Dataset for Reporting of Malignant Mesothelioma of the Pleura or Peritoneum: Recommendations From the International Collaboration on Cancer Reporting (ICCR).

Evidence based structured (synoptic) pathology reports with standardized definitions for each component have been shown to significantly enhance the completeness and quality of data provided to clinicians. 1 – 4 Over the last 2 decades, both the United Kingdom, 5 and United States 6 and others 7 have produced standardized cancer reporting protocols or “datasets” for national use. Other minimum or comprehensive datasets have been developed around the world for a reporting of a variety of cancers on an individual or institutional basis. In 2011, the International Collaboration on Cancer Reporting (ICCR) was convened with a view to reducing the global burden of cancer dataset development and reduplication of effort by the different international organisations engaged in the development of standardized cancer reporting datasets. ICCR datasets are made freely available for use by organisations globally and it is anticipated that in time this will enable the alignment and normalization of pathology cancer data around the world as producers of datasets adopt and incorporate the ICCR datasets. To date, the ICCR collaboration has successfully published five datasets on cancers of the prostate, endometrium, lung and ovarian/fallopian tube/primary peritoneal site, and malignant melanoma. All of the datasets are evidence-based, have been produced by a panel of internationally renowned experts, have been subject to international open consultation, and are freely available for worldwide use at the following website: www.iccr-cancer.org (accessed February 19, 2016). The process of production of each of these datasets has been published in peer-reviewed journals. 8 – 12 ICCR datasets are designed to be as concise as possible so as to encourage uptake, facilitate future translation, limit the burden on reporting pathologists, and avoid jurisdictional pitfalls through the exclusion of non-essential information. Given the interdependence of cancer datasets with the World Health Organization (WHO) Classification of Tumors, the ICCR has undertaken to develop datasets in synchrony with the update of the WHO classifications. In 2015, WHO released the 4th edition of its Classification of Tumors of the Lung, Pleura, Thymus and Heart. 13 As a result the ICCR commenced development of a series of thoracic datasets including a dataset for malignant mesothelioma.

Standardized cancer reporting protocols or “datasets” have been produced for national use for many years. 5 , 6 , 7 Perforce these datasets use the same peer-reviewed evidence as the basis of these protocols, however, each is constructed differently, uses different terminology and similar elements may be based on different methodologies. The adoption of a single international reporting standard avoids duplication of cancer pathology dataset development in many different jurisdictions, which reduces the burden on each country and improves opportunities for interoperability, international comparison and research. The ICCR conducted a pilot project in 2011 to standardize the initial four cancer datasets for lung, melanoma, prostate (radical prostatectomy), and endometrial carcinoma. By using different development processes for this initial collaboration, the ICCR has been able to optimize the development method for all further dataset efforts, as documented in the Guidelines for the Development of ICCR Datasets ( http://www.iccr-cancer.org/datasets/dataset-development ; accessed February 19, 2016) For the construction of the mesothelioma dataset, a Chair was selected by the ICCR Dataset Steering Committee and a Dataset Authoring Committee (DAC) composed of pathologists with expertise in mesothelioma and one clinician was then empaneled. The DAC also included an ICCR representative whose role is to provide guidance and support to the Chair of the DAC regarding ICCR standards and committee participation and to undertake a quality assurance role within the committee. A Project Manager was also appointed to streamline and standardize the dataset development process, reduce individual pathologist’s time and effort, expedite the development timeline and ensure implementation of, and adherence to, ICCR standards. The dataset is composed of two types of elements: Required and Recommended. REQUIRED elements are defined as those which are unanimously agreed by the panel to be essential for the histological diagnosis, clinical management, staging or prognosis of mesothelioma. RECOMMENDED elements are non-mandatory and defined as clinically important and recommended as good practice and should ideally be included in the report, but which are not yet validated or regularly used in patient management. Evidentiary support at Level III-2 or above (based on prognostic factors in the National Health and Medical Research Council (NHMRC) levels of evidence document and defined as “Analysis of prognostic factors amongst persons in a single arm of a randomized controlled trial”) 14 is required to support required (mandatory) elements. Rarely, where Level III-2 evidence is not available, an element can be categorized as required with unanimous agreement of the expert panel. Required elements are mandatory and the sum of these is the minimum information which should be included on the pathology report. Commentary, i.e. explanatory text, diagrams or tables, is added where necessary to clarify the elements: to define the way an item should be reported, to ensure clarity and conformity; to explain why an item is included (e.g. how does the item assist with clinical management or prognosis of the specific cancer); to cite published evidence in support of the element and to state any exceptions or issues that may be encountered by the reporting pathologist. Commentary is designed to provide contextual guidance to the reporting pathologist. As a starting point, a search for all published mesothelioma datasets was undertaken. This scan included review of datasets from existing ICCR members and also of datasets, protocols or checklists published in review articles or other international websites. This information formed the foundation for a comparative review, from which elements that are mandatory/required/core in any one or more of these datasets was extracted for consideration by the DAC along with all responses and commentary. A proposed dataset was developed and circulated to the DAC for their initial thoughts. This initial feedback from the group enabled the chair to determine areas of concordance and dissent prior to a series of web/teleconferences with the DAC to discuss the proposed dataset. Following agreement of the draft datasets by the DAC, the dataset was posted to the ICCR website for a period of 2 months for public comment. All feedback was reviewed and final edits made.

The initial scope of the dataset was for resection specimens for malignant mesothelioma of the pleura. However on consideration, this was expanded to include peritoneal as well as pleural lesions, as peritoneal mesothelioma had more in common with pleural mesothelioma, with respect to behaviour and treatment, than with gynaecological malignancies. After the initial feedback from the DAC, a proposal to expand the scope to include biopsy specimens was discussed and agreed. A detailed review of biopsy for mesothelioma has recently been published. 15 The required elements are shown in Table 1 Documentation of the operative procedure is useful, as correlation of the type of procedure with the material received can be important for patient safety. In resection specimens, the type of surgical procedure is important in determining the assessment of surgical margins. Due to advanced age, clinical status, or extent of disease, few mesothelioma patients are suitable for extrapleural pneumonectomy (EPP) or radical pleurectomy/decortication (P/D) and therefore, diagnosis is usually based upon biopsy alone. Although the volume of tissue sampled is more restricted than for surgical resection specimens, biopsy assessment may contribute significant observations for clinical management and prognosis, in addition to the crucial distinction between secondary tumors affecting the serosal membranes and mesothelioma, and between mesothelioma and benign reactive mesothelial proliferations. The type of biopsy is important as it affects the extent to which a diagnosis may be made with any certainty. Accurate typing of mesothelioma 16 – 19 has been shown to vary by procedure - 83% for open biopsy in comparison to 74% for video-assisted thoracoscopic surgery (VATS) biopsy, and 44% for computed tomography (CT)-guided biopsy, when compared with the subtype assessed in a follow-up series of 83 extrapleural pneumonectomy specimens. 19 Specimen type varies according to the type of operation, and while the nature of the specimens submitted for pathological assessment may be deduced from the procedure, specifying the nature of specimen received provides complementary information and confirmation that entire organ/s have been resected and submitted. The macroscopic site of the tumor is an important component for pathologic staging. The major histological tumor types of malignant mesothelioma as recognized by the WHO classification (4 th edition) 13 are epithelioid, sarcomatoid and biphasic/mixed. By convention a biphasic mesothelioma is diagnosed if the lesser component reaches 10% of the tumor examined. There are a number of histological patterns of malignant mesothelioma which are important to be aware of primarily because of diagnostic confusion. For epithelioid mesothelioma these include common patterns such as solid, tubulopapillary, and trabecular, also less common forms such as micropapillary, adenomatoid (microcystic), clear cell, transitional, deciduoid, small cell and pleomorphic mesothelioma. It should be noted that, at present, there is no uniformity among pathologists for the definition of many of these patterns nor any clear prognostic significance to most of them, and we do not recommend these names be included as part of a diagnosis; their importance lies in recognition by the pathologist that these are patterns seen in mesotheliomas. For sarcomatoid mesothelioma these histological variants may comprise heterologous (osteosarcomatous, chondrosarcomatous and rhabdomyosarcomatous) elements, and desmoplastic mesothelioma. Desmoplastic mesothelioma is characterized by atypical spindle cells and dense hyalinised fibrous stroma, the latter comprising at least 50% of the tumor. 20 The conventional immunohistochemical panel of markers may require modification with some of these patterns to prevent misdiagnosis. Some of these patterns may have prognostic significance; however, until these prognostic patterns are clearly defined and accepted, the current recommendation is to diagnose mesotheliomas as epithelioid, sarcomatoid/desmoplastic, or biphasic/mixed, particularly since radical surgical approaches depend on these general classifications. In some cases, such as small biopsy specimens, a definitive tumor type cannot be assigned and in this situation a value of “mesothelioma not otherwise specified (NOS)” would be used. In extrapleural pneumonectomy specimens the bronchial resection margin status is evaluated by intraoperative frozen section examination. In the surgical pathology specimen, the soft tissue margin status is difficult to assess because the entire pleura represents a margin. Usually in patients with extrapleural pneumonectomy, the surgeon is performing a blind dissection beneath the endothoracic fascia between the pleura and chest wall. Extent of invasion is part of staging for radical pleural surgical specimens. In biopsies the presence of invasion is the most important parameter for separating benign from malignant mesothelial proliferations. Invasion into the endothoracic fascia is a staging parameter and should be determined only by the surgeon or radiologist, since there are no characteristic pathological features appreciable by gross or microscopic examination. The endothoracic fascia represents a connective tissue plane that lies between the parietal pleura and the innermost intercostal muscle. Its histology is not well defined. Sections from parietal pleura that appose the chest wall showing histologic involvement of skeletal muscle is the best surrogate indicator that the endothoracic fascia has been breached. Thoracic or abdominal lymph nodes may be sampled to obtain a diagnosis or for the staging of an already diagnosed tumor. If thoracic, they should be identified by standard station; for abdominal lymph nodes, a suitable specimen identifier or descriptor should be used. A lymph node station should be regarded as positive for mesothelioma regardless of the number of malignant mesothelial cells present or the number of lymph nodes involved provided one node contains malignant mesothelial cells. However, the identification of mesothelial cells in lymph nodes does not necessarily indicate metastasis. Rarely may they represent incidental benign inclusions. 21 , 22 The diagnosis of metastatic mesothelioma should only be made when there is good evidence of a serosa based tumor whether diffuse or, very rarely, localized. Staging by TNM 7 th edition is one of the most important prognostic factors. The recommended elements are shown in Table 2 . Clinical information is essential to proper processing and evaluation of pathological specimens as it can influence pre-test probability of a particular diagnosis. This allows the pathology laboratory to accurately triage processing, including extent of sampling. It also informs the pathologist as to decisions ultimately influencing the number of slides to be examined (serial sections, levels) and potential ancillary studies to be performed 1 , thus avoiding error. For malignant mesothelioma, the radiologic growth pattern and history of previous cancer are important guides to further analysis of a particular specimen. A radiologic nodular growth pattern may prompt correlation with surgical thoracoscopic observations with regard to nodule sampling, while a diffuse growth pattern may lead to a request for deeper or more extensive samples. History of prior cancer could suggest a different panel of immunohistochemical stains to definitively rule out metastasis from a known tumor. A cancer history can prompt a request to review prior outside material or to review an archival in house slide record. 23 Other valuable clinical information includes presence of a pleural effusion and its characteristics (e.g. transudative, bloody, exudative); this can trigger review of and correlation with a concurrent cytological specimen. A history of asbestos exposure is not relevant for the diagnosis of samples in which malignant mesothelioma is a consideration, as this history does not influence sample processing or ultimate diagnosis. 20 A history of neoadjuvant therapy is important in the pathology analysis. Assessment of residual tumor, including nodal status, is critical to staging and prognostication in the neoadjuvant setting. 24 , 25 For pleural mesotheliomas that are received as radical surgical (EPP or P/D) specimens, attempting to measure the dimensions of individual tumor nodules is neither simple (because the distinction between tumor and fibrotic reaction may be difficult to assess) nor informative. Rather, measuring the maximum thickness of tumor appears to be a more useful indicator of tumor burden and can often be compared to radiologic measurements. 6 For peritoneal mesotheliomas, the specimen is normally received in multiple parts and dimensions of the dominant mass should be measured. Where multiple nodules are present, the dimensions of the largest nodule should be recorded. The origin/designation of all tissue blocks should be recorded. This information should be documented in the final pathology report and is particularly important should the need for internal or external review arise. The reviewer needs to be clear about the origin of each block in order to provide an informed specialist opinion. Recording the origin/designation of tissue blocks also facilitates retrieval of blocks for further immunohistochemical or molecular analysis, research studies or clinical trials. In pleural malignant mesothelioma, mitotic count has not been definitively established as an independent parameter in the diagnostic setting or as a determinant of prognosis. However among epithelioid peritoneal malignant mesothelioma, increased mitotic count (greater than 4 in 10 HPF a ) 26 was reported as a poor prognostic indicator, and, more recently, was validated in a multi-observer study of an independent group of patients 27 , establishing a lower cut-off of 5 mitoses in 50 HPF. Ki-67 fraction may also have prognostic significance, but its use as an adjunct to mitotic count has not been investigated. There is no recommended or agreed system for tumor regression grading of mesothelioma that has been treated with neoadjuvant therapy, however a general indication of residual viable tumor 50%, may be useful. It is recommended that pathologists comment upon any coexistent non-neoplastic findings present in the submitted materials. These include, for extrapleural pneumonectomy specimens, such findings as emphysema, small airways disease, respiratory bronchiolitis, asbestosis, asbestos bodies, talc granulomas and pleural plaques. 28 For diagnosing asbestosis, it is recommended that the criteria published by the asbestosis committee of the College of American Pathologists and Pulmonary Pathology Society be used. 29 For peritoneal resection specimens, additional findings such as endometriosis, endosalpingiosis and mesothelial inclusion cysts should be noted. The three most common molecular alterations in malignant mesothelioma are loss of neurofibromin 2 (Merlin, NF2), cyclin-dependent kinase inhibitor 2A (CDKN2A, p16), and BRCA1 associated protein-1 (BAP1). While to date NF2 loss has not been exploited diagnostically, p16 Fluorescence in situ hybridization (FISH) and BAP1 appear to be useful markers for separating benign from malignant mesothelial proliferations. 30 Thus far both these markers have been reported as only lost in malignant mesotheliomas when strict cut-offs are applied. One outcome of the strict cut-off is the major problem of low sensitivity. Overall, studies reporting loss of p16 by FISH in mesotheliomas show a sensitivity around 50%, albeit significantly higher in pleural (67%) than peritoneal mesothelioma (25%). 30 Loss of p16 by FISH in pleural mesothelioma is correlated with adverse survival. 31 , 32 Retention of p16 by immunohistochemistry is a useful prognostic indicator in peritoneal epithelioid malignant mesothelioma, with a significantly prolonged survival in that group. 26 The sensitivity for loss of nuclear expression of BAP1 is not well defined but probably on the order of 50 to 70% for epithelioid mesotheliomas, and very low for sarcomatoid mesotheliomas. 30 But these markers are only useful when lost; positive staining does not rule out a mesothelioma. BAP1 immunohistochemistry in addition is useful as a screening tool for BAP1 germline mutation syndromes, in which there are familial aggregations of mesotheliomas, melanomas including ocular melanomas, renal cell carcinomas, and probably a variety of other tumors. 33 Interestingly, patients with BAP1 germline mutation mesotheliomas are reported to have dramatically better survival rates. 34 However, BAP1 immunohistochemistry is no more than a screening tool in this context, since the vast majority of mesotheliomas that show BAP1 loss only have somatic mutations, and formal genetic analysis is required to confirm germline tumors. Positive immunohistochemistry for EMA b , Glut1 c , IMP3 d and CD e 146 have all been proposed as single markers for malignant mesothelioma when compared to benign proliferations. 30 Since small but significant proportions of benign proliferations are positive for each of these markers, combinations of markers have been proposed, but the correlations are weak. 35 – 38 Therefore in the absence of morphologic invasion (cytology, small biopsy, or cellular atypia alone) these markers should not be relied upon as the sole determinant of malignancy.

A series of web/conference calls were undertaken by the DAC to discuss each of the elements in the proposed dataset for the pathological reporting of mesothelioma. Points of consideration included the following: (1) The publication of the 4th edition of the WHO Classification of Tumours of the Lung, Pleura, Thymus and Heart 13 ; (2) Variation in local practice, both surgical and pathological, amongst the international DAC; (3) The value of biopsy for the diagnosis of mesothelioma; (4) Variation in terminology in use for element names and responses; (5) Whether localized mesothelioma should or should not be included in the scope of the dataset. Localized malignant mesotheliomas are extremely uncommon tumors with the microscopic appearances of diffuse malignant mesotheliomas, but are solitary, and by definition do not show gross or microscopic evidence of diffuse pleural spread. Roughly half of localized malignant mesotheliomas are curable by wide surgical excision. 39 It was agreed that localized malignant mesotheliomas are sufficiently different from ordinary diffuse malignant mesotheliomas that there was little value in including these lesions in the scope of this publication. Each element and its value list or response e.g. “present”, “not assessable”, “not indicated”, was discussed and agreed by the DAC. Standardized terminology with definitions for common terms is used to avoid any ambiguity in the assessment or meaning of the element. The DAC was then assigned tasks for the writing of commentary for those elements requiring further explanatory text taking into account recent and pertinent literature. The goal of the ICCR is to develop a set of data elements which will form the core of any pathology report on the specific cancer around the world. Adoption of the ICCR datasets will help facilitate recording of pathological data in a consistent standardized way. The ICCR datasets need to include any element that is essential to include when reporting a cancer and which has either supporting evidence or unanimous agreement from the DAC to include as best practice. This dataset is therefore the core of any structured pathology report on mesothelioma. It is not, however, the intention that the dataset be restrictive and additional data items may be included when reporting, according to local needs. All structured pathology reports must also include the facility for free text comments to insure that any clarification or nuance in reporting is captured.