Diffuse large B-cell lymphoma (DLBCL)

Epidemiology and Workup

• Most common lymphoma: about 30,000 new US cases per year, roughly 30% of NHL; median age 65; aggressive with symptoms over weeks to months. Overall 60 to 75% are cured.
• Morphology: large cells (4 to 5 times a normal lymphocyte) with a diffuse growth pattern.
• Immunophenotype: CD19, CD20, CD22, CD45, CD79 positive; CD3 negative.
• Workup: excisional biopsy (FNA insufficient, architecture needed); IHC (CD20, CD79a, BCL2, BCL6, MUM1, CD10, MYC, Ki-67); FISH for MYC, BCL2, BCL6 (rule out HGBL double/triple-hit); HBV/HCV/HIV; PET-CT; LDH, beta-2 microglobulin; LP if high risk for CNS; echo before anthracycline.

Cell of Origin (Hans Algorithm)

• Hans criteria use CD10, then BCL6, then MUM1 to separate germinal center B-cell (GCB) from non-GCB (an IHC surrogate for the gene-expression-defined ABC) subtype.
• GCB: CD10+ (regardless of BCL6/MUM1), or CD10−/BCL6+/MUM1−. Non-GCB: CD10−/BCL6−, or CD10−/BCL6+/MUM1+.
• Non-GCB (IHC surrogate for ABC): always CD10−; either BCL6− (any MUM1) or BCL6+/MUM1+.
• GCB survival is better than ABC.
• Biology: in GCB, mutations of BCL6, histone acetyltransferases, and EZH2 drive a repressed transcriptional state. In ABC, mutations in the B-cell receptor pathway lead to unchecked NF-κB activation.
• Gene expression profiling (GEP) is more accurate than IHC for cell of origin; the LymphGen molecular classifier further subdivides DLBCL by genetic clusters. ABC (non-GCB) has historically done worse, but pola-R-CHP is narrowing the gap.

Prognostic Indices

• IPI (mnemonic APELS): Age >60, stage III/IV (Stage), LDH elevated, Performance ECOG ≥2, Extranodal sites >1; score 0 to 5, grouping into low (0 to 1), low-intermediate (2), high-intermediate (3), high (4 to 5).
• NCCN-IPI is better: more granular age and LDH bands and specific extranodal sites.
• R-IPI: collapses IPI into 3 risk groups (very good 0, good 1 to 2, poor 3 to 5) in the rituximab era.

Frontline Treatment

• Historical evolution: 1st-generation CHOP (National High Priority Lymphoma Study) gave CR about 44% and 3-year OS about 54%. Second and third-generation regimens (m-BACOD, ProMACE-CytaBOM, MACOP-B) initially reported OS 55 to 65% but were not superior in randomized comparison.

R-CHOP q3wk x6

• Rituximab 375 mg/m².
• Cyclophosphamide 750 mg/m², an alkylating agent; hemorrhagic cystitis risk (mainly at high doses; mesna is not routinely needed at CHOP doses).
• Doxorubicin (hydroxydaunorubicin) 50 mg/m², a topoisomerase II inhibitor; cardiomyopathy risk, check echo.
• Vincristine (Oncovin) 1.4 mg/m² (max 2 mg), an antimicrotubular agent; neuropathy and constipation.
• Prednisone 100 mg PO days 1 to 5.
• Cures about 60% of patients overall.

Rituximab (anti-CD20)

• Mechanism: activation of the complement cascade generating the membrane attack complex (complement-mediated cytotoxicity); phagocytosis and antibody-dependent cell-mediated cytotoxicity (ADCC); interaction with NK cells via FcRIII and complement receptor 3.
• Infusion reactions: common first-infusion reactions are mainly cytokine-release/infusion-related (worse with high tumor burden), not classic allergy; true hypersensitivity can also occur. Premedicate with acetaminophen and diphenhydramine; have hydrocortisone available (also useful: dexamethasone, loratadine, famotidine).
• Side effects: HBV reactivation (always check before treatment, give prophylaxis if evidence of exposure); TLS with leukocytosis (malignant lymphocytes >25,000); PML from JC virus; hypogammaglobulinemia.
• No maintenance rituximab or autoSCT in patients who achieve CR after initial chemotherapy; observation only.

Pola-R-CHP (new standard for IPI ≥2)

• Polatuzumab vedotin (anti-CD79b ADC with MMAE) plus rituximab, cyclophosphamide, doxorubicin, and prednisone; vincristine is replaced by pola.
• POLARIX (Tilly NEJM 2022): 2-year PFS 76.7% vs 70.2% for R-CHOP, HR 0.73. FDA April 19, 2023 for previously untreated IPI ≥2 DLBCL/HGBL.

Other frontline trials

• ECOG-ACRIN 1412: lenalidomide plus R-CHOP, 3-year PFS 73% vs 61% (HR 0.66), 3-year OS 83% vs 75% (lenalidomide 25 mg days 1 to 10).
• ROBUST (Nowakowski): lenalidomide plus R-CHOP (lenalidomide 15 mg days 1 to 14); did not meet endpoint.
• Negative trials: PHOENIX (ibrutinib plus R-CHOP, Younes), GOYA (obinutuzumab-CHOP vs R-CHOP, Sehn), CAVALLI (venetoclax plus R-CHOP, single-arm phase II vs matched GOYA control; Morschhauser), ALLIANCE/CALGB 50303 (R-CHOP vs R-EPOCH, Bartlett).
• R-EPOCH-DA: dose-adjusted EPOCH plus rituximab is used for select cases, primary mediastinal B-cell lymphoma, HIV-associated DLBCL/HGBL, and double/triple-hit (variable benefit).

Limited-stage (I to II, non-bulky)

• Practical approach: R-CHOP x3, interim PET, if CR give either 1 more cycle of R-CHOP or involved-site RT (ISRT). R-CHOP x3 then ISRT is category 1; R-CHOP x4 lets you omit radiation in a younger patient with disease near the chest.

Elderly and frail

• Mini-CHOP / R-mini-CHOP (roughly 50% dose) for patients over 80 years old; the Peyrade et al phase II study established R-mini-CHOP as tolerable and effective in patients over 80. Use aggressive supportive care and formal geriatric/fitness assessment.

Cardiac dysfunction

• R-GCVP (rituximab, gemcitabine, cyclophosphamide, vincristine, prednisone) replaces the anthracycline with gemcitabine for patients who cannot receive doxorubicin.

CNS Prophylaxis and CNS-IPI

• High risk (CNS-IPI 4 to 6, or kidney/adrenal involvement): give CNS prophylaxis.
• Options: intrathecal methotrexate and/or cytarabine for 4 to 6 cycles during or after initial therapy, or high-dose IV methotrexate (3 to 3.5 g/m² for 2 to 4 cycles) or cytarabine (Schmitz JCO 2016; Puckrin ASH 2020 abstract 477; Orellana-Noia ASH 2020 abstract 478).
• Caveat: recent data (including retrospective analyses and CALGB 50303 experience) suggest limited benefit of prophylaxis; reserve it for the highest CNS-IPI risk plus the specific anatomic sites above.

High-Grade B-Cell Lymphoma (HGBL)

• Spectrum by FISH: (A) HGBL MYC/BCL6+, (B) Burkitt, (C) HGBL MYC/BCL6/BCL2+ (triple-hit), (D) DLBCL MYC/BCL2+.
• Double-hit lymphoma: MYC plus BCL2 rearrangement (WHO 2022: DLBCL/HGBL with MYC and BCL2 rearrangements); MYC plus BCL6 cases, formerly included, are now DLBCL NOS or HGBL NOS (WHO 2022) or provisional HGBL with MYC and BCL6 rearrangements (ICC 2022). About 5 to 10% of DLBCL; CNS disease in about 13%; a disease of the elderly (average age approaching 71); not as sensitive to R-CHOP as to intensive chemotherapy. Triple-hit is MYC plus BCL2 plus BCL6.
• Double-expresser lymphoma: 20 to 25% of new DLBCL. MYC and BCL2 protein overexpression by IHC (MYC ≥40%, BCL2 ≥50% cutoff) without translocations. Worse prognosis than ordinary DLBCL but not as bad as double hit; treated like DLBCL.
• WHO 2022 naming: the formal double-hit entity is "DLBCL/HGBL with MYC and BCL2 rearrangements"; WHO 2022 places MYC/BCL6 cases within DLBCL NOS or HGBL NOS, whereas ICC 2022 keeps HGBL with MYC and BCL2 rearrangements and recognizes HGBL with MYC and BCL6 rearrangements as a provisional entity.
• Treatment of double/triple-hit: R-CHOP does not work; do not consolidate with transplant. Use R-hyperCVAD, DA-EPOCH-R, or R-CODOX-M/IVAC. CNS prophylaxis (HD-MTX or IT). POLARIX included HGBL DH/TH; pola-R-CHP data are emerging.

Relapsed / Refractory DLBCL

• Relapses occur in about 20 to 40%. Confirm with biopsy and assess transplant fitness.

Second-line salvage chemotherapy (about 60% response rate)

• Rituximab plus: DHAP (dexamethasone, cisplatin, cytarabine), DHAX (dexamethasone, cytarabine, oxaliplatin), GDP (gemcitabine, dexamethasone, cisplatin or carboplatin), ICE (ifosfamide, carboplatin, etoposide), ESHAP (etoposide, methylprednisolone, cytarabine, cisplatin), GemOx (gemcitabine, oxaliplatin), MINE (mesna, ifosfamide, mitoxantrone, etoposide).
• Late relapse (>12 months): salvage R-chemo then ASCT if chemosensitive (CR/PR by PET).

ASCT evidence

• Parma trial (NEJM 1995): transplant arm improved 5-year EFS (46% vs 12%) and OS (53% vs 32%) in chemosensitive relapse.
• CORAL (Gisselbrecht JCO 2010): R-DHAP vs R-ICE x3 then ASCT; ORR 63% vs 63.5% with no PFS or OS difference. GCB-DLBCL had improved outcome with R-DHAP vs R-ICE by gene expression profiling (3-year PFS 100% vs 27%).

CAR-T therapy

• Second line (primary refractory or relapse <12 months): CAR-T is preferred. ZUMA-7 (axi-cel) and TRANSFORM (liso-cel) both showed superiority over salvage chemo plus ASCT. Axi-cel and liso-cel are both FDA-approved 2nd line for patients refractory or relapsing within 12 months.
• Third line and beyond (pivotal trials): axicabtagene ciloleucel ZUMA-1 (ORR 82%, CR 54%), tisagenlecleucel JULIET (ORR 53%, CR 40%), lisocabtagene maraleucel TRANSCEND NHL 001 (ORR 73%, CR 53%); ORRs across products 52 to 82% (CR 40 to 54%).
• Products:
  • Axicabtagene ciloleucel (Yescarta): CD28 costim; FDA Oct 2017 for R/R DLBCL after ≥2 lines, expanded to 2nd line Apr 2022 per ZUMA-7. Higher CRS/ICANS.
  • Tisagenlecleucel (Kymriah): 4-1BB costim; FDA May 2018; less CRS but lower CR.
  • Lisocabtagene maraleucel (Breyanzi): 4-1BB costim; FDA Feb 2021; expanded to 2nd line Jun 2022 per TRANSFORM. Better safety profile.

Bispecific antibodies (after ≥2 prior lines, post-CAR-T or ineligible)

• Glofitamab (Columvi): CD20 x CD3 fixed duration (12 cycles); FDA Jun 15, 2023 for R/R DLBCL ≥2 lines (NP30179, Dickinson NEJM 2022). STARGLO (Abramson Lancet 2024) glofit plus GemOx vs R-GemOx improved OS, but FDA ODAC voted 8 to 1 against applicability to the US population (geographic enrollment imbalance); FDA issued a complete response letter July 2025, so the combination is not approved in the US.
• Epcoritamab (Epkinly): CD20 x CD3 subcutaneous; FDA May 19, 2023 for R/R DLBCL ≥2 lines (EPCORE NHL-1, Thieblemont JCO 2023; ORR 63%, CR 39%); label expanded Jun 26, 2024 for R/R FL ≥2 lines. Step-up dosing weeks 1 to 2 to mitigate CRS.

Other R/R agents (transplant-ineligible, generally palliative goal)

• R-GemOx or gemcitabine-based regimens.
• Brentuximab vedotin for CD30+ disease (ORR about 44%, median DOR about 6 months).
• Brentuximab vedotin plus lenalidomide and rituximab: ECHELON-3, OS HR 0.63 (median 13.8 vs 8.5 months) regardless of CD30 status; FDA Feb 2025 for R/R LBCL after ≥2 lines, ineligible for ASCT or CAR-T.
• Lenalidomide monotherapy (ORR about 25%, mainly non-GCB).
• Tafasitamab (anti-CD19) plus lenalidomide: L-MIND, ORR 60%, CR 42.5%; FDA Jul 2020 for R/R DLBCL not transplant candidates (tafasitamab 12 mg/kg with lenalidomide 25 mg days 1 to 21).
• Ibrutinib (BTKi): ORR about 25% overall (37% in ABC, 5% in GCB) (CD79B and MYD88 mutations most likely to respond; CARD11 and TNFAIP3 unlikely).
• Polatuzumab vedotin plus bendamustine and rituximab (BR-Pola): ORR 45% vs 17% for BR, CR 40% vs 17%, median OS 12.4 vs 4.7 months; FDA Jun 2019 (now largely moved to frontline via POLARIX, still usable in R/R).
• Loncastuximab tesirine (Zynlonta): CD19 ADC delivering a pyrrolobenzodiazepine dimer DNA cross-linker; LOTIS-2 ORR 48.3%, CR 24%; FDA Apr 2021 for R/R DLBCL ≥2 lines.
• Selinexor (XPO1/SINE inhibitor): blocks nuclear export, retaining tumor suppressors (p53, p21) in the nucleus and reducing oncoprotein (MYC, BCL2, BCL6) levels by inhibiting XPO1-mediated nuclear export of their mRNA complexes; 60 mg PO days 1 and 3 weekly; SADAL/Kalakonda Lancet Haematol 2020 ORR 27%, CR 12%, median DOR 9.3 months; FDA accelerated approval Jun 2020 for R/R DLBCL ≥2 lines, withdrawn Apr 2026 (confirmatory trial not completed; myeloma indications unaffected).

High-Yield Pearls

• CD3 is negative in DLBCL (a B-cell tumor).
• ABC always CD10−; GCB is CD10+ or BCL6+/MUM1−. GCB does better.
• Pola-R-CHP (POLARIX, HR 0.73) is the frontline standard for IPI ≥2.
• Double/triple-hit: R-CHOP fails; use DA-EPOCH-R, R-hyperCVAD, or R-CODOX-M/IVAC, and do not consolidate with transplant.
• CNS-IPI 4 to 6 or kidney/adrenal: give CNS prophylaxis (HD-MTX or IT).
• Cardiac dysfunction: swap the anthracycline, use R-GCVP.
• Second-line, early relapse (<12 months): CAR-T beats salvage chemo plus ASCT (ZUMA-7, TRANSFORM).