Richter transformation of chronic lymphocytic leukaemia: a British Society for Haematology Good Practice Paper.

Abstract

This Good Practice Paper was compiled according to the BSH process at https://b-s-h.org.uk/media/16732/bsh-guidance-development-process-dec-5-18.pdf and represents best practice in both teaching and district hospitals in the UK. The Grading of Recommendations Assessment, Development and Evaluation (GRADE) nomenclature was used to evaluate levels of evidence and to assess the strength of recommendations. The GRADE criteria can be found at http://www.gradeworkinggroup.org. Recommendations included a systematic review of published English language literature from publication of previous British Society for Haematology (BSH) Management of Chronic Lymphocytic Leukaemia Guidelines (2012) up to 03/2021. In addition, there are some additional pertinent references and a consensus of expert opinion where no published data are available. PubMed, MEDLINE, EMBASE, Cochrane databases and Web of Science were searched using the preliminary search terms: chronic lymphocytic leukaemia OR CLL AND Richter’s transformation OR Richter’s syndrome OR transformed/developed/progressed to aggressive lymphoma/high-grade lymphoma/DLBCL/Hodgkin lymphoma. Systematic reviews, including guidelines from other countries, prospective clinical trials, observational studies, i.e. cohort or case–control studies, expert reviews and opinions and case series were considered and reviewed as appropriate. Review of the manuscript was performed by the BSH Guidelines Committee, Haemato-oncology Task Force, Haemato-oncology sounding board of BSH. Richter transformation (RT) is the development of an aggressive lymphoma arising on the background of chronic lymphocytic leukaemia (CLL).1 RT is uncommon, challenging to treat, very distinct from de novo diffuse large B-cell lymphoma (DLBCL) and requires specific guidance. RT occurs in 2–10% of CLL patients, usually during the disease course rather than at presentation, representing a transformation rate of 0·5–1% per CLL patient per year.2-5 RT should be suspected when a CLL patient develops one or more new ‘B symptoms’, asymmetric, rapidly progressive lymphadenopathy, or a sudden lactate dehydrogenase (LDH) rise. RT presents as DLBCL-RT in ~90%, but can present as Hodgkin lymphoma (HL-RT; ~10%) or rarely as histiocytic/dendritic cell sarcoma or other forms of lymphoma (<1%).6 An important consideration is whether RT is clonally derived from or unrelated to the original CLL, as these two states have distinct clinical and pathological characteristics. Strictly speaking, RT refers to clonally related cases with Richter syndrome, but since the clonal origin is often unknown, the term RT includes all cases. Clonally related RT has an aggressive course, higher rates of treatment resistance and TP53 aberrations compared with clonally unrelated disease, where outcomes are more akin to de novo DLBCL. A clonal relationship is more common in DLBCL-RT (70–80% of cases), compared with HL-RT where it is seen in 30–40%.5 Genetic clonality studies [i.e. sequencing analysis of immunoglobulin heavy chain variable region (IgHV) genes] are not currently routinely performed in practice at most institutions (Fig 1). Histopathological assessment remains the gold standard to confirm the diagnosis of RT. DLBCL-RT is characterised by the presence of large neoplastic B cells with either centroblastic (60–80% of cases) or immunoblastic (20–40% of cases) morphology. Currently, DLBCL-RT is distinguished from ‘accelerated’ CLL with expanded confluent proliferation centres as the management of the entities is distinct.7 Adherence to two key World Health Organisation diagnostic criteria is critical: (i) DLBCL-RT is typified by the presence of sheets of large B-lymphoid cells with a nuclear size equal to or exceeding that of normal macrophage nuclei or more than twice the size of a normal lymphocyte; and (ii) these cells must show a diffuse growth pattern, and not be present in small foci throughout the neoplasm.1 These criteria can be subjective, and review of adequate biopsy specimens by at least two independent pathologists is desirable.7 However, these are subjective criteria, and there is increasing evidence from genetic studies that poor-risk CLL, accelerated CLL and RT are part of a continuum that is driven by underlying genomic instability. Most DLBCL-RT (80%) are classified as activated B-cell type with 20% germinal centre B-cell-like. Many studies have conclusively demonstrated that RT is genetically distinct from de novo DLBCL. TP53 aberrations are seen in ~60%, with alterations in MYC (40%), CDKN2A (30%) and NOTCH1 (30%) also common. Mutations of one or more of these genes are present in 90%.8-10 NOTCH1 mutations are associated with ‘subset 8’ of the B-cell receptor (BCR) in CLL patients, which exhibits autonomous BCR signalling and responsiveness to multiple auto-antigens and other micro-environmental immune stimuli, and higher rates of RT.11 A recent study of paired samples from peripheral blood CLL and tissue RT phases that combined whole genome sequencing and RNAseq identified defects in the DNA damage response (DDR) as the most discriminative feature in RT. Furthermore, pathway-based clonal deconvolution analysis showed that genes in the MAPK and DDR pathways also demonstrated the highest clonal expansion probability. Together, these data point towards disruption of signalling and DDR as dominant drivers of transformation.10 In contrast, HL-RT is characterised by the presence of CD30+/CD15+/CD20− classical Reed–Sternberg cells on a background of small T cells, histiocytes, eosinophils and plasma cells.12 Most HL-RT are clonally unrelated and Epstein–Barr virus (EBV)-positive, representing de novo, EBV-driven lymphoma. Little data exist regarding the genetic hallmarks of HL-RT. Susceptibility to infections is well recognised in CLL patients and can occur in early-stage disease. Therefore, it is important to consider infections that may mimic RT presentation, especially EBV or cytomegalovirus (CMV), in the differential diagnosis. Using an SUVmax cut-off >5, positron emission tomography (PET) detected RT with a high sensitivity (91%) but low specificity (50%) in a retrospective study of 37 patients previously treated with chemotherapy +/− immunotherapy.13 This study demonstrated a high negative predictive value (NPV; 97%) for RT using this cut-off. The same SUVmax cut-off was applied to 332 patients, of whom 95 had histologically-proven RT.14 Sensitivity and NPV for RT detection were 88% and 92% respectively. However, positive predictive value (PPV; 47%) and specificity (38%) remained low: of the 332 patients, 117 were diagnosed with histologically aggressive CLL without RT and 72% of these cases had SUVmax > 5. Using an SUVmax cut-off of >10 improved specificity (95%) with high sensitivity maintained (91%) in a study of 240 patients.15 The sensitivity and specificity of an SUVmax cut-off > 10 may be diminished with targeted inhibitors. In a post hoc analysis of a phase II study of venetoclax in BCR inhibitor-exposed patients, the sensitivity of an SUVmax cut-off > 10 for detecting RT was 71%, with a specificity of only 50%. Fourteen of 19 patients with SUVmax > 10 had CLL with no RT.16 Furthermore, in a Mayo study of BCRi-exposed patients, an SUVmax > 5 again demonstrated high sensitivity of 96% but low specificity.17 PPV of an SUVmax > 5 or > 10 remained low at 51% and 67% respectively. Taken together, histological confirmation remains essential to establish RT. PET may help target the biopsy site to the area with highest 18F-fludeoxyglucose (18F-FDG) uptake and is valuable in excluding RT without biopsy when SUVmax is <5. Two prognostic score systems predict overall survival (OS). First, a clinical RT score was derived from a multivariate analysis of 130 patients who received chemotherapy or chemoimmunotherapy. Five factors independently correlated with shorter survival: performance status > 1, LDH > 1·5 x upper limit of normal, platelets <100 × 109/l, tumour bulk >5 cm, and more than one prior therapy.18 When stratified into four groups according to these factors, median OS ranged from 0·33 to 1·12 years. The score has been validated in other series.2, 19 The second prognostic system5 used Eastern Cooperative Oncology Group performance status, achievement of complete remission (CR) with induction therapy and TP53 status. Median OS was 8 and 25 months for high and intermediate risk patients respectively but the five-year OS was 70% for low-risk patients. This study established that clonally unrelated RT is clinically and biologically distinct from clonally related RT and is characterised by a survival akin to de novo DLBCL (median OS 62·5 vs 14·2 months; P = 0·017). For those with proven RT, 18F-FDG uptake by PET scan may add prognostic information. An SUVmax > 10 was significantly associated with worse OS in a retrospective study (6 vs 21 months for SUVmax < 10, P = 0·015), and patients with advanced stage had poorer OS than limited stage (5·1 vs 13·8 months, P = 0·04).14 The prognostic significance of number of prior treatment lines has been demonstrated in the targeted inhibitor era. Median OS was improved in patients with no prior treatment compared to those previously treated for CLL (46·3 months vs 7·8 months).17 Similar findings were observed in a recent Spanish cohort20 and in the CHOP-OR trial.21 Clonal relatedness of the underlying CLL and DLBCL-RT is a strong prognostic differentiator.5 Patients with RT commonly present in the context of pre-treated CLL and immunosuppression, and given the typical demographics of the CLL population, patients are often older with co-existing comorbidities.22 Treatment has historically involved multi-agent cytotoxic chemotherapy, more recently in combination with an anti-CD20 monoclonal antibody. Although intensive regimens including hyper-fractionated alkylator-based therapy,23, 24 platinum and purine analogue-based therapy25-27 have been studied in small phase II trials, toxicity and low efficacy have limited their broad applicability. CHOP alongside an anti-CD20 antibody form the largest and most contemporary data from prospective phase II trials.28, 29 Outcomes generally remain disappointing with overall response rates (ORR) between 40% and 60% and a median progression-free survival (mPFS) between 6 and 10 months. Given the known activity in other aggressive non-Hodgkin lymphoma, dose-adjusted EPOCH-R (etoposide, prednisolone, vincristine, cyclophosphamide, doxorubicin, rituximab) has also been investigated in a 46-patient single-centre retrospective series;30 however the mPFS was only 3·5 months and toxicity was high (30% died without progression or response). Median OS for RT cohorts studied is ~8–12 months, although potentially lower still in those patients progressing with RT following targeted inhibitor treatment for CLL.31, 32 Given these limited survival outcomes, younger and fitter patients should be considered for consolidation strategies such as autologous (autoSCT) or allogeneic stem cell transplantation (alloSCT). European Society for Blood and Marrow Transplantation (EBMT) data33 (n = 59) suggest that the selected patient population who received an alloSCT [n = 25, 72% reduced-intensity conditioning (RIC)] or autoSCT (n = 34, mostly chemo-sensitive disease) had an improved long-term survival, with outcomes better in patients with chemo-sensitive disease. Relapses were more common following autoSCT (three-year cumulative incidence of relapse 43%) whilst non-relapse mortality (three-year 26%) and chronic graft-versus-host disease were more prevalent post alloSCT. Long-term OS was broadly equivalent with either approach (three-year OS 36–59%). Whilst numbers in this historical series were low, SCT consolidation remains a reasonable approach in otherwise fit patients with chemo-sensitive disease. There is no strong evidence to guide the management of patients with disease sites associated with high risk of central nervous system (CNS) disease or those in whom anthracycline-based therapy is unsuitable. R-CHOP (rituximab-CHOP) is curative in a minority of RT patients receiving the regimen for front-line RT treatment. Previously CLL-treatment-naïve, TP53-intact patients who achieve a complete metabolic response following R-CHOP may have a similar long-term PFS to de novo DLBCL.5, 17, 20 As such, it may be reasonable to observe these patients without consolidation therapy (Fig 1). Patients with TP53 aberrations or those who develop RT having previously received CLL-directed treatment have a poor outcome with R-CHOP alone, although this remains the standard of care and provides at least initial disease control for most patients. There are currently no novel targeted therapies specifically licensed for RT. Optimum treatment for HL-RT is less clear with no prospective trial evidence available. Multiagent chemotherapy is often used, with documented outcomes with ABVD (adriamycin, bleomycin, vinblastine and dacarbazine), CHOP (+/− R) and hybrid regimens from small series.12, 34, 35 Outcomes for 94 patients in a recent multicentre, retrospective series found a two-year OS of 72%.36 Sixty-two patients who received ABVD had a median OS of 13·2 years. This series did not support the use of consolidation SCT in HL-RT, with survival outcomes equivalent. Although the management of R/R RT patients may differ depending on previous therapy, comorbidities and fitness, the outcome is generally poor for all patients. Patients who relapse following cellular therapy or who are not fit for this modality should be offered clinical trials or palliative care. Others should be considered for second-line intensive chemotherapy followed by alloSCT although it is recognised the response rates to second-line chemotherapy remain limited. In light of the poor outcomes described, ongoing clinical and translational research remain critical for progress in RT management. Recent retrospective and phase I/II studies suggest that inhibitors targeting Bruton tyrosine kinase (BTK) (ibrutinib, acalabrutinib, pirtobrutinib),37-41 B-cell lymphoma-2 (BCL2) (venetoclax)42 and the Programmed death-1–Programmed death-ligand-1 (PD1–PDL1) axis — which is upregulated in RT — (nivolumab, pembrolizumab)43, 44 result in an ORR between ~20–50% as monotherapy or in combination. All series are small, heterogeneous, subject to selection bias and challenging to cross-compare. Unfortunately, most responses seen in these trials are not durable or follow up is short. Prospective trials with combination strategies using novel–novel combinations ([e.g. BTK/mammalian target of rapamycin (mTOR) dual inhibition plus immunomodulation45] and targeted inhibitors combined with anthracycline-based immunochemotherapy are ongoing.46, 47 Which strategy will provide the optimum benefit for patients remains unclear. Future selection of novel agents to be tested could be based on targetting the molecular events driving transformation, in particular impaired DDR. Chimaeric antigen receptor-modified T-cell (CAR-T) therapy directed against CD19-positive B-cell malignancies have shown promising results in patients with relapsed or refractory (R/R) DLBCL, leading to international approval of three anti-CD19 CAR-T products.48-50 Owing in part to concerns related to CLL-induced immune T-cell exhaustion, patients with RT were excluded from the pivotal trials of axicabtagene ciloleucel (Axi-cel) and tisagenlecleucel (Tisagen). As a result, there remains an open question about the benefit RT patients may gain from this approach. Recent, small and heterogeneous (each <10 patients) series51, 52 suggest encouraging efficacy, although toxicities observed in larger R/R DLBCL data sets, including immune effector cell–associated neurotoxicity syndrome and cytokine release syndrome, were seen. Detailed response analysis of the CLL versus RT disease components are currently lacking in available data and are necessary in future CAR-T efficacy evaluation. At the time of writing, CAR T-cells are funded through the Cancer Drugs Fund in the UK for DLBCL patients who have failed two or more treatment lines. Patients with a background of CLL (i.e. regarded as having RT) are considered on a case-by-case basis via the UK national panel and must fulfil all other eligibility criteria. Specifically, the two or more lines of prior treatment must be regarded as standard DLBCL regimens e.g., R-CHOP, R-GemOx (rituximab plus gemcitabine and oxaliplatin). The BSH Haemato-Oncology task force members at the time of writing this guideline were Professor Guy Pratt (Chair), Dr Nilima Parry-Jones (Secretary), Dr Simon Stern, Dr Shireen Kassam, Dr Matthew Cullen, Dr Alastair Whiteway, Dr Oliver Miles, Dr Elspeth Payne, Dr Toby A. Eyre. The authors would like to thank them, the BSH sounding board, and the BSH guidelines committee for their support in preparing this Good Practice Paper. The BSH paid the expenses incurred during the writing of this Good Practice Paper. All authors have made a declaration of interests to the BSH and Task Force Chairs which may be viewed on request. TAE has received financial reimbursement, advisory board and/or research funding from Roche, Gilead/KITE, Takeda, Janssen, Abbvie, Beigene, Incyte, Loxo Oncology, Secura Bio and AstraZeneca. RW has received financial reimbursement, advisory board and/or research funding from Abbvie, Janssen, AstraZeneca, and Secura Bio. PH has received financial reimbursement, advisory board and/or research funding from Janssen, Abbvie, Roche, AstraZeneca, Sobi, Beigene, Pharmacyclics, and Gilead Sciences. HM has received financial reimbursement, advisory board and/or research funding from Abbvie, Takeda, Roche and AstraZeneca. PEMP has received financial reimbursement, advisory board and/or research funding from Roche, Gilead, Abbvie, Astra Zeneca, Janssen and Novartis. GF has received financial reimbursement, advisory board and/or research funding from Roche, Abbvie, Janssen, AstraZeneca, Karyopharm, BMS, Takeda, and Centessa Pharmaceuticals. JCR has no conflicts of interest to declare. Members of the writing group will inform the writing group Chair if any new evidence becomes available that would alter the strength of the recommendations made in this document or render it obsolete. The document will be reviewed regularly by the relevant Task Force and the literature search will be re-run every three years to search systematically for any new evidence that may have been missed. The document will be archived and removed from the BSH current guidelines website if it becomes obsolete. If new recommendations are made an addendum will be published on the BSH guidelines website (https://b-s-h.org.uk/guidelines/). While the advice and information in this Good Practice Paper is believed to be true and accurate at the time of going to press, neither the authors, the BSH nor the publishers accept any legal responsibility for the content of this Good Practice Paper. All authors reviewed the literature and contributed to the drafting and editing of this manuscript. TAE co-ordinated and edited the Good Practice Paper and was responsible for the final submission.