Lymphoma has been historically classified into Hodgkin’s lymphoma (HL) and non-Hodgkin’s lymphoma (NHL) [1]. In Japan, HL is less common, whereas NHL accounts for 90%–95% of all lymphoma cases [1]. Lymphoma in the head and cervical regions is categorized into nodal lymphoma (NL) arising from the cervical lymph nodes and extranodal lymphoma (EL) from different body organs, including the oral cavity [2-5]. HL commonly develops as nodal lymphoma, with its most common site being the cervical and mediastinal lymph nodes [6]. In contrast, a lesion develops in the cervical lymph nodes in 20%–30% of all cases of NHL, and the primary extranodal lesion is observed in the head and cervical regions at a similar frequency [6,7]. Although most of the EL in the oral region can be histopathologically diagnosed using incisional biopsy, which is often misdiagnosed as oral cancer or odontogenic inflammation. In contrast, NL in the neck is often diagnosed by fine needle aspiration biopsy (FNAB), but there are cases in which a definitive diagnosis is not reached. Thus, we perform early excisional biopsy as early as possible, according to the Japanese clinical practice guidelines for head and neck cancer [8]. However, in several cases, the patients hesitate to undergo a cervical incision, leading to a delay in the diagnosis of their disease. In this study, we examined the clinical characteristics (clinicopathological parameters, imaging, histopathology, laboratory test and outcome) of lymphoma in which the patients had visited our hospital in the course of the past 11 years for the complaint of oral and cervical symptoms, resulting in the diagnosis of lymphoma, and discussed what is to be observed so as to diagnose this disease in patients in whom diagnosis was difficult.

The current study included 21 cases (8 males and 13 females) diagnosed with lymphoma at the Department of Oral Surgery, Dokkyo Medical University Hospital, over a span of 11 years (2003–2014). The following-up period was from 2003 to 2018. The study design was approved by the Medical Ethical Research Committee of Dokkyo Medical University Hospital (approval ID R49-19-J). The ages of the participants at the initial visit ranged from 14 to 85 years, with a mean age of 61.8 years. The characteristics examined included the site, symptoms, diagnosis at the initial visit, laboratory test values, histopathological diagnosis, time to confirmed diagnosis, as well as stage and outcome. The serum levels of lactate dehydrogenase (LDH), soluble interleukin-2 receptor (sIL2R) and thymidine kinase (TK) were assessed by Center for Clinical Laboratory, Dokkyo Medical University Hospital. The reference values were established as LDH (200-400 U/mL), sIL2R (220-530 U/mL) and TK (< 5.0 U/mL). In the current study, oral and cervical lesions were defined as EL and NL, respectively. To determine whether there is a significant difference, Student’s t-test was performed for the time to confirmed diagnosis, and Fisher’s exact test for the correlation between the outcome and the sIL2R or TK level. P-value <0.05 was considered statistically significant.

Site

Extranodal and nodal lesions existed in 12 and 9 cases, respectively. For the cases of EL, the affected site was the gingiva in 5 cases (2 for the maxilla and 3 cases for the mandible; Figure 1a), the palate in 4 cases (Figure 1b), as well as the mandibular bone, tongue, and buccal mucosa in 1 case each (Table 1). For the cases with a chief complaint of nodal lesions, the affected site was the submandibular node in 7 cases and the superior internal jugular node in 2 cases (Table 2). The sites of lymph nodes were identified on computed tomography (CT) images.

Figure 1: Intraoral picture for EL: (a) Case No. 6; The extraction socket of the left mandibular second molar was recessed, and elastic hard mass formation was observed on the buccolingual side gingiva. (b) Case No. 11; Elastic hard mass formation was identified from the posterior left hard palate to the soft palate. The mass was 20 × 20 mm in size, and the surface was slightly rough.

Symptoms

The initial symptom was elastic hard mass formation in 8 cases, swelling in 2 cases, pain in 2 cases, dysphagia in 1 case, hypoesthesia in 1 case, and healing failure of the extraction socket in 1 case for EL (Table 1), whereas it was swelling for most of NL (Table 2).

Diagnosis at the Initial Visit

In case of EL, the clinical diagnosis at the initial visit was oral cancer in 6 cases, lymphoma in 4 cases, as well as mandibular bone fracture and periodontitis in 1 case each (Table 1). Whereas, in case of NL, it was lymphoma (Figure 2c) in 5 cases, lymphadenitis (Figure 2a) in 3 cases, and cervical lymph node metastasis of submandibular gland cancer (Figure 2b) in 1 case (Table 2).

Figure 2: FDG-PET/CT image for NL: (a) Case No. 9; Lesion was shown slightly abnormal FDG uptake in the left mandibular node (stage I). (b) Case No. 7; Lesions were shown abnormal FDG uptake in the right mandibular and mid internal jugular node (stage II) (c) Case No. 1; Lesions were shown abnormal FDG uptake various parts of the body (bilateral cervical node, right pelvic lymph node etc.) (stage IV).

Time to Confirmed Diagnosis

The time to confirmed diagnosis was defined as the period starting from the initial visit to histopathological confirmation of diagnosis. This time was 14–74 days with a mean of 34.7 days for EL and 14–164 days with a mean of 66 days for NL. No significant difference in the meantime to confirmed diagnosis was observed between EL and NL (P = 0.1).

Histopathological Diagnosis

The histopathological diagnosis was diffuse large B-cell lymphoma (DLBCL) [1] in 6 cases, extranodal marginal zone B-cell lymphoma of mucosa-associated lymphoid tissue (MALTL; Figure 3a-c) [1] in 3 cases, as well as anaplastic large cell lymphoma (ALCL) [1], peripheral lymphoma (PTCL) [1], and Burkitt lymphoma (BL) [1] in 1 case each for EL (Table 1). For NL, it was follicular lymphoma (FL; Figure 4a-c) [1] in 4 cases, DLBCL in 2 cases, as well as MALTL, ALCL, and lymphocyte-rich classical HL (LR) [1] in 1 case each (Table 2).

Figure 3: Extranodal marginal zone B-cell lymphoma of MALT type (MALTL): Case No. 11 (EL); (a) The tumors are composed of numerous nodules (hematoxylin-eosin stain, original magnification x 100). (b) Monocytoid neoplastic cells proliferate around lymph follicles and form lymphoepithelial lesion (hematoxylin-eosin stain, original magnification x 400). (c) Immunostaining for BCL-2 (original magnification x 100).

Figure 4: Follicular lymphoma (FL): Case No. 9 (NL); (a) Dense proliferation of fuzzy large nodules were detected (hematoxylin-eosin stain, original magnification x 40). (b) Tumor cells were composed of predominantly germinal center cells with convoluted and medium nucleus. (hematoxylin-eosin stain, original magnification x 400). (c) Immunostaining for CD10 (original magnification x 40).

Laboratory Test Values

The serum levels of LDH ranged from 142 to 1680 U/mL with a mean value of 383.7 ± 470.7 U/mL for EL (Table 1) and from 127–718 U/mL with a mean of 247.3 ± 178.8 U/mL for NL (Table 2). For the complete blood count, there was no remarkable deviation from the normal ranges in all cases. The serum level of sIL2R ranged from 36.6 to 4275 U/mL with a mean value of 1356.9 ± 1241.4 U/mL for EL (Table 1), whereas it was 137–31215 U/mL with a mean of 3990.6 ± 10212.9 U/mL for NL (Table 2). The serum TK level ranged from 6.6 to 280 U/mL with a mean of 57.5 ± 94.8 U/mL for EL (Table 1), and it was 4.0–230 U/mL with a mean of 38.0 ± 72.6 U/mL for NL (Table 2).

Stage and Outcome

For EL, the stage was I in 2 cases, II in 3 cases, III in a single case, and IV in 6 cases. In terms of outcome, death was observed in no case of stage I disease, 1 of the 3 cases of stage II, 1 of the cases of stage III BL, and all 5 cases of stage IV DLBCL (Table 1). For NL, the stage was I in 2 cases, II in 3 cases, and IV in 4 cases. In terms of outcome, death was observed in only 2 cases of stage IV disease (Table 2). When examining whether the outcome was correlated with the sIL2R or TK level, there was no correlation with the sIL2R level. However, a significant correlation with the TK level was observed on the poorer outcome in both EL and NL above the cutoff value of 40 U/mL (P < 0.001; Figure 5).

Figure 5: Relationship between thymidine kinase (TK) and outcome. Assuming the cutoff value is 40 U/mL, a significant correlation with the outcome was found (Fisher’s exact test; p = 0.0007).

Table 1: Listing of the cases with extranodal lymphoma (EL).

Lymphoma has been historically classified into HL and NHL. In the United States and Europe, HL is 24-54% of the total lymphoma. However, in Japan, the NHL accounts for 90-95% of the total lymphoma, and the HL is not so common [1,5,6]. The current study focuses on the cases of head and neck lymphoma and also demonstrates a common pattern (NHL in 20 of the 21 cases). The national prevalence rate specific to the age group for lymphoma in Japan in 2019 was estimated to be 31.4 and 26.8 per 0.1 million people (male and female, respectively), suggesting that slightly more males develop this condition than females [9]. However, in the previous reports on lymphoma that developed in the head and neck, the male to female ratio was 1:5 [3], 2:3 [10], 3:5 [11] and 9:11 [12]. This indicates a predilection for female patients as shown in the current study. These tendencies were quite different in oral squamous cell carcinoma (OSCC), because the male to female ratio reported in WHO was 11:5 in worldwide [13]. The mean age at the initial visit was 61.8 years (61.4 and 62.3 years for extranodal and NL, respectively) in the current study. This result was consistent with the common age of onset for lymphoma, 50s–60s, and was comparable to the previous reports focusing on lymphoma in the head and neck regions [12-14]. It was a bit lower than the mean age (72.1 years) of OSCC visited in our department (unpublished data). Kawamata et al. investigated the onset sites for EL in Japan and reported approximately 60% of the subjects when combining the palate with the gingiva, whereas 4.3% of the subjects was the tongue [2]. In the current study, the total percentage of cases involving development in the palate and the gingiva was also 75%, while those in the tongue was only a single case. This lower incidence of tongue in EL was quite different from that in OSCC. For examples, in case of patients with OSCC visited in our department, the primary sites were 90 cases (51 cases and 39 cases in mandible and maxilla, respectively), 63 cases in tongue, 18 cases in buccal mucosa, 9 cases in oral floor, 3 cases in lip and 1 case each in palate and mandibular bone (unpublished data). EL was diagnosed as neoplastic disease at the initial visit in most cases. This result was consistent with those of previous reports [2,11,14]. Oral haematolymphoid tumors present as poorly defined or discrete masses, with or without ulceration, as shown in Figure 1 [13]. The surface of the mass in EL was generally smooth (Figure 1b); however, ulcerative or recessive lesion is difficult to distinguish between lymphoma and OSCC (Figure 1a). In contrast, approximately 30% of patients with NL were not diagnosed with neoplastic disease. The reports in which we could read widely also showed that only 50% of the cases were diagnosed with lymphoma at the initial visit [4]. Based on these results, the diagnosis of lymphoma, especially in case of NL, at the initial visit is generally difficult.

At present, imaging examination such as CT, MRI, FDG-PET, and ultrasonography (US) has been used for the diagnosis of lymphoma in the cephalic and cervical regions. Among those, the most common is CT, although Stumpe et al. have raised the issue that the sensitivity and specificity of CT in diagnosing NHL are 81% and 41%: the former is favorable, while the latter is insufficient to achieve confirmed diagnosis [16]. In contrast, the sensitivity and specificity of PET/CT for NHL are 86% and 96%, indicating great diagnostic significance. Although all the patients diagnosed with lymphoma were received PET/CT, the diagnosis of NL is easy when multiple FDG uptake is detected (Figure 2c). However, if FDG uptake are localized only in the neck region, the diagnosis is very difficult by the PET/CT examinations alone (Figure 2a, b). MRI for nodal NHL, as compared with the skeletal muscles, provides iso-intensity signal on T1-weighted images and high intensity on T2-weighted and STIR images [17]. No literature addressing the relative sensitivity and specificity of MRI has been identified. However, Perrone et al. reported that the detection sensitivity of nodal NHL assuming the diffusion coefficient of 1.03 × 10⁻³mm² was 100% [18]. On US, nodal NHL is visualized as a well-circumscribed, spherical-shaped lesion inside which an extremely low echo area is present. On Color Doppler, this lesion is characterized by dendritic blood vessels and a remarkable increase in blood flow within the lymph node. Ahuja et al. reported that the percentage of correct diagnosis with US in malignant cervical lymph nodes, including nodal NHL was 93% [19]. Consequently, diagnostic imaging based on a single imaging modality for nodal NHL is difficult. As it is presently, we recommend the combination of several modalities with a central focus on PET for the diagnosis of NL.

 LDH is an enzyme that converts lactic acid to pyruvic acid through a dehydrogenation reaction. It expresses in most cells constituting the organs in the entire body but it is highly expressed, particularly in the liver, kidney, lung, blood, muscles, and cancer cells. Therefore, in malignant tumors originating in these organs, LDH is released into the blood as a result of tissue destruction and is commonly used as a tumor marker for lymphoma. The cutoff value of serum LDH used for the current study was 200–400 U/mL. The mean serum LDH level for EL and NL was 384 U/mL and 247 U/mL, respectively, and either level was within the normal range. When reviewing individual cases, the positive result was shown in 2 of the 12 cases of EL, as well as 1 of the 9 cases of NL. Additionally, both levels of sIL2R and TK were high in the cases involving positive results. Thus, the sensitivity of LDH in lymphoma developing in the head and neck was found to be low.

The IL2R present on the T cells is composed of alpha, beta, and gamma chains, although the alpha chain is cut by protease and sIL2R is secreted into the serum when T cells are activated [20]. In other words, the sIL2R level rises for not only hematopoietic tumors such as lymphoma and leukemia but also different autoimmune diseases, causing activation of T cells as well as inflammatory diseases accompanied by swelling of the lymph nodes. Thus, the sIL2R level is generally used when determining a response to therapy provided after confirmed diagnosis or an indicator at follow-up visits. However, a report has shown that lymphoma could be achieved for all cases in which histopathological examination was performed due to the presence of swelling in the neck and the sIL2R level of ≥3000 U/mL [21]. The cutoff value for sIL2R used in the current study was 220–530 U/mL, and a negative result was provided in 5 of the 19 cases. Therefore, biopsy is obviously important for the confirmation of the diagnosis of lymphoma, although the measurement of the sIL2R level might be of slight significance in the diagnosis. The mean level of sIL2R for EL and NL was 1377 U/mL and 3991 U/mL, respectively: the mean value for NL is apparently higher. However, the sIL2R level greatly varied according to the cases. The mean sIL2R level calculated excluding 1 case of NL was 587. 5 ± 275.6 U/mL, which was lower than that of EL. Thus, there appeared to be no specificity of the sIL2R level according to the onset sites. Concerning outcome, a report has shown that a sIL2R level of ≥3000 U/mL is associated with a poor outcome [7]. In the current study, sIL2R levels exceeded 3000 U/mL in only 2 cases of EL and NL, leading to unfavorable outcomes.

TK is an enzyme that acts through the salvage pathway involved in DNA synthesis and prolongs the G1/S phase of the cell cycle. The serum level of TKs, i.e., those released from within cells, is determined by the product of intracellular TKs and tumor volume. Therefore, the serum TK level is considered to reflect the quantity and proliferation potential of the tumor [22,23]. The cutoff value for the serum TK level used in the current study was 5.0 U/mL, although a negative result was obtained in only 1 case of the 19 cases in which measurements of the TK level were available with the sensitivity being higher than that of sIL2R. On the other hand, TK levels of ≥40 U/mL has been reported to be associated with poor prognosis in hematopoietic tumors, including lymphoma [24]. The mean TK levels in the current study were 57 U/mL and 38 U/mL for EL and NL, respectively. Three cases of EL and 1 case of NL, in which the TK level was 40 ≥U/mL and the outcome could be followed up, were all associated with unfavorable outcomes. This result was statistically significant, suggesting that measurements of the serum TK levels are important in diagnosing and determining the prognosis of lymphoma in the head and neck.

In the current study, the histopathological diagnosis of B-cell type was made in approximately 80% of cases for both EL and NL. This result was in line with those of previous reports [14]. According to the tissue types, DLBCL accounted for 50% (n = 6) of EL. This result is borderline consistent with those of the reports by Kaibuchi et al. (5 of the 10 cases) [14], Tanaka et al. (7 of the 12 cases) [11], and Kawamata et al. (6 of the 10 cases) [2]. In contrast, the histopathological type was FL in 4 cases and DLBCL in 2 cases for NL; unlike the result for EL, the histopathological type was FL in more cases. Takano et al. reported that the most common was DLBCL, followed by FL in 153 patients with NHL in the head and neck [15]. FL is low-grade lymphoma, which would contribute to a more favorable outcome for EL than for NL in our cases.

The mean time to histopathologically confirmed diagnosis was 34.7 and 66 days for EL and NL, respectively. Although no significant difference was observed (P = 0.1), this time was longer for nodal lesions. This result implies that a biopsy can easily be performed under local anesthesia for EL. However, two repeated incisional biopsies did not lead to a diagnosis of lymphoma, and a subsequent excisional biopsy resulted in a confirmed diagnosis in Case No. 11 of EL, extranodal marginal zone B-cell lymphoma of MALT type (Figure 3a-c). The histopathological diagnosis of lymphoma requires high-level specialty among pathologists, and if the volume of sample collected is small or crushing is present, the diagnostic process might be difficult [25,26]. Therefore, the minimization of tissue crushing, as well as the collection of the larger tissue specimen possible, would be required. In contrast, in case of NL, patients and surgeons hesitate to undergo a biopsy, likely resulting in a delay in histopathological diagnosis for NL. In fact, for case No. 8 of NL, the patient has been already diagnosed as Sjogren’s syndrome (SS) and for case No. 9 of NL (Follicular lymphoma, Figure 4a, b, c), the patient has sialolithiasis, resulting in delays in the definitive diagnosis of lymphoma. NL in the neck is often diagnosed by FNAB, but there are cases in which a definitive diagnosis is not reached. Thus, we are now considering that in case of nodal swelling of the neck, we should perform excisional biopsy as early as possible if the swelling was not disappeared after a fixed observation period. However, different types of imaging examination are necessary before an invasive lymph node biopsy in many cases, since approximately 80% of the patients who visited the department due to a complaint of swelling in the cervical lymph nodes resulted in inflammatory diseases [27]. The results of the current study suggested that imaging examinations (including different modalities) combined with blood tests, including TK for suspected lymphoma, early biopsy without hesitation for NL, and collection of a large biopsy specimen without crushing are important. Moreover, if the diagnosis is indefinite, we must repeat biopsy for EL. Although oral surgeons have few opportunities to treat lymphoma, it is important to suspect lymphoma without hesitation and cooperate with hematologists so that specialized treatments could be initiated as early as possible for lymphoma, particularly in the cases involving cooperation with TK levels of ≥40 U/mL.

The initial diagnosis of EL occurred orally was clinically hard by inspection and palpation. This dilemma could be escaped by the early incisional biopsy because the lesions are visible and easy to approach. Most of NL occurred in the neck regions was possibly diagnosed by the combination with imaging examinations, especially by PET/CT. Early excisional biopsy without crushing the materials should be performed without hesitation, whereas it must be carefully diagnosed in the cases complicated with another neck disease. Measurements of serum TK levels might be useful for both diagnostic and prognostic marker in the head and neck lymphoma.

Acknowledgments

We would like to express our gratitude toward Dr Masaru Kojima at Department of Diagnostic Pathology at Dokkyo Medical University School of Medicine for the support and guidance with this clinical study.

Funding Information

This work was supported in part by Grant-in-Aid for Scientific Research (C) (20K10095).

1. Miyazaki Y. JSH Guideline for Tumors of Hematopoietic and Lymphoid Tissues: Leukemia: 6. Myelodysplastic syndromes (MDS). Int J Hematol. 2017; 106: 151-159.
2. Kawamata A, Mizutani M, Hashida Y, Saito K, Yamashiro M. A clinical study of malignant lymphomas arising in the oral and maxillofacial region with a review of case reports in Japan. J Jpn Soc Oral Tumor. 2015; 27: 119-125.
3. Yamagata K, Onizawa K, Yoshida H. 6 cases of extra-nodal malignant lymphoma arising in the oral and maxillofacial region. Jpn J Oral Diag. 2008; 21: 273-277.
4. Zapater E, Bagan JV, Carbonell F, Basterra J. Malignant lymphoma of the head and neck. Oral Dis. 2010; 16: 119-128.
5. Chihara D, Ito H, Matsuda A, Shibata A, Katsumi A, Nakamura S, et al. Differences in incidence and trends of haematological malignancies in Japan and the United States. Br J Haematol. 2014; 164: 536-545.
6. Picard A, Cardinne C, Denoux Y, Wagner I, Chabolle F, Bach CA. Extranodal lymphoma of the head and neck: A 67-case series. Eur Ann Otorhinolaryngol Head Neck Dis. 2015; 16: 71-75.
7. Urquhart A, Berg R. Hodgkin’s and non-Hodgkin’s lymphoma of the head and neck. Laryngoscope. 2001; 111: 1565-1569.
8. Japanese Society for Head and Neck Center. Japanese Clinical Practice Guidelines for Head and Neck 2022. Tokyo, Jpn: KANEHARA CO., LTD; 2022: 105-106.
9. National Cancer Center Japan Center of Cancer Control and Information Services. https://ganjoho.jp/reg_stat/statistics/stat/cancer/25_ml.html. Tokyo, Jpn. Accessed October 5, 2023.
10. Hanazawa Y, Ito K, Mabashi T, Akizuki K, Uchimiya S, Kanazawa H, et al. Clinical study of malignant lymphomas in the oral and maxillofacial region over the past 6 years. Jpn J Oral Maxillofac Surg. 1992; 38: 1859-1866.
11. Tanaka S, Usui M. Clinical and histopathological study of non-Hodgkin’s lymphoma in the oral and maxillofacial region over the 12 years. Osaka Rosai Byoishi. 1998; 22: 77-86.
12. Walter C, Ziebart T, Sagheb K, Rahimi-Nedjat RK, Manz A, Hess G. Malignant Lymphomas in the Head and Neck Region – a Retrospective, Single Center Study over 41 years. Int J Med Sci. 2015; 12: 141-145.
13. WHO Classification of Head and Neck Tumors. WHO Classification of Head and Neck Tumors. 4th ed. Lyon, France: International Agency for Research on Cancer; 2017; 109-111, 128-129.
14. Kaibuchi N, Fukuzawa S, Momoki Y, Kumasaka A, Kataoka T, Ando T. Clinical study of malignant lymphoma in the oral cavity and neck, especially for FDG-PET. Jpn J Oral Diag. 2013; 26: 289-293.
15. Takano S, Matsushita N, Oishi M, Okamoto S, Teranishi Y, Yokota C. Site-specific analysis of B-cell non-Hodgkins lymphomas of the head and neck: A retrospective 10-year observation. Acta Oto-Laryngologica. 2015; 135: 1168-1171.
16. Stumpe KD, Urbinelli M, Steinert HC, Glanzmann C, Buck A, on Schulthess GK. Whole-body positron emission tomography using fluorodeoxyglucose for staging of lymphoma: effectiveness and comparison with Computed tomography. Eur J Nucl Med. 1998; 25: 721-728.
17. Moritani M, Yamada T, Katase N, Nagatsuka H, Matsumura S, Morikawa M. A case of malignant lymphoma arising in the margin of the tongue. Jpn J Oral Maxillofac Surg. 2011; 57: 269-273.
18. Perrone A?Guerrisi P, Izzo L, D’Angeli I, Sassi S, Mele LL, et al. Diffusion-WEIGHTED MRI in cervical lymph nodes. Differentiation between benign and malignant lesions. Eur J Radiol. 2011; 77: 281-286.
19. huja A.T, Ying M, Ho S.Y, et al. Ultrasound of malignant cervical lymph nodes. Cancer Imaging. 2008; 8: 48-56.
20. Murakami J, Arita K, Wada A, Mihara H, Origasa H, Kigawa M, et al. Serum soluble interleukin-2 receptor levels for screening for malignant lymphomas and differential diagnosis from other conditions. Mol Clin Oncol. 2019; 11: 474-482.
21. Wakasugi T, Sanka T, Takenaga F, Masuda R, Kise Y, Suzuki H. A clinical study of diagnostic biopsies of the cervical lymph node. J Jpn Soc Head Neck Surg. 2014; 24: 101-107.
22. Heaney ML, Golde DW. Soluble cytokine receptors. Blood. 1996; 87: 847-857.
23. Kiran KJ, Martin S. Thymidine kinase 1 as a tumor biomarker: technical advances offer new potential to an old biomarker. Biomark Med. 2018; 12: 1035-1048.
24. Kurokawa K, Kasuga M, Kitamura K, Takaku H. Lab data: test selection and interpretation 2015-2016. Tokyo, Jpn: IGAKU-SHOIN Ltd; 2015: 670-671.
25. Sugiyama S, Iwai T, Oguri S, et al. Clinical study of extra-nodal lymphoma with initial symptoms in oral cavity. Japanese J Head Neck Cancer. 2016; 42: 339-344.
26. Nomura F, Mizuno A, Nakamichi N, Motegi K. A case of malignant lymphoma in the maxilla. J Jpn Stomatol Soc. 1989; 38:277-282.
27. Uchimizu H, Kobayashi T, Mori E, Yamada D, Yanagi K. A 5-year study of patients with neck lymphadenopathy. J Otolaryngology of Japan. 2012; 115: 546-551.