Effusion cytology of malignant mesothelioma enables earlier diagnosis and recognizes patients with better prognosis

A conclusive diagnosis of malignant mesothelioma (MM) can be based on effusion cytology using the guidelines for the cytopathologic diagnosis of epithelioid and mixed‐type MM. Briefly, the diagnosis is obtained when the mesothelial phenotype of malignant cells is established by ancillary techniques. This study is based on the comparison of the overall survival rates of patients with MM when diagnosed by effusion cytology, histopathology, or a combination of both. A total of 144 patients were diagnosed with epithelioid and mixed‐type pleural MM at Karolinska University Hospital between 2004 and 2013. The diagnosis was obtained by histopathology in 74 cases and by cytological examination of pleural effusion in 70 cases. In 29 of the latter cases, a diagnostic biopsy was obtained simultaneously. A total of 104 patients received chemotherapy. All diagnoses were supported by clinical findings, including computer tomography scans. The median time between first symptoms and diagnosis was similar for cytology and histopathology. However, a delay of more than 6 months after first symptoms was seen in many patients in the histopathology group, resulting in late onset of treatment. The overall survival and proportion of long‐term survival were significantly better for cases diagnosed by cytology. Similarly, a better survival, following a cytological diagnosis, was also seen in patients who were only provided the best supportive care. Accurate cytological diagnosis enables conclusive diagnosis of MM. Our finding enables the initiation of treatment as soon as the cytological diagnosis is established, avoiding further delay and deterioration of patient survival and possibilities for treatment.

exposure. 6,7 Treatment options are, however, limited because the diagnosis is typically established late during the course of the disease. The first symptom of the disease is often accumulation of fluid in the pleura, so called effusion, which is withdrawn for therapeutic reasons allowing the expansion of the lungs, at the same time becoming the earliest available biological material for diagnosis.
Studies have shown that early treatment improves patient survival, 8 thus a diagnosis should be reached as quickly as possible, avoiding unnecessary delays.
Traditionally, the diagnosis of MM often depends on histopathology, as has been the gold standard for decades. 9,10 Recent audits, however, showed that effusion cytology provides an accurate option for this diagnosis, with a positive predictive value of 100% and a sensitivity of approximately 70%. 11,12 International guidelines for the cytopathologic diagnosis of epithelioid and mixed-type MM have Once malignancy is established, ancillary analyses will define the tumor phenotype. Diagnostic criteria for immunohisto-and cytochemistry are in fact the same here, considering that some of the stains work better on formalin-fixed and paraffin-imbedded tissue than on the alcohol-fixed exfoliative material and vice versa. Cells of mesothelial origin can be demonstrated by calretinin, mesothelin, HBME-1, podoplanin (D2-40), and WT-1, while the diagnosis can be excluded by CEA, BerEp-4, MOC-31, TTF-1, and Pax8. BerEp-4 and MOC-31 label the same epithelial cell adhesion molecule complex, the former performing somewhat better in nonfixed material and the latter following formalin fixation. Both histology and cytology guidelines state that two markers in favor of mesothelioma and two excluding the diagnosis are sufficient. With disputable staining results, however, additional antibodies must be used. These reactions can also be complemented by analyzing soluble markers such as hyaluronan and mesothelin. 14 In case there is still a suspicion for mesothelioma, electron microscopy can be helpful, providing that the sample contains cell groups and that cells have been fixed properly early in the process.
The aim of the present investigation was to evaluate the clinical outcome of patients with epithelioid and mixed-type MPM, comparing patients with the first conclusive diagnosis obtained by effusion cytology and those with the first conclusive diagnosis obtained by histopathology, including the possible influence of MM phenotype, gender, age, known asbestos exposure, signs of inflammatory response, biomarker levels, and time from first sign of disease to diagnosis and treatment.  (Table 1). In many of these cases, the thoracentesis and biopsy sampling were performed in one session, and for a few others, histology was a follow-up due to reluctance as to the reliability of the

| Pathology
Morphological reassessment verified that all diagnoses met postulated diagnostic criteria. 13,15 All histological diagnoses were based on biopsies obtained during thoracoscopy or by transthoracic needle sampling. The morphology was supported in all cases by the routine immunohistochemistry available during this period as described in published guidelines, principally using the same antibodies as recommended for the histopathological diagnosis. Immunocytochemistry on effusions were in all cases performed cytospin preparations without formalin fixation. Thus, among cytological findings supporting malignancy, there is richness in papillary groups with reactivity to epithelial membrane antigen (EMA) but not desmin ( Figure 2). 16 As described in the guidelines, reactions with desmin perform somewhat better that the more recently recommended BAP-1 when performed on nonfixed cells, 13 while the latter is to prefer formalin-fixed cell blocks. The mesothelial phenotype is shown by the presence of calretinin and mesothelin with simultaneous absence of BerEp4 and CEA ( Figure 3). The cytological diagnoses were based on cells recovered from effusions. The cytomorphology was supported by immunocytochemical demonstration of at least four immunomarkers as recommended by the literature. 13,17 Immunochemical demonstration of additional biomarkers, hyaluronan and mesothelin, 14 was performed in 28 cases. The diagnosis was further supported by electron microscopy of the cell pellet in 34 cases. Ploidy analysis by FISH (UroVysion, Abbot) was used to support the malignant nature of the cytological specimen in 14 cases. 18 The cytological diagnoses were in accordance with radiological findings and clinical course in all cases. The performance of the employed diagnostic routines has been described, 12 showing similar sensitivity and positive predictive values as shown in other experienced laboratories. 11 An abundance of inflammatory cells in an effusion may not only hinder a cytological diagnosis by masking the diagnostic tumor cells but also influence tumor proliferation, potentially influencing the clinical outcome. To evaluate this, the proportion of such cells in 32 nondiagnostic effusions available from the histology-only group were re-evaluated by two cytopathologists (SAO and AH) studying the possible influence on prognosis.

| RESULTS
Based on the average time lag between the first sign of disease and diagnosis, a histological diagnosis was obtained 30 days later than a cytological diagnosis (Table 1; P = .001). This time varied considerably in the histology group, as the diagnosis was obtained after more than MM, the positive predictive value of cytology was 100%, that is, there was no false-positive diagnosis during the 10 years studied. 12 However, the earlier cytological diagnosis did not correlate with earlier treatment in this material. During the first 6 months, the overall survival after diagnosis for patients first diagnosed by effusion cytology was slightly prolonged compared with those diagnosed by histopathology, corresponding to the natural course with an earlier diagnosis ( Figure 4A).
However, the Kaplan-Maier plots deviated significantly after 10 to 12 months. The overall survival was significantly better in patients in whom the diagnosis was obtained with effusion cytology compared with those in whom the diagnosis was based on histopathology alone, with median survival values of 20 months and 14 months, respectively (P = .02).
The 3-year survival was 31% and 11% for the two groups, respectively, and the corresponding 4-year survival was 19% and 5%, respectively.
Better overall survival in the cytology group was also found in patients not receiving chemotherapy (P = .002; Figure   As presented earlier, chemotherapy was given not only earlier, but also slightly more often, to patients diagnosed histologically  F I G U R E 6 A better prognosis in cases recognized by cytology is also seen compared with cases of purely epithelioid morphology on histology; the difference, however, is not being statistically significant due to lower numbers [Color figure can be viewed at wileyonlinelibrary.com] survival and could not explain improved survival rates associated with cytologically obtained diagnoses. Biomarker levels (hyaluronan and mesothelin) 11 were measured mainly in the cytology group; only 2 cases in the histology group had an effusion in which these parameters were analyzed. The registered parameters did not differ significantly between the two methods of diagnosing mesothelioma.
Cox regression (Table 2) showed that the method of diagnosis (P = .034) and sex (P = .040) were significant factors in survival after diagnosis. Thus, patients with cytologically recognizable tumors had better overall survival. Although the cytological diagnosis was obtained 1 month earlier, the difference in median survival after onset of therapy was still 6 months (19 and 13 months for cytology and histology, respectively, P = .055).

| DISCUSSION
The diagnosis of MM has traditionally required histological material. 15,20 However, this implies invasive approaches, such as core needle biopsy or thoracoscopy. The cytological examination of an effusion has also proven to be diagnostic in a substantial proportion of cases. 11,21,22 It is sometimes claimed that cytology is unable to demonstrate invasiveness and therefore insufficient for a conclusive diagnosis of MM. 10 Cytology, however, can recognize malignant conditions based on other criteria such as homozygous deletion of the 9p21 band, biomarker analyses, and electron microscopy. This reliability has previously been repeatedly shown. 11,12 The presence of an effusion is often the first manifestation of this the different outcomes between cytologically and histologically diagnosed tumors, that is, there was no indication that the difference in survival depended on the differences in proportions of sarcomatoid differentiation.
Cytologically recognized MMs often show papillary differentiation. This phenotype is, however, considered more indolent, 26-28 and could not explain the improved survival. In a recent study, integrated molecular pathology could demonstrate that six clusters of genetic aberration in MM showed distinct survival outcomes. 29 Whether any of these subgroups correlate to exfoliation of recognizable mesothelioma cells and better outcome in the present material remains to be shown. It should be noted that none of the included cases were diagnosed with specific entity "Well Differentiated Pleural Mesothelioma" (WDPM). The better outcome could not thus be correlated to morphological or molecular phenotype. A slightly better survival of female patients has also been shown previously 30,31 and may relate to certain differentiated MM types that are less associated with asbestos exposure and, thus, relatively more common in women.
Taken together, the results show that effusion cytology identifies a large subgroup of mesothelioma patients with better prognosis that often respond to chemotherapy. Explanations for this better outcome may correlate to the differentiation of the tumor, as well as to its ability to exfoliate and be recognized, and the molecular background is still not known. As this effusion is often an early sign of the disease, it enables an earlier diagnosis. The diagnostic procedure presented here followed the international guidelines, and the positive predictive value obtained for the cytological diagnosis is sufficient to motivate treatment without further delay. In case the earlier diagnosis obtained by effusion cytology also results in earlier onset of treatment; further, improved survival would be expected. 8 ORCID Sulaf Abd Own https://orcid.org/0000-0001-5431-1232