The prognostic implication of thymoma histologic subtyping. A study of 80 consecutive cases

In this study the authors have investigated the clinicopathologic correlations in 80 consecutive cases of thymoma in order to establish the clinical usefulness of histologic subtyping of these tumours. All cases were histologically examined and classified according to Salyer and Eggleston and to Marino and Müller-Hermelink classifications. Therefore, thymomas were subtyped as predominantly lymphocytic, mixed and predominantly epithelial and cortical, mixed and medullary, respectively. The frequency of the different histologic subtypes was determined, and histologic findings were related to patients' age, surgical stage, and survival. Through the application of Salyer and Eggleston classification, the three histologic subtypes did not correlate with patients' ages at time of diagnosis, surgical stage as determined by local infiltration, and prognosis as determined by survival curves. On the contrary, when Marino and Müller-Hermelink classification was applied, statistically significant relationships between histologic results and age, surgical stage, and prognosis were demonstrated. These results and their implications are discussed, with special reference to the important problem of histogenesis of thymomas and of their clinicopathologic staging.     

Thymic tumours in Denmark. A retrospective study of 213 cases from 1970-1993.

Histological slides of 213 thymic tumours were reviewed twice and classified according to Kirchner and Müller-Hermelink into 122 thymomas (syn. organotypic thymic epithelial tumours (TET)), 58 thymic carcinomas (syn. non-organotypic TET) and 16 lymphomas. Tumour heterogeneity (i.e. features of two subtypes in one tumour) appeared in 38% of the organotypic TET. The overall diagnostic correspondence between the reviews of the 122 organotypic TET was 48%. By reducing the five diagnostic groups to three: organotypic TET benign (medullary and mixed thymomas), organotypic TET low-grade (organoid and cortical thymomas and well-differentiated thymic carcinoma (WDTC)) and non-organotypic TET (usually high-grade thymic carcinomas), and minimising the effect of tumour heterogeneity in this way, the diagnostic correspondence increased to 82%. Correlating histological type with stage, we found that 80% of medullary and 87% of mixed thymomas were stage I, that 85% of cortical and 81% of WDTC were stage II or III, and that non-organotypic TET were stage II or III (86%) or stage IV (14%), respectively. It is suggested to report on the heterogeneity of a given case of thymic epithelial tumour in the pathology reports and give the approximate percentage of each component, telling the clinician which component may determine the prognosis.     

Expression of cortical and medullary thymic epithelial antigens in thymomas. An immunohistological study of 14 cases including a characterization of the lymphocytic compartment

Four monoclonal antibodies against antigens expressed differentially by the normal thymus epithelium, which define the cortical, medullary and subcapsular compartments, were used for immunohistological characterization of the epithelial cells in 14 thymomas. Furthermore, thymoma-associated lymphocytes were studied with monoclonal antibodies directed against T-lymphocyte differentiation antigens (CD1a, CD3, T-cell antigen receptor). Only four of the 14 thymomas could be classified into either medullary or cortical type thymoma based on the immunophenotype of epithelial cells. Ten cases escaped immunophenotypical classification due to co-expression of medullary and cortical antigens by the tumour cells. This aberration from the normal phenotype might indicate the failure of differentiation of such tumours. The immunophenotype of the associated lymphocytes, on the other hand, made it possible to classify the tumours as cortical (5 cases), mixed (2) and medullary (3) thymomas. Four thymomas escaped this classification scheme due to the absence of lymphocytes (2) or to a hybrid immunophenotype (2). Nevertheless, thymocytes of cortical type clearly predominated and were seen in all thymomas with associated lymphocytes. This feature may constitute a good diagnostic tool in differential diagnosis since, in 28 mediastinal or extramediastinal metastasis of tumours not derived from thymic epithelium and associated with various numbers of lymphocytes, none of them were found to contain CD1a positive lymphocytes.     

Thymoma: an integrated clinicopathological and immunohistochemical study

The clinicopathological features of 32 thymomas were reviewed and tumours were staged according to their degree of invasion. Their antigenic profiles were studied using monoclonal antibodies to cytokeratins (CAM 5.2 and DAKO-CK1), HNK-1 (Leu 7), and HLA-DR (TAL-IB5). Stage I (non-invasive) tumours were mainly of the spindle cell (SC) or predominantly lymphocytic (PL) types, whilst all the predominantly epithelial (PE) tumours were either locally invasive (stage II) or showed more extensive spread (stage III). Neoplastic epithelial cells all expressed cytokeratin, but varied in their degree of positivity. CAM 5.2 was more uniformly positive with cells at the periphery of tumour nodules and lining tubulo-cystic areas staining most strongly. DAKO-CK1 gave less uniform positivity but highlighted areas of medullary differentiation. HNK-1 was variably expressed in all tumour groups but was found more often in the invasive tumours (73 per cent stage III, 62 per cent stage II, 50 per cent stage I), particularly those of PE or mixed (M) type. In general, TAL-IB5 expression was lost in the more invasive thymomas. Focal medullary differentiation in tumours suggests a common origin for cortical and medullary epithelium, indicating that sub-division of tumours into cortical or medullary types is not valid. Immunohistochemistry may usefully complement clinical and macroscopic findings in the assessment of malignancy in thymoma.     

Thymoma—a study of 60 cases in Singapore

Of 60 cases of thymomas studied in Singapore between 1988 and 1992, the histogenetic classification proposed by Müller-Hermelink was successfully applied to subtype 58 cases. There were 20 (33%) cortical, six (10%) predominantly cortical, three (5%) medullary and 12 (20%) mixed thymomas. Twelve (20%) cases were well differentiated thymic carcinomas and five (8%) were classified as other thymic carcinomas. The pathological and clinical features are presented in detail. These subtypes showed significant correlation with invasive behaviour (stage) and myasthenia gravis. We conclude that the Müller-Hermelink classification has predictive utility and represents a major step towards the understanding of the biology of thymic epithelial tumours.     

DNA-ploidy analysis correlates with the histogenetic classification of thymic epithelial tumours

The ploidy values of the epithelial component were determined in a series of thymomas and organotypic thymic carcinomas using image cytometry and the results were compared with the histological tumour subtypes according to the histogenetic classification introduced by Marino, Müller Hermelink, and Kirchner (MMHK). Forty-six cases of thymic epithelial tumours were included in the study. After reclassification according to the MMHK classification, the distribution among the subtypes was as follows: three medullary, nine mixed type, five predominantly cortical (organoid), 16 cortical thymomas, and 13 well-differentiated thymic carcinomas. Single cell preparations were made from paraffin-embedded tumour tissue and stained according to Feulgen. Ploidy analysis was performed using an automated image analysis system. In five cases, DNA cytometry could not be performed, for technical reasons. The remaining 41 cases consisted of 11 diploid and 30 non-diploid tumours. The percentage of aneuploid tumours in the different subtypes increased from medullary (0 per cent) through mixed type (44.4 per cent), predominantly cortical (75 per cent), cortical (83.3 per cent) to well-differentiated thymic carcinomas (100 per cent). DNA-ploidy determination using image cytometry correlates with the concept of the MMHK classification of thymomas.     

Immunohistochemical study of 22 cases of thymoma.

The histological and clinical features of 22 thymomas were reviewed. Immunohistochemical studies were performed using antibodies to cytokeratins (CAM 5.2 and S29) and to desmosomal protein. The heterogeneity of staining patterns seen in the epithelial cells supported the concept of separating thymomas into cortical, medullary, or mixed groups. Interdigitating cells were identified by antibody to S100 protein, and these varied in number between different tumours. Clustering of interdigitating reticulum cells occurred in association with foci of mature lymphocytes which were shown by staining of the leucocyte common antigen (CD45). The extent to which this occurred was used to assess the degree of medullary differentiation within the thymomas and this was correlated with the histological and clinical features. The lymphocyte population of six of the thymomas was studied using a range of antibodies to T and B cells; this showed the presence of B lymphocytes in thymomas.     

Histopatholgical spectrum of thymic neoplasms: twelve-year experience at a referral hospital in north India

This study was undertaken to determine the histopathological spectrum and clinical profile of thymic neoplasms at a tertiary referral care centre. A total of 96 thymectomy specimens were received during the study period (1992-2004), which consisted of 54 neoplasms and 42 benign lesions. Among the neoplasms there were 48 thymic epithelial tumors, 3 thymolipomas and 3 thymic carcinoids. The former comprised of 36 male (75%) and 12 female patients (25%) ranging in age from 2-70 years (mean 37 years). Among paraneoplastic syndromes in thymic epithelial tumours, 27 out of 48 (56.25%) cases were associated with myasthenia gravis and one case was associated with pure red cell aplasia. The most frequent histological subtype was cortical thymoma (43.24%) followed by predominantly cortical (24.32%) and well-differentiated thymic carcinoma (18.92%). On staging, all cases of mixed and predominantly cortical subtype were stage 1 whereas one medullary and 2 cortical thymomas and 4 well differentiated thymic carcinoma (WDTC) showed pleural and pericardial invasion (stage III). This study has revealed that half of thymic epithelial tumours presented as myasthenia gravis. The cortical thymoma was the most frequently encountered histologic subtype and most commonly associated with myasthenia gravis.     

Histopathologic changes of thymoma preoperatively treated with corticosteroids

Preoperative treatment of thymoma in advanced stages with corticosteroids may reduce the size of the tumor, but no precise histologic evaluation has been performed. We examined the histopathologic features of pretreatment biopsy and posttreatment surgical specimens of eleven cases of thymoma with such treatment to see the changes of the histologic subtypes based on Muller-Hermelink classification. All specimens were also assessed immunohistochemically for MIB-1 labeling and apoptotic cells to verify the effectiveness of this pretreatment. Seven tumors clinically diminished in size after the treatment with corticosteroids. Fungal infection occurred in three cases postoperatively. The histology of mixed thymomas (two cases) was converted to that of medullary thymoma. Predominantly cortical thymomas (four cases) and cortical thymomas (three cases) changed to show similar histologic features; both became epithelial-rich thymoma with large polygonal tumor cells having indistinct cell borders. In contrast, two well-differentiated thymic carcinomas showed at surgery more prominent squamoid appearance with distinct cell borders. The apoptotic indices of epithelial cells were increased (P = 0.001), and the MIB-1 indices tended to be decreased with corticosteroid treatment. These results suggest that there may be a histogenetic relationship between medullary and mixed thymomas and also between predominantly cortical and cortical thymomas. Corticosteroids may cause degenerative changes in the epithelial cells and lymphocytes and, in thymomas in advanced stages, corticosteroid pretreatment may be warranted, although attention should be paid to infection after surgery.     

Reproducibility of the WHO/IASLC grading system for pre-invasive squamous lesions of the bronchus: a study of inter-observer and intra-observer variation

AIMS: Although many workers have graded pre-invasive squamous lesions arising in the bronchus, there has been no consensus classification system until the latest edition of the WHO/IASLC histological classification of pulmonary and pleural tumours. Because the value of any such system is dependent on its reproducibility, we have circulated a series of such lesions to a panel of histopathologists to assess interobserver and intra-observer variation when the WHO/IASLC classification was applied. METHODS AND RESULTS: Colour transparencies of 28 pre-invasive squamous lesions were assessed by six histopathologists (two with a special interest in pulmonary pathology, two generalists and two trainees) on three separate occasions over a period of 3 months, using the criteria of the WHO/IASLC (mild, moderate and severe dysplasia, and in-situ carcinoma). An additional category of metaplasia was added for those cases that showed no dysplasia. Weighted kappa coefficents of agreement (K(w)) were used to evaluate paired observations with a standard quadratic weighting being employed, such that kappa coefficients corresponded to intra-class correlation coefficients. Wilcoxon's sign-ranked test was used to measure the statistical significance of group trends, when comparing kappa values for the three grading systems. Various 3-point systems were also assessed, through combination of the above groups. Intra-observer agreement was substantially better than interobserver variation (mean: 0.71 vs. 0.55). Between the various pathologist groups, inter-observer variation was relatively minor, although intra-observer variation was higher within the trainee pathologist group. Using weighted kappa values, there was no significant difference in either inter-observer or intra-observer agreement between the five point grading system and a 3-point system of metaplasia/mild, moderate and severe/in-situ grades. However, there was a significant increase in variation when a 3-point system of metaplasia/mild, moderate/severe and in-situ carcinoma was used. CONCLUSION: This study shows levels of interobserver and intra-observer variation similar to those found in other grading systems in histopathology, with no significant decrease in variability found by abridging the system. The WHO/IASLC system is therefore recommended for future use in both clinical and research fields.     

Evaluation of prognostic features in thymic epithelial tumors

Based on the original proposals of Müller-Hermelink [3,8] and the study of 95 tumor specimens from the files of our institute we have established a new concept for the classification of thymic epithelial tumors. Thymomas are related to the structural components of normal thymus and divided in medullary, mixed, predominantly cortical and cortical types. In addition a well-differentiated thymic carcinoma with partial loss of organotypic differentiation is characterized and distinguished from other carcinoma types with total lack of specific thymic features. Prognostic evaluation showed, that medullary and mixed thymomas are always benign tumors, whereas predominantly cortical thymomas, cortical thymomas and well-differentiated thymic carcinomas are low-grade malignant tumors with increasing invasiveness and even metastatic capacity. Moreover the proliferation rate of neoplastic epithelial cells in vitro, which was studied in 12 cases, correlated to the different tumor types and their growth behaviour in vivo.     

Immunohistological evidences of cortical and medullary differentiation in thymoma

Summary The phenotypical characteristics of human epithelial and lymphoid cells have been studied with immunohistochemical methods on frozen sections of 12 thymomas. On the basis of the cytohistological characteristics of thymoma epithelial cells (EC) the thymomas were divided in cortical, medullary and mixed types, according to recently developed light microscopical criteria. When tested with a series of monoclonal antibodies, thymoma EC were all stained by the antibody Ki-M3 (as in the thymus), but reacted with anti-HLA-DR, anti-HLA-A,B,C and with a new monoclonal antibody to cortical EC,21A6, to a lesser extent and with weaker, variable intensity in comparison with the normal thymus. Cortical type thymomas were most reactive and the medullary type almost negative. Thymomas, like normal thymus showed different immunoreactivity patterns with antibodies to prekeratins of different specificities. Cortical type thymomas and areas in mixed thymoma showed an EC staining with the antibody to non-squamous type keratin (35H11) whereas medullary type thymomas and areas showed staining with antibodies to squamous-type keratin (34E12-IV/82) in addition. Lymphoidcellswithcortical(OKT6+,Leu 1 weakly+,Leu2a+,Leu3a+) or mature medullary (OKT6-, Leu 1 strongly+, Leu 2a or Leu 3a+) phenotype were found to colonize tumours with diferent EC types. These immunohistochemical findings largely confirm our earlier cytological distinction of thymoma EC. In addition important differences have been observed in neoplastic cortical EC concerning the HLA-DR and 21A6 immunoreactivity that may be intimately related to the neoplastic process and paraneoplastic immune phenomena.This work has been supported by the Deutsche Forschungsgemeinschaft, SFB 111, project CN5     

Evaluation of a histogenetic classification for thymic epithelial tumours

We reviewed 87 thymic epithelial tumours from Chinese patients and typed them according to the Marino and Müller-Hermelink classification as updated by Kirschner and Müller-Hermelink in 1989. Related categories were grouped for statistical analyses; group 1, medullary thymoma and mixed thymoma; group 2, cortical predominant thymoma; group 3, cortical thymoma and well-differentiated thymic carcinoma; group 4, other thymic carcinomas; and group 5, unclassified. Group 3 tumours were more frequently associated with the myasthenia gravis syndrome compared with group 1 tumours ( P =0.001). They also presented at a more advanced stage. Groups 1 and 2 showed an excellent prognosis (100% survival at 10 years). The 10-year survival for groups 3 and 4 patients was 40% and 30% respectively. Pure medullary thymoma made up a higher proportion of our cases (10.3%) than those of a similar Caucasian study (5.3%). The eight thymic carcinomas (group 4) included two thymic lymphoepitheliomas. We conclude that the histogenetic classification evaluated shows a clear correlation with prognosis and clinical features, even when tested on separate geographic groups, where pathogenetic factors may be different. A common approach to classification of thymic epithelial tumours would greatly facilitate future studies on these possible differences.     

PE-35-related antigen expression and CD1a-positive lymphocytes in thymoma subtypes based on Müller-Hermelink classification

PE-35 monoclonal antibody, detecting a cell-surface antigen of various types of carcinoma and normal epithelium, reacts exclusively with the medullary epithelium in the thymus; therefore, the antigen has been considered as a marker of medullary differentiation in thymomas. Using the catalyzed signal amplification method, which made it possible to apply PE-35 to routinely processed, archival tissues, we examined expression of this antigen, together with CD1a reactivity of lymphocytes, in 40 thymic epithelial tumors subclassified using the Mü1ler-Hermelink system. Medullary thymomas infiltrated with a small number of CD1a-negative lymphocytes were PE-35 positive, although many of the long spindle tumor cells were PE-35 negative. Mixed thymomas and predominantly cortical thymomas, both with prominent CD1a-positive lymphocytes, were also PE-35 positive, although some areas of the latter type were PE-35 negative. Cortical thymomas with decreased numbers of CD1a-positive lymphocytes were largely PE-35 negative. In well-differentiated thymic carcinomas with a few CD1a-positive lymphocytes, two cases were negative, but four cases were at least focally positive with PE-35. All high-grade thymic carcinomas infiltrated with some CD1a-negative lymphocytes were PE-35 positive. These results suggested that medullary thymoma generally possesses the medullary nature, although the latter tends to be lost in the long spindle tumor cells. Mixed and predominantly cortical thymomas may have mixed medullary phenotype and cortical function. Cortical thymoma and many well-differentiated thymic carcinomas may possess the cortical nature, while the large polygonal tumor cells tend to lose immature T-lymphocyte-retaining function. Received: 25 January 1999 / Accepted: 14 July 1999     

Pitfalls in the diagnosis of ectopic hamartomatous thymoma

We have investigated the immunohistochemical expression of p53, mdm2, p21/waf1 and bcl-2 proteins in 31 thymic epithelial tumours comprising five medullary thymomas (MDT), four mixed thymomas (MT), 12 cortical thymomas (CT), eight predominately cortical thymomas (PCT) and two well-differentiated thymic carcinomas (WDTC). We have found p53, mdm2, p21 and bcl-2 protein expression in 25/31, 8/31, 5/31 and 10/31 thymic epithelial tumours, respectively. Coexpression of p53 and mdm2 proteins was found in eight cases (three CT, four PCT and one WDTC). Five of them were also p21 positive and three p21 negative. Discordant p53+/mdm2−/p21− protein expression was found in 19 cases (three MDT, three MT, nine CT, three PCT and one WDTC). Mdm2 and p21 proteins were not expressed in the absence of p53 protein. Coexpression of bcl-2 and p53 proteins was found in seven cases (three MDT, three MT and one WDTC). Eighteen cases were p53+/bcl-2− (10 CT, seven PCT and one WDTC) and three cases (two MDT and one MT) were bcl-2+/p53−. Our findings indicate that in thymomas, p53 expression is more frequently associated with cortical histotypes while bcl-2 expression is strongly associated with medullary and mixed histotypes. In addition, there is an inverse correlation between p53 and bcl-2 protein expression in thymomas. Coexpression of p53/mdm2/p21 proteins may reflect thymomas with wild-type (wt), p53 gene since mdm2 and p21 proteins are inducible by wt p53 gene. However, in view of previous findings that p53 mutation is an early event in thymomas, the possibility of p53 gene mutation with p53-independent mdm2 and p21 expression should be considered in these cases. Discordant p53+/mdm2−/p21− protein expression may represent thymomas with p53 gene mutations unable to activate expression of mdm2 and p21 proteins.     

Immunohistochemical localization of Mcl-1 and bcl-2 proteins in thymic epithelial tumours

Mcl-1 protein is a new member of the bcl-2 protein family. It is believed to be a blocker of apoptosis but might be different from bcl-2 in the control of apoptosis. Using immunostaining of formalin-fixed, paraffin-embedded sections, we investigated the expression of Mcl-1 in 42 thymic epithelial tumours: three medullary thymomas, five mixed thymomas, seven cortical thymomas, eight well-differentiated thymic carcinomas, 14 squamous cell carcinomas, four lymphoepithelioma-like carcinomas and one undifferentiated carcinoma. bcl-2 immunocytochemical localization was also performed for comparison. High-grade thymic carcinomas, especially squamous cell carcinomas, revealed more intense Mcl-1 immunoreactivity as compared to other subtypes ( P < 0.001). In cases that co-expressed Mcl-1 and bcl-2, the less differentiated cells had more intense expression of bcl-2, while the more differentiated cells displayed stronger Mcl-1 immunoreactivity. The differential expression of Mcl-1 and bcl-2 in neoplastic cells provides evidence that these proteins may play different roles in the processes of programmed cell death in thymic neoplasms. The finding that thymic carcinomas have stronger immunoreactivity for Mcl-1 indicates that this protein could be a useful marker to differentiate aggressive thymic epithelial tumours from indolent ones.     

Phenotypical characteristics and proliferative capabilities of thymocyte subsets in human thymoma

The phenotypical and functional characteristics of the lymphoid component present in ten human thymomas have been analyzed. Thymomas were classified according to the predominant epithelial cell type present in the neoplastic gland. In thymomas of the cortical type a large proportion, clearly higher than that present in mixed thymomas, of T6-positive lymphocytes was present. This T6 subset did not proliferate to mitogen but contained almost all thymocytes able to spontaneously proliferate. Higher proportions of T3-positive cells were found in mixed types of thymomas than in cortical-type tumors. This T3+ subset responded to mitogen stimulation and constituted the more mature intrathymic pool. Surprisingly we identified in thymomas a further subset, lacking the cortical and medullary markers T6 and T3, capable of responding to mitogen. The occurrence of higher proliferative responses to phytohemagglutinin (PHA) in the unfractionated population of cells from mixed thymomas than in the cells of cortical thymomas was judged to be attributable to the relatively higher content of both T3-positive and T3, T6-negative thymocytes in the former. Unlike T6+ thymocytes, T3-T6- as well as T3+ cells were practically devoid of spontaneous proliferative capacity. The expansion of the intrathymic lymphoid component in human thymomas should then be considered to be attributable to the "spontaneous" proliferative capacity of the T6+ cell pool. In this respect, cortical thymomas not only contained more T6+ cells than the mixed type but also exhibited a higher lymphocyte/epithelial cell ratio and more frequent mitotic figures.     

An immunohistologic study of the epithelial and lymphoid components of six thymomas

Six thymomas were classified histologically and studied immunohistochemically with a panel of mouse and rat monoclonal antibodies directed against thymic epithelial and lymphoid components. The antibodies included monoclonal antibodies directed against cytokeratin, medullary epithelial cells (ER-TR5), and HLA-DR and HLA-ABC antigens, as well as antibodies with specificity for thymocytes. Histologically, one thymoma was characterized by epithelial predominance (EP type), two showed lymphoid predominance (LP type), and two showed mixed lymphoid/epithelial composition (MLE type); one thymoma was a malignant pure epithelial thymoma (PE type). In the thymomas of the MLE and EP types the major populations of cells consisted of HLA-DR-negative, cytokeratin-positive epithelial cells with large ER-TR5-positive subpopulations (i.e., the phenotype of medullary epithelium). In the thymomas of the LP type, the neoplastic population was composed of cytokeratin-positive, ER-TR5-negative cells that expressed the HLA-DR antigen (i.e., the phenotype of cortical epithelium). The thymoma of the PE type consisted of cytokeratin-positive cells, some of which were ER-TR5- and HLA-DR-positive. Double immunofluorescence studies revealed the presence of varying numbers of additional nonepithelial (nonlymphoid) HLA-DR-positive cells in thymomas of the LP, MLE, and EP types. The intervening lymphoid population in the thymomas of the LP, MLE, and EP types consisted largely of cortical thymocytes, as defined by immunologic characterization. These results suggest that thymomas can be classified as medullary or cortical epithelial neoplasms on the basis of their immunologic phenotypes.     

Can histopathologists reliably diagnose molar pregnancy?

AIMS--To assess the degree of difficulty in diagnosing partial mole by analysing intraobserver and interobserver agreement among a group of pathologists for these diagnoses. METHODS--Fifty mixed cases of partial mole, complete mole, and non-molar pregnancy were submitted to seven histopathologists, two of whom are expert gynaecological pathologists; the other five were district general hospital consultants, one of whom works in Australia. These participants gave each slide a firm diagnosis of either partial mole, complete mole, or non-molar pregnancy. Some 12 months later, the slides were recorded and again submitted for a second diagnostic round to assess intraobserver as well as interobserver agreement. Standard histological criteria for each diagnostic category were circulated with the slides. RESULTS--kappa statistics showed that complete mole could be reliably distinguished from non-molar pregnancy, but neither non-molar pregnancy nor complete mole could be easily differentiated from partial mole. In only 35 out of 50 cases was there agreement between five or more of the seven participants. Agreement between the expert gynaecological pathologists was no better than for others in the group. Interestingly, the intraobserver agreement for each pathologist was good to excellent. CONCLUSIONS--These results imply that the reported histological criteria are either not being applied consistently or that they are lacking in practical use. An atypical growth pattern of trophoblast, rather than the polar accentuation seen in normal first trimester pregnancies, seems to be the important diagnostic histological feature for partial mole. Ploidy studies might also help with problem cases.     

Immunohistochemical metallothionein expression in thymoma: correlation with histological types and cellular origin

Metallothioneins (MTs) are low molecular weight and cysteine-rich intracellular proteins involved in metal homeostasis and detoxication. They are found in certain normal tissues, and are overexpressed in various tumours with correlation to more aggressive behaviour in certain tumours. Since the histopathological types of thymoma have unpredictable invasive potential, MT over-expression was investigated as a possible marker of the invasive potential of thymomas. We studied immunohistochemical MT expression in 27 non-invasive thymomas, 20 micro-invasive thymomas, and 23 macro-invasive thymomas with a mouse monoclonal anti-MT antibody E9 on formalin-fixed paraffin-embedded tissues. MT expression was significantly different among the three groups of thymomas ( P  = 0.02) with a stronger expression in invasive thymomas ( P  = 0.003). However, MT expression was not exclusively limited to invasive thymomas. Therefore, it could not be used as a marker of aggressive potential in individual thymomas. Analysis of MT expression according to the histological types of the thymomas revealed that eight of nine spindle cell thymomas, none of 10 small polygonal cell thymomas, four of 14 mixed thymomas, seven of 29 large polygonal cell thymomas, and seven of eight squamoid thymomas significantly expressed MT. There was a statistically significant difference in MT expression among different histological types of thymomas ( P  = 0.000). The strongest and most consistent expression was observed in spindle cell thymoma and squamoid thymoma. Since spindle cell thymoma was usually non-invasive and squamoid thymoma was more aggressive, MT expression does not correlate with the invasive potential of different histological types of thymomas. But because medullary epithelial cells of the thymus were positive for MT, our results suggest that both spindle cell thymoma and squamoid thymoma might derive from the medullary compartment of the thymus.