High frequency of CagA+ Helicobacter pylori infection in high-grade gastric MALT B-cell lymphomas

A high incidence of Helicobacter pylori infection has been found in patients with gastric MALT (mucosa-associated lymphoid tissue) B-cell lymphoma. Recent studies have indicated that the aggressive strains of the bacterium containing the CagA gene may have direct effects on tumourigenesis. To investigate the involvement of CagA+ strains in MALT lymphomagenesis, a sensitive polymerase chain reaction (PCR)-based detection assay for the gene was developed. DNA extracts from paraffin sections of 123 H. pylori-related gastric biopsies from Italy were analysed, including 56 cases of chronic gastritis, 37 low-grade, and 30 high-grade MALT lymphomas: 30·3 per cent (17/56) of the gastritis cases, 37·8 per cent (14/37) of the low-grade, and 76·7 per cent (23/30) of the high-grade MALT lymphomas were found to contain the CagA gene. The frequency of CagA+ strain infection was signfiicantly higher (P<0·05) in high-grade than in low-grade MALT lymphoma or gastritis. These results suggest that high-grade gastric MALT lymphoma transformation may be more likely to occur following infection by CagA+ strains of H. pylori. © 1998 John Wiley & Sons, Ltd.     

FAS (CD95) mutations are rare in gastric MALT lymphoma but occur more frequently in primary gastric diffuse large B-cell lymphoma

A loss of FAS (CD95) function has been proposed to constitute an important step in early mucosa-associated lymphoid tissue (MALT) lymphoma development and FAS mutations have been recognized in malignant lymphomas, in particular at extranodal sites. Since primary gastric lymphomas frequently exhibit resistance to FAS-mediated apoptosis, we investigated whether FAS is mutated in 18 gastric MALT lymphomas and 28 diffuse large B-cell lymphomas (DLBCL). We detected seven mutations in five lymphomas, one MALT lymphoma and four DLBCL; two DLBCL had two mutations. The MALT lymphoma exhibited a point mutation in the splice donor region of intron 3. Three DLBCL had missense mutations in exon 2, which encodes a signal peptide and a portion of the extracellular FAS ligand-binding domain. One DLBCL carried a point mutation in the splice donor region of intron 8, which would result in exon skipping. Two DLBCL harbored a missense mutation in exon 9, which encodes the intracellular death domain. The two death domain mutations inhibited FAS ligand-induced apoptosis in a dominant-negative mode, when transiently expressed in human T47D breast carcinoma and Jurkat T cells. A signal peptide and an extracellular domain mutation, however, failed to inhibit apoptosis in these transfection assays. They are likely to reduce apoptosis in lymphoma cells solely by a loss of function. In summary, our data show that FAS mutations are rare in primary gastric MALT lymphomas (5.6%) but occur in a subset of primary gastric DLBCL (14.3%) and suggest that these mutations contribute to the pathogenesis of gastric lymphomas by rendering lymphocytes resistant to apoptosis.     

Lymphoma in the BB/E rat: C-myc translocation identified

In a prospective necropsy study involving 257 animals in the BB rat colony in Edinburgh, there was an increased incidence of lymphoma (average 10.9 per cent in all three subgroups: 8 per cent in diabetic, 20 per cent in non-diabetic diabetes-prone, and 3.2 per cent in diabetes-resistant rats). The incidence was significantly increased (P<0.05) in the non-diabetic diabetes-prone subgroup. These results differ markedly from previous results and indicate that the relationship between lymphoma and diabetes is more complex than previously suggested. All the lymphomas bar one involved the ileocaecal nodes and were classified as immunoblastic lymphomas of B-cell origin. There was a striking resemblance both in tissue distribution and in histological classification to the lymphoma seen in the established B-cell lymphoma model, the LOUVAIN rat. Southern blot analysis carried out on the BB rat lymphomas revealed a translocation of variable length involving the c-myc oncogene. Such a translocation has not been demonstrated in the BB rat before.     

Hypermutation of the MYC gene in diffuse large cell lymphomas with translocations involving band 8q24

The sequence of the exon 1/intron 1 boundary region of the MYC gene was determined in two diffuse large cell lymphomas (DLCL), one with t(8;14) (q24;q32) and the other with t(8;22) (q24;q11). Both tumors had multiple mutations in this region. Also, both tumors had mutations in the protein binding site in intron I, which is a frequent target for mutational inactivation in endemic Burkitt's lymphoma (eBL). The translocations at 8q24, and multiple mutations in the exon 1/intron 1 boundary region, are reminiscent of similar findings in eBL. The same underlying oncogenic event that occurs in most eBLs is thus found in some DLCLs. © 1993 Wiley-Liss, Inc.     

Expression of c-myc protein is related to cell proliferation and expression of growth factor receptors in transitional cell bladder cancer

Archival biopsy specimens from transitional cell bladder tumours (n=185) were analysed immunohistochemically for expression of c-myc protein. The results were compared with compared with histopathological and clinical parameters and survival. Forty-three per cent of the tumours were negative for c-myc protein and weak, moderate, or strong cytoplasmic expression was found in 34, 14, and 9 per cent of cases, respectively. Nuclear positivity for c-myc protein was detected in 35 per cent of tumours and nuclear opositivity was related to overexpression of c-erb B-2 (P=0.01) and a high proportion of nuclei were also positive for p53 oncoprotein (p<0.05). Cytoplasmic expression of c-myc protein was related to histological grade (P=0.005), papillary status (P=0.007), the S-phase fraction (P=0.008), the mitotic index (P=0.021), overexpression of epidermal growth factor receptor (P=0.045), and c-erb B-2 (P=0.17). Expression of c-myc protein was not significantly related to the progression of tumours and it had no prognostic value in survival analysis. Independent predictors were the T-category (P<0.001), papillary status. (P=0.001), and S-phase fraction (P=0.061). The results show that while c-myc gene product participates in growth regulation of human bladder cancer cells, it has no independent prognostic significance.     

Lymph nodes in gastric B-cell lymphoma: pattern of involvement and early histological changes

AIMS: This study aims to analyse the histological pattern of nodal involvement in gastric B-cell lymphoma and to detect early involvement of the lymph nodes. METHODS AND RESULTS: Histological findings of 37 resected primary gastric lymphomas with 1313 regional lymph nodes were analysed. The primary tumour was classified into four groups: MALT lymphoma, MALT lymphoma with a minor large B-cell lymphoma (<20%), large B-cell lymphoma with MALT lymphoma, and large B-cell lymphoma without MALT lymphoma. Histological patterns of nodal involvement were divided into sinusoidal, subsinusoidal/marginal, follicular, and diffuse patterns. Semi-nested polymerase chain reaction (PCR) analysis for IgH gene rearrangement was performed. Nodal involvement was found in 2/13 (15%) MALT lymphomas, 5/6 (83%) MALT lymphomas with a minor large B-cell lymphoma, 9/12 (75%) large B-cell lymphomas with MALT lymphoma, and 6/6 (100%) large B-cell lymphomas without MALT lymphoma. The MALT lymphoma and MALT lymphoma with a minor large B-cell lymphoma showed a predominantly sinusoidal and subsinusoidal pattern, whereas diffuse pattern predominated in large B-cell lymphomas without MALT lymphoma and large B-cell lymphomas with MALT lymphoma. The follicular pattern was least common, being observed in 10.2% of large B-cell lymphomas without MALT lymphoma and large B-cell lymphomas with MALT lymphoma. Sinusoidal obliteration with permeation of small monocytoid cells into subsinusoidal zone is a characteristic finding suggesting early nodal involvement of MALT lymphoma. CONCLUSIONS: Histological patterns of nodal involvement in gastric B-cell lymphoma vary according to the histological grade. Immunostaining for CD20 with or without PCR analysis for IgH gene rearrangement would be a useful ancillary method to confirm lymphomatous involvement.     

BCL10 in malignant lymphomas--an evaluation using fluorescence in situ hybridization.

BCL10 is a tumour suppressor gene originally cloned from a t(1;14)(p22;q32) breakpoint in a case of mucosa-associated lymphoid tissue (MALT) lymphoma. Translocations involving this gene, though uncommon, are sometimes encountered in MALT lymphomas. This gene is thought to play an important role in the development of malignant lymphomas. Fluorescence in situ hybridization (FISH) was therefore undertaken on 22 cases of malignant lymphoma of varying histology to establish the incidence of rearrangements involving the BCL10 gene. Initially, one case with a novel t(1;2)(p22;p12) translocation involving the BCL10 gene was identified, in a marginal zone lymphoma of the MALT type, and was reported elsewhere. Seven other cases were subsequently identified with abnormalities in the 1p region, including a translocation with a breakpoint in the 1p22 region in a case of lymphoblastic lymphoma. However, none of these involved the BCL10 gene. Mutation analysis of BCL10 was then performed on 57 cases of malignant lymphoma, including 17 MALT lymphomas, by single-strand conformational polymorphism (SSCP) analysis of tumour DNA. Tissue was obtained for mutation analysis for 12 of the 22 cases analysed by FISH. Selected cases with SSCP band shifts were further studied by direct sequencing. Polymorphisms were identified in eight cases, but no mutations of pathogenic significance were identified. Further RT-PCR and mutation analysis was performed on cDNAs from 12 cases (four MALT, seven diffuse large B-cell lymphoma, one Hodgkin's disease) in which DNA analysis had already been completed. This included the MALT lymphoma with the t(1;2)(p22;p12) rearrangement. Again, no mutations were identified in the coding sequence. This study confirms that rearrangements of the BCL10 gene are uncommon in lymphoma (1/22) and may be limited tothe MALT subtype of non-Hodgkin's lymphomas. It was also found that breakpoints or rearrangements in the 1p22 region do not necessarily involve the BCL10 gene. Moreover, the absence of mutations at both the DNA (0/60) and the mRNA (0/12) level indicates that this gene is not frequently inactivated by mutation, in those tumours in which it is not involved in translocations. Our findings suggest that the BCL10 gene is unlikely to have a frequent or key role in general lymphomagenesis.     

A gene is known by the company it keeps: enrichment of TNFAIP3 gene aberrations in MALT lymphomas expressing IGHV4‐34 antigen receptors

Associations between immunoglobulin (IG) receptors with distinctive immunogenetic features and particular gene mutations are a recurring theme in mature B‐cell lymphomas. Relevant observations have been made in chronic lymphocytic leukemia (CLL), where gene mutations are distributed asymmetrically in cases bearing or not somatic hypermutations within the clonotypic immunoglobulin heavy chain variable region (IGHV) genes (e.g. TP53 mutations predominate in IG‐unmutated CLL, whereas the opposite is seen for MYD88 mutations, enriched in IG‐mutated CLL) and in subsets of cases with stereotyped IG (enrichment for SF3B1 mutations in CLL subset #2). Moreover, similar findings have been reported in splenic marginal‐zone lymphoma, where KLF2 mutations are biased to cases expressing IGHV1‐2*04 IG receptors, and in hairy cell leukemia, where IGHV4‐34‐expressing cases display a low frequency of BRAF mutations but a high frequency of MAP2K1 mutations. The list is now growing with the report of increased frequency of inactivating mutations in the TNFAIP3 gene in MALT lymphomas expressing IG receptors encoded by the IGHV4‐34 gene, particularly of the ocular adnexa. Considering that TNFAIP3 encodes a negative regulator of NF‐κB, this finding further highlights the importance of NF‐κB pathway activation in the natural history of MALT lymphomas. Altogether, these findings allude to selection of genomic aberrations in lymphoma cases with distinctive immune signaling profiles linked to the expression of particular IG receptors. Copyright © 2017 Pathological Society of Great Britain and Ireland. Published by John Wiley & Sons, Ltd.     

No correlation of c-myc overexpression and p53 mutations in liposarcomas

 Although it is well known that oncogenesis is a multistep process involving the activation of normal cellular genes to become oncogenes and/or the inactivation of tumor suppressor genes, this process has seldom been investigated in soft tissue tumours. We screened a group of 36 liposarcomas for genetic abnormalitis in the p53 tumour suppressor gene and c-myc oncogene. Altered c-myc gene expression was examined by differential RT-PCR assay. p53 Gene mutations in exons 4–8 were analysed by using PCR-SSCP analysis and direct sequencing. Elevated c-myc expression was found in 6 of 31 liposarcomas (19.4%). p53 Gene mutations were observed in 5 of 36 liposarcomas (13.9%). Both genetic alterations were associated with the histological subtype of liposarcomas. Whereas c-myc gene expression was a characteristic of myxoid/round cell liposarcomas, p53 gene mutations were found more frequently in pleomorphic variants. Liposarcomas of the well-differentiated subtype showed neither p53 gene mutations nor altered c-myc gene expression. Our results indicate that the c-myc oncogene and the p53 tumor suppressor gene do not seem to cooperate in the oncogenesis of liposarcomas. Received: 22 April 1998 / Accepted: 11 May 1998     

Cytogenetic abnormalities in MALT lymphomas and their precursor lesions from different organs. A fluorescence in situ hybridization (FISH) study

Aims: To analyse the possible activation of distinct molecular pathways in mucosa‐associated lymphoid tissue (MALT) lymphoma, we determined the prevalence of trisomies 3, 12, 18 in MALT lymphomas from different organs, as well as the prevalence of translocations of the MALT1 gene in a subset of primary breast MALT lymphomas. We compared the numerical cytogenetic alterations in lymphomas, precursor lesions and in normal non‐haematolymphoid tissue from the same organs. Methods and results: Forty‐two samples of paraffin‐embedded tissue (29 MALT lymphomas from stomach, breast, parotid and thyroid; two Sjögren's syndrome; two Hashimoto's thyroiditis and nine reactive samples) were studied by fluorescence in situ hybridization (FISH). Analysed together, the cases of gastric, parotid and thyroid MALT lymphomas presented trisomy 3 in 46%, trisomy 12 in 28% and trisomy 18 in 21% of the cases. In contrast to other locations, trisomy 3 was not present in the majority of the cases of primary breast MALT lymphomas. None of the nine breast cases presented MALT1 gene rearrangements. Half of the cases of preneoplastic lesions exhibited trisomy 3 and trisomy 12; none exhibited trisomy 18. Conclusions: Trisomy 3 is the most frequent numerical abnormality in gastric, parotid and thyroid but not in primary breast MALT lymphomas. MALT1 gene rearrangements are also rare in this location, suggesting that distinct molecular pathways may be activated in breast cases.     

bcl-2 and p53 protein expression in follicular lymphoma

Recent studies have shown bcl-2 to be regulated by p53. Other studies have suggested an inverse relationship between p53 and bcl-2 protein expression in breast and colonic cancers and in a variety of subtypes of non-Hodgkin's lymphoma. This study investigates the relationship between bcl-2 and p53 protein expression and the correlation between these findings and the grade and cell type of follicular lymphomas according to the REAL classification. Paraffin-embedded nodal follicular lymphomas (n=37) were subjected to bcl-2 and p53 immunohistochemistry on tissue sections using a three-step ABC system. Positive immunostaining for both oncoproteins was scored using a three-tiered scale: +, <10 per cent cells; ++, 10–50 per cent cells; and +++, >50 per cent cells (<10 per cent was used as a cut-off to define negative tumours). Ninety-seven per cent (36/37) of follicular lymphomas expressed bcl-2 protein in all three grades, manifesting in the small cell (grade 1) through to the large cell (grade 3). p53 protein expression showed a pattern of increasing immunostaining with progression towards the high-grade follicular lymphoma: grade 1=6 per cent (1/16); grade 2=48 per cent (10/21); grade 3=100 per cent (6/6). Five cases comprised varying combinations of grades. This latter finding suggests a role for p53 mutation in the progression/transformation of follicular lymphoma. The mechanism, however, differs from that suggested in breast and colonic cancers, since an inverse relationship between bcl-2 and p53 was not demonstrated in the present study. © 1997 John Wiley & Sons, Ltd.     

WHAT IS BURKITT'S LYMPHOMA?

Burkitt's lymphoma (BL) has been defined on the basis of its characteristic cytomorphology. Although histologically identical, endemic BL and sporadic BL are distinct clinico-anatomical entities. Their morphological identity probably relates to similar chromosomal translocations in both tumours, resulting in c- myc de-regulation and consequent unrestrained proliferation without differentiation. Similar gene rearrangements are found in a proportion of AIDS-related lymphomas that are predominantly extranodal and have the cytomorphology of BL. The term “Burkitt-like lymphoma” (BLL) has been applied to a group of high-grade B-cell lymphomas that appear morphologically intermediate between BL and centroblastic/immunoblastic lymphomas, as detailed in an accompanying paper in this issue. These tumours do not usually show c- myc gene rearrangements. The association of Burkitt's name with such a disparate group of tumours is confusing and new terminology for sporadic BL, AIDS-related BL and BLL is desirable. It is important that clinico-anatomical features, as well as cytomorphology, should be taken into account in the diagnosis of endemic BL. The origin of a case from tropical Africa does not, in itself, imply that it is endemic BL, even more since the AIDS epidemic in that continent. © 1997 John Wiley & Sons, Ltd.     

INVERSE CORRELATION BETWEEN EXPRESSION OF HLA-B AND c-myc IN UVEAL MELANOMA

HLA class I is expressed in 75–85 per cent of uveal melanoma and cytoplasmic c- myc expression has been reported in 78 per cent of uveal melanoma. In skin melanoma, an inverse relationship has been observed between HLA class I expression and c- myc . The purpose of this study was to determine whether a similar correlation occurred between high expression of c- myc and low expression of HLA class I in uveal melanoma. The expression of c- myc , HLA-A, and HLA-B was determined by immunohistochemistry on formalin-fixed and paraffin-embedded sections of 30 uveal melanomas. Cell cultures from four primary uveal melanomas (lines 92-1, MEL 202, OCM-1, and EOM-3) and one uveal melanoma metastasis (line OMM-1) were tested for mRNA levels of c- myc and HLA-A and HLA-B in Northern blot assays. The high level of expression of cytoplasmic c- myc was significantly correlated with low expression of HLA-B ( P =0·03) and vice versa. High expression of HLA-B was significantly correlated with the presence of epithelioid cells ( P =0·004). The inverse correlation observed between c- myc and HLA-B expression is similar to previous observations in cutaneous melanoma. By downregulating HLA-B expression, c- myc may influence the immune response in uveal melanoma. Tumours containing epithelioid cells showed a significantly higher expression of HLA-B than tumours of the spindle cell type. © 1997 by John Wiley & Sons, Ltd.     

Human and mouse cellularmyc protooncogenes reside on chromosomes involved in numerical and structural aberrations in cancer 1

A molecular clone of viralmyc (v-myc, the oncogene of avian myelocytomatosis virus, MC29, detected homologous human, mouse, and Chinese hamster cellularmyc (c-myc sequences by Southern filter hybridization. A v-myc probe, containing sequences from the 3′ domain of the gene, hybridized to single human HindIII and mouse EcoRI genomic DNA fragments of the cellular myc genes whose segregation could be followed in interspecies somatic cell hybrids. Human c-myc segregated concordantly with the enzyme marker glutathione reductase and with a karyotypically normal chromosome 8. A rearrangement of human c-myc was observed in Burkitt's lymphoma cells possessing the t(8;14) translocation. These results suggest that human c-myc is located close to the breakpoint on chromosome 8 (q24) involved in the t(8;14) translocation. The mouse c-myc gene segregated concordantly with chromosome 15 in mouse-Chinese hamster cell hybrids. These gene assignments are noteworthy, as structural and numerical abnormalities of human chromosome 8 and mouse chromosome 15 are associated frequently with B-cell neoplasms.     

Epstein-Barr virus-associated gastric lymphomas are distinct from mucosa-associated lymphoid tissue-type lymphomas: genetic abnormalities of p53 gene.

The authors studied 46 primary gastric lymphomas for expression of the p53 gene by immunohistochemistry and screened for mutations in p53 exon 5-8 by polymerase chain reaction-single strand conformation polymorphism. Twenty-five specimens cases were also analyzed for loss of heterozygosity (LOH) of chromosomal region 17p12-13.1. In 36 lymphomas negative for Epstein-Barr virus (EBV) infection, of which 29 were of mucosa-associated lymphoid tissue (MALT) type, p53 genetic changes were found in 47.2% but correlated poorly with overexpression. Only 20% of the mutations involved exon 7. There were recurrent mutations of intron 7, intron 6, and exon 6. In contrast, the 10 EBV-positive cases, none of MALT type, had a much higher rate of mutation, and all showed both p53 overexpression and p53 mutation and/or LOH, and 87.5% had mutations involving exon 7. Four of these involved codon 242, not seen in the EBV-negative group. Splicing mutations of intron 8 were seen in three specimens, two involving the same nucleotide position. In four of five specimens, LOH analysis identified microsatellite instability, allelic loss, or both. The Helicobacter pylori infection rate in the EBV-positive group (20%) was much lower than in the EBV-negative group (91.7%). These differences between the two groups suggest involvement of different carcinogens. Mutation of codon 242 has not been specifically associated with other tumors and may represent a mutational hot spot in the EBV-positive lymphomas.     

Lack of Bcl10 mutations in malignant cartilaginous tumors

The Bcl10 gene was recently isolated from the breakpoint region of t(1;14)(p22;q32) in mucosa-associated lymphoid tissue (MALT) lymphomas. Somatic mutations of Bcl10 were found in not only t(1;14)-bearing MALT lymphomas, but also a wide range of other tumors. To clarify the actual frequency and spectrum of Bcl10 mutations in primary malignant chondrogenic tumors, we examined 89 cases of malignant chondrogenic tumors comprising 17 conventional chondrosarcomas, 33 mesenchymal chondrosarcomas, and 39 clear cell chondrosarcomas. Polymerase chain reaction single-stranded conformation polymorphism and sequencing analyses were done. No Bcl10 mutations were found in our series of malignant chondrogenic tumors. While screening for mutations, we also found three polymorphisms at codons 8 exon 1 of the Bcl10 gene. Our results strongly suggest that somatic mutations of Bcl10 are extremely rare in malignant cartilaginous tumors and do not commonly contribute to their molecular pathogenesis.     

Mutational analysis of the 5' noncoding region of the bcl-6 gene in primary gastric lymphomas.

Bcl-6 mRNA and protein are frequently expressed in the transformed counterparts of the germinal center B-cells, diffuse large B-cell lymphoma and follicular lymphoma, irrespective of the gene rearrangements. Most of the primary gastric lymphomas are thought to be of mucosa-associated lymphoid tissue (MALT) origin, and neither bcl-6 gene rearrangement nor protein expression is found in low-grade gastric lymphomas of the MALT type as in normal marginal zone cells. However, bcl-6 protein expression was identified in high-grade gastric lymphomas, suggesting its role in high-grade transformation. In this study, polymerase chain reaction-single strand conformation polymorphism (PCR-SSCP) analysis for bcl-6 primer was performed in order to ascertain the molecular mechanisms of bcl-6 protein expression in primary gastric lymphomas. A total 31 cases of gastric lymphoma were classified into low-grade gastric lymphomas of MALT type (n = 13), high-grade gastric lymphomas of MALT type (n = 6) and gastric diffuse large B-cell lymphomas (n = 12). Bcl-6 mutations were observed in 11 of 13 (84.6%) low-grade gastric lymphomas of the MALT type and in 8 of 12 (66.7%) diffuse large B-cell gastric lymphomas. In 6 cases of the high-grade gastric lymphomas of the MALT type, both the low- and high-grade components demonstrated the same frequency (3/6, 50%) of mutations. The tissue obtained from the marginal zone of Peyer's patch by microdissection technique revealed no bcl-6 mutations by the PCR-SSCP analysis. These findings suggest that the acquisition process of bcl-6 mutations by the marginal zone cells may be involved in the lymphomagenesis of the stomach, but our data does not explain the reason why bcl-6 protein is expressed only in high-grade gastric lymphomas.     

Mucosa-associated lymphoid tissue (MALT) Iymphoma

Isaacson et al. defined MALT lymphoma as a neoplasm that mimics MALT lymphocytes, which are normally present in the small intestine. However, there are various problems with this definition of MALT lymphoma. First, the incidence of MALT lymphoma is not high in sites where MALT lymphocytes are normally present. Lymphoepithelial lesions are most common in the small intestine and the tonsil, but MALT lymphoma is rare in these sites. In contrast, MALT lymphoma is frequent in the tissue in which MALT lymphocytes are not normally present, such as the stomach, the breasts, the lungs, and orbital tissue. In these tissues or organs, chronic infections, such as chronic gastritis, lymphocytic mastopathy, Hashimoto's thyroiditis, and myoepithelial sialoadenitis (or Sjögren's syndrome), can be a precursor condition of lymphoma. These findings suggest that in pathological conditions MALT lymphoma arises, not from MALT tissue normally present, but from the lymphoid tissue. Also evidence is lacking that lymphoepithelial lesions imitate the normal function of MALT lymphocytes. Arber et al. Found no difference in the frequency of lymphoepithelial lesions between primary and secondary breast lymphomas, suggesting that lymphoepithelial lesions may merely reflect the aggressiveness of the tumor. We have not observed proliferating monotypic plasma cells in the dome epithelium merging with the underlying CCL cells. Isaacson et al. reported that MALT lymphoma is characterized by a good prognosis. However, extranodal lymphomas generally have a good prognosis and it remains to be determined if MALT lymphoma has a better prognosis than other extranodal lymphomas. There are reports indicating that the stage is a more important prognostic indicator than the grade in thyroid and stomach lymphomas. Hyjek et al. identified the presence or absence of capsular invasion as the most important prognostic indicator of thyroid lymphoma. We suggest that the relative prognoses of extranodal lymphomas should be determined by comparing various types of lymphomas at the same stage. Monocytoid B cell lymphoma in the salivary gland does not have a good prognosis. If MALT lymphoma and monocytoid B cell lymphoma are the same disease, then the prognosis of MALT lymphoma would be expected not to be very good. Despite the unresolved issues concerning Isaacson's definition of MALT lymphoma, Isaacson et al.'s characterization of MALT lymphoma should be highly evaluated in that it has disproved the theory that all B cell lymphomas are of FCC origin and suggested the existence of B cell lymphomas associated with an interfollicular pattern of spread. In addition to the stomach, the thyroid, the lung, and orbital tissue, MALT lymphoma can occur in the kidney, the breast, and the urinary bladder. However, it is doubtful that immune mechanism of MALT exists in these tissues under normal conditions. Tissues adjacent to the epithelium are exposed to massive antigenic stimulation when the epithelium is disrupted. MALT lymphocytes and monocytoid B cells may be B lymphocytes responding to such massive antigenic exposure. Further studies are needed to clarify the nature and histogenesis of MALT lymphoma.