Guideline for diagnosis and management of hairy cell leukaemia (HCL) and hairy cell variant (HCL‐V)

Abstract

This guideline was compiled according to the BSH process (https://b-s-h.org.uk/guidelines). 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. A literature search was undertaken by Niche Science and Technology Ltd on June 30, 2018, using PubMed, with a further check by the authors on March 1, 2020. The following search terms were used: hairy cell leukaemia AND; BRAF V600E mutation, vemurafenib, dabrafenib, immunotherapy, rituximab, immunotoxin conjugate, moxetumomab, minimal residual disease, hairy cell leukaemia variant. Review of the manuscript was performed by the British Society for Haematology (BSH) Guidelines Committee Haemato-Oncology Taskforce, the BSH Guidelines Committee and the Haemato-Oncology Sounding Board of BSH. It was also placed on the members section of the BSH website for comment, and has been reviewed by the Leukaemia Care patient advocacy group. Hairy cell leukaemia (HCL) is an uncommon, chronic B cell leukaemia, first reported as a distinct entity in the 1950s.1, 2 HCL accounts for 2% of lymphoid leukaemias, with a male predominance and median age at diagnosis of 58 years. Classical HCL and its variant form (HCL-V) are now regarded as separate entities,3 with different cytological, haematological and immunophenotypic features. BRAF V600E mutation, present in virtually 100% of cases of classical HCL,4 is regarded as a disease-defining event, and is absent in HCL-V. Advances in management of classical HCL, from use of interferon in the 1980s, to purine analogues in the 1990s, monoclonal antibodies in the 2000s and BRAF inhibitors in the current decade, have resulted in excellent prognosis, with the majority of patients achieving long-lasting remissions and prolonged survival.5, 6 There remains no clear optimal treatment for HCL-V, which is one-tenth as common as classical HCL, with a 5 year survival rate of 57%.7 Advances in diagnostics and treatment necessitate an update of the 2012 BSH guidelines.8 Some patients with HCL are asymptomatic, with incidental finding of pancytopenia. Others present with lethargy or infection. Symptomatic splenomegaly is unusual, although the spleen is frequently palpable. Cytopenias, normally affecting at least two lineages, are a consistent laboratory feature. Leucopenia is a frequent feature in classical HCL; monocytopenia is virtually always present. By contrast, monocytopenia is not a feature of HCL-V, where there is often a leucocytosis. Classical hairy cells may be seen in the peripheral blood film (see description in Table I). The peripheral blood film may show characteristic medium-sized lymphoid cells with an oval or indented (kidney-shaped) nucleus with homogenous, ground-glass chromatin, slightly less clumped than that of a normal lymphocyte. The nucleolus is typically absent, or inconspicuous. Cytoplasm is abundant and pale blue, with circumferential ‘hairy’ projections.9 Discrete cytoplasmic vacuoles or rod-shaped inclusions (ribosome lamellar complexes) may be seen.3 Bone marrow aspiration is frequently unsuccessful, reflecting fibrosis induced by the hairy cell infiltrate. The bone marrow trephine biopsy often shows patchy infiltration, making it important to obtain a good-sized specimen. The hairy cell infiltrate is characterised by widely-spaced lymphoid cells, and the pattern of marrow involvement is commonly interstitial, becoming diffuse, and creating a ‘honeycomb’ appearance. ‘Blood lake’ pseudo-sinus formation may be seen, with extravasation of red cells into involved areas. Reticulin fibrosis may be present, but collagen deposition is not seen. There may be a minor intra-sinusoidal component. Of note, there are cases displaying a hypocellular marrow, where loss of haemopoietic elements (in particular of the granulocytic lineage) can lead to an incorrect diagnosis of aplastic anaemia. Immunohistochemical demonstration of an abnormal B cell infiltrate is essential in these cases, in which the hairy cell infiltrate may be almost invisible on routine stains.3 The classic immunophenotypic profile of HCL consists of bright surface immunoglobulin of multiple clonally-related isotypes;10, 11 bright co-expression of CD20, CD22 and CD11c and expression of CD103, CD25, CD123, TBX21 (TBET), annexin A1, FMC7, CD200 and cyclin D1.3 Annexin A1 is the most specific marker; it is not expressed in any other B cell lymphoma12 and can be useful to distinguish classical HCL from HCL-V, and from splenic marginal zone lymphoma. If liquid marrow is available for FC, the ‘Hairy Cell Panel’ of CD11c, CD25, CD103 and CD12313 can be useful to distinguish HCL (score 3–4) from other B cell disorders (score 0–1). Further markers, such as CD27, CD43, CD81 and CD200, may be differentially expressed in HCL and HCL-V, reflecting their differing disease biology, and may be informative when included in flow cytometry panels.14 CD27 and CD38 are typically negative in classical HCL, which distinguishes it from other lymphoproliferative disorders.15 IHC panels for use on the marrow biopsy may include CD20, DBA44, Annexin A1, CD25 and Cyclin D1.16 Cytochemistry for tartrate-resistant acid phosphatase (TRAP) has been replaced by TRAP IHC, and IHC for the V600E-mutant BRAF protein is also available. Both cyclin D1 and SOX11 are expressed in a significant proportion of HCL cases, which appears highly specific for HCL outside of mantle cell lymphoma.17 Need for diagnostic or therapeutic splenectomy is infrequent, but, when undertaken, histology shows diffuse expansion of the red pulp, with disruption of normal architecture and extreme atrophy of the white pulp. ‘Blood lakes’ surrounded by neoplastic cells may be seen. These appearances are shared by HCL-V and splenic diffuse red pulp small B cell lymphoma (SDRPL), included as a provisional entity in the current WHO Classification.3 The differential diagnosis of HCL includes other lymphoid malignancies which present with splenomegaly, including the B cell neoplasms: hairy cell leukaemia variant (HCL-V), lymphoplasmacytic lymphoma (LPL), splenic marginal zone lymphoma (SMZL), B- cell prolymphocytic leukaemia (B-PLL) and splenic diffuse red pulp small B cell lymphoma (SDRPL), and the T cell neoplasms: T cell large granular lymphocytic leukaemia (T-LGL) and hepatosplenic T cell lymphoma. Useful differentiating factors are described in Table II. Gender M:F Age 4:1 Median: 58 years M> F Middle aged- elderly 1:1 >40 years 1:1 >60 years CD20 bright+ CD103+ CD25+ CD27− CD11c+ CD123+ DBA44+ Annexin A1+ Cyclin D1+ (weak) CD20 bright+ CD103+ CD25- CD27+ CD11c+ CD123- DBA44+ Annexin A1− Cyclin D1− CD20 bright+ CD103−/+ CD25- CD27+ CD11c-/+ CD123- DBA44+ Annexin A1− Cyclin D1− CD20+ CD103- CD25-/+ CD27+ CD11c+/− CD123- DBA44+ Annexin A1− Cyclin D1− Myeloid disorders presenting with pancytopenia, splenomegaly and bone marrow fibrosis also form part of the clinical differential diagnosis. The BRAF V600E mutation was identified in 2011 using whole-exome sequencing as the causal genetic event in the pathogenesis of HCL.4 The mutation confers constitutive BRAF-MAPK pathway activation and favours tumour survival. BRAF V600E has also been proposed to explain the ‘hairy’ morphology, possibly via overexpression and constitutive activation of the RHO family of small GTPases and upregulation of GAS7 (a growth arrest-specific molecule).18 Although BRAF V600E mutation testing may not always be necessary for the diagnosis of HCL, mutation analysis is useful when faced with diagnostic uncertainty, and early identification of BRAF V600E allows more precise monitoring when disease-eradicating therapies are used. Various methodologies are in use for detection.19 Next Generation Sequencing (NGS) may become the preferred methodology in the longer term.20 Classical HCL typically has no biased use of specific IGHV and expresses IGHV with mutated conformation.10, 11, 21 However, a minor subset with classical HCL phenotype and unmutated IGHV has been described. This subset often has leucocytosis at presentation, is refractory to single-agent cladribine, has a more aggressive behaviour and can harbour TP53 mutations.22 Conversely, HCL-V frequently uses IGHV4-34 in unmutated conformation (approximately 40% of all HCL-V). This group typically has higher disease burden at diagnosis, poor response to single-agent cladribine and shorter overall survival.23 Whole-exome sequencing of these HCL-V cases has identified BRAF V600E mutation to be lacking, with MAP2K1 the most frequently mutated gene, encoding MAP2K1/MEK1 downstream from BRAF,24 a potential future therapeutic target.25 Additional recurrent mutations have been identified in HCL, including CDKN1B (p. 27).26 There is no widely agreed upon staging system for HCL. The disease affects mainly bone marrow and spleen, with lymphadenopathy in a minority. Approximately 9% of patients have abdominal lymphadenopathy at presentation, and the incidence increases at relapse. Lymphadenopathy has been associated with inferior response to treatment and reduced overall survival (OS).27 While computerised tomography (CT) at presentation is not considered essential, it may provide some prognostic information. Where lymphadenopathy has been demonstrated, response assessment should include a repeat CT. In cases where cytopenias are minimal and the patient largely asymptomatic, an initial period of watchful waiting is usually appropriate. But the majority of patients are symptomatic, with anaemia, infections or bleeding, or symptomatic splenomegaly, justifying treatment. The purine nucleoside analogues (PAs), pentostatin and cladribine, remain the standard front-line treatment for HCL,28 with an overall response (OR) of 90–100% [complete response (CR) in 70–90%] and no significant difference between the two agents. Remissions are durable (median 15 years),5, 29 and survival is close to that for an age-matched general population. A small number of asymptomatic patients with well-preserved blood counts may not need immediate therapy. In patients presenting with infection, this should be treated before commencing PA therapy. Pentostatin is administered as a short intravenous (IV) infusion every 2–3 weeks until remission is achieved (usually 8–10 injections). The dose may need to be modified based on renal function. Patients can experience nausea for up to 72 h after the infusion. Cladribine (CDA) can be administered in a number of ways, including as a 7-day continuous IV infusion (which may require a hospital admission), daily or weekly IV infusions or as a subcutaneous (SC) injection. There is no evidence that these are not all equally effective30-32 and the preferred choice is usually 5 consecutive days SC administration for convenience of delivery. Both pentostatin and cladribine cause temporary myelosuppression and a more prolonged immunosuppression. GCSF and prophylactic anti-infective agents (cotrimoxazole, aciclovir) may be initiated during treatment and continued for up to six months, or until adequate neutrophil and lymphocyte recovery. In the case of CDA, they should be commenced after the 5-day course of treatment, since rashes can occur when the drugs are given concurrently. Allopurinol is not required. Blood transfusion after purine analogue therapy should be with irradiated blood, indefinitely.33 Splenectomy is rarely undertaken now, since PA therapy is effective in reducing the size of the spleen. Interferon is now rarely used, being poorly tolerated and less effective than PAs, but may occasionally still be useful in patients who present with serious infection and severe pancytopenia. Rituximab monotherapy has not been formally tested in the first-line setting, but, based on the relatively low response rates seen in relapsed HCL,34 would not be recommended as an alternative to PA other than in special circumstances. Its use is reserved for patients unable to tolerate PAs, or who present with an active infection. There is evidence that improved remissions may be achieved with the combination of rituximab and a PA.35-38 A recent publication comparing concurrent versus delayed rituximab (by ≥6 months after detection of MRD in blood) showed MRD-free complete remission to be greatly enhanced by concurrent rituximab.39 Concurrent cladribine and rituximab (CDAR) resulted in increased transient grade 3–4 thrombocytopenia, but higher neutrophil counts and platelet counts at 4 weeks, compared to delayed rituximab. Response assessment will normally include bone marrow trephine biopsy, which should be deferred for at least 4–6 months after cladribine therapy.8 CR is defined as the absence of hairy cells from the peripheral blood and bone marrow, along with resolution of organomegaly and cytopenias. Partial response (PR) is defined as a normalisation of cytopenias, along with a minimum 50% improvement in both organomegaly and bone marrow infiltration, with no circulating hairy cells. With pentostatin, the practice has been to perform bone marrow trephine biopsy after 8–9 injections, or when the blood count has normalised (apart from lymphopenia), offering 2–3 further pentostatin injections if CR is demonstrated.5 In response assessment, CD20 staining in conjunction with morphological assessment is recommended, with use of DBA44 to identify subtle residual infiltration. Use of CD11c and annexin A1 is not recommended, these being expressed in myeloid cells, and use of CD79a and CD19 can over-estimate residual disease by staining of plasma cells.40 In CR, immunohistochemistry reveals no clustering (≥3 cells) of CD20-positive or DBA44-positive cells. If there is PR with significant residual HCL, a second course of cladribine may be required to achieve CR, usually given at least 6 months after the end of therapy. In a single-centre follow-up study of 242 patients,41 18 patients treated with cladribine (12 as first-line and six as second-line therapy) who remained in PR after bone marrow reassessment received a repeat treatment with cladribine 4–7 months after initial treatment, leading to CR in 14 patients, of whom all but one remained in CR at a median follow-up of 6 years. Despite the durable remissions achieved with front-line therapy, approximately 50% of patients will relapse. Younger patients (<40 years) may have shorter remissions (median 63 months vs. 145 months for those >40 years age).42 In the relapse setting, PAs induce lower CR rates and shorter remissions than when used in front-line therapy.5 Addition of rituximab (6–8 doses given concurrently or sequentially) to a PA can improve this response, but there is limited data on optimum schedule.43, 44 A poor response (remission lasting <3 years) or primary resistance to one PA may be overcome by switching treatment to the alternative PA.45 Bendamustine plus rituximab has been used in a small study (12 patients) of relapsed HCL with promising results (100% OR, 60% CR).46 Moxetumumab pasudotox is a recombinant immunotoxin directed against CD22 and linked to a truncated Pseudomonas exotoxin. CD22 is strongly expressed on hairy cells. Initial studies showed promising results (OR 86%) in heavily pre-treated patients with relapsed/refractory HCL.47 In the pivotal international phase II trial, durable CRs were seen in 30% of patients with relapsed HCL, with bone marrow minimal residual disease (MRD) eradication in 85% of responding patients.48 Administration is by IV infusion on days 1, 3 and 5 of each 28-day cycle. Very little myelo- or immunosuppression is seen, but there is a risk of haemolytic uraemic syndrome (HUS) and/or capillary leak syndrome (CLS). The latter are reversible and can be minimised by good pre-hydration. HCL is a rare disease, and it is important for patients with relapsed or refractory disease to be considered for a clinical trial, particularly those who may benefit from a novel agent. Identification of the BRAF V600E mutation and a constitutively-active BRAF-MEK-ERK pathway in HCL cells provided the scientific basis for therapeutic use of BRAF inhibitors (vemurafenib, dabrafenib) and MEK inhibitors (trametinib) in patients with relapsed or refractory HCL.4 After the first clinical description of a patient with HCL successfully treated with vemurafenib, two separate multicentre phase II clinical and were performed using in relapsed/refractory The was administered at a dose of daily for a median of weeks in the and 18 weeks in the response rate was in the patients) and 100% in the a median follow-up of months, median survival in the was months in patients who achieved CR months for those achieving in the the survival and the overall survival at year were and and of was also in 21 consecutive HCL patients treated on outside clinical with median treatment were in all patients, with CR in 40% and median survival Although were small for at of hairy cells in and response rates and of is with reversible and grade with minimal and and usually to cell and cell occur in have been used in more than 50% patients to although the benefit of dose remains with at lower BRAF dabrafenib, has been given in combination with the MEK to relapsed/refractory HCL patients in an phase II multicentre After a median of months of treatment, was no from is also in combination with rituximab in an phase II single-centre for patients with relapsed/refractory with daily given for weeks, and with rituximab every 2 weeks for analysis of cases showed CR with CRs was grade and to that when was used This combination has no B cell is in and the survival and of hairy cells in In a phase II of cases of relapsed/refractory HCL, was 60% CR and after prolonged administration of the case of in patients with HCL-V show promising and further There is little specific data on the treatment of HCL-V, but it is that the response rates and of response to PAs is inferior to that seen in classical HCL. this it is recommended that cladribine plus rituximab be used as first-line therapy, based on small improvement in with the 58 of patients with HCL-V have been treated with pasudotox with less good results than seen in classical HCL, but this may be to the of a of HCL-V patients have been treated on the phase II trial, with the as the classical HCL BRAF is not given that the variant cases not harbour BRAF However, mutations are seen in the cases, and has been to have Splenectomy may be considered in refractory cases to and cytopenias. HCL is an chronic leukaemia, but in treatment the two that classical HCL patients can have a normal Identification of BRAF V600E as to the pathogenesis of classical HCL has resulted in in particularly for relapsed and refractory HCL-V, regarded as different from classical HCL, is now identified as a distinct and treatment of HCL-V remains an of The authors to and of Niche Ltd for in the initial literature The BSH Haemato-Oncology members at the of this guideline were and The authors would to the BSH and the BSH for their in this The BSH the during the of this authors have a of to the BSH and which may be on has received from and and from has received from or from and from and from has been on for for has no of to of the group will the group if any evidence available that would the strength of the in this or it The will be and from the BSH current website if it If are an will be on the BSH website While the and in this is to be and at the of to the the BSH the any for the of this

Topics & Concepts