discussed below. (See 'Selection of salvage chemotherapy' below.)
●Response to salvage chemotherapy – Subsequent care is guided by PET/CT response, using the five-point scale (also called the Deauville score) (table 1), according to Lugano criteria (table 2):
•Complete response – Patients with CR (Deauville 1 to 3) and selected patients with Deauville 4 proceed to autologous HCT. The acceptable threshold for proceeding to HCT in a patient with Deauville 4 varies between institutions. (See 'Autologous transplantation' below.)
•Lesser responses – For patients with a less robust PR, no response, or progressive disease (ie, Deauville 5 and most patients with Deauville 4), we generally treat with anti-CD19 CAR-T cell therapy. Management of patients with an inadequate response to intensive salvage therapy (secondary refractory disease) is described separately. (See "Diffuse large B cell lymphoma (DLBCL): Second or later relapse or patients who are medically-unfit", section on 'No prior CAR-T cell therapy'.)
A systematic review and meta-analysis of 12 studies of autologous HCT, which included 313 patients with relapsed DLBCL, examined the predictive value of PET after salvage therapy [19]. Compared with chemotherapy-sensitive disease, patients with chemotherapy-resistant disease (ie, positive PET after salvage therapy) had shorter PFS (HR 4.3; 95% CI 3.1-6.0). Other prospective and retrospective studies also reported more favorable outcomes after autologous HCT in patients with chemotherapy-sensitive disease [19-24]. Patients with PR prior to transplantation have a 30 to 60 percent probability of disease-free survival (DFS) at three to five years; by comparison, patients with DLBCL that is resistant to salvage chemotherapy have a DFS <10 to 20 percent [7,25]. Other studies also reported the prognostic importance of negative PET after salvage therapy [26-28].
Older age (≥50 years), response less than CR, and primary refractory disease were associated with inferior survival in a registry study of patients with DLBCL undergoing autologous HCT [29]. The International Prognostic Index (IPI) identified groups with different OS and PFS among 80 patients undergoing autologous HCT for r/r DLBCL [8]. Analysis of 150 patients from three separate studies also reported that the IPI distinguished groups with different outcomes in this setting [7].
●Post-HCT monitoring – Surveillance for relapse and late AEs is described below. (See 'Monitoring' below.)
Our approach to autologous HCT for relapsed DLBCL is consistent with guidelines of the United States National Comprehensive Cancer Network (NCCN) [30], European Society of Medical Oncology (ESMO) [31], and the American Society for Blood and Marrow Transplantation (ASBMT) [32].
ANTI-CD19 CAR-T CELL THERAPY — Three commercial anti-CD19 chimeric antigen receptor (CAR)-T cell products: lisocabtagene maraleucel (liso-cel), axicabtagene ciloleucel (axi-cel), and tisagenlecleucel (tisa-cel) have been approved by the US Food and Drug Administration (FDA) for treatment of relapsed/refractory (r/r) DLBCL after ≥2 prior treatments.
●Manufacture – CAR-T cells are generated from the patient's own T lymphocytes, which are genetically modified (transfected) ex vivo with a gene that encodes a CAR to direct the patient's T cells against the lymphoma. The T cells are expanded in a production facility and then infused back into the patient as therapy.
●Products – Commercially available CD19-directed CAR-T cell products differ modestly in molecular design, but they also differ regarding manufacturing time, preferred bridging therapy, and adverse effects (AEs) [33].
●Outcomes – Approximately half of patients with refractory B cell lymphoma achieve complete response (CR) with CAR-T cell therapy and some patients have had remissions >3 years, but the duration of response (DoR) remains to be better-defined [34,35]. Outcomes with the individual CAR-T cell products are presented below.
●Toxicity – All CAR-T cell products are associated with serious, potentially fatal complications, but the rates and spectrum of AEs vary with the individual products. Presently, it appears that liso-cel is associated with less toxicity than the other commercially available CAR-T cell products. The most severe complications of treatment include:
•Cytokine release syndrome (CRS) is a severe systemic response to the activation and proliferation of CAR-T cells that is typically manifest as high fever, flu-like symptoms, hypotension, and mental status changes. Some degree of CRS is observed in nearly all treated patients and may be life-threatening for some, but it typically responds to treatment with aggressive supportive care that includes tocilizumab and corticosteroids, as described separately. (See "Cytokine release syndrome (CRS)".)
•Immune effector cell-associated neurotoxicity syndrome (ICANS) can be severe or life-threatening, as described separately. (See "Immune effector cell-associated neurotoxicity syndrome (ICANS)".)
•Other AEs include hypersensitivity reactions, serious infections, prolonged cytopenias, prolonged hypogammaglobulinemia, and second malignancies, including treatment-related myeloid neoplasms (eg, myelodysplastic syndrome or acute myeloid leukemia).
●Mitigation strategy – The US FDA labels carry a boxed warning for CRS and neurologic events. In the US, CAR-T cell products are only available through a risk evaluation and mitigation strategy (REMS). Facilities that dispense these agents require special certification, staff must be trained to recognize and manage AEs, and tocilizumab (a humanized monoclonal antibody against the interleukin-6 receptor) must be available for immediate administration.
Liso-cel — Lisocabtagene maraleucel (liso-cel) is an anti-CD19 CAR-T cell product with a 4-1BB (CD137) co-stimulatory domain that is administered as sequential infusions of two components (CD8+ and CD4+ CAR-T cells), which are selected from the leukapheresis material and independently activated, transduced, and expanded [36]. Presently, reports suggest that liso-cel may be better-tolerated than the other commercially available anti-CD19 CAR-T cell products, but the individual products have not been directly compared.
In the multicenter TRANSCEND NHL001 study, liso-cel was administered to 256 evaluable patients with r/r DLBCL; median age was 63 years, patients received a median of three lines of prior systemic therapy (including transplantation in one-third), and those with moderate kidney or cardiac dysfunction were included [36]. The objective response rate (ORR) was 73 percent, including 53 percent CR; median DoR was 17 months and one-year DoR was 55 percent. Grade ≥3 CRS was reported in 2 percent and grade ≥3 neurological events in 10 percent of treated patients; 3 percent of patients died with treatment-associated AEs.
Preliminary results of the phase 3 TRANSFORM trial (presented as an abstract) reported that for 184 patients with early relapse or primary refractory DLBCL, liso-cel had an acceptable safety profile and it achieved more favorable outcomes than standard care (ie, salvage chemotherapy followed by autologous hematopoietic cell transplantation [HCT]) [15]. Patients treated with liso-cel achieved superior median event-free survival (EFS; 10 versus 2 months, respectively), median progression-free survival (PFS; 15 versus 6 months), CR (61 versus 36 percent), ORR (86 versus 48 percent), and median overall survival (OS; not reached versus 16 months, with six-month median follow-up). Liso-cel was also associated with improved patient-reported quality of life, cognitive function, fatigue, and pain compared with standard care [37].
Axi-cel — Axicabtagene ciloleucel (axi-cel) comprises a murine anti-CD19 single chain variable fragment (scFv) linked to CD28 and CD3 zeta co-stimulatory domains. Axi-cel achieved superior EFS and ORR compared with autologous HCT in a phase 3 trial for patients with early relapse or primary refractory DLBCL [14].
In the phase 3 trial for early relapse or refractory DLBCL, axi-cel achieved superior median EFS compared with standard care [14]. With median follow-up of 25 months, axi-cel was associated with 61 percent estimated two-year OS, 8.3 month median EFS, and 41 percent 24-month EFS. Grade ≥3 CRS was reported in 6 percent and grade ≥3 neurologic events in 21 percent; there were no deaths related to CRS or neurologic events. In the multicenter ZUMA-1 study for patients with refractory DLBCL, axi-cel was associated with >2 year median OS, 82 percent ORR, six month median PFS, and 11 month median DoR [34,38].
Tisa-cel — Tisagenlecleucel (tisa-cel) comprises a murine anti-CD19 scFv linked to a CD8 hinge and a transmembrane region fused to intracellular signaling domains for 4-1BB and CD3 zeta [39]. In a phase 3 trial for patients with early relapse or primary refractory DLBCL, outcomes with tisa-cel were similar to those with salvage chemotherapy followed by autologous HCT [16]. As described above, we suggest not using tisa-cel for primary refractory DLBCL or first relapse.
In the phase 3 trial, 322 patients with early relapse or primary refractory DLBCL were randomly assigned to tisa-cel versus standard care [16]. Median EFS was 3 months for both groups and ORR was 46 percent for tisa-cel versus 43 percent for standard care. Deaths from treatment-related AEs were reported in 10 of 161 patients treated with tisa-cel.
AUTOLOGOUS TRANSPLANTATION — Autologous hematopoietic cell transplantation (HCT) is generally reserved for patients with late first relapse of DLBCL (eg, ≥12 months from last treatment) in medically-fit patients who have a complete response (CR) or very good partial response (PR) to intensive salvage chemotherapy, as discussed above. (See 'Relapse ≥12 months after treatment' above.)
Note that in treating relapsed DLBCL, we distinguish between:
●Salvage therapy – Intensive treatment that reduces the burden of disease, while determining if it is sensitive to chemotherapy (and thus amenable to transplantation).
●Conditioning therapy – High-dose chemotherapy (and occasionally immunotherapy or radiation therapy) administered immediately prior to HCT.
Selection of salvage chemotherapy — For patients with relapsed or refractory (r/r) DLBCL who will pursue autologous HCT, we suggest intensive salvage therapy rather than lower-intensity salvage therapy, because intensive treatment is more likely to achieve a robust response that enables transplantation [9,40-42]. No studies have directly compared intensive salvage regimens with lower-intensity regimens; comparisons across studies are difficult because of different inclusion criteria, pathologic features, patient composition, and outcome measures.
We select a salvage regimen based on the presence of DLBCL involving the central nervous system (CNS) or a higher-risk for CNS involvement, prior therapy, toxicity, comorbid conditions (eg, kidney dysfunction), institution/clinician experience, and patient convenience. We include rituximab in the salvage regimen, unless there is documented intolerance to rituximab. Patients are generally treated with two or three cycles of salvage therapy.
●CNS involvement
•Documented CNS involvement – For patients with CNS involvement, documented by imaging, biopsy, or cytology, we generally treat with a high-dose cytarabine-based salvage regimen, as described below. (See 'CNS involvement' below.)
•Higher risk for CNS involvement – Some experts administer high-dose cytarabine-based salvage therapy to selected patients at higher risk for CNS disease. There is no consensus about the selection of such patients, but in patients who initially present with DLBCL, higher risk for CNS involvement includes ≥2 of the following features: age >60 years, elevated lactate dehydrogenase (LDH), Eastern Cooperative Oncology Group (ECOG) performance status >1, extranodal involvement at >1 site, and kidney or adrenal involvement. (See "Initial treatment of advanced stage diffuse large B cell lymphoma", section on 'Central nervous system (CNS) evaluation'.)
●No CNS involvement – For patients who do not have CNS involvement or high risk for CNS involvement, selection of a salvage regimen is discussed below. (See 'No CNS involvement' below.)
Prophylaxis for tumor lysis syndrome (TLS) should be considered for patients who have a high tumor burden (eg, large tumor masses or markedly elevated LDH). Prophylaxis for TLS is described separately. (See "Tumor lysis syndrome: Prevention and treatment".)
No CNS involvement — For patients with r/r DLBCL with no CNS involvement, we choose salvage therapy according to comorbidities, toxicity, pathologic features, clinician experience, and suitability for outpatient administration; some experts also consider cell-of-origin (COO) analysis in selecting a salvage regimen. The following regimens have comparable efficacy (and potential for proceeding to HCT), but adverse effects vary and the preferred regimen varies among institutions.
Common salvage regimens include:
●R-GDP – (See 'R-GDP (Rituximab, gemcitabine, dexamethasone, cisplatin)' below.)
●R-ICE – (See 'R-ICE (Rituximab, ifosfamide, carboplatin, etoposide)' below.)
●R-DHAP – (See 'R-DHAP (Rituximab, dexamethasone, high dose cytarabine, cisplatin)' below.)
●R-GEMOX – (See 'R-GEMOX (Rituximab, gemcitabine, oxaliplatin)' below.)
Factors that may influence the choice of a salvage regimen are:
●Outpatient administration – R-GDP and R-ICE can be administered in the outpatient setting.
●Prior etoposide therapy – We generally avoid R-ICE for patients who previously received R-EPOCH (rituximab, etoposide, prednisone, vincristine, cyclophosphamide, doxorubicin) or other etoposide-containing regimen.
●Germinal center B cell (GCB) pathology – Some experts favor treatment with R-DHAP for patients with GCB immunophenotype, according to COO analysis. (See "Prognosis of diffuse large B cell lymphoma", section on 'Cell of origin studies'.)
●Kidney function – For patients with limited kidney function, carboplatin-based salvage therapy may be chosen over cisplatin-based therapy.
CNS involvement — For patients with documented CNS involvement and for selected patients at higher risk for CNS involvement, we suggest a high-dose cytarabine-based regimen, rather than other salvage regimens. No studies have directly compared high-dose cytarabine-based salvage therapy with other intensive regimens for r/r DLBCL; this suggestion is based on activity of such regimens against primary and secondary CNS lymphoma. (See "Secondary central nervous system lymphoma: Treatment and prognosis", section on 'Relapsed DLBCL, secondary CNS disease'.)
Higher-risk features for CNS involvement are described above. (See 'Selection of salvage chemotherapy' above.)
High-dose cytarabine-based salvage regimens include:
●R-DHAP – (See 'R-DHAP (Rituximab, dexamethasone, high dose cytarabine, cisplatin)' below.)
●R-ESHAP – (See 'R-ESHAP (Rituximab, etoposide, methylprednisolone, cytarabine, cisplatin)' below.)
Some experts consider methotrexate-containing regimens (eg, MATRix-RICE) acceptable, based on treatment of patients with secondary CNS lymphoma involvement. (See "Secondary central nervous system lymphoma: Treatment and prognosis", section on 'Relapsed DLBCL, secondary CNS disease'.)
Salvage regimens — Common intensive salvage regimens for r/r DLBCL follow:
R-GDP (Rituximab, gemcitabine, dexamethasone, cisplatin)
●Administration – R-GDP includes rituximab (375 mg/m2 intravenously on day -1), gemcitabine (1000 mg/m2 per day on days 1 and 8), dexamethasone (40 mg by mouth daily on days 1 through 4), and cisplatin (75 mg/m2 on day 1), administered at 21-day intervals [40,43,44].
●Adverse effects – Hematologic toxicity is universal and febrile neutropenia is seen in approximately 15 percent of patients.
●Outcomes – R-GDP was less toxic but had comparable efficacy compared with R-DHAP in the phase 3 LY.12 trial [40]. R-GDP has not been compared head-to-head with other salvage regimens.
In the LY.12 trial, 619 patients (69 percent had r/r DLBCL) were randomly assigned to R-GDP versus R-DHAP [40]. There was no difference in overall survival (OS), event-free survival (EFS), overall response rate (ORR; 45 versus 44 percent), or ability to proceed to autologous HCT (52 versus 49 percent), but GDP was associated with fewer episodes of grade ≥3 toxicity (47 versus 61 percent), less hospitalization (47 versus 99 percent), and better patient-reported quality of life (QoL). There were eight treatment-related deaths (two patients receiving R-GDP, six with R-DHAP). However, some experts favor R-DHAP in patients with GCB features based on a retrospective study [45], as described below. (See 'R-DHAP (Rituximab, dexamethasone, high dose cytarabine, cisplatin)' below.)
R-ICE (Rituximab, ifosfamide, carboplatin, etoposide)
●Administration – R-ICE includes rituximab (375 mg/m2 on day -1), ifosfamide (5000 mg/m2 continuous infusion for 24 hours on day 2), mesna (100 mg/m2 on days 1, 2, and 3), carboplatin (area under the curve 5 [maximum dose 800 mg] on day 2), and etoposide (100 mg/m2 per day on days 1, 2, and 3) (table 7) and is generally given every 21 days [9,46,47].
●Adverse effects – Hematologic toxicity is universal, with one-third of patients requiring transfusions and grade ≥3 nonhematologic adverse effects, including infection (in up to one-quarter of patients) and occasional nephrotoxicity.
●Outcomes – In the phase 3 CORAL trial, there was no difference in outcomes among 396 patients who were randomly assigned to three cycles of salvage therapy using R-ICE versus three cycles of R-DHAP, but R-DHAP caused more kidney toxicity [9]. R-DHAP and R-ICE achieved comparable ORR (64 versus 63 percent), three-year EFS (26 and 35 percent), and three-year OS (47 and 51 percent). In both arms, 16 percent of patients had febrile neutropenia, but R-DHAP was associated with grade 4 renal toxicity in 11 patients and more patients needed platelet transfusions (57 versus 35 percent).
R-GEMOX (Rituximab, gemcitabine, oxaliplatin)
●Administration – R-GemOx includes rituximab (375 mg/m2 on day -1), gemcitabine (1000 mg/m2 on day 2), and oxaliplatin (100 mg/m2 on day 2) [48,49].
●Adverse effects – Severe hematologic toxicity occurs in half of patients and neuropathy can occur.
●Outcomes – R-GemOx is associated with ORR in up to half of patients and CR in up to one-third of patients with r/r DLBCL [41,48-51].
Oxaliplatin has not been approved by the US Food and Drug Administration (FDA) for treatment of r/r DLBCL.
R-DHAP (Rituximab, dexamethasone, high dose cytarabine, cisplatin)
●Administration – R-DHAP includes rituximab (375 mg/m2 on day -1), dexamethasone (40 mg/d on days 1 to 4), cisplatin (100 mg/m2 on day 1 by continuous infusion), and cytarabine (2 g/m2 in a three-hour infusion on day 2) every three weeks [9].
For patients with pre-existent kidney insufficiency, some experts replace cisplatin with carboplatin or oxaliplatin to lessen nephrotoxicity, but there are limited outcomes data with these regimens [52].
●Adverse effects – Hematologic toxicity is universal, with one-third of patients requiring transfusions, and grade ≥3 nonhematologic adverse effects include infection (in up to one-quarter of patients) and occasional nephrotoxicity.
●Outcomes – In the phase 3 CORAL trial, outcomes were similar among 396 patients who were randomly assigned to three cycles of salvage therapy using R-DHAP versus R-ICE, but R-DHAP caused more kidney toxicity [9]. The two trial arms achieved comparable ORR (64 versus 63 percent), three-year EFS (26 and 35 percent), and three-year OS (47 and 51 percent). In both arms, 16 percent of patients had febrile neutropenia, but R-DHAP was associated with grade 4 renal toxicity in 11 patients and more patients needed platelet transfusions (57 versus 35 percent). The LY.12 and ORCHARRD phase 3 trials reported similar outcomes with D-HAP for r/r DLBCL [40,53].
Retrospective analysis of the CORAL study reported that outcomes varied according to COO status and that for patients with GCB-like DLBCL, treatment with R-DHAP was associated with better outcomes than patients treated with R-ICE (three-year PFS: 100 percent versus 27 percent) [45].
R-ESHAP (Rituximab, etoposide, methylprednisolone, cytarabine, cisplatin)
●Administration – R-ESHAP includes rituximab (375 mg/m2 on day 1), etoposide (40 mg/m2/day as a one-hour infusion on days 1 to 4), methylprednisolone (250 to 500 mg/day as a 15-minute infusion on days 1 to 5), cisplatin (25 mg/m2/day as a continuous infusion from day 1 to 4), and cytarabine (2 g/m2 as a two-hour infusion on day 5), every three or four weeks [42].
●Adverse effects – Hematologic toxicity is universal, with significant rates of neutropenic fever (30 percent) if growth factors are not used. Other adverse effects (eg, nausea, vomiting, diarrhea, nephrotoxicity, electrolyte disturbances) are generally mild.
●Outcomes – A retrospective study of 163 patients reported that ESHAP for relapsed DLBCL was associated with 75 to 86 percent ORR and 41 to 50 percent CR, while for primary refractory DLBCL, ORR was 33 percent and CR was 8 percent [42].
Autologous HCT — Autologous HCT consists of myeloablative conditioning chemotherapy followed by infusion of autologous hematopoietic stem/progenitor cells. Results from trials that directly compared autologous HCT versus CAR-T cell are described above. (See 'Relapse <12 months or primary refractory DLBCL' above.)
Transplantation is generally reserved for medically-fit patients with relapsed DLBCL who achieve a CR or a very good PR to salvage chemotherapy. There is no age above which the benefits of autologous HCT are clearly outweighed by toxicity [32], but some institutions reserve transplantation for patients ≤75 years. A bone marrow examination should be performed prior to transplantation to exclude involvement of the graft with myelodysplastic syndrome or acute myeloid leukemia. Factors that determine suitability for autologous HCT are discussed separately. (See "Determining eligibility for autologous hematopoietic cell transplantation".)
Approximately half of patients who undergo autologous HCT for r/r DLBCL are alive five years later. In a population-based study, autologous HCT achieved 51 percent four-year OS; this was superior to 28 percent four-year OS for the entire population of 239 patients with relapsed DLBCL [54]. A registry study reported 63 percent five-year OS and 48 percent five-year disease-free survival (DFS) after autologous HCT in 470 patients; notably, DFS after transplantation was considerably longer than the duration of the first remission (median 51 versus 11 months, respectively) [55].
Most relapses occur during the first two years after transplantation; nonrelapse mortality (NRM) surpasses relapse as the main cause of death beginning eight years after transplantation [56]. The most common causes of NRM in the first two years are respiratory failure (31 percent), infection (13 percent), cardiac toxicity (15 percent), and secondary malignancy (15 percent); after two years, secondary malignancies, including myelodysplastic syndrome, acute leukemia, and solid tumors, are the most common cause of NRM. (See "Survival, quality-of-life, and late complications after hematopoietic cell transplantation in adults".)
Stem cell source — We suggest peripheral blood stem-progenitor cells (PBSPCs) as the graft source, rather than a bone marrow (BM)-derived graft, because PBPCs yield faster engraftment, less potential for contamination with tumor cells, improved quality-of-life, and lower transplantation-associated costs [57]. However, BM is an acceptable option if there is an inadequate collection of PBSPCs.
PBSPC grafts were associated with better outcomes compared with BM grafts in a phase 3 trial of autologous HCT in 95 patients with chemotherapy-sensitive r/r non-Hodgkin lymphoma (NHL) [57]. Compared with BM, PBSPCs were faster to achieve >500 granulocytes/microL (15 versus 26 days, respectively) and >20,000/microL platelets (13 versus 18 days); furthermore, patients receiving PBSPCs required fewer platelet transfusions (median 4 versus 8 units), had fewer somatic complaints, and were more physically active. Another study, which included 59 patients with r/r NHL, reported that non-mobilized PBSPC grafts led to more rapid neutrophil recovery, less transfusion-dependence, shorter hospital stays, and no significant difference in relapse-free survival (RFS; 70 versus 53 percent after median 460 day follow-up) compared with BM [58]. A retrospective study of 150 patients also reported faster hematopoietic recovery with PBSPCs [59].
Sources and collection of autologous grafts are discussed separately. (See "Hematopoietic support after hematopoietic cell transplantation", section on 'Autologous PBPC transplantation' and "Sources of hematopoietic stem cells", section on 'PBPC mobilization'.)
Conditioning regimen — Patients who undergo autologous HCT for r/r DLBCL should receive myeloablative conditioning therapy. No conditioning regimen has proven to be superior and the preferred protocol varies between institutions. The most common conditioning regimens used for r/r DLBCL are:
●BEAM (carmustine [BCNU], etoposide, cytarabine, melphalan) [60]
●CBV (cyclophosphamide/BCNU/etoposide) [61]
BEAM is generally associated with more gastrointestinal toxicity, while CBV has more pulmonary toxicity [61,62]. Other conditioning regimens have been used less-commonly in this setting [61,63,64], but bendamustine-containing conditioning regimens [65] should be avoided because they may impair collection of T cells, if CAR-T cell therapy is a future consideration.
There is no clear benefit to including rituximab in the conditioning regimen, based on a large registry analysis and retrospective studies [29,66,67]. A clinical trial that randomly assigned 224 patients to conditioning with rituximab plus BEAM (R-BEAM) versus I-131 tositumomab plus BEAM (B-BEAM) reported no difference in OS, PFS, or early (100 day) treatment-related mortality, but B-BEAM was associated with more severe mucositis (52 versus 18 percent) [66].
Myeloablative conditioning regimens are discussed in greater detail separately. (See "Preparative regimens for hematopoietic cell transplantation", section on 'Myeloablative (MAC) regimens'.)
No role for post-HCT rituximab maintenance therapy — We suggest not administering rituximab maintenance therapy after autologous HCT, because of a lack of demonstrated benefit [32].
The multicenter CORAL trial that randomly assigned patients to R-ICE versus R-DHAP also included a second randomization to no maintenance therapy versus maintenance rituximab (375 mg/m2 every two months for a year) after autologous HCT [68]. Compared with no maintenance therapy, rituximab maintenance resulted in similar rates of estimated four-year EFS (52 versus 53 percent), PFS (52 versus 56 percent), and OS (61 versus 65 percent), but it was associated with more adverse events after day 100 (mostly infections).
MONITORING — Patients should be monitored for relapse and for treatment-related toxicities.
●After CAR-T cell therapy – We generally repeat positron emission tomography (PET)/computed tomography (CT) at day 30 and, if positive, repeat it at day 100; after that, we repeat PET/CT only for symptoms.
We monitor IgG levels and CD4 counts for at least one year. We provide acyclovir prophylaxis for one year and trimethoprim-sulfamethoxazole prophylaxis until CD4 counts are >200/microL.
●After autologous hematopoietic cell transplantation (HCT) – We schedule follow-up three to six months after transplantation, but there is no consensus for the protocol and schedule of further follow-up and practices vary among institutions. To limit radiation exposure and because the likelihood of detecting an asymptomatic relapse is small, we limit routine CTs and we do not perform regularly scheduled PETs. (See "Long-term care of the adult hematopoietic cell transplantation survivor", section on 'Autologous HCT'.)
SPECIAL CONSIDERATIONS DURING THE COVID-19 PANDEMIC — The coronavirus disease 2019 (COVID-19) pandemic has increased the complexity of cancer care. Important issues include balancing the risk from treatment delay versus harm from COVID-19, ways to minimize negative impacts of social distancing during care delivery, and appropriately and fairly allocating limited health care resources. Additionally, immunocompromised patients are candidates for a modified vaccination schedule (figure 1), other preventive strategies (including pre-exposure prophylaxis), and the early initiation of COVID-directed therapy. These issues and recommendations for cancer care during the COVID-19 pandemic are discussed separately. (See "COVID-19: Considerations in patients with cancer".)
CLINICAL TRIALS — Often, there is no better therapy to offer a patient than enrollment onto a well-designed, scientifically valid, peer reviewed clinical trial. Additional information and instructions for referring a patient to an appropriate research center can be obtained from the United States National Institutes of Health (www.clinicaltrials.gov).
SOCIETY GUIDELINE LINKS — Links to society and government-sponsored guidelines from selected countries and regions around the world are provided separately. (See "Society guideline links: Management of diffuse large B cell lymphoma".)
INFORMATION FOR PATIENTS — UpToDate offers two types of patient education materials, "The Basics" and "Beyond the Basics." The Basics patient education pieces are written in plain language, at the 5th to 6th grade reading level, and they answer the four or five key questions a patient might have about a given condition. These articles are best for patients who want a general overview and who prefer short, easy-to-read materials. Beyond the Basics patient education pieces are longer, more sophisticated, and more detailed. These articles are written at the 10th to 12th grade reading level and are best for patients who want in-depth information and are comfortable with some medical jargon.
Here are the patient education articles that are relevant to this topic. We encourage you to print or e-mail these topics to your patients. (You can also locate patient education articles on a variety of subjects by searching on "patient education" and the keyword(s) of interest.)
●Basics topics (see "Patient education: Diffuse large B cell lymphoma (The Basics)" and "Patient education: Autologous bone marrow transplant (The Basics)" and "Patient education: Allogeneic bone marrow transplant (The Basics)")
●Beyond the Basics topics (see "Patient education: Diffuse large B cell lymphoma in adults (Beyond the Basics)" and "Patient education: Hematopoietic cell transplantation (bone marrow transplantation) (Beyond the Basics)")
SUMMARY AND RECOMMENDATIONS
●Refractory diffuse large B cell lymphoma (DLBCL) is disease that did not achieve a complete response (CR) with initial treatment; a repeat biopsy should be performed unless there was clear progression of size and metabolic activity of disease. (See 'Refractory DLBCL' above.)
●Relapsed DLBCL is disease that recurs after a CR; biopsy is required to confirm the diagnosis and exclude other conditions. (See 'Relapsed DLBCL' above.)
●Pretreatment evaluation includes (see 'Pretreatment evaluation' above):
•Medical fitness based on performance status (table 3) and comorbid conditions (table 4). (See 'Fitness categories' above.)
Management of less medically-fit patients is described separately. (See "Diffuse large B cell lymphoma (DLBCL): Second or later relapse or patients who are medically-unfit".)
•Restaging includes positron emission tomography (PET)/computed tomography (CT). (See 'Restaging' above.)
•Prognosis according to the International Prognostic Index (IPI) (table 6). (See 'Prognosis' above.)
●Preferred treatment varies with the setting (algorithm 1):
•Refractory DLBCL or early relapse (<12 months after initial therapy) – For refractory disease or early relapse, we recommend CD19-directed chimeric antigen receptor (CAR)-T cell therapy using lisocabtagene maraleucel (liso-cel) or axicabtagene ciloleucel (axi-cel), rather than autologous hematopoietic cell transplantation (HCT) (Grade 1B). (See 'Relapse <12 months or primary refractory DLBCL' above.)
•Late relapse (≥12 months) – For late first relapse, we suggest intensive salvage chemotherapy followed by autologous HCT, rather than CAR-T cell therapy (Grade 2C). (See 'Relapse ≥12 months after treatment' above.)
●CAR-T cell therapy can cause severe adverse effects, including cytokine release syndrome (CRS) and immune effector cell-associated neurotoxicity syndrome (ICANS), as discussed separately. (See "Cytokine release syndrome (CRS)" and "Immune effector cell-associated neurotoxicity syndrome (ICANS)".)
●Autologous HCT consists of myeloablative conditioning therapy and infusion of an autologous graft in patients who achieve CR or very good partial response (PR) to intensive salvage chemotherapy, based on PET/CT response (table 2).
•Salvage therapy – (See 'Salvage regimens' above.)
-No CNS involvement – For no CNS involvement, choice of salvage therapy is informed by prior therapy, toxicity, comorbid conditions (eg, kidney dysfunction), institution/clinician experience, patient convenience, and may consider pathologic features. (See 'No CNS involvement' above.)
-CNS involvement – For documented CNS involvement and selected patients at higher-risk for CNS involvement, we suggest a high-dose cytarabine-based regimen, rather than other intensive salvage regimens (Grade 2C). (See 'CNS involvement' above.)
•Autologous HCT – (See 'Autologous HCT' above.)
-Stem cell source – We suggest peripheral blood stem-progenitor cells (PBSPCs) as the graft source, rather than a bone marrow (BM)-derived graft (Grade 2B). (See 'Stem cell source' above.)
-Conditioning – Myeloablative conditioning therapy is used for autologous HCT. (See 'Conditioning regimen' above.)
-No maintenance therapy – There is no role for rituximab maintenance therapy in this setting. (See 'No role for post-HCT rituximab maintenance therapy' above.)
●Monitoring after CAR-T cell therapy and after autologous HCT is described above. (See 'Monitoring' above.)
