On the diagnosis of malignant pleural mesothelioma: a necropsy-based study of 171 cases (1997–2016)

Dear Editor:

Pira et al. discussed the potential of immunohistochemical (IHC) phenotypization for the differential diagnosis between malignant mesothelioma (MM) and other entities. We welcome their effort as a reasonable synthesis of previous reviews.^1–4 It is unclear to us, however, what relevance their remarks have to our work. Out of 171 autopsy cases considered in our article, 126 received biopsies and 91 also underwent IHC phenotypization: a large part of the cases examined with IHC had a selection of staining consistent with, or very close to, current standards (Table 1).

OPEN IN VIEWER

Table 1.

Distribution of staining among cases with histologic and immunohistochemical assessment of biopsy samples, 1997–2016, from Barbieri et al. ⁵

Status	Number of cases	%
At least 2 markers of mesothelioma and 2 of carcinoma or other neoplasms	60	65.9
At least 2 markers of mesothelioma and 1 of carcinoma or other neoplasms	23	25.3
1 Marker of mesothelioma and at least 2 of carcinoma or other neoplasms	1	1.1
At least 2 markers of mesothelioma but 0 of carcinoma or other neoplasms	4	4.4
Immunophenotypization carried out but detailed results unavailable	3	3.3
Total number of cases with immunophenotypization	91	100.0

Mesothelioma markers: calretinin, CD 10, CK AE1-AE3, CK 5, CK 7, CK MNF, desmin, mesothelin, p53 pan CK, podoplanin, vimentin, WT1. Markers of nonmesothelial neoplasms: actin, B 72.3, Ber EP4, CAM 5.2, CD 15, CD 31, CD 34, CD 56, CD 5, CD 68, CD 99, CD X2, CEA, CK 8, CK 18, CK 8-18, CK 20, c-KIT, EMA, factor VIII, HMB-45, HMFG1, MAC387, anti-melanomaoligoclonal antibodies, miogenin, p63, S100, TTF-1.

Based on the results of our observations, we do not consider immunophenotypization to be a strong requirement for diagnosis confirmation. We are not alone. For instance, the 2012 International Mesothelioma Interest Group (IMIG) recommendations stated that “A definitive diagnosis of malignant mesothelioma requires a workup including immunohistochemistry” and suggested including 2 positive and 2 negative markers for the initial workup, but recognized also that “Most MMs are readily identified or strongly suspected on routine hematoxylin-eosin staining” and that results from the pathology laboratory must be put “in the context of appropriate clinical, radiologic and surgical findings”. ⁶ The 2014 Helsinki consensus recommended the use of 2 positive and 2 negative IHC markers in the diagnosis of epithelioid MM, but what it considered “critical” for the diagnosis was “the clinical correlation with the gross distribution of the tumor.” ¹ The 2015 Italian consensus affirmed that “IHC is the most helpful ancillary technique in integrating the diagnosis of MPM” and that it “plays a limited role for the diagnosis of sarcomatoid MPM.” ² The 2017 revision of IMIG recommendations confirmed the 2013, above-mentioned statements. ³ They were unpublished when we submitted our manuscript and we could not include them in our discussion.

It appears from such experts’ opinions that IHC staining does not determine the diagnosis by itself but must be put first in the context of other pathologic findings and then in that of all other clinical and imaging findings. If anything, central to MM diagnosis is the distribution of the neoplasm, and it is difficult to see which technique could be better than autopsy for such a purpose. We were not, therefore, surprised that the 2018 British Thoracic Society guidelines, while confirming the recommendations to carry out IHC staining and to use a panel of 2 positive plus 2 negative markers, assessed both as grade D recommendations, i.e., those based on the lowest levels of evidence, such as case reports and case series, or expert opinion. ⁴

Our results were not obtained through experimental design and assessment at necropsy was not blind to the clinical and pathologic history of cases. Our report summarized the results of a consecutive case series, observed during a long period of real-life clinical and pathologic practice, which requires pathologists carrying out necropsies not to blind. As such it entails limitations (lack of some methodologic rigor) but bears also advantages (showing real-life results).

Finally, the context in which our study originated—a criminal trial for manslaughter, where doctor Barbieri served as expert witness for the public prosecution office—was stated as a potential source of conflicting interest in our article. ⁵