Introduction

According to the World Health Organization (WHO) classification primary bone lymphoma (PBL) or primary non-Hodgkin lymphoma of bone (PB-NHL) is defined as a single bone lesion with or without regional lymph node involvement, or as multiple bone lesions without visceral or lymph node involvement [1]. PB-NHL is a rare type of lymphoma, representing < 1% of all lymphomas and < 5% of all extranodal lymphomas [2, 3]. Several different histologic lymphoma types have been described in the category of PB-NHL where diffuse large B-cell lymphoma (DLBCL) is, by far, the most common histologic type and represents over 95% of all cases [2,3,4]. The initial diagnosis of PB-NHL can be challenging owing to unusual clinical presentation, difficulties in obtaining sufficient diagnostic material, and the frequent necrosis and crush artifact in the biopsy specimen [2]. PB-NHL (PB-DLBCL) usually presents in a relatively young population (55 years) and have a good prognosis with a 5-year overall survival (OS) rates ranging from 60% to 95% (mean OS, 82%) [4,5,6,7]. Previously it was shown that most of PB-DLBCL (90%) have Germinal Center B-Cell (GCB) phenotype by Li et al. which is usually associated with better prognosis compared with Non-Germinal Center B-Cell (Non-GCB) imminophenotype [4].

The excellent survival rate is similar or even superior to that observed for patients with non-osseous DLBCL originating from germinal center B cells (i.e., GCB type of DLBCL) treated with rituximab-combined chemotherapy [4]. However, the biological explanation for the favorable prognosis of patients with PB-DLBCL compared to non-osseous DLBCL remains mainly unknown [4].

Herein, we present a rare case of PB-NHL (PB-DLBCL) in a 45-year-old male with a right mandibular mass who was originally thought to have post-tooth extraction osteomyelitis. After the patient was diagnosed with PB-NHL (PB-DLBCL), he showed an excellent response to two cycles of rituximab with cyclophosphamide, vincristine, prednisone (R-CVP) therapy. We stress on importance of flow cytometrical evaluation in establishing PB-NHL diagnosis and to facilitate therapeutic approach in such cases. Furthermore, we performed a systemic literature search and review via PubMed (Medline), Ovid (Medline), CINAHL, Clinical Key, Scopus, Google Scholar database and identified and discussed 42 similar cases published in the English language literature [8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39].

Case Presentation

An otherwise healthy 45-year-old male was admitted to the hospital presenting with a large right jaw mass and pain. A complete blood count with differential and complete metabolic panel were within normal limits. The patient was seen 4 months prior for the extraction of an infected right lower molar tooth. Three months after the initial tooth extraction, the patient was seen by the oral and maxillofacial surgery service for persistent pain and swelling of the jaw since the initial tooth extraction (Fig. 1).

Fig. 1
figure 1

A Initial presentation of right sided mandibular mass. B Blue arrow heads indicate borders of intra-oral lesion observable on physical examination in the right mandibular region

Full size image

A computerized tomography (CT) scan of the head and neck was performed and revealed ill-defined heterogeneous soft tissue with fluid attenuation in the right oromandibular location with cortical disruption in the adjacent mandible with local bone involvement and no regional lymphadenopathy (Fig. 2). A differential diagnosis of osteomyelitis, atypical infection, and soft tissue tumor was considered and was followed by incisional biopsy and culture sampling of the lesion. Unfortunately, histological findings failed to reveal the neoplastic process and showed mainly necrotic tissue with severe crushed artifact, reactive bone changes and lymphoplasmacytic inflammation with fibrosis without evidence of fungal infection. The patient received 1 week of intravenous antibiotics and was discharged with 3 weeks of oral antibiotics. At 3-week follow-up, due to the lack of clinical improvement on antibiotics, a fine needle aspiration (FNA) with rapid on-site evaluation (ROSE) was performed. Atypical large lymphoid cells with a moderate amount of cytoplasm, irregular nuclei, and prominent nucleoli were seen and material was obtained for flow cytometric evaluation and core biopsies of the lesion were taken (Fig. 3).

Fig. 2
figure 2

Axial (A) and coronal (B) images from the CT scan of the neck, in bone windows, showing bone expansion and trabecular sclerosis involving the right mandibular body (solid arrows) with some periosteal reaction, all consistent with bone involvement with lymphoma. There is also an overlying soft tissue mass (dashed arrows) which is consistent with extraosseous extension of lymphoma

Full size image
Fig. 3
figure 3

Fine needle aspiration cytology smears (A and B) show single scattered large atypical lymphoid cells with high nucleus-cytoplasm ratio and dysmorphic nucleuses (1000 ×)

Full size image

Histologic sections showed soft and bone tissue diffusely infiltrated by partially necrotic atypical lymphoid cells and histiocytes (Fig. 4A). Immunohistochemical analysis of lymphoid cells were positive for CD20 (Fig. 4B), CD79a, PAX5, cMYC, BCL2, BCL6 (Fig. 4C) and negative for CD3P, CD5, CD10, MUM1, TdT, and CD68. Proliferation rate was assessed by the expression of Ki-67, which was high at approximately 80% (Fig. 4D). In situ hybridization (ISH) for EBV (EBER) was negative and Kappa and Lambda light chains showed a slight predominance of Kappa light chain in viable areas. Special stains for an acid-fast bacteria (AFB) and Grocott methenamine silver (GMS) were performed and showed negative results.

Fig. 4
figure 4

A Histopathological examination of lesion biopsy showed predominantly necrotic and focally fibrotic diffuse lymphoid infiltrate surrounding bone trabecular (H&E, 200 ×). Immunohistochemistry showing strongly positive staining for CD20 (B, 200 ×), BCL6 (C, 200 ×), and high proliferation index, approximately 80%, by Ki-67 (D, 200 ×)

Full size image

Flow cytometric immunophenotyping identified an aberrant B-lymphocyte population with Kappa light chain restriction (Fig. 5). These morphologic and immunohistochemical studies were suggestive of focally necrotic diffuse large B-cell lymphoma involving bone.

Fig. 5
figure 5

Flow cytometric immunophenotyping identified an aberrant B-lymphocyte population with Kappa light chain restriction

Full size image

Consequent staging imaging studies of the chest, abdomen, and pelvis were performed and failed to identify any evidence of other sites of lymphoma involvement or lymphadenopathy elsewhere in the body. Furthermore, flow cytometric immunophenotyping of blood samples was negative for any aberrant lymphocyte populations. The diagnosis of PB-NHL of the mandible was established.

Fluorescence in situ hybridization (FISH) for c-MYC, BCL2, and BCL6 rearrangements was performed and showed negative results.

Chromosome microarray analysis (CMA) was performed at the University of Texas Medical Branch and identified the gains of chromosome 7 and 21, loss of 15q11.1q22.2, and copy neutral loss of heterozygosity (cn-LOH) of 1p36.33p36.21 and 6p25.3p21.2. The gain of chromosome 7 has been previously reported in large B cell lymphomas of bone (Fig. 6) [40].

Fig. 6
figure 6

Chromosome microarray analysis was performed at the University of Texas Medical Branch and identified the gains of chromosome 7 and 21, loss of 15q11.1q22.2, and copy neutral loss of heterozygosity (cn-LOH) of 1p36.33p36.21 and 6p25.3p21.2. The gain of chromosome 7 has been previously reported in large B cell lymphomas of bone

Full size image

Prior to chemotherapy initiation, a cardiologic evaluation was performed and revealed a reduced ejection fraction of 35–40% on echocardiography. Follow-up nuclear medicine myocardial perfusion testing found no evidence of ischemic changes and confirmed a reduced ejection fraction of 35%. R-CVP regimen, where doxorubicin was excluded due to reduced ejection fraction, was chosen, and promptly initiated. After receiving three cycles of R-CVP chemotherapy, the patient was doing well, showing almost near resolution of the jaw mass and pain (Fig. 7). The patient will continue to receive therapy and follow-up after completion of all chemotherapy cycles.

Fig. 7
figure 7

Reduction of initial mass size after 2 cycles of R-CVP treatment (see Fig. 1A and B for comparison)

Full size image

Materials and Methods

Hematoxylin and eosin (HE) and immunohistochemical (IHC) stains were performed on formalin-fixed, paraffin-embedded tissue sections at our institution. The following antibodies were used for IHC analysis: CD1a, CD2, CD4, CD5, CD7, CD8, CD10, CD21, CD34, CD43, CD79a, CD99, CD117, cyclinD1, PAX-5, MUM-1, Myeloperoxidase, and TdT (Agilent/Dako, Carpinteria, California). CD3, CD15, CD20, CD30, AE1/AE3, BCL-2, Ki-67 (Ventana/Roche, Tucson, Arizona) and CD45, CD56, BCL-6 and c-MYC (Cell marque, Rocklin, California). In situ hybridization (ISH) was performed using a probe for Kappa and Lambda light chains and Epstein-Barr virus (EBV)-encoded viral RNAs (Ventana/Roche, Tucson, Arizona).

Eight-color flow cytometric analysis (BD FACS Canto II, San Jose, California) was performed on the bone marrow tissue, BD FACS Diva software (San Jose, California) according to standard procedures at our institution. T-, B-, and myeloid lineage antigens were analyzed, using a comprehensive panel of monoclonal antibodies.

Fluorescence in situ hybridization (FISH) analysis with probes (Abbott Molecular/Vysis and Cytocell) specific for MYC (8q24), CDKN2A (p16, 9p21), MLL (11q23), TCR (14q11.2), IGH (14q32.3), E2A (TCF3;19p13.3), ETV6/RUNX1 (TEL/AML1, 12p13/21q22.3), ABL1/BCR (9q34.1/22q11.2), centromere 4 (D4Z1), centromere 10 (D10Z1) and centromere 17 (D17Z1) regions was performed at the reference laboratory on fresh tissue.

Chromosomal microarray analysis (CMA) was performed using Affymetrix CytoScan HD microarray. This microarray and associated software (Chromosome Analysis Suite) are manufactured by Affymetrix (Santa Clara, CA) and used by our institution’s Molecular Diagnostics Laboratory (MDL) for the purpose of identifying DNA copy number gains and losses associated with large chromosomal imbalances [41].

The systemic literature search was performed via PubMed (Medline), Ovid (Medline), CINAHL, Clinical Key, Scopus, Google Scholar database with specific keywords: “Primary bone lymphoma”, “primary bone lymphoma and jaw”, “primary bone lymphoma and mandible”, “primary bone lymphoma and maxilla”, “Primary Non Hodgkin Lymphoma of Bone”.

Discussion

Within known cases of extranodal lymphomas, PB-NHL is exceptionally rare, representing < 1% of all lymphomas and < 5% of all extranodal lymphomas [2, 3]. Where only 0.6% of all PB-NHL presents in the mandible [17, 22].

The clinical presentation of our patient, when he presented with severe pain and jaw mass after recent tooth-extraction with resistance to antibiotic treatment was commonly seen in the literature review. This sequence of events in disease presentation could be potentially recognized as a sufficient clinical clue found for PB-NHL as was commonly seen in previously reviewed cases of PB-NHL. Interestingly, the only mantle cell lymphoma case was different in clinical presentation from most of reviewed cases, where the patient presented with the absence of inflammation, purulent discharge, and pain. However, the therapeutic approach in this case utilized CHOP-based chemotherapy, like most other cases of PB-NHLs, and showed a good clinical response after 3 cycles with the disappearance of the mandibular swelling (Table 4) [29].

In order to look for other potential signs which could be helpful in recognized PB-NHL early we performed an extensive review of previously published cases and summarized the findings in Table 1. Of reviewed cases, 71% (30) of patients presented with periodontal disease, where 80% (24) of these patients presented with periodontal disease before the developing other symptoms. Regional paresthesia was also a common symptom reported in cases of periodontal PB-NHL and likely represents an early presentation of disease. The presentation of regional paresthesia was reported in 40% (17) of patients. Beside the previously mentioned symptoms, most of reviewed PB-NHL cases (95%, 40 cases) presented without regional lymphadenopathy (LAD) or classic B-symptoms (98%, 41 cases), such as fever, chills, or weight loss. Interestingly, initial imaging studies were helpful in recognizing bone involvement in 76% (32) of cases (see Table 1).

Table 1 Summary of patient demographics and initial clinical presentations of periodontal PB-NHL collected from cases in literature
Full size table
Table 2 Summary of differential diagnoses considered, and final diagnoses made based on pathologic findings from cases in literature
Full size table

Despite the relatively classic clinical presentation of periodontal PB-NHL, according to our literature review only in 7% (3) of cases physicians considered lymphoma as a potential differential diagnosis (Table 2) [26, 31]. That is why we believe that awareness of these characteristic features may better facilitate the rate of diagnosis and consequent treatment. Even though patients with PB-NHL generally have relatively good 5-year progression free survival (PFS) and overall survival (OS), the delay of correct diagnosis and consequent chemotherapy could potentially poorly impact patient outcomes [2, 42]. Therefore, the development of reliable diagnostic criteria and features in recognizing periodontal PB-NHL is of considerable importance.

Table 3 Summary of histopathology findings and performance of flow cytometry, EBV testing, and HIV testing collected from cases in literature
Full size table

Although necrosis and reactive fibrosis were previously described as the histopathologic findings in PB-NHL, our literature review failed to identify any additional characteristic features. Approximately 74% (31) of cases of periodontal PB-NHL did not provide detailed histologic evaluation and lacked information describing necrosis or reactive connective tissue changes (Table 3). However, it does not conclude that these findings were truly absent, but rather they were likely overlooked. Only 26% (11) of cases described some degree of necrosis and fibrosis on histological evaluation. The sensitivity and specificity of IHC can be negatively impacted by poor quality tissue samples containing extensive fibrosis, necrosis, or crush artifact [43,44,45,46].

Beside characteristic clinical presentation, certain diagnostic studies can be helpful and essential for PB-NHL recognition, especially taking in consideration diagnostic difficulties: very necrotic material and severe crushed artifact [2, 5, 47]. Flow cytometry analysis has the characteristic ability to recognize even small viable populations of clonal or aberrant cells on the background of severe necrosis. We believe that after initial imaging evaluation, patients should be sent for FNA with consequent onsite cytology and flow cytometry assessment as an essential step for diagnosing PB-NHL. Additionally, this diagnostic sequence could potentially prevent unnecessary invasive procedures associated with aesthetic alteration of the jaw and surgical complications [5]. Interestingly, that as per our literature review, flow cytometry evaluation was not widely used and only 12% (5) of cases utilized flow cytometry in the diagnostic process [26, 31]. After FC testing and discovery of an aberrant, clonal population, IHC and molecular analyses should be performed to further lymphoma categorization.

There are three main differential diagnoses which should be ruled out before making a diagnosis of PB-NHL: plasmablastic lymphoma (PBL), B-lymphoblastic lymphoma (B-LBL), and Burkitt lymphoma (BL).

One of the main differential diagnoses for periodontal lymphomas, especially in HIV positive patients, is plasmablastic lymphoma (PBL) [48,49,50]. It is a very aggressive CD20-negative large B-cell lymphoma with plasmacytic immunophenotype. It was originally described in the oral cavity and frequently occurs in association with HIV infection, even though it was described in other locations and with other causes of immunodeficiency [51, 52]. It is important to rule out PBL since it responds poorly to commonly used chemotherapy regimens for PB-NHL and more than 75% of patients die of the disease, with a median survival of 6–11 months [48,49,50,51,52,53,54]. This lymphoma type can be ruled out by using our diagnostic approach with characteristic flow cytometric and consequent IHC tissue evaluations: (usually CD20 negative, CD138 or CD38 positive, MYC positive, and EBV positive).

Another important differential diagnosis is B-lymphoblastic lymphoma (LBL). It is a clinically aggressive disease, representing approximately 2% of all non-Hodgkin lymphoma cases. About 6 cases of B-LBL were described in the oral and maxillofacial region [48, 55]. The treatment of B-LBL is based on B-cell acute lymphoblastic leukemia (B-ALL) protocol: modified hyper-CVAD plus rituximab regimen. The addition of rituximab has improved 3-year overall survival from 35% to 68% compared with hyper-CVAD alone in B-ALL/LBL patients. This lymphoma type can be easily ruled out by using our diagnostic approach, with characteristic flow cytometric and consequent IHC tissue evaluations (usually CD10 positive, CD34 positive, TdT positive, and light chain clonality negative).

Burkitt lymphoma (BL) is a highly aggressive but curable lymphoma that often presents in extranodal sites or as an acute leukemia [48]. The endemic type of BL is known to classically occur in children (especially in equatorial Africa and in Papua New Guinea) and can present with involvement of the jaw and other facial bones (e.g., orbital bones) in about 50–70% of cases. The incidence peaks among children aged 4–7 years and a male-to-female ratio of 2:1 [48, 56, 57]. BL can be a highly aggressive but potentially curable tumor; intensive chemotherapy leads to long-term overall survival in 70–90% of cases, with children doing better than adults. In all patients with endemic BL, the EBV genome is present in > 95% of the neoplastic cells, which is why performance of EBV in situ hybridization could be critical for diagnosis. This lymphoma type can be ruled out by using our diagnostic approach as well, with characteristic flow cytometric and consequent IHC tissue evaluations (usually BCL2 negative, Ki-67 ~ 100%, MYC positive, and EBV evaluation). Molecular studies (FISH) for common BL translocations should be performed as well: t(2;8), t(8;14), and t(8;22) [48].

Every DLBCL with high grade features should be evaluated with the Hans algorithm in order to distinguish germinal center B-cell like (GCB) versus non-germinal center B-cell like (non-GCB) immunophenotype using IHC for CD10, BCL6, and MUM1 for prognostic purposes [58]. In situ hybridization studies for Epstein-Barr virus (EBV)-encoded viral RNAs should be performed as well to rule out EBV-positive DLBCL, NOT OTHERWISE SPECIFIED (NOS) [48]. Additionally, we need to remember that EBV can be positive in PBL and BL, which should be ruled out as well. Immunohistochemical evaluation for BCL2 and MYC could be performed for prognostic reasons and identify if it is double expressor lymphoma. This IHC characterization is especially helpful when material is not available to be sent for FISH or NGS. Double expression of BCL2/MYC is more frequently observed in non-GCB immunophenotype (66% vs 39%) and it was found to be predictive of poor survival [59]. The next step in lymphoma evaluation would be FISH studies to rule out high grade B-cell lymphoma with MYC and BCL2 and/or BCL6 rearrangements, since this type of lymphoma has a very poor prognosis [48].CMA analysis was performed at our institution to better characterize the specific neoplastic entity identified in our patient. However, whole genome sequencing and gene expression profiling (GEP) was not able to be performed due to non-viable tissue samples. GEP is at the frontier of diagnostic and prognostic testing in the evaluation of lymphoma. In a study by de Groen et al., they analyzed PB-DLBCL, polyostotic-DLBCL, disseminated-DLBCL, and non-osseous DLBCL (NO-DLBCL) through histologic, immunohistochemical, FISH, GEP, and targeted next generation sequencing (NGS) to evaluate trends between genotypic expression and phenotypic presentation [60]. Their findings revealed that stage I/II PB-DLBCL has significantly greater PFS/OS rates versus NO-DLBCL. They were able to identify a significant difference in the number of key mutations in B2M, EZH2, IRF8, TNFRSF14 between PB-DLBCL and NO-DLBCL, where PB-DLBCL was associated with at least one mutation in these genes [60]. Furthermore, PB-DLBCL was found to have significantly different clusters of expression regarding immune surveillance genes (e.g., CTLA4 and CXCL12) versus NO-DLBCL, possibly owing to the unique presentation of PB-DLBCL and favorable outcomes [60]. The conclusion of this study outlines the importance of molecular studies in proper identification of osseous DLBCL and offer new avenues for future studies to better characterize and manage this unique identity.

In our review of literature, we identified a significant variety of treatment regimens, where chemotherapy monotherapy and chemotherapy with radiation were the most common treatment methods; all findings were combined and organized in Table 4 for readers’ convenience. These regimens were similarly used, representing 31% (13) and 40% (17), respectively, and for a total of 71% of all treatments (Table 4). Cyclophosphamide, doxorubicin, vincristine, and prednisone (CHOP) based therapy (with or without rituximab) was the most common form of chemotherapy. Roughly 27% of cases used CHOP based therapies without radiation and roughly 29% of cases used CHOP based therapies with radiation. Of all studied cases, 76% (32) achieved complete remission, without any clear superiority of any regimen. Roughly 24% (10) of cases did not report any outcomes from treatment, further adding to the uncertainty of regimen superiority in our review. Only 13% (4) of cases relapsed after initial treatment.

Table 4 Summary of treatments used and consequent outcomes in patients from cases in literature
Full size table

Since the advent of anti-CD20 therapy (e.g., rituximab), the need for radiation therapy in combination with standard chemotherapy regimens has been a topic of active discussion [61]. Ramadan and colleagues demonstrated that patients who received a combination therapy with R-CHOP and consolidative radiation therapy experienced poorer outcomes compared with R-CHOP alone [62]. In another study by Batoo et al. they demonstrated that there was no clear benefit of radiation therapy in combination with R-CHOP therapy versus R-CHOP therapy alone [63]. While these studies have suggested that the addition of radiation therapy either demonstrates poorer outcomes or provides no statistical benefit, other investigations have revealed alternative findings. Tao et al. performed an extensive analysis of outcomes in patients treated with CHOP, R-CHOP, and combination therapy with radiation therapy [7]. They concluded that consolidative radiation therapy after chemotherapy had the greatest 5-year OS and PFS rates, compared to other treatment regimens. The 5-year OS rate for patients who received consolidative radiation therapy was 91% versus 68% for those who did not receive it, and the 5-year PFS rates were 88% versus 63%, respectively. Additionally, they demonstrated that consolidative radiation therapy was associated with improved OS (97% vs 67% at 5 years) and PFS (94% vs 74% at 5 years) in patients with stage I-II disease [7]. They found that patients treated with CHOP-based therapies experienced better 5-year OS and PFS rates (82% and 80%, respectively) compared to other chemotherapy regimens [7]. Their analysis did not find significant benefit for patients with the addition of rituximab or intrathecal therapy [7].

In consideration of the ongoing debate regarding the standard of care in PB-NHL, specifically in patients with large B-cell subtype, it is difficult to assess which treatment regimen will provide the greatest benefit. Furthermore, there have yet to be any treatment recommendations in PB-NHL specific to each different bone location (i.e., femur vs. mandible vs. pelvis, etc.). As such, the inclusion of rituximab and/or consolidative radiation therapy in treatment regimens should be tailored to disease stage and site of involvement for each patient.

Herein we would like to share our experience in treating such a rare type of lymphoma as PB-NHL of mandible. Our patient showed significant benefit from initial rituximab infusion prior to CVP therapy, evidenced by a decreased size of the mass 2 days prior to CVP therapy.

Furthermore, the benefit of including consolidative radiation therapy for periodontal PB-NHL should be weighed carefully against the long-term periodontal bone health, considering that outcomes in early stage PB-NHL are greater than in stage III–IV PB-NHL or in non-osseous lymphomas [4, 7, 61].

Conclusion

In this manuscript we summarize our experience with managing mandibular PB-DLBCL, provide comprehensive review of the literature on this topic and propose the University of Texas Medical Branch diagnostic approach for PB-NHL (see Fig. 8).

Fig. 8
figure 8

A Diagnostic approach in present case; B Diagnostic approach of periodontal primary bone lymphoma with flow cytometry as an essential diagnostic component

Full size image

The presented case illustrates the diagnostic challenges of periodontal PB-NHL and shows the characteristic presentation: the history of recent tooth extraction, development of local paresthesia, osteolytic findings on imaging, and rapid mass enlargement. Additionally, we would like to stress the importance of flow cytometry as an essential diagnostic test during evaluation of these patients because of the high risk of misdiagnosis due to unusual clinical and histological presentations.