Diagnosis of myelomonocytic and macrophage neoplasms in routinely processed tissue biopsies with monoclonal antibody KP1.

A new monoclonal antibody, KP1, against the CD68 antigen, which labels macrophages and other members of the mononuclear phagocyte lineage in routinely processed tissue sections, has been used to stain a range of lymphoid, histiocytic, and myelomonocytic proliferations. All 20 neoplasms of myeloid, myelomonocytic, and presumed macrophage derivation reacted with antibody KP1. None of the 22 cases of T cell neoplasia had positive reactions. Although 14 of 41 B lineage lymphomas and leukaemias were stained by antibody KP1, staining was usually confined to small dots of reactivity, in contrast to the strong and extensive cytoplasmic staining seen in the neoplasms of myeloid and macrophage/monocyte origin. Furthermore, positive B cell neoplasms were almost all small cell proliferations, which are unlikely to be confused with myelomonocytic malignancies. It was concluded that antibody KP1 is a valuable addition to a panel of monoclonal antibodies for phenotyping lymphomas, particularly in routinely fixed tissues. It should assist the pathologist in the recognition of extramedullary presentation of leukaemia, aid in the diagnosis of suspected cases of true histiocytic neoplasia, and allow for quantitation of macrophages infiltrating lymphomas and other solid tumors.     

Immunohistochemical characterization of Ki-M6 monoclonal antibody in Bouin-fixed,paraffin-embedded sections of normal and neoplastic human tissues

Summary A monoclonal antibody (Ki-M6) against the CD 68 antigen, which labels cells of the monocyte/macrophage system, was tested on Bouin-fixed, paraffin-embedded samples of normal, reactive and neoplastic tissues by an avidin-biotin-peroxidase complex method, with the aim of establishing its value in diagnostic pathology. In normal human tissues, Ki-M6 reactivity was confined to the so-called resident macrophages populating normal organs under physiological conditions. Moreover, restricted reactivity against cells of macrophage lineage was observed in reactive and inflammatory lesions. Granulocytes, monocyte/macrophage-related immune accessory cells, and other analysed normal tissue structures did not reveal any reactivity. Ki-M6 was strongly reactive with the cases of benign (4/4) and malignant (15/15) fibrous histiocytomas, in addition to the true histiocytic lymphomas (3/3). Cases of granular cell tumour (2/3) showed strong reactivity with Ki-M6, whereas only few immunoreactive cells, with weak staining, were seen in the other Ki-M6-positive neoplasms [neurofibroma (3/3), benign schwannoma (1/2), ganglioneuroma (1/1), malignant schwannoma (5/9), melanoma (9/28), dermatofibrosarcoma protuberans (1/1), myelomonocytic leukaemia (3/3)]. Among the epithelial malignancies tested (47 cases), Ki-M6 was positive only in renal cell carcinoma (11/14). Malignant lymphomas of the Hodgkin (56 cases) and non-Hodgkin type (67 cases) were uniformly non-reactive. From these data, Ki-M6 appears to be an excellent marker of monocyte/macrophage-related cells and appears to be a reliable indicator for fibrous histiocytomas and true histiocytic malignancies. The availability of this additional antibody capable of staining routinely processed tissue is of practical interest.     

Tumours classified as "malignant histiocytosis" in children are T-cell neoplasms.

During the last five years increasing evidence has accumulated that many tumours classified as 'histiocytic' in the past do not originate from macrophages, but from transformed (or anaplastic) large lymphoid cells. Most of these studies have focused upon adult neoplasms. Knowledge concerning the lineage of 'histiocytic' tumours in the paediatric age group is more limited. In this study we have examined the clinical, morphological and immunophenotypical features of six childhood malignancies originally diagnosed as being of histiocytic origin. Three patients showed an aggressive course with involvement of internal organs and very short survival times. Two patients were brought into remission: one is alive without active disease after seven years; the other died after seven years due to treatment-related cardiomyopathy. The remaining patient had a protracted course for two and a half years, but subsequently deteriorated and died three years after diagnosis. The histomorphological features in five cases were those of anaplastic large cell lymphomas. The remaining case consisted of pleomorphic (rather than anaplastic) large lymphoid cells. In all cases the immunophenotypical examination showed features characteristic of activated T lymphocytes. All cases were positive for Ki-1 (CD30), and three were positive for epithelial membrane antigen (EMA). Histiocyte-associated markers were positive in residual reactive macrophages, but nowhere could unequivocal positivity for macrophage-associated markers be seen in the neoplastic cells. It is concluded that most childhood malignancies in the past classified as 'histiocytic' are examples of anaplastic large cell (Ki-1) lymphomas of T-cell type and that true histiocytic malignancies are exceedingly rare in the paediatric age group.     

Surface marker studies in chronic lymphocytic leukaemia and non-Hodgkin's lymphoma.

Immunological surface marker techniques were applied in a study of 29 cases of chronic lymphocytic leumaemia and 22 of non-Hodgkin's lymphoma. Surface marker characteristics distinguished 2 subtypes of B lymphocytes. Chronic lymphocytic leukaemia was a monoclonal proliferation of B lymphocytes which produced spontaneous rosettes with mouse erythrocytes and had faintly immunofluorescent surface immunoglobulin. The majority of non-Hodgkin's lymphomas also had their origin from B lymphocytes but in contrast, this subtype did not show receptors for mouse erythrocytes and their surface immunoglobulin was brightly staining and demonstrated "capping". The clonal origin of nodular lymphomas could also be demonstrated on frozen sections stained for surface immunoglobulin. Two cases of true histiocytic lymphoma were identified. The current information available on surface marker characteristics of the leukaemias and lymphomas is reviewed.     

Pathologic and immunologic characterization of malignant lymphoma in Taiwan. With special reference to retrovirus-associated adult T-cell lymphoma/leukemia

One hundred four cases of malignant lymphomas, including 90 cases of non-Hodgkin's lymphoma, 5 cases of histiocytic malignancy, and 9 cases of Hodgkin's disease were analyzed pathologically and immunologically using a panel of monoclonal and conventional antibodies for T-, B-, histiocyte, and Hodgkin's neoplastic cells. Our results revealed a high frequency of T-cell lymphoma (42.3%), a low percentage of follicular lymphoma (10.5%), and Hodgkin's disease (8.7%) in Taiwan. More than half of the malignant lymphomas belonged to the high-risk unfavorable group. Peripheral T-cell lymphomas (33 cases) showed characteristic clinical and histologic features, which can sometimes be confused with Hodgkin's disease. Monoclonal antibodies Leu-M1 and 2H9 were an important aid for their differential diagnosis. Five of the 33 peripheral T-cell lymphomas were positive for antibody to adult T-cell lymphoma/leukemia (ATL) virus associated antigen (ATLA). Four patients were from the northeast coast of Taiwan, I-Lan county. Five (4.8%) were diagnosed as true histiocytic malignancies, including two true histiocytic lymphoma and three malignant histiocytosis. Two cases each of large cell lymphoma and immunoblastic lymphoma showed no identifiable marker expression. The distribution of lymphoproliferative disorders in Taiwan is similar to that in Japan but much different from western countries.     

Clonality of T cell and phenotypically undefined lymphoid neoplasms: the value of genotypic analyses.

The value of genotypic analysis for routine assessment of leukaemia and lymphoma was shown by the findings in a selected series of 30 cases. T cell receptor (TcR) gene rearrangements were observed in six out of nine cases of CD3+ CD8+ lymphocytoses and provided clear evidence for clonality in this group. The T cell proliferations in two of the remaining cases masked B cell lymphocytic leukaemia and hairy cell leukaemia, while in the third case no cause was found for the polyclonal proliferation. Heterogeneity of phenotype and genotype were observed in peripheral T cell lymphomas: one out of six cases showed TcR gene rearrangement, one case retained its germline configuration, a further case masked B cell lymphoma and the remainder were polyclonal. Genotypic analysis was helpful in the analysis of a tumour of mixed T cell and myeloid phenotype which was shown to be germline for TcR and immunoglobulin genes, consistent with a myeloid origin. Two histiocytic tumours were found to have clonal rearrangement of TcR genes. Nine out of 11 B cell tumours showed immunoglobulin gene rearrangement. It is concluded that genetic analyses are useful in the analysis of T cell, histiocytic, and B cell tumours in which an immunoglobulin phenotype cannot be defined.     

Analysis of large-cell lymphomas using monoclonal and heterologous antibodies.

Fifteen cases of large-cell lymphoma, diagnosed as centroblastic (5), B-immunoblastic (5) or true histiocytic (5). lymphoma and one case of malignant histiocytosis were studied with monoclonal antibodies. Each diagnosis was based on morphological as well as marker studies. A panel of monoclonal and heterologous antibodies against T lymphocyte differentiation antigens (Leul, Leu2a, Leu3a, OKT4, OKT8, TA1), B lymphocyte subsets (BA1, BA2, HLA-DR, alpha C3b receptor antiserum, surface immunoglobulins), the common acute lymphoblastic leukaemia antigen (CALLA), monocytes/macrophages (OKM1, anti-human monocyte 1, TA1, Mac1, HLA-DR, anti-C3b receptor), myeloid cells (VIM-D5, elastase, OKM1) and the cells of the Langerhans cell/interdigitating reticulum cell series (OKT6, NA1/34). The results show a specific staining pattern for true histiocytic lymphoma (histiocytic sarcoma). Centroblastic and B-immunoblastic lymphomas showed gradual differences with mostly strong staining for HLA-DR and weak with anti C3b receptor for B-immunoblastic lymphomas in contrast to centroblastic lymphomas. Staining with BA1 and BA2 indicated immunological heterogeneity in these lymphomas. The number of admixed cells was usually low with few B cells and a shift in the ratio helper/inducer to suppressor/cytotoxic T cells in favour of the suppressor/cytotoxic subset.     

Immunocytochemical characterisation of cutaneous lymphomas other than mycosis fungoides

The immunophenotypic properties of 25 cutaneous non-Hodgkin lymphomas other than mycosis fungoides or Sezary syndrome were investigated and correlated with clinical and histopathological data. The 11 low grade lymphomas were all of B cell origin, whereas the 14 high grade lymphomas comprised B and T cell tumours, true histiocytic proliferations, and one "nul" cell lymphoid neoplasm. For the high grade lymphomas correct prediction of the immunological phenotype based on morphological criteria was only possible in three cases. In contrast, all of the low grade lymphomas showed the non-epidermotropic infiltration pattern considered to be characteristic of cutaneous B cell tumours. For these conditions, however, immunophenotypic investigations provided a convenient means of improving discrimination between benign (polyclonal) and malignant (monoclonal) lesions, and also showed similarities with nodal lymphomas in terms of expression of lymphoid subset markers and composition of the non-neoplastic white cell infiltrate. No differences were identified between primary and secondary or concurrent cutaneous and extracutaneous lymphomas. Cutaneous non-Hodgkin lymphomas other than mycosis fungoides or Sezary syndrome constitute a heterogeneous group of neoplasms and most of these disorders are likely to represent cutaneous equivalents of nodal malignancies. Immunophenotypic investigations form a useful supplement to their histogenetic characterisation and may provide a common conceptual basis for their classification.     

Frequent immunoglobulin and T-cell receptor gene rearrangements in "histiocytic" neoplasms

The authors have analyzed the DNA of immunoglobulin and T-cell receptor genes in a series of 6 malignancies which were judged to be of histiocytic derivation on the basis of morphologic criteria. They found that 4 of these cases showed rearrangements of the beta T-cell receptor genes in spite of the lack of any specific immunohistochemical markers for B or T cells. One case showed rearrangements of both heavy and light chain immunoglobulin genes and probably represents either a sinusoidal large cell lymphoma or a B-cell lymphoma with activation of histiocytes simulating malignant histiocytosis. A single case lacked both immunoglobulin and T-cell receptor rearrangements consistent with immunologic analyses that suggested its origin from an interdigitating reticulum cell. The result of this study in conjunction with the authors' previous immunologic observations suggests that many presumed histiocytic malignancies actually represent T-cell lymphomas. Alternatively, beta T-cell receptor rearrangement may be a common feature of tumors that show monocyte/histiocytic differentiation.     

PG-M1: A New Monoclonal Antibody Directed against a Fixative-Resistant Epitope on the Macrophage-Restricted Form of the CD68 Molecule

A new anti-macrophage monoclonal antibody (PG-M1) was produced by immunizing BALB/c mice with fresh spleen cells from a patient with Gaucher's disease. PG-M1 reacts strongly with a fixative-resistant epitope of an intracytoplasmic molecule, selectively expressed by virtually all macrophages of the human body. Although attempts to immunoprecipitate the molecule recognized by PG-M1 have failed so far, the reactivity of the antibody with COS-1 and WOP cells transfected with a human complementary DNA clone encoding for the CD68 antigen suggests that PG-M1 is a new member of the CD68 cluster. However, unlike other CD68 antibodies (KP1, EBM11, etc.), which react with both macrophages and myeloid cells, PG-M1 detects a fixative-resistant epitope on the macrophage-restricted form of the CD68 antigen. In 957 routinely fixed, paraffin-embedded samples, PG-M1 showed a more restricted reactivity with elements of the monocyte/macrophage lineage than the previously described monoclonal antibodies MAC-387 (anti-calgranulins), KP1 (CD68) and Ki-M1P. Among hematological malignancies, PG-M1 only labels acute leukemias of M4 and M5 type and rare examples of malignant histiocytosis/true histiocytic sarcoma. In contrast, acute leukemias of the M1, M2, M3, M6, M7, and L1-L3 types, non-Hodgkin's lymphomas, and Hodgkin and Reed-Sternberg cells of Hodgkin's disease are consistently PG-M1-negative. In the daily diagnostic practice, PG-M1 seems to be particularly valuable for the diagnosis of myelomonocytic or monocytic leukemia and neoplasms of true histiocytic origin in routine paraffin sections.     

Macrophages in non-hodgkin lymphomas and hairy cell leukaemia

Methods used in the study of human mononuclear phagocytes in vitro were applied to surgical specimens from 49 patients with non-Hodgkin lymphomas and eight patients with hairy cell leukaemia. Two of the tumours (both classified as “true histiocytic” neoplasms by the Kiel criteria) were distinguished by the presence of atypical macrophages in the in vitro system. In one the atypical cells were adherent; In the other example they were non-adherent. These tumours were the only examples of mononuclear phagocyte neoplasia identified in this series. All the remaining 47 cases of non-Hodgkin lymphoma were judged to be of lymphoid origin. While initial observations on hairy cell leukaemia-derived spleen cells suggested macrophage neoplasia, this impression does not stand up to more detailed analysis. The findings are more in keeping with a B lymphoid cell lineage. In hairy cell leukaemias and low grade lymphoma the proportion of macrophages per gram weight of tissue is diminished. This suggests a deficiency of macrophage functional activity compared with normal; the nature of this defect is not clear.     

Malignant lymphomas of true histiocytic origin. A clinical, histological, immunophenotypic and genotypic study

The clinical, histological, immunophenotypic and genotypic properties of four cases of lymphoma of true histiocytic origin are described. The cases were identified by typing 925 non-Hodgkin's lymphomas by immunophenotypic and/or genotypic techniques, and they all presented with skin lesions. The histological and immunophenotypic examination showed dense, diffuse infiltrates of markedly pleomorphic mononuclear cells that were positive for macrophage-associated markers, and negative for B-cell, T-cell and myeloid cell-associated antigens. Staining for Ki-1 and epithelial membrane antigen was also negative. Gene rearrangements studies were performed in three cases, and all of these showed germline configuration of both T-cell receptor and immunoglobulin genes. In all cases, the clinical course was aggressive with rapid and widespread dissemination to internal organs, poor response to conventional chemotherapy, and short survival times (0.5 to 14 months). This suggests that although true histiocytic tumours are very rare, their recognition may be important for clinical and/or prognostic reasons.     

Immunophenotyping of hematopoietic malignancies in paraffin sections

Immunophenotyping of hematopoietic malignancies is usually accomplished in frozen sections or cell suspensions. To determine whether this procedure was also feasible in paraffin sections, we performed a double-blind immunoperoxidase study of 65 hematopoietic tumors whose phenotypes had been determined previously in fresh tissue. A selected antibody panel was used, including anti-LN2, UCHL-1, anti-cathepsin B, anti-Leu M1, anti-MB2, and anti-MT1. A correct phenotype was obtained on paraffin sections in 95% of cases. All 31 B-cell malignancies were properly classified, showing reactivity for LN2 and MB2. In 14 of 15 T-cell hematopoietic malignancies, all cells reacted with anti-MT1 and/or UCHL-1; the 1 case negative for these antigens was misdiagnosed as a B-cell tumor because of misinterpreted LN2 reactivity in benign histiocytes. Four of 5 true histiocytic neoplasms were positive for cathepsin B and LN2 but lacked other antigens; the fifth case was wrongly considered a B-cell proliferation because only bland histiocytes displayed cathepsin B. Only 1 of 7 Hodgkin's lymphomas was misdiagnosed (as a T-cell tumor); in the other 6 cases, Reed-Sternberg cells were reactive for LN2 and LEU M1. Five of 6 extramedullary myeloid leukemias also stained for LN2, MT1, and LEU M1. One showed LN2, MB2, and MT1; this case was classified as a B-cell neoplasm and indeed represented a pre-B-cell transformation of chronic myelogenous leukemia. These results show that the specified panel of antibodies may be useful for immunophenotyping of hematopoietic neoplasms when only paraffin sections are available for analysis. However, it cannot supplant traditional cell-marker studies of hematopoietic tumors because of its lesser accuracy.     

Human malignant lymphomas in vitro. Characterization of biopsy cells and establishment of permanent cell lines.

A series of 55 biopsies from different types of malignant lymphomas were characterized in short-term culture experiments and during prolonged growth in vitro. The majority of the lymphocytic lymphomas and half of the histiocytic lymphomas expressed surface immunoglobulin, either in monoclonal or polyclonal form, indicating B-lymphocyte derivation. No lysozyme production was noted in either type of lymphoma, giving further support to the notion that histiocytic lymphomas are not truly histiocytic. Production of beta2-microglobulin was higher in histiocytic than in lymphocytic lymphoma and Hodgkin's disease but did not significantly differ from the production observed in non-neoplastic lymph node disorders. Incorporation of 3H-thymidine varied greatly within each category of lymphoma; the highest mean labelling index was noted in histiocytic lymphoma, possibly reflecting the generally more malignant course in such cases. Epstein-Barr virus-associated nuclear antigen was observed in one case of Hodgkin's disease. Attempts to establish permanent tumor cell lines were successful only from two explants of lymphocytic lymphoma and one pleural effusion from histiocytic lymphoma. The two cell lines derived from lymphocytic lymphomas both exhibited B-lymphocyte characteristics. The histiocytic lymphoma line lacked lymphocyte markers, produced lysozyme and was found to be rich in cytoplasmic esterases. These features are consistent with a "true" histiocytic derivation of this line. Lymphoblastoid cell lines representing non-neoplastic EBV-carrying lymphocytes contaminating the biopsies were derived from 19 biopsies, with the highest frequency noted in cultures of biopsies from Hodgkin's disease. The tumor lines were all EBV-genome negative.     

Immunohistochemistry of markers of histiomonocytic cells in malignant fibrous histiocytomas. A monoclonal antibody study

Nine cases of malignant fibrous histiocytomas (MFH) were examined immunohistochemically in frozen sections with six different monoclonal antibodies to histiomonocytic and related cells (EBM11, HAM-56, KB90, antibodies to dendritic reticulum cells, HLADR and LCA). Ten other soft tissue sarcomas, two desmoid tumors, twelve carcinomas, three seminomas and four lymphomas were studied for comparison. All cases of MFH showed positivity for histiomonocytic cell antigens. In six cases, the positive cells could be clearly interpreted to be infiltrating non-neoplastic cells. However, immunoreactivity for multiple histiocytic markers (EBM11, HAM-56, KB90, HLADR) was seen in tumor cells in three cases of MFH. In one of these cases, the positivity could be verified with KP1, an antibody to histiomonocytic cells applied in formalin fixed and paraffin embedded tissue. None of the tumors was positive with the antibody to dendritic reticulum cells or LCA. In the series of non-histiocytic tumors, no cases showed widespread positivity for multiple histiocytic markers. Our results suggest that in relation to true histiomonocytic differentiation MFH might be a heterogeneous group of tumors. The widespread immunoreactivity for multiple histiocytic markers in some cases may indicate a true histiomonocytic differentiation in some MFHs.     

Non-HTLV-1-associated primary gastric T-cell lymphomas show cytotoxic activity: clinicopathological,immunohistochemical characteristics and TIA-1 expression in 31 cases

AIMS: Most primary gastrointestinal lymphomas are of B-cell origin and T-cell origin is very rare. Recent studies have suggested that human T-cell lymphotrophic virus type 1 (HTLV-1) may be involved in the development of primary gastric T-cell lymphoma. We analysed 31 patients with primary gastric T-cell lymphoma in south-west Japan, an area endemic for HTLV-1, and determined their phenotypes, genotypes, and HTLV-1 status. METHODS AND RESULTS: Here we present 31 cases of primary gastric T-cell lymphoma in a HTLV-1-endemic area in Japan and analyse the clinical status, histology, phenotype and virus status. The median age at onset of primary gastric T-cell lymphoma was 57 years with a gender ratio of M:F = 1.58:1. Six of the 31 primary gastric T-cell lymphoma cases had HTLV-1 proviral DNA (five males, one female), nine of the 31 cases were positive for anti-adult T cell leukaemia antibody, without examination of HTLV-1 proviral DNA (five males, four females), eight were non-HTLV-1-associated primary gastric T-cell lymphoma (four males, four females) and the other eight cases were unknown. Primary gastric T-cell lymphoma usually presented as a large ulcerated tumour at the corpus to the antrum and histologically consisted of anaplastic large cell type (n = 2), pleomorphic large cell type (n = 3), pleomorphic medium and large cell type (n = 14), pleomorphic medium cell type (n = 11), and angioimmunoblastic T-cell lymphoma type (n = 1). There were no clear macroscopic and microscopic differences between HTLV-1-associated and non-HTLV-1-associated primary gastric T-cell lymphoma. Most patients died within 2 years of diagnosis, and both types of primary gastric T-cell lymphoma (with and without HTLV-1) were associated with poor prognosis. Cytotoxic marker analysis showed that HTLV-1-associated lymphomas were negative for TIA-1, while non-HTLV-1-associated lymphomas were positive for TIA-1. CONCLUSIONS: Our results suggest that in HTLV-1-endemic areas, patients with HTLV-1-associated primary gastric T-cell lymphoma should be managed carefully and that TIA-1 seems to be useful for identifying the aetiology of this lesion.     

Immunohistochemistry of primary gastrointestinal lymphomas: a study of 76 cases

A retrospective study of 76 primary gastrointestinal lymphomas utilizing an avidin: biotinylated horseradish peroxidase complex (ABC) technique demonstrated 22 B-cell lymphomas, including two associated with alpha-heavy chain disease. Seven cases were classified as true histiocytic lymphomas based on a positive reaction for one or more of three histiocytic enzyme markers utilized, predominantly alpha-1-antitrypsin and alpha-1-antichymotrypsin. However, in 20 cases, an intense admixture of reactive histiocytes was noted and these cells stained preferentially for the enzyme, lysozyme. Twenty cases, which stained for both kappa and lambda light chains and positively or negatively for albumin, could not be classified and 27 cases failed to stain with any of the antisera utilized.     

Expression of interleukin-1 in Reed-Sternberg cells and neoplastic cells from true histiocytic malignancies.

Interleukin-1, a peptide produced by monocytes, histiocytes, and interdigitating reticulum cells, plays an important role in the regulation of immune function. In this styde, we examined the production of interleukin-1 in 115 patients with a variety of human lymphomas by using a rabbit anti-interleukin-1 antibody and the immunoperoxidase technique. Interleukin-1 was detected in Reed-Sternberg cells from 20 patients with Hodgkin's disease as well as in neoplastic cells from 9 patients with true histiocytic lymphoma or malignant histiocytosis. In the other 86 cases, which included T- and B-cell lymphomas, no interleukin-1 could be detected. This result indicates a close relationship between Hodgkin's disease and true histiocytic malignancies and provides additional evidence to support our hypothesis that Reed-Sternberg cells are related to interdigitating reticulum cells.     

Primary lymphomas of the gastrointestinal tract I. Plasma cell tumours

The histology of 125 cases of primary gastrointestinal lymphomas arising in the stomach and small and large intestine has been reviewed. The material was gathered from the Bland-Sutton Institute of Pathology at the Middlesex Hospital and from the Westminster Hospital. Of the initial total of 143 cases diagnosed, 18 were rejected. Of the acceptable 125 cases, 51 lymphomas were arising in stomach, 53 in the small intestine and 21 in the large intestine including rectum. Excluding the four children in the series, ages ranged from 18 to 82 and were fairly evenly distributed across the decades. There was no significant sex difference in the Middlesex Hospital cases but in the Westminster Hospital series the male to female ratio was approximately 2.6 to 1. One significant finding to emerge from this histological survey, and which forms the basis of this communication, is the proportion of lymphomas considered to be predominantly of plasma cell type. These plasma cell tumours, or extramedullary plasmacytomas, accounted for 49 out of the 125 cases (39%) of gastrointestinal lymphomas. They were less common in stomach and most common in the intestine, the majority occurring in the ileocaecal region. Conversely, Hodgkin's disease, in contrast to some series, was not encountered. Of the non-Hodgkin's lymphomas, grade I tumours were uncommon and true histiocytic lymphomas were distinctly rare. The high incidence of plasma cell tumours were uncommon and true histiocytic lymphomas were distinctly rare. The high incidence of plasma cell tumours in our series is in keeping with the morphological findings of a previous study carried out in patients with alpha-chain disease and in a small series of primary gastrointestinal lymphomas.