Challenges and Opportunities in the Management of Diffuse Large B‐Cell Lymphoma in Older Patients

Abstract

Most patients with diffuse large B‐cell lymphoma (DLBCL) are diagnosed at age 60 years or older. Challenges to effective therapy among older individuals include unfavorable biologic features of DLBCL, geriatric vulnerabilities, suboptimal treatment selection, and toxicities of cytotoxic chemotherapy. Wider application of geriatric assessments may help identify fit older patients who benefit from standard immunochemotherapy without unnecessary dose reductions. Conversely, attenuated regimens may provide a better balance of risk and benefit for selected unfit or frail patients. Supportive care with the use of corticosteroid‐based prephase, prophylactic growth factors, and early institution of supportive and palliative care can help maximize treatment tolerance. Several novel or emerging therapies have demonstrated favorable toxicity profiles, thus facilitating effective treatment for elderly patients. In the relapsed or refractory setting, patients who are not candidates for stem cell transplantation can benefit from newly approved options including polatuzumab vedotin‐based combinations or tafasitamab plus lenalidomide, which may have higher efficacy and/or lower toxicity than historical chemotherapy regimens. Chimeric antigen receptor T‐cell therapy has been successfully applied to older patients outside of clinical trials. In the first‐line setting, emerging immunotherapy options (bispecific antibodies) and targeted therapies (anti‐CD20 antibodies combined with lenalidomide and/or B‐cell receptor inhibitors) may provide chemotherapy‐free approaches for DLBCL. Enrolling older patients in clinical trials will be paramount to fully examine potential efficacy and toxicity of these strategies. In this review, we discuss recent advances in fitness stratification and therapy that have expanded curative options for older patients, as well as future opportunities to improve outcomes in this population.

Implications for Practice

Management of diffuse large B‐cell lymphoma in older patients poses challenges due to aggressive disease biology and geriatric vulnerability. Although R‐CHOP remains standard first‐line treatment, geriatric assessment may help evaluate patients' fitness for immunochemotherapy. Corticosteroid prephase, prophylactic growth factors, and early palliative care can improve tolerance of treatment. Novel salvage options (polatuzumab vedotin‐based combinations, tafasitamab plus lenalidomide) or chimeric antigen receptor T‐cell therapy should be considered in the relapsed or refractory setting for patients ineligible for stem cell transplantation. Emerging immunotherapies (bispecific antibodies) and targeted therapies provide potential first‐line chemotherapy‐free approaches, which need to be rigorously assessed in clinical trials that involve geriatric patients.

Short abstract

Diffuse large B‐cell lymphoma (DLBCL) is an aggressive but potentially curable cancer that mainly affects older individuals. This review presents recent advances in fitness stratification and therapy that have expanded curative options for older patients with DLBCL and future opportunities to improve outcomes in this population.

Introduction

Diffuse large B‐cell lymphoma (DLBCL) is an aggressive, yet potentially curable, cancer. It predominantly affects older individuals, as median age at diagnosis is 66, and a third of patients are diagnosed at age >75 (Fig. 1A, B) [1, 2]. Age over 60 (and further >75) is an unfavorable prognostic factor in DLBCL and is included in the International Prognostic Index (IPI) [3, 4] Worse overall survival (OS) among older individuals continues to be observed even among patients who receive full‐dose rituximab, cyclophosphamide, doxorubicin, vincristine and prednisone (R‐CHOP; Fig. 1C) [4, 5]. In the U.S., 5‐year relative survival in DLBCL is 78% for patients younger than 55 but only 54% for those age older than 65 [1]. Achieving event‐free survival at 2 years from diagnosis leads to normalization of life expectancy among younger patients, but older patients experience excess mortality even after 5 years [6]. Moreover, up to 25% of older Americans and Europeans do not receive any immunochemotherapy after DLBCL diagnosis; this proportion is even higher for patients aged over 75 or those with functional impairment [7, 8, 9]. Therefore, worse outcomes among older patients with DLBCL may result from an interplay of unfavorable biology of the disease, baseline health status, suboptimal management, and late toxicities of chemotherapy, underscoring significant unmet treatment needs in this population [10].

Figure 1.

Epidemiology of diffuse large B‐cell lymphoma (DLBCL) according to age in data from the SEER program, 2010–2017. (A): Incidence of DLBCL according to age and sex. (B): Proportions of DLBCL cases diagnosed in each age group. (C): Overall survival by age at diagnosis. Molecular differences according to age in the large cohort of DLBCL by Reddy et al. The p values are from Mantel–Haenszel score test for trend. (D): Proportion of patients with ABC, GCB, or unclassified DLBCL; cell of origin was determined by RNA sequencing (n = 744). (E): proportion of patients with high, medium, or low risk genomic mutation profile according to the classifier by Reddy et al. (n = 733). (F): Prevalence of mutations in the 10 most commonly mutated genes in DLBCL (n = 950), stars indicate uncorrected p < .05, but no difference was statistically significant after correction for multiple testing [16].Abbreviations: ABC, activated B‐cell; GCB, germinal center B‐cell.

In this review, we will discuss newly recognized clinicopathologic features that characterize DLBCL in older patients, challenges in the management, current treatment strategies, and emerging opportunities for improved therapy.

Biology of DLBCL in Older Patients

In the 2017 World Health Organization (WHO) classification, DLBCL encompasses many entities with variable biology and clinical behavior that are classified using morphologic, gene expression profiling (GEP), and molecular criteria. Specific subtypes include T‐cell/histiocyte‐rich large B‐cell lymphoma; Epstein‐Barr virus (EBV)–positive DLBCL; primary DLBCL of the central nervous system (CNS); primary cutaneous DLBCL, leg type; primary mediastinal (thymic) large B‐cell lymphoma (LBCL); and intravascular LBCL [11]. Other cases are grouped into DLBCL, not otherwise specified (NOS), further distinguishing the germinal center B‐cell (GCB) and activated B‐cell (ABC) subtypes. In clinical practice, the GCB/ABC subtypes are approximated to GCB and non‐GCB phenotypes using immunohistochemistry for CD10, BCL6, and MUM1/IRF4 [12]. Furthermore, WHO distinguishes two categories of high‐grade B‐cell lymphoma (HGBL), historically often diagnosed and managed as DLBCL: HGBL with MYC and BCL2 and/or BCL6 rearrangements (“double‐” or “triple‐hit” lymphoma; DHL) and HGBL, NOS. As subtype‐specific treatment options begin to emerge, advanced molecular testing will gain importance regardless of patient age to help identify the most efficacious therapeutic approaches.

(a) ABC DLBCL is distinguished from GCB by GEP indicative of postgerminal center origin; it is driven by chronic active B‐cell receptor (BCR) signaling and NF‐kB deregulation [12, 13]. The prevalence of ABC DLBCL increases continuously with age, reaching 40%–50% after age 60 (Fig. 1D) [14, 15, 16]. This skewed distribution may hypothetically be influenced by progressive immune senescence and aging‐related shifts in the B‐cell repertoire [17]. The ABC subtype is characteristic of extranodal lymphomas, including primary CNS, testicular, and cutaneous leg‐type DLBCL, which are enriched in the elderly population and harbor molecular features associated with immune evasion [18]. In many studies, ABC DLBCL displays inferior survival after immunochemotherapy [13, 14]. The emerging genomic data reveal that it encompasses a poor‐prognosis subtype characterized by MYD88 ^L265P and CD79B mutations (termed MCD or cluster C5) with frequent extranodal invasion [19, 20, 21, 22]. ABC lymphomas may respond to novel agents targeting the BCR (ibrubinib, acalabrutinib, zanubrutinib) or immunomodulatory mechanisms (lenalidomide), opening the possibility for less toxic treatment approaches suitable for older patients [23, 24].

(b) EBV‐positive DLBCL is an aggressive lymphoma that occurs more frequently in elderly individuals and does not display a pathognomonic morphology or immunophenotype. Its diagnosis requires demonstration of EBV‐encoded small RNA by in situ hybridization [25, 26]. Median age at diagnosis is 71, and extranodal disease is present in 70% of patients [27]. EBV‐positive DLBCL has inferior survival in all IPI categories, with median OS of 24 months [27]. Currently, no specific therapy improves outcomes of EBV‐positive DLBCL over standard immunochemotherapy. However, EBV‐positive tumors express CD30 in about half of cases, providing a potential opportunity for targeted therapy using brentuximab vedotin [28]. Furthermore, nearly all EBV‐positive DLBCL strongly express programmed death ligand‐1 (PD‐L1), offering a possibility of immune checkpoint inhibitor‐based therapy [29].

(c) HGBL is a newly defined group of aggressive lymphomas that include the DHL, as well as many prior cases of Burkitt‐like or aggressive immunoblastic variants. They have worse outcomes with R‐CHOP immunochemotherapy and a higher risk of CNS invasion [30, 31, 32]. HGBL has not yet been thoroughly characterized, but advanced age correlates with increased prevalence of high‐risk molecular DLBCL subtypes (Fig. 1E) and of MYC rearrangement, although not with any single gene mutations (Fig. 1F) [16, 33]. Older patients have a higher incidence of other aggressive histologies like plasmablastic or primary effusion lymphoma without HIV infection. The dose‐adjusted etoposide, prednisone, vincristine, cyclophosphamide, doxorubicin, and rituximab (DA‐EPOCH‐R) combination has been used in patients even over age 80, providing a realistic option for fit older individuals with HGBL [34]. However, other intensive regimens applied in HGBL are not feasible in older patients; the optimal approach remains an area of major clinical need.

Challenges in the Management of DLBCL in Older Patients

Geriatric patients have decreased tolerance of chemotherapy because of limited bone marrow reserve, altered drug metabolism, presence of comorbid conditions, or impaired physical function and/or cognition. Resulting concerns may lead to application of attenuated regimens or withholding curative therapy based on age alone, which may not always be well‐justified clinically [7, 35, 36]. In one retrospective study, 23% of elderly patients received abbreviated or attenuated chemotherapy despite good performance status [37].

Aging is a heterogeneous phenomenon, as same‐age individuals may markedly differ in physiological or psychosocial robustness and routine oncology assessments are insufficient to identify vulnerabilities. Traditional measures of performance status do not accurately predict treatment‐related complications [38, 39]. Without accurate characterization of global fitness, geriatric patients are both at risk of undertreatment (receiving an attenuated regimen despite ability to tolerate a more effective therapy) and overtreatment (receipt of standard treatment despite high vulnerability to complications). Both may lead to inferior outcomes.

Comprehensive Geriatric Assessment in DLBCL

Comprehensive geriatric assessment (CGA) is a strategy to characterize the fitness of an elderly patient using multiple domains: physical function, cognitive function, comorbidities, socioeconomic, nutritional, and psychological status, geriatric syndromes (e.g., fall risk), and polypharmacy. CGA captures age‐specific vulnerabilities beyond those identified by routine assessments [39, 40, 41]. In hematologic malignancies, CGA correlates with treatment tolerance [42, 43, 44, 45, 46], survival [43, 44, 45, 46, 47, 48, 49, 50, 51], and health care use [45, 52, 53]. CGA is often combined with personalized supportive care (e.g., nutritional counseling), and it may impact treatment strategy [54].

In DLBCL, studies have used age in conjunction with validated instruments evaluating activity of daily living, instrumental activity of daily living, and comorbidities to define categories of “fit,” “unfit,” and “frail” (Table 1) [55, 56, 57, 58, 59, 60, 61]. Simplified CGA (sCGA), as described by Tucci et al. using criteria shown in Table 2, can be a predictor of treatment and outcomes in DLBCL, consistently showing worse survival for frail individuals compared with fit ones [55, 56, 57, 58, 59, 60, 61]. In one prospective observational study, curative‐intent treatment provided a survival benefit over palliative therapy for fit patients, whereas differences in OS could not be demonstrated in the unfit and frail groups [56]. However, in a randomized study, no difference between R‐CHOP and an attenuated regimen was observed in fit patients, with similar rates of event‐free survival (48% vs. 46%, respectively) and OS (62% and 63%, respectively) [62]. In a further study of 1,353 patients, even the unfit and frail groups showed a survival benefit from attenuated (<70% dose intensity) yet curative (anthracycline‐based) approaches, although treatment selection was not prespecified and may have been influenced by confounding factors [63]. One Australian study found sCGA categories to be associated with dose reductions during the first chemotherapy cycle but not with subsequent dose reductions, hospitalizations, or toxicity, concluding that these adverse events may have been mitigated by preemptive management [57]. Other prospective studies applied geriatric assessments to personalize chemotherapy (including dose reductions or drug replacements) and balance effectiveness against safety, but control groups were universally lacking [58, 59, 64].

Table 1.

Studies of CGA in older patients with diffuse large B‐cell lymphoma

Study, country	Design	N	Age, median	Domains	Category	Follow‐up, mo	Outcomes	Results
Tucci 2015, Italy [56]	P	173	77	Age, ADL, IADL, CIRS‐G	Fit 46%, unfit 16%, frail 38%	24	OS	2 yr OS in curative vs. palliative: fit: 88% vs. 25%, b unfit: 75% vs. 45%, frail: 44% vs. 39%; frail + unfit vs. fit: HR, 3.69 (95% CI, 1.09–12.5) a
Merli 2014, Italy [61]	P	318	Frail: 78, fit: 72	Age, ADL, IADL, CIRS‐G	Fit 70%, frail 30%	36	OS, SMR, relative survival	OS HR frail vs. fit: HR: 2.85 (1.76, 4.61); SMR: frail 12.5, fit HR: 4.7; 5 yr relative survival: frail 23.8%, fit 71.5%
Spina 2019, Italy [63]	P	1353	72	Age, ADL, IADL, CIRS‐G	Fit 41%, unfit 25%, frail 34%	29	OS	HR frail vs. fit: 2.74 (95% CI, 2.13–3.51); HR palliative vs. curative: unfit: 2.95 (95% CI, 1.50–5.83), frail: 2.97 (95% CI, 2.05–4.30)
Ong 2019, Australia [57]	R	205	73	Age, ADL, IADL, CIRS‐G	Fit 41%, unfit 21%, frail 38%	29	DR, toxicity at C1, unplanned admission, OS, PFS	CGA associated with preemptive DR; OS frail vs. fit: HR:2.89 (95% CI, 1.28–6.56) a ; Other comparisons not statistically significant
Spina 2012, Italy [58]	P, single‐arm	100	75	ADL, IADL, comorbidity	Fit 55%, unfit 32%, frail 13%	64	CRR, toxicity, OS, DFS, CSS	5 yr OS: 60%, DFS 80%, C74%, CRR 81%, CGA not associated with toxicity; OS frail vs. fit: HR: 2.55 (95% CI, 1.14–5.73) a ; OS unfit vs. fit: HR: 1.96 (95% CI, 1.04–3.70) a
Olivieri 2012, Italy [59]	P, single‐arm	91	74	Age, ADL, CIRS‐G, geriatric syndrome	Fit 59%, unfit 24%, frail 17%	57	CRR, OS, DFS, EFS	CRR: fit 82%, unfit 64%, frail 60%, 5y OS 46%, EFS 31%, DFS 41%; OS frail vs. fit: HR: 2.89 (95% CI, 1.39–6.13); OS unfit vs. fit: HR: 2.36 (95% CI, 1.20–4.64)
Marchesi 2013, Italy [60]	R	73	78	Age, ADL, CIRS‐G, geriatric syndrome	Fit 29%, unfit 38%, frail 33%	NA	OS, PFS	2 yr OS: fit and unfit better than frail b curative better than palliative in fit and unfit b

^aMultivariable Cox model.

^bStatistically significant.

Abbreviations: ADL, activities of daily living, CGA, comprehensive geriatric assessment; CI, confidence interval; CRR, complete response rate; CSS, cause‐specific survival; DFS, disease‐free survival; DR, dose reduction; EFS, event‐free survival; HR, hazard ratio; IADL, instrumental activities of daily living; NA, not available; OS, overall survival; p, prospective; PFS, progression‐free survival; R, retrospective; SMR, standardized mortality rate.

Table 2.

sCGA by the Fondazione Italiana Linfomi [56]

sCGA group	Domain
	Age	Comorbidities (CIRS‐G)	Residual ADLs	Residual IADLs
Fit	65–79	None scored 3–4 or <5 scored 2	6	8
Unfit	≥80	None scored 3–4 or 5–8 scored 2	5	6–7
Frail	≥80	≥1 scored 3–4 or >8 scored 2	≤4	≤5

CIRS‐G categories include (a) cardiac, (b) vascular, (c) hematologic, (d) respiratory, (e) ear, nose, throat, (f) upper GI, (g) lower GI, (h) hepatobiliary, (i) renal, (j) genitourinary, (k) musculoskeletal and skin; (l) neurologic, (m) endocrine/breast, and (n) psychiatric. Score 2 indicates need for daily treatment or moderate disability/morbidity; score 3 indicates poor control with first‐line therapy, significant disability, severe problem or organ failure. ADLs include independence in (a) bathing, (b) dressing, (c) toileting, (d) getting in/out of bed, (e) continence, (f) feeding. IADLs include independence in (a) using telephone; (b) shopping; (c) preparing own meals; (d) housework; (e) laundry; (f) using transportation; (g) taking own medication; (h) managing money.

Abbreviations: ADL, activity of daily living; CIRS‐G, Cumulative Illness Rating Score for Geriatrics; IADL, instrumental activity of daily living; sCGA, simplified comprehensive geriatric assessment.

Based on the above evidence, formal geriatric assessments may be useful to guide treatment decisions for older patients with DLBCL. However, a proper comparative randomized study against usual care, with direct treatment assignment based on CGA and sCGA, is necessary to establish whether its addition translates into survival or quality of life improvements. One barrier to CGA implementation is the requirement for time and multidisciplinary resources, which are not easily accessible to community oncology practices [65, 66, 67]. Some centers have established multidisciplinary geriatric oncology clinics to integrate CGA into a team‐based approach [66, 68]. This care model requires additional visits that may be challenging in DLBCL, which often requires urgent staging and treatment due to escalating burden of symptoms. As an alternative, self‐administered CGA has been proposed [68]. Screening tools, easily completed within minutes, can help identify vulnerabilities, thus limiting the number of patients needing a CGA, but their performance characteristics are inconsistent [38, 69, 70, 71]. Although these tools can be used according to the Society of International Geriatric Oncology [69], the 2018 American Society of Clinical Oncology (ASCO) guideline recommended integration of CGA into routine evaluations for all geriatric patients [38].

Management Strategies for Older Patients with DLBCL

Minimizing Early Toxicity: Prephase and Growth Factors

Prephase with prednisone and vincristine prior to full‐dose chemotherapy was first introduced in the randomized German High Grade Non‐Hodgkin's Lymphoma Study Group NHL‐B2 trial and then routinely used in the RICOVER‐60 trial and in other recent trials [72, 73, 74]. Prephase was associated with a lower treatment‐related mortality during cycle 1 (<2% vs. 5% without prephase) and lower risk of tumor lysis or febrile neutropenia [75, 76]. Prephase may allow patients to overcome lymphoma‐related functional impairment, facilitating subsequent therapy. We recommend prednisone (1 mg/kg) administered orally for up to 7 days with or without a single dose of vincristine to all older patients with significant symptomatic or radiographic burden of DLBCL (Fig. 2).

Figure 2.

The authors’ preferred management for older patients with DLBCL.Abbreviations: ASCT, autologous stem cell transplantation; CAR, chimeric antigen receptor; CGA, comprehensive geriatric assessment; DHL, double or triple‐hit lymphoma; DLBCL, diffuse large B‐cell lymphoma; EBV, Epstein‐Barr virus; EOL, end of life; FISH, fluorescence in situ hybridization (; GCB, germinal center B‐cell; HDT/ASCT, high‐dose therapy/autologous stem cell transplantation; HGBL, high‐grade B‐cell lymphoma; iPET, interim positron emission tomography; R±B, rituximab with or without bendamustine; R‐CEOP, rituximab, etoposide, vincristine, and prednisone; R‐CGVP, rituximab, cyclophosphamide, gemcitabine, vincristine, and prednisone; R‐CHOP, rituximab, cyclophosphamide, doxorubicin, vincristine, and prednisone; R‐COMP, rituximab, cyclophosphamide, vincristine, liposomal doxorubicin, and prednisone; RCOP, rituximab, cyclophosphamide, vincristine, and prednisone; sCGA, simplified comprehensive geriatric assessment; WHO, World Health Organization

Prophylactic white blood cell growth factors can decrease the risk of potentially life‐threatening neutropenic infections. According to the ASCO guidelines, they should be considered for patients aged 65 and over who receive immunochemotherapy for DLBCL, as the risk of febrile neutropenia in this group is up to 37% [77]. However, observational data indicate that one‐third of older U.S. adults treated with R‐CHOP‐like immunochemotherapy do not receive first‐cycle prophylaxis [78].

Initial Treatment: Fit Patients

R‐CHOP administered every 3 weeks for six cycles has been the cornerstone of first‐line therapy for fit older patients with DLBCL [79]. Extending treatment to eight cycles or dose‐intense administration every 14 days did not improve results and increased toxicity [73, 80, 81]. Recently, a randomized Chinese trial replaced doxorubicin 50 mg/m² with epirubicin 70 mg/m² for patients over age 60 showing no impact on survival [82]. However, doxorubicin was associated with a decrease in cardiac function at 3 years after treatment (p = .0009), which was not observed after epirubicin (p = .08), making epirubicin a potentially safer anthracycline for older patients. There is currently no evidence that modifying treatment according to the GCB/ABC phenotype or EBV status can alter prognosis. For fit patients with HGBL (particularly DHL), the DA‐EPOCH‐R can be considered, although comparative data against R‐CHOP are lacking, and this regimen did not provide any advantage over R‐CHOP in unselected DLBCL cohort [34, 83].

The strategy of abbreviated chemotherapy (3 cycles of R‐CHOP) with consolidative radiation therapy (RT) can minimize cumulative chemotherapy toxicity in early‐stage DLBCL [84]. In contrast to younger patients, who often prefer to avoid RT exposure, the overall balance of risk and benefit may favor the combined modality approach in older individuals. In one population‐based study, survival outcomes were identical between shorter chemotherapy (≤4 cycles) with RT and full‐course R‐CHOP (≥6 cycles), but RT recipients experienced less acute toxicity [85]. Recently, four cycles of R‐CHOP without RT for stage 1/2 DLBCL was evaluated in two randomized clinical trials, but patients over age 60 were not eligible for abbreviated strategy because of higher risk [86, 87]. As an alternative, the National Clinical Trials Network S1001 trial used an interim positron emission tomography (iPET)–directed strategy in 132 patients with median age 62 and nonbulky (<10 cm) DLBCL/DHL [88]. Those with negative (Deauville 1–3) iPET scans after three cycles of R‐CHOP received one additional dose of chemotherapy, whereas those with positive (Deauville 4–5) iPET received RT followed by ibritumomab tiuxetan (a radioimmunoconjugate that links rituximab with yttrium isotope ⁹⁰Y). Overall, 89% of patients were iPET‐negative, and 5‐year progression‐free survival (PFS) was 87% in the entire cohort (OS was 89%), without significant difference for iPET‐negative and iPET‐positive groups. Application of the S1001 strategy may thus decrease chemotherapy or radiation exposure and toxicity for many older patients with early‐stage DLBCL/HGBL.

Options for Unfit and Frail Patients

High‐level evidence for management of unfit and frail patients with DLBCL is lacking, as most studies to date had a single‐arm design. Table 3 summarizes efficacy and safety of attenuated regimens examined in phase II trials: R‐mini‐CHOP (with attenuated doses of vincristine: 1 mg, doxorubicin: 25 mg/m², and cyclophosphamide: 400 mg/m²) [89]; doxorubicin replaced by liposomal doxorubicin [90, 91]; doxorubicin replaced by mitoxantrone [92]; rituximab with cyclophosphamide, vincristine, and prednisone (R‐CVP) [92]; R‐CVP plus gemcitabine (R‐GCVP) [93]; rituximab with gemcitabine and oxalipatin (R‐GemOx) [94]; ofatumumab‐mini‐CHOP [74]; dose‐adjusted cyclophosphamide, doxorubicin, vincristine, and prednisone with rituximab [95]; and bendamustine and rituximab (BR) [96, 97, 98]. Their comparative effectiveness versus R‐CHOP is unknown, but retrospective data in unfit and frail patients suggest similar outcomes with various attenuated approaches [99]. R‐mini‐CHOP, R‐CNOP, or R‐GCVP have been recommended by guidelines as viable curative options [100]. Replacing rituximab with the third‐generation antibody obinutuzumab did not improve outcomes [101]. BR is often viewed as palliative, although in a prospective sCGA‐guided trial focused on frail patients, it resulted in 51% 2‐year OS, comparable to 59% observed after R‐mini‐CHOP [89]. Other studies reported less‐favorable results for bendamustine combinations, with median OS ≤1 year [98, 102].

Table 3.

Attenuated regimens as first‐line therapy for unfit or frail patients with DLBCL

Study and country	Regimen a	n	Age, median	Comorbidities	ECOG PS ≥2	Follow‐up, mo	RR/CR,%	Survival	Adverse effects
Visani 2008, Italy [90]	R‐COMP (every 21 d)	20	73	NA	45%	24	90/65	Not reported	G3/4: neutropenia 26% of cycles, febrile neutropenia 5%
Corazzelli 2011, Italy [91]	R‐COMP (every 14 d)	41	73	CKD 12%, high impact cardiac conditions 56%	32% iADL<4: 53%	27	73/68	4 yr DFS 72%, OS 67%	Therapy discontinuation 15%, cardiac G3–5: 17%
Peyrade 2011, France, Belgium [89]	R‐mini‐CHOP	150	83	NA	34%	20	73/62	2 yr PFS 47%, median 21 mo, 2 yr OS 59%, median 29 mo	G3/4 febrile neutropenia 7%, neutropenia 40%, thrombocytopenia 7%, anemia 9%
Peyrade 2017, France, Belgium [74]	Ofatumumab + mini‐CHOP	120	83	CCI: high 24%, intermediate 29%	29% iADL<4: 53%	27	68/56	2 yr PFS 57%, OS 65%	G3/4: 53%, neutropenia 21%, serious adverse events: 43%
Merli 2020, Italy [101]	Obinutuzumab + mini‐CHOP	34	82	sCGA unfit	6%	16	66/42	2 yr PFS 49%, OS 68%	G3/4 neutropenia 49%, infections 4%, thrombocytopenia 4%
Musolino 2011, Italy [95]	DA‐POCH‐R, 6–8 cycles	23	77	Preexisting cardiac diseases 18%	74%	22	90/57	3 yr EFS 54%, OS 56%	G3/4 neutropenia 48%, febrile neutropenia 13%, anemia 13%
Hainsworth 2010, U.S. [92]	R‐CNOP/R‐CVP 3 cycles + R 24m	51	78	R‐CVP: cardiac 57%, frailty/poor PS 43%	37%	48		2 yr PFS 71%, OS 72%	G3/4 neutropenia 41%, thrombocytopenia 12%
Fields 2014, U.K. [93]	R‐GCVP	62	77	LVEF<50%: 44%, comorbidities ≥2: 66%	50%	25	61/39	4 yr OS 67%	G3/4 neutropenia 48%, thrombocytopenia 31%, anemia 11%, infection 28%, cardiac 16%
Weidmann 2011, Germany [96]	BRq21 4–6 cycles	14	85	Ineligible for R‐CHOP: 71%	29%	NA	69/54	Median PFS 8 mo, OS 8 mo	G3/4 neutropenia 23%, G3: infection 10%
Park 2016, U.S. [98]	BRq21 8 cycles	23	80	NA, ineligible for R‐CHOP: 100%	52%	29	78/52	Median PFS 5 mo, OS 10 mo	G3/4 neutropenia 17%, anemia 26%, thrombocytopenia 17%, sepsis 4%
Storti 2018, Italy [97]	BRq28 4–6 cycles + R 2 cycles	45	81	Cardiac 40%, renal 6%; 100% ineligible for R‐CHOP	36%: frail 22%, unfit 78%	33	62/53	2 yr PFS 38%, median 10 mo, 2 yr OS 51%, median 30 mo	G3/4 51%, neutropenia 38%
Flinn 2019, U.S. [102]	Bendamustine + ofatumumab 6 cycles q21 d	21	83	NA	NA	NA	91/33	Median PFS 9 mo, OS 12 mo	G3/4 thrombocytopenia (14%), neutropenia (10%), diarrhea (10%), vomiting (10%), dehydration (10%)
Shen 2018, China [94]	R‐GemOx	60	75	Common comorbidities: 38%, CCI>:3: 100%	45%	45	75/47	3 yr PFS 49%, OS 65%	G3/4 neutropenia 15%

^aSix cycles unless specified.

Abbreviations: BR, bendamustine and rituximab; CCI, Charlson comorbidity index; CHOP, cyclophosphamide, vincristine, doxorubicin, and prednisone; CKD, chronic kidney disease; CNOP, cyclophosphamide, mitoxantrone, vincristine, and prednisone; COMP, cyclophosphamide, vincristine, liposomal doxorubicin, and prednisone; CR, complete response; CVP, cyclophosphamide, vincristine, and prednisone; DA‐EPOCH, dose‐adjusted etoposide, prednisone, vincristine, cyclophosphamide, doxorubicin; DFS, disease‐free survival; ECOG‐PS, Eastern Cooperative Oncology Group performance status; G, grade; GCVP, gemcitabine, cyclophosphamide, vincristine, and prednisolone; GemOx, gemcitabine and oxaliplatin; iADL, instrumental activities of daily living; LVEF, left ventricular ejection fraction; NA, not available; OS, overall survival; PFS, progression‐free survival; R, rituximab; RR, response rate.

Mitigating the Risk of Relapse

Approximately 40% of older patients with DLBCL will experience either primary refractory disease or a recurrence after initial therapy, leaving substantial room for improvement. So far, combinations of R‐CHOP with other agents, including lenalidomide [103], bortezomib [104, 105], and ibrutinib [5], have not improved outcomes, even selecting patients by cell of origin. In particular, adding ibrutinib to R‐CHOP resulted in worse PFS and OS among patients older than 60 with an excess of serious adverse events, illustrating the challenge of intensifying R‐CHOP even with rationally selected targeted agents [5].

Consolidation (after achieving a response to chemotherapy) and maintenance have been investigated as strategies to mitigate the risk of recurrence in DLBCL, but so far these efforts have not changed the standard of care. Maintenance rituximab does not provide any benefit after R‐CHOP in DLBCL [106]. Ibritumomab tiuxetan was examined as consolidation after four cycles of R‐CHOP in a single‐arm study for patients older than 60, with 85% PFS at 2 years [107]. However, this approach has not moved forward to a randomized setting and is rarely applied because concerns about secondary leukemia, even though the incidence of treatment‐related myeloid neoplasms in this setting is low (about 4% over 10 years) [108, 109].

Maintenance lenalidomide is one approach that was examined in a randomized study. The phase III REMARC trial demonstrated improved PFS for patients age 60–80 who achieved complete or partial response to R‐CHOP and received 2 years of maintenance lenalidomide (hazard ratio [HR], 0.77; 95% confidence interval [CI], 0.54–0.93) [110]. However, the benefit did not translate into a better OS (HR, 1.22; 95% CI, 0.86–1.72), and grade 3/4 adverse effects were higher with lenalidomide than placebo (56% vs. 22%). Surprisingly, the PFS benefit was only observed in the GCB subgroup, challenging the concept of specific efficacy of lenalidomide in the ABC‐DLBCL.

Novel Options for Relapsed or Refractory Disease

Only a minority of elderly patients with relapsed or refractory DLBCL are eligible for the standard curative approach using salvage chemotherapy followed by high‐dose therapy and autologous stem cell transplantation (ASCT). Even with ASCT, older patients experience higher mortality and higher risk of second relapse, with long‐term disease‐free survival of less than 20% [111, 112, 113]. Therefore, choosing the intensive salvage therapy should be preceded by a thorough discussion of patient's individualized risks, values, and goals of care. When this route is not an option, many low‐intensity strategies can provide a palliative benefit. The most common historical cytotoxic regimens include R‐GemOx and BR, which result in response rate (RR) of 50%–60% and median PFS of 5–7 months [114, 115, 116, 117].

Recently, more efficacious options have emerged. In a randomized phase II trial, BR in combination with the CD79b‐targeting antibody–drug conjugate polatuzumab vedotin (1.8 mg/kg intravenously on day 2 of each 21‐day cycle) resulted in higher RR (45% vs. 18%) and complete response (CR) rate (40% vs. 18%) and longer PFS (9.5 vs. 3.7 months) and OS (12 vs. 5 months) than BR among patients with ASCT‐ineligible relapsed or refractory DLBCL and median age 69 (nearly half of whom had received ≥3 lines of prior therapy), at the expense of modestly increased cytopenias and no significant increase in infections [118]. Polatuzumab vedotin at 2.4 mg/kg every 3 weeks also showed single‐agent activity (or in combination with rituximab) in a phase I study, with RR of 56% and median PFS of 5 months [119]. Neuropathy is a common adverse effect of this agent but typically presents at grade 1/2 and reverses upon discontinuation. The second novel option is the combination of anti‐CD19 monoclonal antibody tafasitamab with lenalidomide [120]. Although both drugs have modest efficacy as monotherapy, in a phase II trial of 81 patients with median age 72 (range, 62–76), their combination produced a 60% RR, 43% CR rate, and median PFS of 12 months (median OS not reached) [120]. Toxicity was limited, including febrile neutropenia in 12%, pneumonia in 6%, and pulmonary embolism in 4%. Importantly, 93% of patients who achieved CR remained in remission at 18 months. Prolonged (>5 years) remissions were also observed in about 23% of older patients with relapsed or refractory DLBCL receiving lenalidomide and rituximab in a phase II trial, although overall RR was only 35% [121]. Single‐agent selinexor (approved by the U.S. Food and Drug Administration) as well as (off‐label) lenalidomide or BCR inhibitors can be offered as subsequent lines of therapy, but their efficacy is limited and the risk/benefit ratio should be examined carefully, ideally in the context of emerging molecular biomarkers [23, 24, 122].

Immune effector cell therapy using autologous chimeric antigen receptor (CAR) T‐cells has transformed the field of lymphoma therapy, with two products (axicabtagene ciloleucel and tisagenlecleucel) currently approved for treatment of relapsed or refractory DLBCL after at least two lines of therapy. CAR T‐cell infusions provided high rates of initial response (60%–80%) as well as durable (>18 months) complete remissions for about a third of all patients enrolled in the registration trials [123, 124]. Data on participants in the pivotal ZUMA‐1 trial of axicabtagene ciloleucel who were age 65 or older (up to 76) indicated similar in vivo expansion of CAR T cells compared with younger participants and numerically higher RR (92% vs. 61%, respectively) and CR (75% vs. 53%) [125]. Incidence of treatment‐related mortality (4%), grade ≥3 cytokine release syndrome (CRS; 7% and 12%, respectively) infections, and cytopenias were similar in both age groups, but older patients experienced more grade ≥3 neurotoxicty (44% and 28%, respectively). Early real‐world experience indicates that cellular therapy can be delivered to older patients with comorbidities, who would not have been eligible for the registration trials [126, 127]. In one case series from a tertiary institution, geriatric patients receiving commercial CAR T‐cell therapy were evaluated using a multidimensional geriatric assessment that did not appear to be prognostic with this treatment modality [128]. Outcomes and toxicities were not majorly different from younger patients, but survival was better than for geriatric patients not treated with CAR T cells. The need for administration in specialized centers, lymphodepleting chemotherapy potentially leading to prolonged cytopenias, several weeks required for the CAR T‐cell production, significant rates of the CRS and neurotoxicity that are difficult to tolerate for older individuals, and costs and reimbursement issues constitute barriers to widespread application of this potentially curative modality for relapsed or refractory DLBCL.

Emerging Approaches to First‐Line Therapy and Palliative Care

Despite lower toxicity of attenuated regimens, many patients older than 80 or in the unfit and frail categories are just unable to receive, or do not accept, cytotoxic chemotherapy. Currently, these patients are offered supportive care alone or rituximab monotherapy. Rituximab is associated with longer OS than no treatment at all and anecdotally can produce durable remissions in combination with RT for patients with localized disease [7]. The need for noncytotoxic options in elderly patients provides an opportunity to investigate chemotherapy‐free approaches, of which strategies using either combinations of targeted agents or immunotherapy alone appear to be most promising. Researchers from MD Anderson have administered 2 months of rituximab, lenalidomide, and ibrutinib (RLI) to 60 patients (28% age ≥70) with newly diagnosed non‐GCB DLBCL, followed by a combination of RLI with CHOP or EPOCH chemotherapy [129]. After RLI alone, 86% of patients obtained a response, and 36% were in CR, raising a possibility that this combination might provide a noncytotoxic approach for the non‐GCB subtype. In a preliminary experience from a phase II trial of 24 frail (by sCGA) patients treated with first‐line rituximab and lenalidomide (20 mg per day for 21 days, every 28 days) for 6 months, followed by lenalidomide 10 mg per day up to total 12 months, RR was 50%, including 17% CR, although four treatment‐related deaths were observed [130]. An international phase I/II trial (NCT03677154) is specifically focused on patients with newly diagnosed DLBCL who are older than 80 or those over age 60 with functional impairments or comorbidities precluding R‐CHOP. Patients receive single‐agent mosunetuzumab, a bispecific antibody that concurrently engages CD20 on malignant B‐cells and CD3 on cytotoxic T cells to effect cellular immune attack against DLBCL. The initial results indicate that up to 58% of patients achieve response, and 42% achieve CR on mosunetuzumab monotherapy. Toxicity was limited (no grade >2 CRS or neurotoxicity, no treatment‐related deaths) and appears lower than after CAR T‐cells [131]. If sufficient rates of durable CR are confirmed, mosunetuzumab may emerge as an important novel chemotherapy‐free approach in DLBCL. This potential was recognized by the “breakthrough therapy” designation for mosunetuzumab by the U.S. Food and Drug Administration in July 2020.

Another important emerging aspect is early institution of supportive and palliative care for patients with aggressive lymphomas. Studies suggest that 33% of older patients in the U.S. who die from lymphoma pass away in an acute care hospital, 37% are admitted to the intensive care unit within a month before death, and 43% never access hospice services [132]. Barriers to improvement in the end‐of‐life care in this population exist from the perspective of clinicians (difficulties with recognizing the terminal phase of the disease or addressing palliative aspects of therapy), patients (prognostic discordance or unrealistic expectations based on data from younger populations), and systems (limited availability or inadequate reimbursement for specialty palliative care) [133]. However, in recent years, early palliative care and attention to the quality of end‐of‐life care in hematologic malignancies have been more emphasized and subject to rigorous research [134, 135, 136, 137].

Conclusion

Although the management of DLBCL in older individuals continues to present a challenge, it may paradoxically motivate the lymphoma community to develop novel, less toxic approaches and to personalize therapy according to the disease biology, patients' baseline health, and individualized goals of care. Clinicians should make use of the supportive care components delineated above: corticosteroid‐based prephase, growth factor support, palliative care, and the newly approved treatment options for relapsed or refractory disease. Higher‐level evidence of benefit and value of geriatric assessments is needed to facilitate changes at the health policy level that would support routine use of CGA with adequate reimbursement. Of critical importance is enrolling older patients in clinical trials to thoroughly examine the efficacy and toxicity of emerging modalities. Chemotherapy‐free targeted or immunotherapy‐based strategies that encompass bispecific antibodies like mosunetuzumab, small molecules, antibody–drug or radioimmunoconjugates, and autologous and allogeneic CAR T‐cells are likely to advance to the first‐line setting, where their true benefit over standard approaches will need to be ultimately tested in randomized trials. Success of these approaches may depend on identification of accurate biomarkers. Future studies should also incorporate evaluation of quality of life and patient‐centered outcomes to better understand the impact of new therapies on older patients, as they may assign different relative value to the possibility of cure versus toxicities and burden of therapy.

Author Contributions

Conception/design: Mengyang Di, Scott F Huntington, Adam J. Olszewski

Manuscript writing: Mengyang Di, Scott F Huntington, Adam J. Olszewski

Final approval of manuscript: Mengyang Di, Scott F Huntington, Adam J. Olszewski

Disclosures

Scott F. Huntington: Celgene, Bayer, Genentech, Pharmacyclics, Novartis, AbbVie (C/A); DTRM Biopharm, Celgene, TG Therapeutics (RF); Adam J. Olszewski: TG Therapeutics, Genentech, Spectrum Pharmaceuticals, and Adaptive Biotechnologies (RF). The other author indicated no financial relationships.

(C/A) Consulting/advisory relationship; (RF) Research funding; (E) Employment; (ET) Expert testimony; (H) Honoraria received; (OI) Ownership interests; (IP) Intellectual property rights/inventor/patent holder; (SAB) Scientific advisory board