BACKGROUND

Lymphoplasmacytic lymphoma (LPL) is a neoplasm with small clonal B lymphocytes, plasmacytoid lymphocytes, and plasma cells, usually involving the bone marrow (BM) and, sometimes, the spleen and the lymph nodes.1 LPL is often associated with monoclonal immunoglobulin (MI) in serum, and for the diagnosis of Waldenstrom macroglobulinemia (WM) a MI of immunoglobulin (Ig)M-type of any concentration is required as well as BM engagement. In LPL not fulfilling the diagnostic criteria for WM (here named non-WM LPL) the MI is of IgA- or IgG-type, or missing, or if of IgM-type the disease does not infiltrate the BM. The diagnosis of WM/LPL can, however, be challenging because of some morphological and immunophenotypic overlap with other low grade B cell lymphomas, especially marginal zone lymphoma (MZL) and plasma cell myeloma.1

MYD88 L265P is found in about 90% of WM cases,1 but is neither specific nor required for a WM diagnosis. It is also found in non-WM LPL and in other subtypes of lymphoma.2-7

Many of the clinical manifestations of WM and non-WM LPL are similar and often include symptoms such as anaemia, lymphadenopathy, and organomegaly. In addition, some patients present with B symptoms, including night sweats, recurrent fever, and weight loss.8, 9 Other manifestations of the disease are different in WM and non-IgM LPL patients. Symptoms in WM patients are often related to IgM MI such as hyperviscosity, cryoglobulinemia, or antibody mediated disorders, while non-WM LPL patients more often have symptoms due to nodal disease.

Compared to WM, non-WM LPL is uncommon, around 5% of all LPL cases, and only a few systematic studies have been published.2, 6, 10-12

Here, we present information from an analysis of data from a nation-wide population-based lymphoma registry that includes 1511 WM/LPL patients. The focus was on patient characteristics and outcome of 33 patients in whom we could confirm a non-WM LPL diagnosis.

PATIENTS AND METHODS Swedish cancer and lymphoma registries

The Swedish Cancer Registry (SCR), established in 1958, is a compulsory registry based on reports from both diagnostic and clinical units. In 2000, the Swedish Lymphoma Group initiated the Swedish Lymphoma Registry (SLR) to gather complementary information specific to lymphoma. The SLR includes data on all subtypes of lymphoma in patients ≥18 years old, reported by the responsible physician. The SLR’s validity and coverage is high and more than 95% of the patients registered in the SCR are also reported to SLR, with an opportunity to cross-link the registries. Between 2000 and 2006, data in the SLR were restricted to clinical characteristics and prognostic factors at the time of diagnosis, time for first-line treatment, and time of death.

The following data from the time of diagnosis were collected from the SLR: age, B-symptoms, World Health Organization performance status (WHO PS), sex haemoglobin level, lymphocyte counts, albumin level, and lactate dehydrogenase (LDH) level. If these data were missing in the SLR, the data were extracted from medical records as well as the reports from serum protein electrophoresis, including the type of M-component and pathology analyses. The following values were dichotomized: haemoglobin > or ≤115 g/l, platelet counts > or ≤100 × 10⁹/l, albumin > or ≤35 g/l, and WHO PS 0/WHO PS1-4. Because data on CT scan were missing in nearly half of the patients in the registry, spleen and lymph node size were not included in the analyses. Furthermore, because data regarding the type of first-line treatment were not mandatory to report to the SLR between 2000 and 2006, these data were missing in more than half of the patients.

Study population

In total, 1511 patients included in the SLR with a broad diagnosis of WM/LPL between 1 January 2000 and 31 December 2014 were evaluated for the sub-diagnoses WM and non-WM LPL by extracting data from the medical records (Figure 1). For 1139 patients (75%), a serum protein electrophoresis (SPE), including immunofixation, and type and level of MIs was available as well as a pathology report of a biopsy/aspiration of the bone marrow and/or lymph node at the time of diagnosis.

Flowchart of patient selection process

In Sweden, the WHO diagnostic criteria for lymphoma were used during the period and, as in the WHO 2017 classification, we defined non-WM LPL as a disease without a serum MI of IgM type, but with an IgA or IgG MI and infiltration of LPL cells in lymph nodes or BM, or LPL with an IgM MI, but without BM infiltration. After evaluation of the SPEs and the pathology reports, criteria for WM were fulfilled in 981 of 1139 patients (86.1%) (fully described in a previous publication13); 124 of the 1139 patients (10.9%) seemed to fulfil the non-WM LPL criteria. Details of the patient selection process are shown in Figure 1.

To further verify the diagnosis of non-WM LPL, an experienced haematopathologist reviewed all 124 pathology reports according to the WHO 2017 criteria. In 57 cases with any uncertainty, the same haematopathologist reviewed the diagnostic tissue samples. The formalin-fixed, paraffin-embedded blocks could be retrieved and used for complementary immunohistochemically staining in all 57 cases and for MYD88 L265P mutation analysis in 11 cases.

Analysis of MYD88 L265P mutation

DNA was extracted using the QIAamp DNA FFPE tissue kit (Qiagen®) from archived bone marrow aspirates, lymph nodes or spleen. Semi-quantitative real-time PCR analysis was performed in triplicates using a Rotorgene-Q real-time polymerase chain reaction (PCR) Cycler (Qiagen®) and a cell line heterozygous for the L265P mutation (RPC1-WM1) as a positive control. Primers were specific for exon 5 of MYD88 (NM_002468.4), including one primer specifically targeting the c794T>C, the Leu265Pro (L265P) mutation. The analyses were performed at Karolinska University Hospital, Department of Pathology.

Statistical analysis

All statistical analyses were performed in the free R software, version 3.5.3 and the packages survival,14 relsurv15 and prodlim.16 To estimate overall survival (OS), the Kaplan–Meier method was used and the subsequent curves were compared using a log-rank test.17 Relative survival (RS) was estimated using the Pohar-Perme estimator18 and the subsequent survival curves were compared using a log-rank type test.19 To measure RS, life tables of the population of Sweden were downloaded from the Human Mortality Database.20 The median follow-up time was estimated using the reverse Kaplan–Meier method.21 The Pearson’s chi-square test was used for frequency tabulation. Univariate Cox regression was performed in SPSS Statistical software, version 24 (IBM Corp., Armonk, NY, USA).

This study was approved by the regional ethics committee of Umeå (2011-44-31 M + 2017-106-32 M). The study was conducted in accordance with the Declaration of Helsinki.

RESULTS

Initially, 124 (10.9%) of the patients in the SLR did not fulfil the diagnosis of WM LPL. (For details, see Figure 1). After reviewing all available pathology reports for these 124 patients and the BM/tissue samples in 57 of 124 patients (46%), we could confirm the diagnosis of non-WM LPL in only 33 patients. In 62 cases we found a misdiagnosis with the following final diagnoses: small B-cell lymphoma Not Otherwise Specified (NOS) (n = 6); marginal zone lymphoma (MZL) (n = 33); chronic lymphocytic leukaemia (CLL)/ small lymphocytic lymphoma (n = 7); multiple myeloma /plasmacytoma (n = 6); AL-amyloidosis (n = 1); monoclonal gammopathy of unknown significance (MGUS) (n = 2); aggressive lymphoma (transformed indolent lymphoma?) (n = 2); T-cell lymphoma (n = 1); multicentric Castleman disease (n = 1) and follicular lymphoma (n = 1). In two cases, a biclonal paraprotein (both with MIs of IgM and IgG type) was identified, fulfiling the WM diagnosis and, therefore, excluded from further analysis. In total, 29 cases had insufficient information or lost/absent diagnostic material.

Results for the patients with non-WM LPL (n = 33) Morphological findings

Patients with non-WM LPL showed BM involvement in 30 of 33 patients (91%) with a high cellularity (>75%) in about half of the patients. The lymphoma cells showed plasmacytic differentiation and were composed of a mixed population of small lymphocytes, lymphoplasmacytic cells and plasma cells in varying frequencies. The pattern of infiltration varied: nodular two of 18 (11%), diffuse eight of 18 (44%), paratrabecular two of 18 (11%), interstitial one of 18 (6%) and mixed five of 18 (28%). In eight of 23 patients (35%) Dutcher bodies were identified.

Immunophenotypical findings

In all patients with non-IgM LPL, flow cytometry or immunohistochemistry was performed. The B-cells markers CD20, CD19, CD22 and CD79a were positive in all cases. The cells expressed the light chain kappa in 22 of 33 patients (67%) and lambda in eight of 33 patients (24%). In three of 33 patients (9%), the data on the light chain were missing. CD5 was positive in two of 18 (11%), CD10 negative in 17 of 17 (100%), and CD23 positive in six of 17 (35%) analysed cases.

Genetic findings

Conventional cytogenetic investigations were not performed. MYD88 L265P was analysed in 11 of 33 patients (33%) as part of the review and was positive in eight of 11 cases (73%). We could not show any difference in the clinical presentation between MYD88 L265P positive and negative patients, but the number of patients was small. Out of the eight patients positive for MYD88 L265P, three showed an IgG LPL and one IgA LPL with lymphocytosis. Two patients had non-secretory disease and one had free light chains in serum. One patient had an IgM MI and lymph node infiltration of LPL cells, but no BM involvement.

Two of the three MYD88 L265P negative patents had an IgG MI, both with bone marrow involvement and one had an IgA LPL with involvement of the bone marrow, lymph nodes and thymus.

Clinical and MI characteristics in non-WM LPL patients

The median age was 69 years at the time of diagnosis. There were 18 males (55.5%) with a median age of 67 years (range 38–89 years) and 15 females (45.5%) with a median age of 69 years (range 49–84). Of the males, nine of 19 (47.4%) were 66 years or older, as were eight of 15 (53.3%) of the females.

In most of the patients with non-WM LPL, the lymphoma cells secreted a paraprotein, mostly IgG (18/33 patients, 54.5%) or IgA MI (5/33 patients, 15.6%). The median concentration of the paraprotein IgG and IgA was 28.4 g/l (range 4–72) and 29.8 (range 16–45), respectively. Non-secretory LPL was seen in four of 33 cases (12.2%). Although the median age was higher in patients with IgG LPL compared with those with IgA LPL (69.5 and 57 years respectively), this was not significant, nor were other patient characteristics, such as sex, WHO PS, elevated LDH, B symptoms, lymphocytosis ≥5 × 109 or albumin level, significantly different between the two groups (Pearson’s chi-square test).

Comparison of non-WM LPL and WM

Comparison of the non-WM LPL patients and the WM patients, who were fully described in a previous article (Brandefors et al.13), is shown in Table 1. Patients with non-WM LPL were younger, had more adverse prognostic factors, such as elevated LDH, anaemia and lymphocytosis at diagnosis, than WM patients. In Cox univariate analysis, age was the only prognostic factor for OS in non-WM LPL patients [hazard ratio (HR) 1.054, 95% confidence interval (CI) 1.002–1.108, p = 0.04], while in WM patients the following prognostic factors for OS were identified: age, WHO PS, elevated LDH, B symptoms, Hb ≤115g/L, platelet count ≤100 × 109/L, lymphocytes ≥5 × 109/L and albumin level. Non-WM LPL patients received treatment at diagnosis more often. Between 2007 and 2014, 13 of 19 (68.4%) of the non-WM LPL patients and 153 of 575 (26.6%) of the WM patients needed therapy at the time for diagnosis. Most of the non-WM LPL patients received therapy regimens containing rituximab (10/13, 77% patients). The most used chemotherapies were chorambucil (4/13, 31% patients) and bendamustine (4/13, 31% patients) or different cyclophosphamide containing regimens (4/13, 31% patients). Only one patient received single rituximab. In WM patients, about half of the patients received rituximab containing therapies and 16% received single rituximab. Further, one third of the therapies were chlorambucil, 21% fludarabin containing regimens, 22% cyclophosphamide containing regimes and only 5% contained bendamustine.

TABLE 1. Patient and disease characteristics at diagnosis in non-WM LPL patients (n = 33) and WM patient (n = 981) Patients with a diagnosis of non-WM LPL (n = 33) Patients with a diagnosis of WM (n = 981) p -value Median age (range), years 69 (38–89) 73 (29–94) Sex 17 (51.5%) 721 (73.5%) 0.01 Male 0.59 Female 18 (54.5%) 596 (60.8%) 15 (45.5%) 385 (39.2%) WHO PS 1 22 (66.7%) 620 (64.0%) 11 (33.3%) 328 (34.6%) LDH – elevated (n = 32a) 12 (37.5%) 118 (13.7%) <0.001 B-symptoms – yes (n = 30a) 7 (23.3%) 143 (15.6%) 0.38 Haemoglobin ≤115 g/L (n = 28a) 20 (71.4%) 393 (48.2%) 0.03 Platelets ≤100 × 109/L (n = 21a) 1 (4.8%) 50 (7.8%) Lymphocytosis ≥5 × 109/L (n = 21a) 8 (38.1%) 40 (6.5%) <0.001 Albumin level, median 32 35 Albumin ≤35 g/L (n = 27a) 17 (63.0%) 405 (51.1%) 0.31 MI IgM 2 981 (100%) MI IgM kappa 2 642 MI Ig M lambda — 339

MI IgG (kappa/lambda)

18 (54.5%)

0

0

0

MI IgA (kappa/unknown)

5 (15.6%)

0

0

0

Bence Jones proteinuria - kappa 1 (3.0%) — Free light chains/s - kappa 1 (3.0%) — Non-secretory LPL 4 (12.2%) — No bone marrow involvement, MI of IgM type, both kappa 2 (6.1%) — No bone marrow involvement, MI unknown, enlarged lymph node - kappa 1 (3.0%) — Treatment at diagnosis (2007–2014) 13/19 (68.4%) 153/575 (26.6%) <0.001 Abbreviations: Ig, immunoglobulin; LDH, lactate dehydrogenase; MI, monoclonal immunoglobulin; WHO PS, World Health Organization performance status.

The median follow-up times from diagnosis for non-WM LPL and WM were 75.2 and 79.9 months, respectively. By the end of the observation period, 17 (50%) of the non-WM LPL and 449 (45.8%) of the WM patients had died. Neither the OS nor the RS differed significantly between the two groups (p = 0.249 and p = 0.137 respectively), (Figure 2a). However, the median survival time was shorter in non-WM LPL compared to WM (71 vs. 96 months respectively), as was the 3-year and 5-year survival (71% and 55% respectively) (Figure 2b) with corresponding figures for WM 78% and 66% respectively (Brandefors et al.13). Between the periods 2000–2006 and 2007–2014 the median OS and RS did not significantly increase for non-WM LPL patients (p = 0.161 and p = 0.233 respectively). (Figure 2b), while the patients with WM had significantly increased median OS and RS between these two periods, p < 0.001 and p = 0.013, respectively.13

Kaplan-Meier curves for overall survival (OS) and relative survival (RS); (A) non-WM LPL and WM for the years 2000–2014, (B) non -WM LPL for two time periods; 2000–2006 and 2007–2014, (C) males and females with non-WM LPL for the years 2000–2014

Females with non-WM LPL had a significantly superior survival than males, with the sex difference in OS (p = 0.002) and RS (p = 0.004) (Figure 2) more pronounced in the first period, 2000–2006. This sex difference was not found in WM patients (OS: p = 0.18; RS: p = 0.463). When comparing males and females, no significant differences were found regarding patient characteristics, such as age, WHO PS, elevated LDH, B symptoms, lymphocytosis ≥5 × 109, albumin level or type of treatment.

DISCUSSION

In a national, population-based lymphoma registry, including 1511 patients with the diagnosis of WM/LPL, we identified 124 patients who did not fulfil the diagnostic criteria for WM. After review of all 124 pathology reports and in 57 cases, also the diagnostic tissue samples, 33 patients were confirmed to fulfil the diagnostic criteria of non-WM LPL; the remaining patients had other types of small B-cell lymphomas or insufficient diagnostic material. We aimed to include non-WM LPL with a distinct certified diagnosis, therefore we excluded 29 patients because of insufficient diagnostic material, e.g. only fine needle biopsies, insufficient or poor diagnostic tissue, and lost diagnostic material. Of the excluded patients 65% had the diagnosis before 2007, but since then the SLR has been updated with more defined variables and improved diagnostic accuracy. It is well known that the diagnosis of WM/LPL can be challenging and sometimes difficult to distinguish from other small B-cell lymphomas. In our cohort, six of the patients registered as WM/LPL received the final diagnosis of small B-cell lymphoma UNS, in agreement with the WHO classification that ‘the distinction between the LPL and other small B-cell lymphomas is not always clear-cut.’1

The morphological and immunophenotypical features are considered the same for WM and non-WM LPL, also confirmed in our study, but the finding of CD23 expression in our non-WM LPL (6/17, 35%) is higher than in 244 WM patients described from the University Hospital of Salamanca, 11%–15%.22

A valuable diagnostic tool for differentiating WM from other subtypes of lymphoma with lymphoplasmacytic differentiation is the MYD88 L265P mutation.23, 24 We detected MYD88 L265P in eight of 11 non-WM LPL patients (73%); in other studies, the corresponding figure ranges from 40%–71%.2, 6, 10, 12 MYD88 L265P was analysed in three patients who eventually had another diagnosis. One of these, a patient with the final diagnosis of small B-cell lymphoma NOS, showed a MYD88 L265P mutation.

Data on nodal and extranodal involvement are incomplete in the SLR, but from medical records and pathology reports, we found 11 patients with nodal, two with extranodal involvement, and one with central nervous system infiltration (Bing-Neel syndrome). A recent study showed that non-WM LPL patients more often had enlarged lymph nodes, splenomegaly and extranodal involvement.2

Relative survival may better reflect disease-related survival, especially in patients of older age and when the cause of death is unknown. The RS for non-WM patients at 5 years was only 60% versus 79% in WM patients. However, the long-term RS and OS did not show any significant difference. The survival in earlier studies is divergent, for example, the study from Varettoni et al. showed no difference in survival between non-WM LPL and WM patients, while another study showed a worse OS for non-IgM LPL.2, 10 In our cohort, female non-WM LPL patients had a significantly longer median OS and RS compared to male patients. We cannot explain this as our females were not younger, nor had better prognostic factor, nor were treated differently than the males. The sex difference in survival has also been seen in other subtypes of lymphoma.25 Pfreundschuh et al. have shown that elderly females with diffuse large B-cell lymphoma had greater benefit of treatment with rituximab than elderly males, due to slower rituximab clearance in women.26 The number of patients with treatment at diagnosis were too small for statistical analysis.

Our study has several strengths. To our knowledge, this is the only study based on an unselected population from a national registry covering all lymphoma diagnoses during a period of 15 years. The analyses includes review of the diagnostic pathological reports and the tissue samples from all non-WM LPL patients. Compared with other small observational studies, we had access to a large amount of specific clinical data.

This study also has some limitations. Because of the disease rarity, only 33 non-WM LPL patients could be included. Some important data in SLG were not available or were incomplete such as results from CT scans, treatments before 2007, and all second line treatments.

In summary, our population-based study pointed out the challenge to correctly diagnose patients with non-WM LPL. We could show differences in clinical features between WM and non-WM LPL, but no significant differences in long-term survival. To better understand this disease entity more studies are needed.