Malignant pleural mesothelioma (MPM) is a rare tumor

Transcription

1 Reproducibility of Malignant Pleural Mesothelioma Histopathologic Subtyping Luka Brcic, MD, PhD; Gregor Vlacic, MD; Franz Quehenberger, PhD; Izidor Kern, MD Context. Malignant pleural mesothelioma (MPM) is a rare tumor with poor prognosis. Several studies have analyzed potential prognostic markers, but histologic type remains the single most important prognostic factor. Histologic subtypes of epithelioid MPM seem to have prognostic and therapeutic implications. Interobserver agreement in histologic pattern classification should be high. Objective. To assess interobserver and intraobserver reproducibility in histologic differentiation between the main types of MPMs, and in further subtyping of epithelioid-type mesothelioma. Design. One representative hematoxylin-eosin stained slide was selected from the archive for each of 200 patients with MPM. They were reviewed independently by 3 pathologists and classified according to the current World Health Organization classification of pleural tumors. After the first round of evaluations, a consensus meeting was organized where problems were addressed and representative images for each histologic category were selected. Two months later, cases were reevaluated by all 3 pathologists. Results. After the first round, overall interobserver agreement for histologic subtyping of mesothelioma was fair (j, 0.36). The agreement was increased to substantial (j, 0.63) in the second round. Improvement was found in interobserver agreement for all types of MPM and for most epithelioid subtypes. Conclusions. Moderate to substantial agreement in histologic typing and subtyping of MPM can be achieved. However, training with additional clarification of diagnostic criteria, their strict application, and help from consensus-based illustrative images is needed. (Arch Pathol Lab Med. 2018;142: ; doi: / arpa oa) Malignant pleural mesothelioma (MPM) is a rare tumor but is also the most common primary malignant tumor of the pleura and has a dismal prognosis. Classical cases are caused by asbestos exposure and present after a long latency period. The incidence of MPM in the United States is decreasing owing to earlier asbestos use control. 1 However, in many other developed and developing countries, it is expected to reach a peak in the next 5 to 10 years. 2 Although there are occasional reports of patients living more than 5 years after the first diagnosis, most commonly, patients are diagnosed in an advanced stage and have a median survival of up to 13 months regardless of treatment modalities. 3 5 Previous studies 6,7 have analyzed different clinical, radiologic, and pathologic variables as prognostic factors. Accepted for publication September 6, Published as an Early Online Release March 6, From the Institute of Pathology (Dr Brcic) and the Institute for Medical Informatics, Statistics and Documentation (Dr Quehenberger), Medical University of Graz, Graz, Austria; and Cytology and Pathology Laboratory, University Clinic of Respiratory and Allergic Diseases, Golnik, Slovenia (Drs Vlacic and Kern). The authors have no relevant financial interest in the products or companies described in this article. Presented in part as an oral presentation at the 13th International Conference of the International Mesothelioma Interest Group (imig 2016); May 1 4, 2016; Birmingham, United Kingdom. Corresponding author: Luka Brcic, MD, PhD, Institute of Pathology, Medical University of Graz, Neue Stiftingtalstraße 6, 8010 Graz, Austria ( luka.brcic@medunigraz.at). Histologic type is still the most important single prognostic factor. The epithelioid type of MPM is associated with significantly better survival than the sarcomatoid and biphasic types. 8,9 Even more, histologic subtyping is currently used to make treatment decisions. 10 According to the latest World Health Organization (WHO) Classification of Tumors of the Lung, Pleura, Thymus and Heart, 11 the epithelioid type is very heterogeneous. It presents with different histologic patterns. Among these, the most common are solid, tubulopapillary, and trabecular, followed by micropapillary, adenomatoid, clear cell, transitional, deciduoid, and pleomorphic. 11 The prognostic significance of different histologic patterns of epithelioid mesotheliomas has been analyzed in a few studies, which suggest poorer survival for individuals with the pleomorphic pattern In contrast, the myxoid/microcystic pattern was shown to be a favorable prognostic marker. 15 Furthermore, Kadota et al 13 have reported better prognosis for individuals with mesotheliomas presenting with trabecular or tubulopapillary patterns than for those with other epithelioid mesothelioma patterns. Proper histologic subtyping of epithelioid MPM may be a potentially useful tool for prognostic stratification of patients. It is therefore essential to have uniform diagnostic criteria and terminology, as well as good interobserver reproducibility among pathologists. However, although there are few studies regarding interobserver reproducibility in the diagnosis of MPM, mainly focused on immunohistologic reproducibility, there is only 1 Arch Pathol Lab Med Vol 142, June 2018 Reproducibility of Pleural Mesothelioma Subtyping Brcic et al 747

2 published article analyzing the reproducibility of histologic subtyping in epithelioid MPM. 12 The purpose of this study was to assess interobserver and intraobserver reproducibility in the histologic differentiation between the main types of MPMs, and in further subtyping of epithelioid MPM. The impact of a consensus meeting, during which a representative set of slides was agreed upon, was evaluated. MATERIALS AND METHODS Mesothelioma Samples We identified 200 consecutive samples from patients diagnosed with MPM in the archives of the Institute of Pathology University of Zagreb School of Medicine, Zagreb, Croatia (100) and the Cytology and Pathology Laboratory at the University Clinic of Respiratory and Allergic Diseases, Golnik, Slovenia (100). Most of the patients were diagnosed by video-assisted thoracoscopic biopsies (137 of 200 patients, 68.5%), while 39 (19.5%) had pleurectomy and 24 (12%) underwent percutaneous pleural needle core biopsies. The original diagnoses of all samples were made by using at least 2 mesothelial cell markers (calretinin, cytokeratin 5/6, WT-1, D2-40, or thrombomodulin), and at least 2 negative markers (Ber-EP4, MOC31, carcinoembryonic antigen, or thyroid transcription factor-1). Only 1 representative hematoxylin-eosin (HE) stained slide from each case was selected for this study. All slides were digitized with an Aperio scanner (Leica Biosystems, Wetzlar, Germany) with an 340 objective and reviewed independently by 3 pathologists (L.B., G.V., I.K.). Two pathologists (I.K., L.B.) are specialized in pulmonary pathology and have 20 and 5 years of experience, respectively. The third pathologist (G.V.) is a young pathologist with 1 year of experience in pulmonary pathology. Whole slide images from patients with MPM were classified as epithelioid, sarcomatoid, or biphasic according to the 2015 WHO criteria for the classification of pleural tumors. 11 Epithelioid MPM cases were further subtyped according to the predominant pattern as acinar, adenomatoid (microglandular), micropapillary, solid, tubulopapillary, trabecular, pleomorphic, clear cell, deciduoid, signet ring cell, or small cell variant, as defined previously. 13,19,20 After the first round of evaluations, a consensus meeting was organized. During this meeting, diagnostic problems were addressed and slides with discordant diagnoses were jointly reviewed. Reference images (Figure 1, A through H) illustrating typical patterns for each histologic category were selected and later shared among all of the reviewers. All cases were reevaluated by all 3 pathologists after an interval of 2 months following the consensus meeting. The pathologists were allowed to synchronously review the reference images. Reproducibility of the subtyping between the first and the second round for each pathologist was assessed as well (intraobserver reproducibility). To evaluate the influence of the sample size on the reproducibility, overall interobserver agreement, after both rounds, was analyzed for the (small) group of samples obtained by percutaneous pleural needle biopsy. Statistical Analysis To evaluate the reproducibility of histopathologic mesothelioma subtyping, interobserver variability was calculated. Furthermore, the effect of the consensus meeting on the diagnoses of each individual pathologist was assessed. Cohen j was used to assess interobserver and intraobserver agreement. Values of 0.20 to 0.40, 0.41 to 0.60, 0.61 to 0.80, and greater than 0.80 were taken to reflect fair, moderate, substantial, and excellent agreement, respectively. 21 Analyses were performed with R ( accessed November 3, 2016) and the irr 0.84 package. RESULTS Assessment of overall interobserver agreement of histologic subtyping of mesothelioma during the first round of evaluation led to a j value of only 0.36 (fair agreement), with j values between each pair of pathologists equal to 0.34, 0.35, and 0.38 (Figure 2). The highest overall interobserver agreement was observed for the sarcomatoid type of MPM, followed by the epithelioid type (0.60 and 0.50, respectively). Classification of MPM as biphasic, based on the HE slides only, appeared to be very difficult, as reflected by a j value of In further subtyping epithelioid MPM, the best agreements were detected for tubulopapillary, pleomorphic, and trabecular patterns (0.54, 0.49, and 0.42, respectively) (Table). At the consensus meeting, several problems were addressed, possible solutions were discussed, and consensus was defined. As expected, since our approach included only HE-stained slides, the criteria used to identify spindle cells as malignant mesothelial cells were slightly different for each reviewer. We agreed upon strict criteria for the identification of sarcomatoid and biphasic types of MPM at the meeting. Namely, all samples with tumors consisting of spindle cells with atypia and polymorphous appearance or mesenchymal-appearing cells were to be classified as sarcomatoid. At least 10% of each, epithelioid and sarcomatoid components, were required in order to classify MPM as biphasic. However, we are aware that without additional analysis, the decision of whether a spindle-cell component represented cellular reactive stroma or a genuine sarcomatoid component is not possible. Another issue discussed was the assessment of morphologically striking pseudodominating patterns, such as micropapillary or papillary patterns. It was sometimes difficult to accurately determine the percentage of area covered by these patterns when they were present. There was a tendency to allocate higher percentages of area to these striking patterns. It was stressed at the meeting that very careful analysis of the whole slide is mandatory. Some cases with solid patterns were classified as deciduoid on the basis only of the larger amount of cytoplasm even if pronounced eosinophilia or glassy cytoplasm was not observed. Interestingly, and not at all expected, one of the largest problems was defining and classifying the trabecular pattern. By definition, this pattern consists of small cells in thin cords or single files. However, when there was a slitlike space appearing in between 2 cell lines, sometimes wider and sometimes narrower in the same analyzed area, it was difficult not to refer to this pattern as a crushed acinar pattern. It was decided to classify this pattern as trabecular only when a narrow space was present or when there was no space between the cords of cells. Other problems were noted with the acinar pattern, which is defined as a pattern with glandlike structures. Namely, it was unclear when to identify small glandlike structures as glands, and when to refer to them as microcysts. As a result, the adenomatoid pattern was sometimes difficult to define. Two months after the consensus meeting, a second evaluation round yielded a higher j value of 0.63 (substantial agreement) for overall interobserver agreement, with improved reproducibility for each pair of pathologists (j values of 0.56, 0.64, and 0.70) (Figure 3). Significant improvement in interobserver agreement was detected for all major types of MPM (Table). For epithelioid type, j values in the first and second rounds were 0.50 and 0.74, respectively. These values were similar to those for sarcomatoid type (0.60 and 0.90). Interestingly, the concordance for the biphasic group showed the strongest improvement, with j increasing from 0.28 to 0.64 (fair to 748 Arch Pathol Lab Med Vol 142, June 2018 Reproducibility of Pleural Mesothelioma Subtyping Brcic et al

3 Figure 1. Representative images of selected histologic patterns. Solid (A), deciduoid (B), pleomorphic (C), micropapillary (D), acinar (E), trabecular (F), tubulopapillary (G), and signet ring cell (H) patterns are presented (hematoxylin-eosin, original magnifications 320 [A, B, C, D, E, F, and H] and 310 [G]). Arch Pathol Lab Med Vol 142, June 2018 Reproducibility of Pleural Mesothelioma Subtyping Brcic et al 749

4 Interobserver Agreement According to the Histologic Types and Subtypes of Malignant Pleural Mesothelioma After the First and Second Rounds of Slide Viewing Histologic Type Histologic Subtype First-Round j Value Second-Round j Value Epithelioid Acinar Adenomatoid Deciduoid Micropapillary Trabecular Tubulopapillary Pleomorphic Rhabdoid Signet ring cell Solid Sarcomatoid Biphasic Figures 2 and 3. General interobserver reproducibility. Interobserver reproducibility between pairs of pathologists (A, B, and C) after the first (Figure 2) and second (Figure 3) rounds of slide review demonstrate an increase from fair (0.34, 0.35, and 0.38) to moderate (0.56) and substantial (0.64 and 0.70) agreement. substantial agreement). In subanalysis of histologic subtypes after the first and second rounds, consensus regarding a diagnosis of micropapillary, deciduoid, and solid patterns showed the highest improvements. For the diagnosis of acinar subtype, however, interobserver agreement remained low, with j values of 0.29 and 0.32 (fair agreement). Analysis of intraobserver reproducibility between the 2 rounds revealed moderate agreement for all pathologists (j values of 0.56, 0.56, and 0.50), with the lowest value found for the least experienced pathologist. Overall interobserver agreement for the samples obtained by percutaneous pleural needle biopsy after the first round showed j value of 0.33 (fair agreement), which was the same as for the whole study group (j, 0.36). However, after the second round, in this small group, j value of 0.56 was reached (moderate agreement), which was poorer than 0.63 (substantial agreement) found for the whole study group. DISCUSSION Although many reports have been published regarding new prognostic markers for mesothelioma, their results are still controversial and histologic classification remains the most important prognostic factor. Histologic criteria for typing and subtyping of MPM have been developed 11,19 and are applied in daily diagnostic practice. Some relatively recent studies 8,9 have shown that survival significantly differs depending on specific histologic types of mesotheliomas, notably with poorer overall survival for sarcomatoid and biphasic types than for epithelioid type. As importantly, treatment decisions today are also based on histologic subtyping. 10 To use histologic subtyping, which is fast and inexpensive, as a clinical prognostic marker, the use of welldefined criteria and high reproducibility of the diagnoses are essential. Although immunohistochemistry is reliable when assessing epithelioid types of MPM, it can be less effective when used to classify sarcomatoid and biphasic MPM. Sarcomatoid MPM is defined as a malignant tumor composed of spindle and mesenchymal-appearing tumor cells of mesothelial origin. These tumors are almost always positive for cytokeratins and often lack detectable expression of mesothelial markers. 11 Biphasic MPM requires at least 10% of sarcomatoid and epithelioid patterns for diagnosis. Because of these difficulties with immunohistochemistry in sarcomatoid MPM, it might be difficult to differentiate reactive stroma of epithelioid MPM from biphasic MPM. In mesothelial proliferations, the combination of loss of BAP1 expression and p16 deletion, as determined by fluorescence in situ hybridization, has been shown to be a reliable marker of malignancy. However, the reverse staining pattern cannot rule out mesothelioma. 22 According to Righi et al, 23 BAP1 expression might also be useful for distinguishing biphasic MPM from epithelioid MPM with reactive stroma. The problems associated with differentiating the major MPM types on HE-stained slides only were also apparent in our study. This was reflected as fair (biphasic) and moderate (sarcomatoid and epithelioid) interobserver reproducibility after the first round. After the consensus meeting, selection of representative images, and application of strict morphologic criteria, the reproducibility was improved. Interobserver reproducibility was now moderate for biphasic, substantial for epithelioid, and excellent for sarcomatoid type. During the consensus meeting, representative slides with desmoplastic stromal reaction were compared to sarcomatoid and biphasic mesotheliomas. We decided to classify atypical and pleomorphic spindle cells, especially those with high cellularity, as sarcomatoid mesothelial cells. We are aware of the fact that this might also be only reactive stroma, but the above approach was the only possible solution in our pure morphologic approach. This may 750 Arch Pathol Lab Med Vol 142, June 2018 Reproducibility of Pleural Mesothelioma Subtyping Brcic et al

5 limit our study to a certain degree. However, the more important aim of this study was to assess the reproducibility of the patterns in the epithelioid type of MPM. Like adenocarcinoma of the lung, epithelioid MPM is histologically very heterogeneous. It is very often composed of more than 2 different growth patterns. As previously mentioned, there is prognostic importance in the further subtyping of epithelioid MPM ,24 This is even more important if histologic subtyping is reproducible, as it might also be correlated with radiologic, biological, and molecular/ genetic findings, similar to lung adenocarcinoma In the first round of this study, we found interobserver reproducibility of histopathologic subtyping of epithelioid malignant mesothelioma to be fair, which was lower than expected. This clearly reflects the difficulty with such diagnoses. At the consensus meeting, we carried out detailed analyses of problematic cases, discussed problems, and defined consensus. Furthermore, a set of representative images for each pattern was compiled and later used as a tool for decision support when deciding on MPM subtype. This educational effort led to appreciable improvement in the second round interobserver agreement, although after a rather short period of 2 months. Interobserver reproducibility increased appreciably, as reflected in higher j values (substantial agreement). Interestingly, the acinar pattern led to the most problems in both rounds. Based on our experience, the problem is not so much in the definition of the acinar pattern, but in the definition of the trabecular and adenomatoid patterns. When are glandlike structures to be considered microglandular structures? Should there be a slit in between the cords of cells defining the trabecular pattern? If yes, when does this slit become a glandlike structure ( crushed acinar pattern ) (Figure 4, A through C)? There are no answers to these questions in the available literature. We have demonstrated that the sample size is another limiting factor: in samples obtained by percutaneous pleural needle biopsy, overall interobserver agreement increased from fair (after the first round) to moderate agreement (after the second round), which was lower when compared to the whole study group (from fair to substantial agreement). The importance of this finding is even greater, since this difference was obtained in a rather small group of only 24 samples (which is at the same time the limiting factor of this subanalysis). Another limitation of our study may stem from the differences in the levels of experience among the pathologists. However, from the intraobserver reproducibility (comparison between the first and second round results for each pathologist), which showed moderate agreement for all, we conclude that this is not in fact a limitation. Even more, it was clearly shown that even the most experienced pathologist changed the manner in which he subtyped the slides in approximately the same number of cases as the other 2 pathologists after the consensus meeting. This also demonstrates the need for better definitions and clear criteria for specific patterns of epithelioid MPM. Earlier analogous approaches in breast carcinoma 28 and soft tissue sarcomas 29 have led to similar results. Even when experienced pathologists were questioned, histologic pattern reproducibility (in lung adenocarcinoma) proved to be problematic, but increased after education and training The German group also showed that the increase in reproducibility is different for different patterns of lung adenocarcinoma. 30 Similar findings have now been presented for MPM. To our knowlwdge, only 1 study has been Figure 4. Discrepant cases. Samples for which interobserver reproducibility was not achieved in the second round are displayed: trabecular or acinar pattern (A and B); signet ring cell, adenomatoid/ microglandular, or acinar pattern (C) (hematoxylin-eosin, original magnifications 35 [A], 320 [B], and 310 [C]). conducted regarding mesothelioma subtyping, wherein one of our authors (L.B.) participated. 12 However, the mesothelioma cohort was smaller, only 2 pathologists from a single institution participated, and intraobserver reproducibility was not addressed. CONCLUSIONS Our study demonstrated that moderate to substantial agreement in histologic typing and subtyping of MPM can be achieved. Size of the sample seems to be an important limiting factor in achieving this agreement. Furthermore, consensus-based illustrative images improve learning and help in achieving higher consensus in the classification of MPM. In spite of all this, even among experienced Arch Pathol Lab Med Vol 142, June 2018 Reproducibility of Pleural Mesothelioma Subtyping Brcic et al 751

6 pulmonary pathologists excellent agreement was not achieved for most of the histologic types/subtypes. This clearly points to the need for additional clarification of diagnostic criteria at least for some patterns, which would improve diagnostic reproducibility. References 1. The Surveillance, Epidemiology, and End Results (SEER) Program of the National Cancer Institute. SEER cancer statistics review , mesothelioma. mesothelioma.pdf. Accessed December 28, Robinson BW, Lake RA. Advances in malignant mesothelioma. N Engl J Med. 2005;353(15): Beebe-Dimmer JL, Fryzek JP, Yee CL, et al. Mesothelioma in the United States: a Surveillance, Epidemiology, and End Results (SEER)-Medicare investigation of treatment patterns and overall survival. Clin Epidemiol. 2016;8: Bille A, Belcher E, Raubenheimer H, et al. Induction chemotherapy, extrapleural pneumonectomy, and adjuvant radiotherapy for malignant pleural mesothelioma: experience of Guy s and St Thomas hospitals. Gen Thorac Cardiovasc Surg. 2012;60(5): Gill RR, Richards WG, Yeap BY, et al. Epithelial malignant pleural mesothelioma after extrapleural pneumonectomy: stratification of survival with CT-derived tumor volume. AJR Am J Roentgenol. 2012;198(2): Panou V, Vyberg M, Weinreich UM, Meristoudis C, Falkmer UG, Re OD. The established and future biomarkers of malignant pleural mesothelioma. Cancer Treat Rev. 2015;41(6): Hoda MA, Dong Y, Rozsas A, et al. Circulating activin A is a novel prognostic biomarker in malignant pleural mesothelioma a multi-institutional study. Eur J Cancer. 2016;63: Rusch VW, Giroux D, Kennedy C, et al. Initial analysis of the international association for the study of lung cancer mesothelioma database. J Thorac Oncol. 2012;7(11): Meyerhoff RR, Yang CF, Speicher PJ, et al. Impact of mesothelioma histologic subtype on outcomes in the Surveillance, Epidemiology, and End Results database. J Surg Res. 2015;196(1): NCCN Clinical Practice Guidelines in Oncology (NCCN Guidelinest). Malignant Pleural Mesothelioma, version July 7, nccn.org/professionals/physician_gls/f_guidelines.asp. Accessed July 9, Galateau-Salle F, Churg A, Roggli V, et al. Diffuse malignant mesothelioma. In: Travis WD, Brambilla E, Burke AP, Marx A, Nicholson AG, eds. WHO Classifiction of Tumours of the Lung, Pleura, Thymus and Heart. Lyon, France: International Agency for Research on Cancer; 2015: World Health Organization Clssification of Tumours; vol Brčić L, Jakopović M, Brčić I, et al. Reproducibility of histological subtyping of malignant pleural mesothelioma. Virchows Arch. 2014;465(6): Kadota K, Suzuki K, Sima CS, Rusch VW, Adusumilli PS, Travis WD. Pleomorphic epithelioid diffuse malignant pleural mesothelioma: a clinicopathological review and conceptual proposal to reclassify as biphasic or sarcomatoid mesothelioma. J Thorac Oncol. 2011;6(5): Ordóñez NG. Pleomorphic mesothelioma: report of 10 cases. Mod Pathol. 2012;25(7): Alchami FS, Attanoos RL, Bamber AR. Myxoid variant epithelioid pleural mesothelioma defines a favourable prognosis group: an analysis of 191 patients with pleural malignant mesothelioma. J Clin Pathol. 2017;70(2): Betta PG, Andrion A, Donna A, et al. Malignant mesothelioma of the pleura: the reproducibility of the immunohistological diagnosis. Pathol Res Pract. 1997;193(11 12): Andrion A, Magnani C, Betta PG, et al. Malignant mesothelioma of the pleura: interobserver variability. J Clin Pathol. 1995;48(9): Skov BG, Lauritzen AF, Hirsch FR, Skov T, Nielsen HW. Differentiation of adenocarcinoma of the lung and malignant mesothelioma: predictive value and reproducibility of immunoreactive antibodies. Histopathology. 1994;25(5): Husain AN, Colby T, Ordonez N, et al. Guidelines for pathologic diagnosis of malignant mesothelioma: 2012 update of the consensus statement from the International Mesothelioma Interest Group. Arch Pathol Lab Med. 2013;137(5): Allen TC. Recognition of histopathologic patterns of diffuse malignant mesothelioma in differential diagnosis of pleural biopsies. Arch Pathol Lab Med. 2005;129(11): Landis JR, Koch GG. The measurement of observer agreement for categorical data. Biometrics. 1977;33(1): Sheffield BS, Hwang HC, Lee AF, et al. BAP1 immunohistochemistry and p16 FISH to separate benign from malignant mesothelial proliferations. Am J Surg Pathol. 2015;39(7): Righi L, Duregon E, Vatrano S, et al. BRCA1-associated protein 1 (BAP1) immunohistochemical expression as a diagnostic tool in malignant pleural mesothelioma classification: a large retrospective study. J Thorac Oncol. 2016; 11(11): Galateau-Salle F, Churg A, Roggli V, Travis WD; World Health Organization Committee for Tumors of the Pleura. The 2015 World Health Organization Classification of Tumors of the Pleura: advances since the 2004 classification. J Thorac Oncol. 2016;11(2): Kadota K, Yeh YC, D Angelo SP, et al. Associations between mutations and histologic patterns of mucin in lung adenocarcinoma: invasive mucinous pattern and extracellular mucin are associated with KRAS mutation. Am J Surg Pathol. 2014;38(8): Choi ER, Lee HY, Jeong JY, et al. Quantitative image variables reflect the intratumoral pathologic heterogeneity of lung adenocarcinoma. Oncotarget. 2016;7(41): Russell PA, Barnett SA, Walkiewicz M, et al. Correlation of mutation status and survival with predominant histologic subtype according to the new IASLC/ ATS/ERS lung adenocarcinoma classification in stage III (N2) patients. J Thorac Oncol. 2013;8(4): Robbins P, Pinder S, de Klerk N, et al. Histological grading of breast carcinomas: a study of interobserver agreement. Hum Pathol. 1995;26(8): Coindre JM, Trojani M, Contesso G, et al. Reproducibility of a histopathologic grading system for adult soft tissue sarcoma. Cancer. 1986; 58(2): Warth A, Cortis J, Fink L, et al. Training increases concordance in classifying pulmonary adenocarcinomas according to the novel IASLC/ATS/ERS classification. Virchows Arch. 2012;461(2): Warth A, Stenzinger A, von Brünneck AC, et al. Interobserver variability in the application of the novel IASLC/ATS/ERS classification for pulmonary adenocarcinomas. Eur Respir J. 2012;40(5): Wang C, Durra HY, Huang Y, Manucha V. Interobserver reproducibility study of the histological patterns of primary lung adenocarcinoma with emphasis on a more complex glandular pattern distinct from the typical acinar pattern. Int J Surg Pathol. 2014;22(2): Thunnissen E, Beasley MB, Borczuk AC, et al. Reproducibility of histopathological subtypes and invasion in pulmonary adenocarcinoma: an international interobserver study. Mod Pathol. 2012;25(12): Arch Pathol Lab Med Vol 142, June 2018 Reproducibility of Pleural Mesothelioma Subtyping Brcic et al