New developments in the diagnosis and characterization of Waldenström’s macroglobulinemia

ABSTRACT

Introduction

Waldenström’s macroglobulinemia (WM) is defined as a lymphoplasmacytic lymphoma (LPL) with immunoglobulin M (IgM) monoclonal gammopathy and morphologic evidence of bone marrow infiltration by LPL. Immunophenotyping and genotyping provide a firm pathological basis for diagnosis and are particularly valuable in differential diagnosis between WM and related diseases. Emerging technologies in mutational analysis present new opportunities, but challenges remain around standardization of methodologies and reporting of mutational data across centers.

Areas covered

The review provides an overview of the diagnosis of WM, with a particular focus on the role of immunophenotyping and genotyping.

Expert opinion

Demonstration of LPL with a bone marrow biopsy is essential to reach a definitive diagnosis of WM. However, MYD88^L265P and a typical WM immunophenotypic profile are valuable for the differential diagnosis of WM and related diseases, such as marginal zone lymphoma, multiple myeloma, and chronic lymphocytic leukemia. These methodologies must be utilized across centers and with appropriate standards followed in the evaluation and reporting of sensitivities and specificities. The diagnostic and/or prognostic value of mutations in genes such as CXCR4 and TP53 that are currently not routinely evaluated in the diagnosis of WM should be explored.

KEYWORDS:

• CXCR4
• diagnosis
• p53 genetics
• immunophenotyping
• macroglobulinemia
• Waldenstrom
• MYD88
• paraproteinemias

1. Introduction

Waldenström’s macroglobulinemia (WM) is a lymphoplasmacytic lymphoma (LPL) with immunoglobulin M (IgM) monoclonal gammopathy and morphologic evidence of marrow infiltration by LPL [1–3]. WM is the most common of two LPL subtypes recognized by the World Health Organization (WHO), accounting for approximately 95% of cases [3]. The other subtype, non-WM LPL, includes cases with immunoglobulin G (IgG) or immunoglobulin A (IgA) monoclonal protein, non-secretory LPL, and IgM LPL without bone marrow involvement [3]. Non-WM LPL cases are rare and poorly defined. WM comprises approximately 2% of all cases of non-Hodgkin’s lymphoma and has an incidence of ~4 cases per 1 million person-years [4–6]. The disease is more common in males than in females, and in Caucasians compared with other races [5]. Most patients with WM are aged ≥65 years and many have non-lymphoma-related comorbidities typical of the elderly population [6,7].

WM is associated with a heterogenous presentation, with symptoms attributable to overall disease burden as well as symptoms attributable to the concentration and/or immunological/physicochemical properties of plasma cell–secreted IgM (Figure 1). Unlike other lymphomas, the malignant B-cell clone in WM undergoes plasmacytic differentiation and the extent of differentiation varies between patients. The level of IgM paraprotein correlates with the number of plasma cells rather than with the overall extent of bone marrow infiltration [8,9]. Hyperviscosity syndrome related to high levels of IgM is a hallmark symptom of WM, but IgM levels do not have a linear association with serum viscosity [10]. Similarly, patients with low levels of IgM may experience symptoms that depend on the immunological properties of secreted IgM, such as neuropathy, hemolytic anemia, and cryoglobulinemia.

Figure 1. Symptoms of WM can be attributed to the overall disease burden and/or the IgM paraprotein. IgM: immunoglobulin M, WM: Waldenström’s macroglobulinemia.

Outcomes for patients with WM have improved over time with the introduction of new drugs and treatment regimens [11–14]. Accurate and timely diagnosis is key to ensuring patients gain access to the most appropriate treatment at the right time. This review summarizes the current approach to WM diagnosis, with a particular focus on molecular diagnosis. Methods for the detection of the highly prevalent MYD88^L265P mutation are considered, as well as approaches to testing for mutations in other relevant genes, such as CXCR4.

2. Clinical presentation of Waldenström’s macroglobulinemia

WM is an indolent lymphoma and about a quarter of patients with WM are asymptomatic at the time of diagnosis [15]. Anemia is the primary reason for patients with WM to present to primary care or to be referred to internal medicine, being present in ~40% of symptomatic patients (Figure 2) [15,16]. B symptoms, especially weight loss, are also relatively common, reported in about a quarter of symptomatic WM patients at presentation [15,16]. A similar proportion of patients may present with symptoms related to hyperviscosity (arterial hypertension, bleeding, headache, cardiac insufficiency, or reduced level of consciousness) or neuropathy [15,16].

Figure 2. The main indication for initiation of therapy in 595 patients with WM from the Greek Myeloma Study Group database (adapted from Dimopoulos et al.) [16].

CAGG: cold agglutinemia, WM: Waldenström’s macroglobulinemia.

3. Overview of Waldenström’s macroglobulinemia diagnosis

Differentiation of WM from other B-cell disorders (Table 1) is usually achievable through correlation of clinical features, laboratory tests, bone marrow morphology, B-cell immunophenotype, and genomic profile (Table 2). Hemograms often reveal anemia in patients with WM, and more rarely leukopenia/neutropenia and/or thrombocytopenia; abnormal hemograms are often the first indication of disease in patients with asymptomatic WM. Biochemistry is usually normal, but a complete metabolic profile may reveal abnormal kidney function. While the latter is rare, with a cumulative incidence of approximately 5% it can reflect disease-specific etiologies such as amyloidosis, cast nephropathy, light chain deposition, cryoglobulinemia, and direct tumor infiltration [17,18]. Increased bilirubin and lactate dehydrogenase may also be found, which can be an indication of hemolysis [16].

Table 1. WM differential diagnosis (adapted from Dimopoulos et al.) [16].

Disease	IgM MP	BM biopsy	Cytogenetic	MYD88 mutation	CXCR4 mutation	Immunophenotype (by IHC or FC in BM)	Clinical presentation and other
WM	100%	Morphology: lymphoplasmacytes or cells with lymphoplasmacytic differentiation, together with a small population of clonal plasma cells ≥10% LPL	Del6q (30–50%) Trisomy 4 (10–20%) Del17p (<10%)	~90%	30–40% (half of cases with subclonal pattern)	B-cell population: CD20^+, sIgM^+, CD22^+ (weak), CD79^+, CD25^+, CD27^+, FMC7^+, BCL2^+, CD52^+, CD5^+/−, CD10^+/−, CD23^+/−, CD103^−; plasma cell population: CD138^+ CD38^++, CD19^+, CD45^+, CD56^−, k/λ restricted	Hyperviscosity, lymphadenopathy, splenomegaly, neuropathy
IgM MGUS	100%	<10% LPL in the BM	?	30–60%	10–20%	Usually few cells found	No symptoms or only IgM-related; <3 g/dL IgM
Myeloma	0.5% [39]	Plasma cells	IgH translocations (60%) Trisomies (70%) Del13 (50%)	0	Rare	CD138^+, CD38^+, CD19^−, CD56^+, CD45^−, CD20^−	Lytic bone disease Cyclin D1 staining positive in 75% [usually associated with t(11;14)]
SMZL	10–30%	Intrasinusoidal infiltration by CD20^+ cells	Del7q (19%), +3q (19%), +5q (10%)	0	Rare	CD19^+, CD20^+, CD22^+, CD79a^+, CD79b^+, FMC7^+ IgM^+ CD5^− (weakly + in 10–25%), CD10^−, CD43^−, BCL6^−, cyclin D1^−CD103^−, but occasionally+ CD11c^−/+ CD25^−/+CD11c^+	Splenomegaly more common; circulating cells of characteristic morphology may be found
FL	0%*	Small cleaved lymphocytes, paratrabecular localization in the BM	Translocations involving BCL-2 (70–90%)	0	Rare	CD5^−, CD10^+/−, CD11c^−/+, CD103^−, CD25^−, CD138^−, CD38^+, CD45^+, BCL2^+, BCL6^+	Lymphadenopathy predominates
MCL	0%*	Monotypic, medium- to small-sized lymphocytes with abnormal nucleus	t(11;14)(q13;q32)	0	Absent	CD5^+, CD10^−, CD23^−, CD25^−, CD45^+, CD103^−, CD138^−	Lymphadenopathy and extranodal involvement common
CLL	5% [40]	Small lymphocytes with clumped chromatin and scant cytoplasm	Del13 (60%) +12q (20%) Del11q (15%) Del17p (7%)	2%	Absent/rare	CD19^+, CD5^+, CD10^−, CD23^+, CD45^+, CD103^−, CD138^−	BM and PB involvement predominance Splenomegaly and lymphadenopathy frequent

*Single cases are reported in the literature.

BM: bone marrow, CLL: chronic lymphocytic leukemia, FC: flow cytometry, FL: follicular lymphoma, IgH: immunoglobulin H, IgM: immunoglobulin M, IHC: immunohistochemistry, LPL: lymphoplasmacytic lymphoma, MCL: mantle cell lymphoma, MGUS: monoclonal gammopathy of undetermined significance, MP: monoclonal protein, PB: peripheral blood, SMZL: splenic marginal zone lymphoma, WM: Waldenström’s macroglobulinemia.

Table 2. Essential evaluation of patients with WM (adapted from Castillo et al.) [21].

History and physical examination	Comments
Include familial history of WM and other B-cell lymphoproliferative disorders	There is a familial component in ~10% of cases
Include funduscopic examination	Early retinal changes are associated with hyperviscosity; recommended for patients with a serum IgM >30 g/L
Review of systems	Heterogeneous manifestations
Laboratory studies
Complete blood count	Anemia is frequent; other cytopenias are rare
Complete metabolic panel	Relevant alterations are not very frequent
Serum immunoglobulin levels (IgA, IgG, IgM)	IgM monoclonal protein
Serum and urine electrophoresis with immunofixation	Monoclonal demonstration
Serum beta-2-microglobulin level	Prognostic value
If clinically indicated
Coombs test	Indicated if the patient has anemia with evidence of hemolysis
Cryoglobulins	10–20% of cases
Cold agglutinin titer	~10% of cases; ensure proper sample handling
Serum viscosity	~30% of WM cases have hyperviscosity; recommended for patients with a serum IgM >30 g/L
Screening for von Willebrand disease	~14% of WM cases have vWF inhibitor activity by lab tests
24-h urine protein quantification	Demonstrable in 30% of cases
Bone marrow aspiration and biopsy
Immunohistochemistry	Clonal infiltration must be demonstrated
Flow cytometry	Variable percentage of clonal cells
Testing for MYD88 L265P gene mutation	>90% of cases; other MYD88 mutations present in ~1–2% of patients with WM
Computerized tomography scans of the chest, abdomen, and pelvis with intravenous contrast
In patients being considered for therapy	Lymphadenopathy usually not very relevant

IgA: immunoglobulin A, IgG: immunoglobulin G, IgM: immunoglobulin M, vWF: von Willebrand factor, WM: Waldenström’s macroglobulinemia.

Demonstration of monoclonal IgM by serum electrophoresis with immunofixation is essential for the diagnosis of WM, and evaluation of serum IgM levels is critical for prognosis and monitoring of treatment responses [19,20]. Funduscopic examination is recommended for patients with a serum IgM >30 g/L and for all patients with symptoms of hyperviscosity syndrome, such as oronasal bleeding and visual disturbances. A 24-hour urine analysis with urine electrophoresis and immunofixation may also be considered at diagnosis. Bence-Jones proteinuria (free monoclonal light chains in the urine) is rarer in WM than in multiple myeloma but may still be present in about a third of patients [21,22].

Serum free light chain (sFLC) levels and the κ/λ ratio of sFLCs may be useful indications of disease burden in WM [23,24]. In a study by Leleu et al., sFLC levels showed earlier response and progression than IgM or M-spike measurements in 48 patients with WM. Median κ/λ ratio was 13.5, but the 95% confidence interval (0.01–665) crossed the normal reference range (0.26–1.65) [23]. There are insufficient data to recommend a role for free light chain assessment in routine response evaluation, but it is essential for patients with suspected renal complications and amyloidosis [18,25].

It is also important to determine serum β2 microglobulin levels because elevated levels predict poorer patient outcomes. A serum β2 microglobulin level of more than 3 mg/L is a component of the International Prognostic Scoring System for WM (IPSSWM), along with age (>65 years), hemoglobin level (≤11.5 g/dL), platelet count (≤100 × 10⁹/L), and serum monoclonal protein concentration (>70 g/L) [19].

Imaging studies are not recommended in the evaluation of asymptomatic patients, although computed tomography of the chest, abdomen, and pelvis with intravenous administration of contrast is recommended to document organomegaly and/or lymphadenopathy in patients who are considered for treatment. If organomegaly and/or lymphadenopathy are detected prior to treatment initiation, imaging during or after completion of therapy is also advised [18,21]. There are limited data on the utility of fluorodeoxyglucose positron emission tomography (FDG-PET) imaging in WM, although it may be applicable in some patients (for example, those with suspected histologic transformation) [26,27].

3.1. Histology

A diagnosis of WM requires the demonstration of morphologic marrow infiltration by LPL and this is best achieved with assessment of an adequate trephine biopsy [2,3,18]. A trephine biopsy also provides for a better assessment of the extent of marrow disease. Arbitrary values for levels of marrow disease are no longer considered appropriate, as some symptomatic patients will have disease burdens of less than 10% [25]. The value of bone marrow assessment for asymptomatic patients with IgM monoclonal gammopathy of undetermined significance (MGUS) is not well established, but patients with suspected symptomatic WM should always undergo bone marrow biopsy, including immunophenotyping and genetic assessments, independently of the level of the IgM protein [18,21]. IgM MGUS, which is defined as IgM monoclonal gammopathy with no evidence of bone marrow infiltration, rarely progresses to WM when no risk factors are present [28–31]. Bone marrow biopsy and closer monitoring may be warranted in patients with IgM MGUS with IgM levels ≥15 g/L and/or altered free light chain assessments, because they have a greater risk of progression [30,31].

LPL in the bone marrow can show an interstitial, nodular, or diffuse pattern [32]. Although a purely paratrabecular pattern is unusual, a combined paratrabecular and interstitial infiltration pattern is common [32]. LPL cells can show differentiation from small lymphocytes with clumped chromatin, inconspicuous nucleoli, and sparse cytoplasm, to mature plasma cells, which are usually present in small numbers and may contain Dutcher bodies [16,21].

Plasma cell differentiation may be evident morphologically but can be further highlighted using plasma cell–specific markers, such as CD138, IRF4, and VS38c. Plasmacytoid lymphocytes are frequently detected and typically have intermediate cytologic features between small lymphocytes and plasma cells. Reactive mast cells, which may be present in the marrow microenvironment, are a characteristic feature of WM [21,33].

The presence of plasma cell differentiation, monotypic plasma cells, immunoglobulin inclusions, and reactive mast cells appear to be the most important morphologic features in discriminating between WM and marginal zone lymphoma (MZL) [34–36].

4. Immunophenotype and other applications of flow cytometric analysis

Immunophenotyping is a powerful and potentially underutilized tool in the diagnosis of WM. In the past decade, a distinct B-cell immunophenotype for WM was identified using flow cytometric analysis [37]. The WM clone expresses pan-B-cell markers (CD19, CD20, CD22, Sig) but can be distinguished from normal B cells by weak CD22 expression and homogeneous positivity for CD25 [37]. Fewer than 20% of WM cases are positive for CD5, CD10, CD11c, or CD23 [37,38].

The B-cell compartment changes with progression from IgM MGUS through asymptomatic WM to symptomatic WM, with an increasing number of total B cells and light chain isotype positive B cells. Complete light chain restriction of the B-cell compartment appears to predict for progression in asymptomatic WM, but also overall survival in symptomatic patients [37].

The plasma cell compartment is less easily distinguished from normal bone marrow plasma cells. Plasma cell clones have, by definition, restricted immunoglobulin heavy and light chain isotypes but otherwise normal antigenic expression, including positivity for CD19 and CD45, which contrasts with myeloma clones [37]. In common with the B-cell compartment, the plasma cell clone changes with progression from IgM MGUS to WM [37].

Immunophenotyping can be particularly useful in differential diagnosis between WM and related lymphomas (Table 1). The distinction between WM and splenic MZL is especially challenging. A WM clone will typically show homogenous CD25 expression with weak CD22, whereas patients with MZL will typically possess the opposite phenotype (CD25 negative with strong CD22), although variations can clearly occur in some patients. Other important differentiating markers include surface membrane IgM and CD79b, which are often over-expressed in WM, and CD305, which is upregulated in MZL but not in WM [16,41]. These phenotypic differences are demonstrable only by flow cytometry, suggesting that flow cytometry plus MYD88 analysis has more value than immunohistochemical assessment of trephine biopsies for differentiation of WM from splenic MZL.

5. Genotype

5.1. Cytogenetic abnormalities

The are no disease-defining cytogenetic abnormalities associated with WM, and a lack of highly recurrent translocations is also a distinguishing feature [42,43]. However, a range of cytogenetic abnormalities have been reported that appear to affect clinical features and outcomes. 6q deletion has been reported in 40–50% of WM cases and is associated with poor prognosis, as well as progression to symptomatic disease and a shorter time to progression to symptomatic disease [44–46]. A copy number abnormality of 18q (+18q11-18q23) may also be common in WM and involved in malignant transformation [45].

Conventional karyotyping is no longer conducted in WM diagnosis because of the difficulty in collecting samples of tumor cells in metaphase and without blood contamination. However, the use of fluorescence in situ hybridization may be considered to support a differential diagnosis. For example, t(11;14) is common in IgM multiple myeloma but absent in WM [47,48].

5.2. MYD88

The 5th edition of the WHO Classification of Haematolymphoid Tumours describes WM as a disease with two molecular subsets defined by the presence or absence of the MYD88 p.L265P (NP_002459.2) mutation [3]. Several studies have reported MYD88^L265P in >90% of patients with proven well-defined WM and other MYD88 mutations in ~4%, with only a small minority being wild-type for MYD88 [46,49–56]. Some studies from Japan and South Korea have reported a lower incidence of the MYD88^L265P mutation [57–59], but there is at this time no definitive evidence for significant differences according to ethnicity.

L265P and other activating mutations in MYD88 result in spontaneous assembly of protein complexes that lead to multiple pro-survival signaling cascades [60]. Given its prevalence, mutational analysis of MYD88 is strongly recommended as part of the diagnostic procedure for WM [18]. Testing for MYD88^L265P can also be used to confirm WM involvement in complex extramedullary cases such as pleural effusions and Bing–Neel syndrome (LPL invasion of the leptomeningeal tissue and/or the central nervous system) [61,62].

It is also appropriate to consider whether patients with suspected WM and MYD88^WT may have an alternative underlying pathology. In one study, 18 (28%) of 64 patients previously diagnosed with MYD88^WT WM were considered to have an alternative clinicopathological diagnosis following review, with IgM multiple myeloma being the most common (n = 7, 11%) [63]. IgM multiple myeloma may be distinguished from WM with careful assessment of all clinicopathological features. IgM multiple myeloma will be characterized by an infiltrate composed entirely of plasma cells and will lack monotypic B cells. The plasma cells will frequently express an aberrant or myeloma-like phenotype. WM by definition will be composed of both monotypic B cells and plasma cells. Genomic assessment has further utility, as IgM myeloma will frequently possess t(11;14) translocation, whereas MYD88^L265P is typically absent [48,63]. Furthermore, IgM multiple myeloma patients will lack lymph node disease but may have lytic bone disease.

MYD88 mutational status has some prognostic implications for the clinical course and treatment response. MYD88^WT is an independent risk factor for progression in asymptomatic WM patients [64] and is associated with a greater risk of transformation than MYD88^L265P [63,65,66]. There are conflicting data regarding the impact of MYD88^WT on overall survival; some studies demonstrate inferior survival outcomes compared with patients with the MYD88^L265P mutation, but others do not [63,65–68]. However, it may be notable that Abeykoon et al., who did not detect a survival difference according to MYD88 status, reported a higher proportion of MYD88^WT patients (21%) within their WM patient population than is typical [65].

MYD88^WT appears to be associated with a poorer response to the Bruton’s tyrosine kinase (BTK) inhibitor ibrutinib compared with MYD88^L265P (Table 3), which has led some to propose genotype-based treatment strategies [46]. However, the data to support such an approach are limited, and these algorithms may soon be outdated with the availability of next-generation BTK inhibitors. Higher response rates have been reported with acalabrutinib [69] and zanubrutinib [70] in MYD88^WT WM patients. Indeed, zanubrutinib was prospectively evaluated in MYD88^WT WM and is the first BTK inhibitor for which very good partial responses have been reported in this patient subset [70] (Table 3).

Table 3. Patient responses to BTK inhibitors by MYD88 and CXCR4 status.

BTK inhibitor	Publication	Comments (patient status, phase, design)	MYD88^L265P	MYD88^WT
Ibrutinib	Treon et al. [74,75]	R/R, phase 2, prospective	At 59 months MYD88^L265P× CXCR4 ^WT (n = 36) PFS: not reached MRR: 97.2% MYD88^L265P × CXCR4^MUT (n = 22) PFS: not reached MRR: 68.2%	At 59 months MYD88^WT ×CXCR4^WT (n = 4) PFS: 0.4 years MRR: 0%
	Dimopoulos et al. [76]	Rituximab-refractory, phase 3, open label	At 18 months MYD88^L265P× CXCR4^WT (n = 17) PFS: 94% MRR: 82% MYD88^L265P × CXCR4^MUT (n = 7) PFS: 86% MRR: 71% Median time to response: 1.7 months	1 patient who had progressive disease died 93 days after ibrutinib discontinuation
	Treon et al. [77]	TN, phase 2, prospective	MYD88^L265P × CXCR4^WT (n = 16) ORR: 100% MRR: 94% Median time to response: 0.9 months MYD88^L265P × CXCR4^MUT (n = 14) ORR: 100% MRR: 71% Median time to response: 1.7 months	NA
Zanubrutinib	Trotman et al. [89]	TN and R/R, phase 1/2, open label	At 30 months MYD88^L265P/CXCR4^WT (n = 39) ORR: 97% MRR: 87% VGPR/CR: 59% MYD88^L265P/CXCR4^WHIM (n = 11) ORR: 100% MRR: 91% VGPR/CR: 27%	n = 8 ORR: 100% MRR: 63% VGPR/CR: 25%
Zanubrutinib vs. ibrutinib (ASPEN trial)	Tam et al. [90]	TN and R/R, phase 3, randomized, open label	Zanubrutinib ORR: 94% MRR: 77% VGPR: 28% Ibrutinib ORR: 93% MRR: 78% VGPR: 19%
Zanubrutinib (ASPEN trial extension)	Dimopoulos et al. [70]	TN and R/R, phase 3, open label (ASPEN study single-arm extension)		At 17.9 months ORR: 81% MRR: 50% VGPR: 27%
Acalabrutinib	Owen et al. [69]	TN and R/R, phase 2, open label	ORR: 94%	ORR: 79% (no VGPRs)

BTK: Bruton’s tyrosine kinase, CR: complete response, MRR: major response rate, NA: not available, ORR: overall response rate, PFS: progression-free survival, R/R: relapsed/refractory, TN: treatment-naïve, VGPR: very good partial response.

5.3. CXCR4

Activating mutations in the C-terminal domain of chemokine receptor gene CXCR4 are present in up to 40% of patients with WM who almost always also carry MYD88 mutations. CXCR4 mutations are rare in other lymphoproliferative diseases, with only sporadic cases reported in patients with MZL and the activated B-cell subtype of diffuse large B-cell lymphoma (DLBCL) [46].

The mutational landscape of CXCR4 is complex, with >50 nonsense and frameshift mutations described. Interestingly, somatic CXCR4 mutations observed in WM tumor cells are similar to a germline mutation found in the immunodeficiency disorder warts, hypogammaglobulinemia, infections, and myelokathexis (WHIM) syndrome [71]. Preclinical studies have shown that cells with the most common CXCR4^S338X mutation are resistant to ibrutinib-triggered apoptosis and show sustained signaling of Ak strain transforming (AKT) and extracellular signal-regulating kinase (ERK) compared with wild-type CXCR4 cells [72]. In contrast to MYD88^L265P, CXCR4 mutations in WM are usually subclonal and a patient can harbor multiple mutations within separate clones [46].

Mutational analysis of CXCR4 has limited utility in diagnosis and is not commonly performed in routine clinical practice as a consequence of this and the technical challenges in detection and interpretation. Beyond diagnosis, CXCR4 mutations are associated with increased bone marrow disease, serum IgM levels, symptomatic hyperviscosity, thrombocytopenia, and acquired von Willebrand factor deficiency at presentation [67,73]. CXCR4 mutations also appear to have prognostic utility in the context of BTK inhibitor–based therapy. Delayed and poorer responses to ibrutinib, as well as inferior progression-free survival, have been reported in patients with the MYD88^L265P/CXCR4^MUT genotype compared with MYD88^L265P/CXCR4^WT patients (Table 3) [74–77]. Long-term follow-up of the ASPEN head-to-head trial of zanubrutinib versus ibrutinib in WM indicates that the same is true for zanubrutinib, although there was a trend toward higher rates of complete and very good partial responses with zanubrutinib versus ibrutinib for MYD88^L265P/CXCR4^MUT patients (21% vs. 5%, P = 0.15) [78].

There are limited data on the impact of CXCR4 in the context of other therapies [18]. A study conducted by the French Innovative Leukemia Organization reported that CXCR4 mutational status had no impact on disease responses or survival outcomes of 69 patients treated with first-line bendamustine plus rituximab [79]. Similarly, proteasome inhibitors, such as bortezomib and carfilzomib, appear to act independently of CXCR4 [80,81], while survival outcomes were not impacted in patients treated with idelalisib and obinutuzumab [82].

5.4. Other somatic mutations

A range of somatic mutations in genes other than MYD88 or CXCR4 have been reported in patients with WM (Figure 3) [83]. The number of detectable genetic abnormalities in IgM monoclonal gammopathies increases with the aggressiveness of the disease. In a study evaluating the 12 most frequently mutated genes in WM, the percentage of patients with alterations varied from 21% IgM MGUS, 35% asymptomatic WM, and 50% symptomatic WM [84].

Figure 3. The frequency of somatic mutations occurring in WM (adapted from García-Sanz et al.) [83].

*Corresponding to 57 IgM-MGUS and 62 WM.

^†Corresponding to 14 IgM-MGUS, 23 asymptomatic WM, and 24 symptomatic WM.

IgM: immunoglobulin M, MGUS: monoclonal gammopathy of undetermined significance, NA: not assessed, WM: Waldenström’s macroglobulinemia.

Frequency and distribution of the mutations in the main genes of WM according to the different studies using next-generation sequencing. The total number of mutated patients in each study, as well as the global mutation frequency (%) considering the five studies, are displayed.

TP53 aberrations, including mutations or deletion of the TP53 locus on chromosome 17 (17p13.1), have been observed in ~10% of patients with newly diagnosed WM and 25% at more advanced stages, and may be associated with mutated MYD88 and CXCR4 [83,85]. In common with other lymphomas, TP53 aberrations in WM appear to predict more aggressive disease [85,86]. While TP53 mutations predict for poor outcomes with immunochemotherapy in chronic lymphocytic leukemia [87], definitive data in WM are lacking at this time.

Somatic mutations in ARID1A have been detected in 3–17% of patients with WM, including nonsense and frameshift variants, although no prognostic value has been demonstrated to date [42,56,84]. ARID1A may modulate TP53 and is thought to act as an epigenetic tumor suppressor in ovarian cancer [46]. In patients with WM, ARID1A mutations are associated with increased bone marrow infiltration [88]; deletion of its homolog ARID1B (located on 6q) may contribute to the poor prognosis associated with 6q deletion.

Mutations in CD79A and CD79B can be found in ~10% of patients with WM [46,56,61,84,91]. Both are components of the B-cell receptor (BCR) pathway and can form heterodimers with each other, so activating mutations of these components could contribute to the chronic BCR signaling observed in WM cells [46]. In one study, mutations in CD79A and CD79B were found to be nearly exclusive of CXCR4 mutations, suggesting that CD79A/B mutations may have an independent role in facilitating mutated MYD88-directed progression in WM [71]. In addition, CD79B mutations have been associated with histologic transformation in some WM patients [92]. An association of CD79B and MYD88 mutations has also been described in activated B-cell-like DLBCL [60].

6. Methodologies for MYD88 and CXCR4 mutational analysis

There are no standardized protocols for mutational analysis of MYD88 or CXCR4 in patients with WM (Table 4). DNA for mutational analysis of MYD88 or CXCR4 can be extracted from bone marrow, peripheral blood, or plasma [18,21]. However, the use of peripheral blood can give false-negatives and is not recommended in routine diagnosis [18]. Several studies indicate that mutational analysis of cell-free DNA from plasma can be reliably used to identify MYD88^L265P and CXCR4^S338X [93–95]. However, further confirmation is needed before this approach can be recommended in clinical practice.

Table 4. Key genes and methodologies in WM diagnosis and prognosis.

Gene (mutation)	Sample	Selection	Methodology	Comments
MYD88 (L265P)	Bone marrow	No selection	AS-PCR or ddPCR	Easy to perform routinely Highly sensitive Superior quantification with ddPCR
MYD88 (other mutations)	Bone marrow	CD19^+	Sanger or next-generation sequencing If CD19^+ selection is not performed, a limit of detection of variants of 1% for next-generation sequencing	Lower sensitivity than AS-PCR or ddPCR
CXCR4	Bone marrow	CD19^+	Sanger or next-generation sequencing	Lower sensitivity than AS-PCR or ddPCR for CXCR4 (S338X) hot spot; require analysis and intepretation; CXCR4 mutations likely to be in C-terminal domain and subclonal
TP53	Bone marrow	CD19^+	Sanger or next-generation sequencing	TP53 mutations have a frequency of 7.5%, are associated with CXCR4 mutations, and negatively impact prognosis

AS-PCR: allele-specific polymerase chain reaction, ddPCR: droplet digital polymerase chain reaction, WM: Waldenström’s macroglobulinemia.

Allele-specific (AS) polymerase chain reaction (PCR) is used by most laboratories to detect MYD88^L265P because it is relatively simple and highly sensitive. More recently, droplet digital PCR (ddPCR) has emerged as an alternative that allows absolute quantification of DNA and provides higher sensitivity than AS-PCR [96,97]. Drandi et al. validated the use of ddPCR for the detection of MYD88^L265P and reported a sensitivity of 5.0 × 10⁻⁵ for ddPCR versus 1.0 × 10⁻³ for AS-PCR [96]. The sensitivity offered by ddPCR will reduce the risk of false-negative assessments and has clear utility in the assessment of patients with low-level marrow disease, such as IgM MGUS and IgM-related disorders. The sensitivity of ddPCR might also allow for peripheral blood diagnosis in elderly and unfit patients.

Drandi et al. explored the potential for MYD88^L265P mutation detection by ddPCR to be used for minimal residual disease (MRD) monitoring, finding that the assay correlated with an IGH-based MRD ddPCR assay, which is widely used but not validated for WM [96]. In this small study, 6/22 (27%) patients with WM were MRD-negative (MYD88^L265P <5.0 × 10⁻⁵) in the bone marrow and 17/35 (49%) in the peripheral blood after first-line therapy. Fludarabine- and bendamustine-containing regimens showed a high potential for MRD negativity, and MRD monitoring could potentially be used to determine the duration of therapy and whether maintenance therapy is required following treatment [96]. Such MRD-based strategies may provide a more detailed evaluation of response and survival outcomes given the low incidence of complete responses, particularly in the context of conventional rituximab-based therapies. It is, however, important to be aware that MYD88^L265P has been detected in healthy mature B cells [98].

While MYD88 profiling can be performed on whole bone marrow samples, CD19⁺ cell selection using magnetic beads or flow sorting is recommended for detection of subclonal mutations (e.g. CXCR4 and TP53). Ideally, these mutations should be evaluated by next-generation sequencing (NGS) approaches, but targeted sequencing could have a role in mutational analysis of CXCR4. Patients with WM may have CXCR4 frameshift or nonsense mutations, but only nonsense mutations have been associated with an increased risk of progression and/or death versus CXCR4^WT, and 80–90% of CXCR4 nonsense mutations in WM are located in S338X [99,100]. Therefore, an AS-PCR or ddPCR assay targeting CXCR4^S338X may be sufficient in a routine diagnostic setting.

In patients who are negative for MYD88^L265P where a diagnosis of WM is strongly suspected, Sanger sequencing or, ideally, NGS may be used to identify other MYD88 mutations or mutations associated with other lymphoproliferative syndromes, such as KLF2 or NOTCH2 in MZL [101]. CD19⁺ selection is again recommended in this setting [46].

If CD19⁺ selection is not feasible, it is important that laboratories are aware that there is a high risk of false-negative results with targeted NGS of MYD88 or CXCR4, particularly when a sample is contaminated with blood. A detection limit of 1% is required when using NGS to detect MYD88 or CXCR4 variants. In a study comparing the sensitivity of a clinically validated and targeted NGS assay (Rapid Heme Panel) with AS-PCR, the targeted NGS of unselected bone marrow samples was associated with 69% sensitivity for MYD88^L265P and 16% sensitivity for CXCR4^S338X compared with AS-PCR [102].

7. Conclusion

Accurate and timely diagnosis of WM is essential to ensure appropriate treatment, optimize outcomes, and preserve quality of life for patients. The heterogeneity of patient presentations in WM poses a particular challenge in diagnosis, but advances in clinical understanding and technology in the past decade have provided a much stronger basis for pathological diagnosis. The identification of MYD88^L265P as a differentiating feature of WM is particularly useful, but in the absence of standardized testing protocols it is essential that laboratories adopt appropriate assays and determine their sensitivities.

Genotyping in WM outside of clinical trials is usually limited to MYD88, but data suggest that CXCR4 and TP53 mutations can support prognosis and treatment decisions. More research is needed to determine the value of these genes to diagnosis, prognosis, and treatment decisions.

8. Expert opinion

WM is defined by the presence of morphologically detectable LPL on an adequate trephine biopsy in the context of IgM monoclonal gammopathy [1–3]. The use of arbitrary levels of bone marrow disease for disease definition are inappropriate, given that symptomatic disease requiring therapy can occur in patients with less than 10% infiltration [25].

Diagnostic guidelines for WM are consistent regarding the need for a bone marrow biopsy to achieve a definitive diagnosis [18,103]. A biopsy is essential to establish LPL in the bone marrow and quantify the degree of infiltration. However, accurate noninvasive alternative paths to diagnosis are highly desirable in a disease of elderly patients. Immunophenotyping is a powerful diagnostic tool that has the potential to provide the basis for such an approach. The accuracy of diagnosing WM based on flow cytometry immunophenotyping combined with MYD88 testing from bone marrow aspirates should be evaluated compared with bone marrow biopsy–based diagnosis. However, there are significant barriers to such an approach. Identifying the hallmark WM immunophenotype is difficult and requires considerable technical expertise. As such, flow cytometry is more likely to be used by most centers to establish the presence of clonal B cells and to differentiate between related lymphoid diseases, such as chronic lymphocytic leukemia.

The recent WHO classifications of WM describe two molecular subsets based on the presence or absence of MYD88 p.L265P [3]; it would be more accurate to define molecular subtypes based on the presence or absence of all activating MYD88 mutations and not exclusively L265P. Studies indicate that patients with WM and wild-type MYD88 have a higher risk of transformation, shorter overall survival, and respond poorly to certain therapies compared with patients with mutated MYD88 [63–68]. However, the interpretation of the data is complicated by a lack of standardization in the methodologies used to test for MYD88 mutations across trials. A comparison of outcomes for patients without MYD88 mutations from the iNNOVATE trial of ibrutinib plus rituximab versus rituximab monotherapy and the pivotal ibrutinib study is illustrative.

MYD88^WT patients in the iNNOVATE trial treated with ibrutinib plus rituximab had much higher major response rates (73% vs. 0%) and much longer progression-free survival (>4 vs. 0.4 years) compared with MYD88^WT patients treated with ibrutinib monotherapy in the pivotal study. However, the iNNOVATE trial used an NGS assay with either unselected bone marrow aspirate or formalin-fixed paraffin-embedded specimens to identify MYD88 mutations, whereas the pivotal ibrutinib trial used AS-PCR and Sanger sequencing with CD19-selected bone marrow aspirate. The relative percentage of MYD88^WT patients in the iNNOVATE trial was much higher than in the ibrutinib monotherapy study (16% vs. 6%) and likely reflects the lower sensitivity of the methodology and the presence of patients misclassified as MYD88^WT [74,75,104–106].

Related to the need for standardization, it is essential to report the sensitivity and specificity of assays used to test for MYD88 status. These parameters must also be defined outside of clinical trials. MYD88 status can determine the differential diagnosis of WM from related conditions and the potential for false-negatives must be considered, as well as the potential for activating MYD88 mutations other than L265P.

The importance of an accurate assessment of MYD88 status was repeatedly stressed during the 11th International Workshop on Waldenström’s Macroglobulinemia (IWWM). The panel recommended the adoption of ddPCR as the gold standard for MYD88 testing. The higher sensitivity of ddPCR compared with AS-PCR reduces the risk of false-negatives and misdiagnosis [96]. Its higher sensitivity also increases the potential to test for mutations using non–bone marrow sources of DNA, such as cell-free DNA [96]. Noninvasive approaches to molecular testing will be particularly important if/when MRD monitoring is incorporated into post-treatment evaluation in WM.

More research is needed to determine the value of evaluating the mutational status of CXCR4, TP53, and other genes that are mutated at a lesser frequency than MYD88. Routine analysis of CXCR4 is difficult and costly, but CXCR4 mutations are considered essentially unique to WM and impact response rates and survival outcomes with BTK inhibitors [46]. Accordingly, measures to include CXCR4 mutations in diagnosis should be investigated. Accurate detection of CXCR4 mutations will require NGS coupled with CD19⁺ cell selection. Similarly, TP53 requires further study in WM using similar approaches.

Molecular-based treatment algorithms have been proposed for WM, and we would encourage detailed genomic assessments in all clinical trials so that these can be further refined and applied to routine clinical management.

Article highlights

• Waldenström’s macroglobulinemia (WM) is an indolent B-cell lymphoplasmacytic lymphoma (LPL) with immunoglobulin M (IgM) monoclonal gammopathy and morphologic evidence of bone marrow infiltration by LPL.

• About a quarter of patients with WM are asymptomatic at the time of diagnosis; anemia is the primary reason for patients with WM to present to primary care.

• Bone marrow biopsy is essential for a definitive diagnosis of WM.

• A typical WM immunophenotype is a powerful and potentially underutilized tool in the differential diagnosis of WM from related conditions, such as marginal zone lymphoma.

• The WM clone expresses pan-B-cell markers (CD19, CD20, CD22, Sig) but can be distinguished from normal B cells by flow cytometry based on weak CD22 expression and homogeneous positivity for CD25.

• The MYD88 p.L265P (NP_002459.2) mutation is found in >90% of patients with WM and is strongly recommended in the diagnosis of patients with suspected WM.

• CXCR4 and TP53, which are mutated at a lesser frequency than MYD88, have limited relevance in establishing a diagnosis of WM but have some prognostic value.

• There are no standardized protocols for mutational analysis of MYD88; allele-specific (AS) polymerase chain reaction (PCR) is used by most laboratories to detect MYD88^L265P, but droplet digital PCR has emerged as an alternative that allows absolute quantification of DNA and provides higher sensitivity than AS-PCR.

Declaration of interest

RGS declares honoraria from Astellas, BeiGene, Incyte, Janssen, Novartis, the Spanish Public National System, and Takeda; consulting services to BeiGene and Takeda; and research funding from Astellas, Gilead, Incyte, IVS Technologies (patent partial use), Janssen, the Spanish Society of Hematology and Hemotherapy, and Takeda.

RO has received honoraria from Janssen, BeiGene, Celgene, and AstraZeneca; and has provided advisory services to BeiGene and Janssen.

MV has received honoraria from AbbVie, AstraZeneca, BeiGene, and Janssen; and provided advisory services to AbbVie, AstraZeneca, BeiGene, and Janssen.

The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed.

Reviewer disclosures

Peer reviewers on this manuscript have no relevant financial relationships or otherwise to disclose.

Acknowledgments

Editorial support was provided by Luke Smith, PhD, of Porterhouse Medical, to collate and incorporate author input on the manuscript, all carried out under the authors’ direction.

Additional information

Funding

Editorial support was funded by BeiGene.