Genotypic analysis of large cell lymphomas which express the Ki-1 antigen

The monoclonal antibody Ki-1 reacts with Reed-Sternberg cells in Hodgkin's disease and with the tumour cells in a minority of large cell non-Hodgkin's lymphomas. This study describes the results of immunophenotypic and DNA analysis in 30 cases of non-Hodgkin's lymphoma, all of which expressed the Ki-1 antigen. The genotypic analysis has been undertaken using both immunoglobulin and T-cell receptor gene probes. Sixteen cases were shown by this method to be of monoclonal T-cell origin, six of B-cell origin, while in eight cases there was no evidence of either T- or B-cell lineage. This confirms previous immunohistological data indicating that non-Hodgkin's lymphomas which express the Ki-1 antigen may be of either T-cell or B-cell origin.     

CD26/dipeptidyl peptidase IV expression in human lymphomas is restricted to CD30-positive anaplastic large cell and a subset of T-cell non-Hodgkin's lymphomas

CD26 is identical to the cell surface ectoenzyme dipeptidyl peptidase IV (DPPIV). is associated with T-cell activation and proliferation and also may function as an auxiliary adhesion factor. Although has been previously studied on lymphoid populations and on leukemias/lymphomas of B- and T-cell phenotype, little is known about its expression and functional role in some specific types of lymphomas, such as CD30-positive anaplastic large cell (ALC) lymphomas and Hodgkin's disease (HD). A series of 81 lymphoma samples, including 23 cases of HD, 17 cases of CD30-positive ALC lymphomas, 41 cases of other non-Hodgkin's lymphomas (NHL), and a panel of HD- or ALC lymphoma-derived human cell lines were evaluated for expression by enzyme histochemistry and immunohistochemistry on frozen sections and cell smears. protein was expressed on neoplastic cells in 12 of 17 (71%) ALC lymphomas irrespective of their antigenic phenotype and in seven of 15 (47%) T-cell NHLs. In contrast, we did not detect expression in tumor cells from 26 cases of B-cell NHL other than ALC lymphomas or in Reed Sternberg (RS) cells and variants of 21 of 23 HD cases. Accordingly, expression was maintained on the CD30-positive ALC lymphoma cell line Karpas 299, but the molecule was not detected on HD-derived cell lines of B, T, or non-B non-T phenotype. These results may support a new potential tool for the phenotypic separation of ALC lymphomas from HD based on the differential expression of the molecule. Moreover, given the demonstration that is identical to the human adenosine deaminase (ADA) binding protein, it could be speculated that also may function by interacting with ADA to regulate the growth of expressing neoplastic cells in ALC lymphomas.     

Genotypic analysis of diffuse, mixed cell lymphomas. Comparison with morphologic and immunophenotypic findings

Malignant lymphoma, diffuse, mixed small and large cell type, as defined in the Working Formulation, is heterogeneous both morphologically and immunophenotypically and, in some cases, clonality may be difficult to determine. Because gene rearrangement analysis has been shown to be a sensitive method for determining clonality, the authors genotyped 20 cases and compared the results with histologic and immunophenotypic findings. Immunophenotypic studies demonstrated that all lesions were composed predominantly of T cells. In addition, in eight cases either a monoclonal B-cell population (five lesions) or an aberrant immunophenotype (two T, one B) was detected, supporting a malignant diagnosis. In seven of these eight lymphomas, genotypic analysis confirmed the presence of a population of clonal cells. One case with an abnormal T-cell phenotype was germline. In 12 cases the immunophenotypic results were uncertain (i.e., no clonal population or abnormal immunophenotype was identified). Genotypic analysis provided evidence of clonality in eight. In four cases with uncertain immunophenotypic results, a clonal population also could not be identified with the use of Southern blot analysis. Thus, the authors conclude that gene rearrangement analysis is a valuable tool in the study of diffuse mixed cell lymphomas and is complementary to immunophenotypic studies. In addition, the authors analyzed the major breakpoint region of the bcl-2 protooncogene on chromosome 18, either by Southern blot analysis and/or with the polymerase chain reaction. The authors identified the t(14;18)(q32;q21) translocation in seven B-cell lymphomas, five of which were not considered to be of follicular center cell type on the basis of morphologic findings. These results suggest that the histologic spectrum of follicular center cell lymphomas is greater than is appreciated in the literature.     

Expression of killer cell inhibitory receptors is restricted to true NK cell lymphomas and a subset of intestinal enteropathy-type T cell lymphomas with a cytotoxic phenotype

BACKGROUND/AIMS: Killer inhibitory receptors (KIR) have a modulating effect on the cytotoxic functions of natural killer (NK) cells and T cells. Because lymphoma cells often have the same receptors as their non-neoplastic counterparts, this study investigated the expression of KIR on well defined groups of NK and T cell lymphomas, with and without a cytotoxic phenotype, from different sites of origin. METHODS: Nine CD56+/CD3- NK cell lymphomas, 29 CD3+/CD56- T cell lymphomas with a cytotoxic phenotype, and 19 T cell lymphomas without a cytotoxic phenotype were stained for KIR using monoclonal antibodies specific for CD94, CD158a, and CD158b. In addition, the expression of KIR was studied on normal lymphoid tissues. RESULTS: KIR expression was seen in five of nine true NK cell lymphomas including three of four nasal, one of four cutaneous, and one of one intestinal lymphoma nasal type. Double staining for CD56 and CD94 in normal lymphoid tissues revealed that KIR was predominantly expressed by CD56+ NK cells and sporadically on CD8+ T cells. Moreover, enteropathy-type T cell lymphomas with a cytotoxic phenotype showed KIR expression (three cases expressing CD94 and one case expressing CD158a). All nodal and extranodal nonintestinal T cell lymphomas with or without a cytotoxic phenotype lacked expression of KIR. CONCLUSIONS: These results show that KIR expression is restricted to CD56+/CD3- true NK cell lymphomas originating from the nose, gut, and skin, as well as in a subset of extranodal T cell lymphomas originating from the small intestine, which possessed a cytotoxic phenotype. Thus, the presence of KIR on NK/T cell lymphomas seems to mimic the distribution of KIR found on NK and T cells in normal lymphoid tissue.     

Genotypic and immunophenotypic analysis of anaplastic large cell lymphoma (Ki-1 lymphoma)

Genotypic and immunological analysis was performed in 10 patients with anaplastic large cell lymphoma (Ki-1 lymphoma), 4 were male and 6 female. Immunophenotypically, 9 of these patients expressed some T cell markers; CD 2 in 9, CD 4 in 9, CD 3 in 6, CD 8 in 4, and UCHL-1 in 8. For B cell markers, 4 patients expressed B 1 and one expressed L 26; these patients also expressed T cell markers. In genotypic analysis, 9 of 10 cases displayed some incidence of T cell receptor gene (TCR) rearrangement or deletion, which was found in 3 patients for C beta 1, 7 for C beta 2, 8 for J gamma, 5 for C gamma (Hind III), 5 for C gamma (EcoRI), 3 for J delta 1 (Hind III), 3 for J delta 1 (BamHI), 3 for J delta 2, and 2 for C delta. One patient with rearrangement of TCR also showed rearrangement of immunoglobulin heavy chain. Another patient exhibited rearrangement of both TCR and immunoglobulin chain. These results indicate a common T cell lineage for this type of lymphoma.     

Primary large cell lymphomas of the mediastinum: an analysis of 20 cases

Malignant lymphomas originating primarily in the mediastinum consist predominantly of Hodgkin's disease of the nodular sclerosis type, lymphoblastic lymphomas, and large cell non-Hodgkin's lymphomas of diffuse growth pattern (DHL). This analysis of 20 cases of primary mediastinal DHL presents the clinical and pathologic findings in nine patients with T-immunoblastic sarcoma (T-IBS), six with sclerosing variants of follicular center cell lymphoma (FCCL), and five with B-immunoblastic sarcoma (B-IBS). T-IBS patients were predominantly young adult women (mean age 31 years) presenting with relatively well confined mediastinal tumors; four of nine manifested the SVC syndrome. The immunomorphologic findings in T-IBS were similar to those of node-based peripheral T-cell lymphomas. Patients with FCCL and B-IBS were predominantly men, exhibited a broad age range, and presented with larger tumors with a high incidence of contiguous involvement of intrathoracic structures (83% in FCCL, 60% in B-IBS). Chemotherapeutic intervention attained CR in 19 of 20 patients, with 14 of 20 remaining alive in relapse-free CR a median of 26 months after completion of therapy. Durable CR was attained in eight of nine T-IBS patients, in four of six patients with FCCL, and in three of five patients with B-IBS. The morphologic features of these lymphoma subtypes are presented in detail and discussed in relation to the complex differential diagnosis of mediastinal neoplasms.     

Logical analysis of diffuse large B-cell lymphomas

OBJECTIVE: The goal of this study is to re-examine the oligonucleotide microarray dataset of Shipp et al., which contains the intensity levels of 6817 genes of 58 patients with diffuse large B-cell lymphoma (DLBCL) and 19 with follicular lymphoma (FL), by means of the combinatorics, optimisation, and logic-based methodology of logical analysis of data (LAD). The motivations for this new analysis included the previously demonstrated capabilities of LAD and its expected potential (1) to identify different informative genes than those discovered by conventional statistical methods, (2) to identify combinations of gene expression levels capable of characterizing different types of lymphoma, and (3) to assemble collections of such combinations that if considered jointly are capable of accurately distinguishing different types of lymphoma. METHODS AND MATERIALS: The central concept of LAD is a pattern or combinatorial biomarker, a concept that resembles a rule as used in decision tree methods. LAD is able to exhaustively generate the collection of all those patterns which satisfy certain quality constraints, through a systematic combinatorial process guided by clear optimization criteria. Then, based on a set covering approach, LAD aggregates the collection of patterns into classification models. In addition, LAD is able to use the information provided by large collections of patterns in order to extract subsets of variables, which collectively are able to distinguish between different types of disease. RESULTS: For the differential diagnosis of DLBCL versus FL, a model based on eight significant genes is constructed and shown to have a sensitivity of 94.7% and a specificity of 100% on the test set. For the prognosis of good versus poor outcome among the DLBCL patients, a model is constructed on another set consisting also of eight significant genes, and shown to have a sensitivity of 87.5% and a specificity of 90% on the test set. The genes selected by LAD also work well as a basis for other kinds of statistical analysis, indicating their robustness. CONCLUSION: These two models exhibit accuracies that compare favorably to those in the original study. In addition, the current study also provides a ranking by importance of the genes in the selected significant subsets as well as a library of dozens of combinatorial biomarkers (i.e. pairs or triplets of genes) that can serve as a source of mathematically generated, statistically significant research hypotheses in need of biological explanation.     

The expression of myeloid antigens CD13 and/or CD33 is a marker of ALK+ anaplastic large cell lymphomas

We retrospectively studied the immunophenotype by flow cytometry of 20 anaplastic large cell lymphomas (ALCLs) (9 anaplastic lymphoma kinase [ALK]+ and 11 ALK-) with a particular emphasis on the aberrant expression of the myeloid associated antigens CD13 and/or CD33. All ALCLs expressed CD45, HLA-DR, and CD30. Most (8/9) ALK+ ALCLs expressed at least 1 surface T-cell antigen (CD4, 6/9 [67%]; CD7, 6/9 [67%]; CD2, 5/9 [56%]; CD5, 2/9 [22%]; CD8, 2/9 [22%]; CD3, 1/9 [11%]). All ALK-ALCLs expressed at least 1 surface T-cell antigen (CD3, 7/11 [64%]; CD4, 6/11 [55%]; CD2, 6/11 [55%]; CD7, 2/11 [18%]; CD5, 1/11 [9%]; CD8, 1/11 [9%]). CD13 and/or CD33 were expressed in all (9/9) ALK+ ALCLs compared with 1 of 11 ALK-ALCLs (9%) (P < .0001). Surface CD3 was more likely expressed in ALK-ALCLs (7/11) compared with ALK+ ALCLs (1/9) (P .03). The myeloid-associated antigens CD13 and/or CD33 are sensitive but not entirely specific markers of ALK+ ALCLs and should not be misinterpreted as indicating myeloid sarcoma.     

Nuclear antigens in neoplastic lymphocytes of B cell and T cell non-Hodgkin''s lymphomas.

Gross nuclear morphology is a major diagnostic feature in the identification of subtypes of non-Hodgkin's lymphoma (NHL). The authors have shown that the size, shape, and chromatin distribution of the lymphocyte nuclei vary extensively both within and between samples of a subtype, and have proposed that the variations may reflect qualitative and quantitative differences in extrachromatinic components. To test this hypothesis, the organization of individual nuclear antigens in NHL and in reactive hyperplasia biopsies was examined by immunofluorescence labeling of frozen sections with previously characterized monoclonal antibodies. The results have been correlated with observations of the staining patterns produced by the antibodies in mitogenically stimulated human peripheral blood lymphocytes. Labeling pattern and intensity with each antibody were consistent between preparations of blood lymphocytes, and all four antibodies labeled all blood lymphocyte samples tested. In contrast, only 15% of the 53 biopsies were labeled by all four antibodies, although all were stained by anti-peripherin, nearly 80% by I1, and almost 60% by PI1. Antibody PI2 labeling was detected in only 20% of the samples. Variation in labeling intensity was equally extensive both within and between biopsy samples. In general, there was little homogeneity between samples of an NHL subtype as to which antigens were detected, their labeling intensity, or their pattern of intranuclear distribution. These observations are consistent with earlier reports of significant diversity in the morphology of nonchromatin components in such samples. The data support the proposition that the heterogeneity of gross nuclear morphology in nuclei of NHL biopsies may be due in part to disordered expression or abnormal organization of nuclear proteins.     

Genotypic analysis of pulmonary Langerhans cell histiocytosis

Reported studies show that the systemic form of Langerhans cell histiocytosis (LCH) is a clonal expansion of Langerhans cells (LC) associated with aberrant expression of several oncogenes or tumor-suppressor genes. LCH of the lung is a heterogenous group of lesions thought to be a reactive rather than neoplastic process. The histogenesis of the LCH of the lung is uncertain, and to date there are no studies investigating its underlying molecular abnormalities. We performed comparative genotypic analysis by using allelic loss (LOH) of polymorphic microsatellite markers associated with tumor suppressor genes. Fourteen cases of formalin-fixed, paraffin-embedded LCH of the lung were studied. Microdissection of a total of 26 nodules from 14 patients and paired reference lung tissue was performed under stereomicroscopic visualization. To evaluate allelic loss, we used a panel of 11 polymorphic microsatellite markers that were situated at or near tumor suppressor genes on chromosomes 1p, 1q, 3p, 5p, 9p, 17p, and 22q. The PCR products were analyzed by using capillary electrophoresis to identify germline heterozygous alleles and LOH. Allelic loss at 1 or more tumor suppressor gene loci was identified in 19 of 24 nodules. The total fractional allelic loss (FAL) ranged from 6% (1q) to 41% (22q), with a mean of 22%. The FAL in individual cases ranged from 0 (7 nodules) to 57% (1 nodule). Fifteen discordant allelic losses at 1 to 3 chromosomal loci were identified in 8 patients with multiple synchronous nodules. Our results show that LOH of tumor suppressor genes is present in the LCH of the lung, and they indicate that the putative tumor suppressor genes situated on chromosomes 9p and 22q may play a role in the development of a subset of the LCH of the lung.     

Cadherins in reactive lymph nodes and lymphomas: high expression in anaplastic large cell lymphomas

Using a polyclonal pan-cadherin antibody and a monoclonal E-cadherin antibody (HECD-1) we have investigated cadherin expression in lymphomas and reactive lymph nodes. Routinely processed tissue from nine reactive lymph nodes and 48 lymphomas (six T-cell, six high-grade B-cell, 15 low-grade B-cell, 13 anaplastic large cell and eight Hodgkin's disease) were immunostained. The reactive cases showed pan-cadherin membrane associated staining of endothelium and epithelioid granulomas. No staining of lymphoid cells was seen. Pan-cadherin immunostaining was present in three of six T-cell lymphomas, two of six high-grade B-cell lymphomas, 12 of 13 anaplastic large cell lymphomas and three of eight cases of Hodgkin's disease. No staining of low-grade B-cell lymphomas was identified with the pan-cadherin antibody. E-cadherin was not detected in any of the lymphomas that showed pan-cadherin expression. The frequent and strongest cadherin expression in anaplastic large cell is noteworthy. The tumour cells of this lymphoma subtype are characterized by copious cytoplasm and a cohesive appearance, features which impart a superficial resemblance to carcinoma cells. Since cadherin molecules are known to have major morpho-regulatory functions our data suggests that the expression of cadherin molecules by anaplastic large cell lymphomas may play an important role in determining their characteristic epithelioid phenotype.     

Plasmacytic differentiation in follicular center cell (FCC) lymphomas

Although follicular centers are the sites of production of plasma cell precursors, plasmacytic differentiation in follicular center cell (FCC) lymphomas is rarely seen, presumably because of a "block" in differentiation of the large noncleaved FCC. The authors reviewed a large number of these cases to determine the frequency of plasmacytic differentiation in FCC lymphomas. In one hundred ninety-eight, consecutive FCC lymphomas with a follicular pattern from a two-year period, 17 (9%) cases were found in which there were large numbers of plasma cells. Immunoperoxidase studies of paraffin-embedded sections (PIP) for cytoplasmic immunoglobulin showed polytypic marking in ten of these and a monotypic plasma cell population in seven. In this latter group, isotypically identical marking of the plasma cell and FCC populations could be demonstrated in three cases with immunoperoxidase (where the FCCs showed cytoplasmic marking) and in one case (of one tested) with surface typing studies. In addition, three patients had serum paraproteins identical to the plasma cell cytoplasmic immunoglobulins. These findings indicate that a small minority of FCC lymphomas contain sufficient plasma cells to be a diagnostic problem, and that in some of these cases, plasma cells are a differentiated component of the FCC lymphomas.     

Cytokeratin expression and vimentin content in large cell anaplastic lymphomas and other non-Hodgkin''s lymphomas.

The immunophenotypes of 74 malignant lymphomas (9 Hodgkin's disease, 19 low-grade B-cell, 20 high-grade B-cell, 8 T-cell, and 18 large cell anaplastic lymphomas [LCAL]) have been characterized with antibodies against leucocyte differentiation antigens, keratin, and vimentin. All the non-LCAL were CD45 positive and keratin negative. The LCALs had a more varied immunophenotype, with CD45 present only in 11 of 18 cases and keratin present in 5 of 18 of these rare lymphomas. The lymphoid origin of these latter cases was proven by gene rearrangement studies. All LCALs were CD30+, and, where tested, vimentin positive. Of four different vimentin monoclonal antibodies tested, V9 and MVI stained the highest number of lymphomas. Positive staining of tumor cells was seen in 61 of 71 cases. Vimentin-negative cases included Burkitt's as well as some follicular lymphomas.     

Human T cell differentiation antigens characterizing a cytotoxic/suppressor T cell subset

Monoclonal antibodies were raised against the leukemic T cells from a patient with chronic lymphocytic leukemia. Two antibodies, termed T411 and T811, were obtained which were reactive by indirect immunofluorescence only with cells of the T cell lineage. The T411 antibody recognized a polypeptide chain of 100,000 dalton apparent molecular weight which was present on the surface of 94 +/- 4% of peripheral blood T lymphocytes, but only on 20 +/- 8% of thymus cells. The antibody T811 reacted with a surface molecule composed of 2 poly-peptide chains of 32,000 and 34,000 dalton apparent molecular weight, which was expressed only on 25 +/- 10% of blood T lymphocytes and on 90 +/- 4% of thymus cells. Functional analysis of the T811+ and T811- T cell subsets isolated by rosetting with anti-mouse Ig coated ox erythrocytes revealed that both subpopulations were able to mount a proliferative response to allo-antigens, whereas allo-antigen induced cytotoxic cells and their precursors were only found in the T811+ subset. The pokeweed mitogen induced in vitro differentiation of B lymphocytes into immunoglobulin secreting cells was dependent on the presence of the T811- subset, whereas the T811+ T cells efficiently suppressed this differentiation.     

Immunohistochemical analysis of human lymphomas with monoclonal antibodies to B cell and Ia antigens reactive in paraffin sections

Monoclonal antibodies (MoAbs) to B cell- and T cell-specific antigens uniformly have been restricted in their use to cell suspension or frozen section techniques because the antigens that they identify are either masked or lost in the fixation or paraffin-embedding processes. Antiimmunoglobulin antisera, although readily identifying cytoplasmic immunoglobulin in paraffin sections, has not been as useful as originally hoped since the majority of B cell lymphomas express surface immunoglobulins, which requires cell suspensions or frozen sections for detection. Because cell suspension procedures disrupt tissue architecture and frozen section techniques grossly distort morphology, neither method allows the combination of optimal morphologic and immunologic classification of lymphomas. We recently reported two MoAbs, LN-1 and LN-2, that react with B cells in paraffin sections. LN-1 reacts with the surface membrane and cytoplasm of germinal center B cells. LN-2 reacts uniquely with the nuclear membrane and cytoplasm of mantle zone and germinal center B cells and interdigitating histiocytes. We have also identified a new MoAb, LN-3, that reacts with the HLA-DR antigen in paraffin sections. We now report the use of LN-1, LN-2, and LN-3 in the analysis of paraffin sections from 58 non-Hodgkin's lymphomas and 15 cases of Hodgkin's disease. The types of cells reactive with these MoAbs in neoplastic lymphoid proliferations largely recapitulate their benign morphologic and immunologic counterparts. As a panel, LN-1, LN-2, and LN-3 were reactive with 98% of B cell lymphomas, and LN-1 and LN-2 were negative on all T cell lymphomas. In addition to identifying the cell of origin of these malignant proliferations, these MoAbs were also useful for identifying architectural features in neoplastic lymph nodes. Thus, these reagents provide the ability to assess the immunologic phenotype of neoplastic lymphocytes in conjunction with the critical morphologic criteria requiring paraffin embedding.     

Most null large cell lymphomas are B lineage neoplasms

DNA of immunoglobulin and the beta T cell receptor genes was analyzed for rearrangements in 34 diffuse large cell lymphomas that failed to express immunoglobulins or T cell antigens. Twenty-eight cases had both heavy and light chain immunoglobulin rearrangements, two cases had only heavy chain gene rearrangements, three cases had only light chain gene rearrangements, and one case failed to show rearrangements for any of the immunoglobulin genes. None of the cases showed rearrangements for the beta T cell receptor gene. These results indicate that the vast majority of diffuse large cell lymphomas that lack definitive B or T cell phenotypic markers are actually B cell in origin.     

Immunohistochemical analysis of human MHC class II antigens in B-cell non-Hodgkin's lymphomas

Twenty cases of B-cell non-Hodgkin's lymphoma (NHL) were evaluated by immunohistochemical staining of frozen sections with a panel of anti MHC class II monoclonal antibodies (MCA). These studies showed marked differences in the expression of class II antigens both between different cases and within the population of cells of individual cases. In all of the cases studied the majority of the tumour cells reacted with MCAs directed against determinants common to the products of SB and DR loci. However, MCAs specific for DC determinants failed to react with 3/20 cases and in several other cases stained only a minority of cells. Absent or reduced expression of DC antigens was most marked in lymphocytic lymphoma; however, in centroblastic-centrocytic and centroblastic lymphomas, DC antigens could be detected on the majority of cells.     

Endoglin: a 180-kD endothelial cell and macrophage restricted differentiation molecule.

A MoAb RMAC8 was generated by immunizing Balb/c mice with cultured human umbilical vein endothelial cells (HUVE). The molecule recognized had a mol. wt of 180 kD non-reduced, 95 kD after reduction and 66 kD in its reduced and N-deglycosylated form. Sequential immunoprecipitation studies with the MoAb 44G4, which recognizes the O- and N-glycosylated homodimer endoglin, showed that both MoAbs recognize the same molecule on HUVE and phorbol myristate acetate (PMA)-stimulated U937 cells. The distribution of the RMAC8-recognized molecule was the same as that described for endoglin, i.e. arterial and venous endothelium, myelomonocytic and pre-B leukaemia cells and cell lines; however, unlike 44G4, RMAC8 also reacted weakly with monocytes and strongly with in vitro differentiated macrophages as well as peritoneal and alveolar macrophages. This distribution of endoglin was confirmed by Northern blot analysis using a full length endoglin cDNA probe. These studies suggest that endoglin is a differentiation marker on macrophages.     

Histological and immunohistochemical characteristics of diffuse large cell B-cell lymphomas

In the REAL classification, diffuse large B-cell non-Hodgkin's lymphomas are grouped together. We investigated histological variants and immunohistochemical profile of diffuse large B-cell lymphomas in 53 patients. Accuracy of the diagnosis was 73.6% without immunohistochemistry. The usefulness of immunophenotyping in making the correct diagnosis depended on a specific histological variant of diffuse large B-cell lymphoma. Variants with polymorphic and anaplastic morphology or massive reactive component have been diagnosed by routine histological methods with poor validity.