Material and methods

A retrospective and multicenter study was carried out by 8 Argentinian pathology laboratories (6 from Buenos Aires City and 2 from Buenos Aires Province).

We reviewed cases with previous DPM diagnosis since 2009 to 2018, obtained from surgical biopsies or small biopsies representative of diagnostic histologic criteria, with immunohistochemistry slides available to confirm DPM diagnosis. Small biopsies were included only if they have clear mesothelial cells stromal invasion. In cases where there was both a biopsy and resection of a patient, the sample corresponding to resection was chosen.

The slides were reviewed by two or more pathologists from each laboratory without centralized review, because histopath ologic parameters have been shown to be reproducible^5-7,16.

All the potential transitional type cases and problematic or controversial cases were jointly reviewed and discussed in face-to-face meetings.

We described histological types, main and second pre dominant E-DPM subtypes according the 2015 WHO criteria. E-DPM subtypes analyzed were tubular, papillary, acinar, trabecular, solid, micropapilar, adenomatoid, clear cells, de ciduoide, small cells, pleomorphic and rhabdoid. In S-DPM cases we indicated if desmoplastic or giant cells were present. Biphasic DPM were considered in cases with at least 10% epithelioid component in S-DPM or sarcomatoid component in E-DPM. Transitional mesothelioma was diagnosed in cases with cohesive large epithelioid cells and well-defined border without sarcomatoid characteristics. High nuclei-cytoplasmic ratio, and prominent nucleoli were present. We arbitrarily considered whether the presentation was extensive or focal (≥ 5%) in combination with other subtypes.

We followed IMIG 2018 histological features recommenda tions: necrosis (absent, present) nuclear atypia and mitosis/10 HPF (or fields needed to reach 2 mm²). Nuclear atypia was evaluated considering nuclear size and irregularity. It is estab lished by the areas of the highest grade, present in > 5% of the tumor, as follows: mild (nuclei uniform in size and shape), moderate (nuclei intermediate in size between mild and severe atypia, with slight irregularity in shape), and severe (bizarre, large nuclei of variable size with some nuclei twice as large). Scoring schemes were performed considering nuclear atypia (1 for mild, 2 for moderate, and 3 for severe atypia) and mitotic count: 1 for low (0-1 mitosis/10 HPF), 2 for intermediate (2-4 mitosis/10 HPF), and 3 for high (≥ 5 mitosis/10 HPF). Nuclear scoring scheme: total score was computed as the sum of the two parameter scores, ranging from 2 to 6. Grade I for total scores 2 or 3, grade II for total scores 4 or 5, and grade III for a total score 6. Overall tumor grade was calculated as low grade when it had nuclear score I, grade II without necrosis, high grade when it had nuclear score II with necrosis, and III with or without necrosis. We decide to evaluate nuclear score and necrosis not only in E-DPM but also in other histological types.

The stroma was categorized as desmoplastic, myxoid, with heterologous differentiation, fibrotic or without anything to specify (NOS).

Immunohistochemistry was reviewed to confirm mesothelial neoplasia and to conduct a survey of the antibody panel used in our region when the diagnosis of DPM is suspected. We considered calretinin, WT-1, AE1AE3, cytokeratin 5, CEA, TTF-1, and other antibodies used in some special cases.

To perform statistical analysis, we grouped S-DPM, transi tional and biphasic-DPM in no E-DPM. We calculated frequen cies and Chi-squared test or Fisher exact test and Pearson test to compare categorical variables with IBM SPSS Statistics program.

Limitations of the study: it was difficult to obtain clinical data, environmental exposure history, smoking status, staging, treatments, and evolution.

The confidentiality of the data was respected, in accor dance with the Declaration of Helsinki and all its modifica tions and the Guide for Human Health Research (Ministry of Health of the Nation, Argentina, 2016) and Patient Rights Manual for the Health Team (Ministry of Health of the Nation, Argentina, 2021).

Results

We reviewed 398 cases corresponding to 238 (60%) men and 160 (40%) women with a median age of 66 years (24-91). Samples were obtained from surgical biopsies (78%), small biopsies (16.5%) and 5.5% from surgical resections (pleurectomy or pleural- pneumonectomy).

E-DPM was the most frequent histological type (306/398, 77%), followed by biphasic (47/398, 12%) and S-DPM (41/398, 10%). In 4/398 cases transitional DPM was found in extensive form, and in 8/398 cases focally (in 4 biphasic, 3 S-DPM and one solid E-DPM).

E-DPM main pattern was tubular in 111 (36%), solid in 101 (33%), trabecular in 37 (12%), papillary in 35 (11%), pleomorphic in 9 (3%), micropapillary in 6 (2%), adeno matoid in 6 (2%) and deciduoide in one E-DPM case.

We found a second pattern in 179 (58%) E-DPM cases. Tubular-papillary, tubular-solid and solid-tubular were the most frequent combinations (Table 1). When we considered together the main and second patterns, 48% of the E-DPM presented tubular subtype and another 48% solid subtype.

Table 1 Main and second pattern combinations frequencies presented in 106 epithelial-diffuse pleural mesothelioma cases 

Stroma, necrosis, and nuclear score frequencies and differences between E-DPM and no E-DPM are presented in Table 2.

Table 2 Histopathological characteristics of diffuse pleural mesothelioma: comparison between epithelioid-diffuse pleural mesothelioma and non-epithelioid diffuse pleural mesothelioma shows statistically significant differences 

When overall tumors were categorized, we found low grade in 233 (76%) E-DPM, in 28 (60%) B-DPM, in 26 (63%) S-DPM, and in only one transitional extensive form. We applied the same criteria to categorize more frequent E-DPM main patterns (Fig. 1).

Fig. 1 Overall tumor grade in more frequent diffuse epithelioid mesothelioma main patterns (n = 293) 

The immunohistochemical antibody panel used was variable depending on availability at the time of diagnosis in each laboratory. In E-DPM cytokeratin 5/6, calretinin and WT-1 were employed in 84%, 91% and 79% cases, respectively. CEA was used in 66% cases, with negative results, and TTF-1 resulted negative in 250 (82%) cases performed. Pancytokeratin (AE1AE3) was used in 90% of S-DPM cases, all were positive. Cytokeratin 5/6, cal retinin and WT-1 were employed in 85%, 98% and 76% of S-DPM, respectively. B-DPM diagnosis was done using pan-cytokeratin in 70% of these cases, cytokeratin 5/6 in 68%, calretinin in 85%, WT-1 in 79%, CEA in 64%, and TTF-1 in 74% of the cases. Expression results are de scribed in Table 3. Extensive transitional DPM expressed cytokeratin 5 in 3/4 of the cases, calretinin and WT-1 in all 4 cases. Other antibodies, MOC-31, EMA, HBME, vimentin, thrombomodulin, and S100 protein, were also used with variable frequency, therefore their results could not be analyzed.

Table 3 Antibody expression frequency according to total cases made for each antibody. Positivity was evaluated in relation to number of each marker performed since the immunohistochemical panel performed was not homogeneous 

Discussion

This study shows DPM pathological diagnosis feasibility in Argentina, and the possibility of applying prognostic histological factors in routine practice.

Most samples were obtained by open or video-thora coscopy biopsy which allowed good tumor representativity to analyze histology subtypes and grading, that also we could analyze in small biopsies.

It is well known that E-DPM has better prognosis than S-DPM and B-DPM, so the latter are not selected for surgery. But less is known about prognosis implications of E-DPM patterns and even less about the meaning of a second pattern. It was described tubular with a better overall survival than others, especially than pleomorphic E-DPM, but also than solid and micropapillary patterns^17-21. In this way Paajanen et al¹⁷ proposed to classify these subtypes into low-grade (trabecular, tubule-papillary) or high-grade (solid, micropapillary, pleomorphic). In the present study E-DPM was the most frequent type, with tubular and solid as the most frequent main patterns. But more than half of cases had a second pattern, showing up E-DPM tumor heterogeneity. Even though tubular/ papillary combination was the most frequently found, in 48% of the E-DPM reviewed cases we found solid pattern either as main or second (combined with tubular, papillary and trabecular cases). We consider that these features can add useful data to better understand mesotheliomas biological behavior.

We found that desmoplastic stroma was present mostly in no E-DPM, but not exclusively, E-DPM had it also in a good proportion of cases (97/306). Myxoid stroma was found almost exclusively in E-DPM. Stroma character istics should be considered in the differential diagnosis with other tumors. It also could have prognosis impact: in general patients with myxoid E-DPM seem to have a better overall survival when compared to those with epithelioid mesotheliomas^22,23.

Kadota et al⁸ studied seven nuclear features in 232 E-DPM and found that nuclear atypia and mitotic count were independent prognostic factors. These were utilized to create a three-tier nuclear grade score that correlated with clinical outcome. These results were confirmed in other studies⁹ and were incorporated to recommendations in EURACAN/IASLC Proposals for Updating the Histo logic Classification of Pleural Mesothelioma²⁴, in National Intersociety Consensus on Mesothelioma (Academia Nacional de Medicina de Buenos Aires)²⁵, and recently in WHO classification adapted to two-tier overall tumor grade including nuclear score and necrosis⁴. The objec tive of our study was not to analyze reproducibility, but rather a survey on the possibilities of performing the score in routine samples, therefore doubts and interpretations were discussed in face-to-face meetings even if in no case the differences would have changed the level in the score (data not shown). We were able to apply the criteria to all samples. Tubular E-DPM and solid E-DPM had necrosis and nuclear score differences which indicates their value in discriminating groups with greater aggressiveness. Bilecz et al²⁶ observed an association between solid and trabecular pattern with necrosis and high nuclear score. Moreover, they found in univariate analyses worst overall survival (OS) than tubular and better than pleomorphic but could not confirmed in multivariate analyses. Only pleomorphic features were able to predict OS in a study of Zhang et al²⁷. Additionally, if they considered it as a func tion of nuclear features and included 2-level nuclear grade rather than growth pattern as a covariate, pleomorphic features remained independently prognostic. As men tioned, in this series solid pattern was frequently present, not only as the main pattern but also together with other ones. There were a few pleomorphic cases, but mostly of high grade, so we consider important to identify and report them, because at least indirectly they can suggest forms of greater aggressiveness. Although nuclear grad ing and necrosis were not recommended to evaluate the prognosis in non-E-DPM, we extended their application to non-E-DPM and found significant differences. Low grade overall tumor was present in most of the E-DPM while the frequency of decrease in the non-E-DPM highlighting their different biological behavior. Perhaps grading should be explored in this group.

All potential transitional cases were discussed in face-to-face meetings, and there we found that diagnostic criteria here described are useful and possible to follow after training^4,7,28,29. Extensive form transitional cases presented necrosis, nuclear score II or III and although they are isolated cases point out their high degree of aggressiveness as described.

The immunohistochemical analysis shows up panel variability between laboratories. Although all the par ticipants belong to the same region of Argentina, the availability of antibodies is not always the same even in a single laboratory. Pan cytokeratin (AE1 AE3) was specially used in S-DPM, but we also find it valuable to identify the sarcomatoid component in B-DPM and differentiate it from fibroblasts as previously described^29,30 specially taking into consideration Barbieri et al³¹ notice about misclas sification particularly high for biphasic DPM (three-fourths of biphasic DPM at necropsy had been classified as epi thelioid at VATS or surgery). Claudin-4 is considered a marker with the highest sensitivity and specificity for the diagnosis of carcinomas versus E-DPM^4,32, however it was not included in any of the panels here analyzed and to our knowledge it is not a marker used in Argentina. We believe that immunohistochemistry standardization is required in our region to improve diagnosis, preserve tissue, and optimize resources.

This study has the limitation of lacking histopathological correlation with clinical history, treatments, and survival, confirming its prognostic value, but nevertheless we be lieve morphological prognostic factors can be established routinely and they contribute to selecting and adapting treatments in each patient. They are critical also in clinical trials design and their results interpretation³³.

Our work remarks tumor heterogeneity of DPM and the feasibility of evaluating prognostic histopathological criteria on pleural biopsies in E-DPM, but also in B-DPM and S-DPM. Main and second pattern description have shown value in discriminating E-DPM groups with greater aggressiveness. We were able to recognize the transi tional variant and given its clinical importance we must call the attention on it.

We consider that comprehensive pathological study of mesotheliomas provides useful information that al lows establishing prognoses and guiding treatments with greater precision. In addition, it allows to know in detail the histological characteristics of DPM in Argentina.