Upsala Journal of Medical Sciences (Jan 2020)

The value of desmosomal plaque-related markers to distinguish squamous cell carcinoma and adenocarcinoma of the lung

  • Inmaculada Galindo,
  • Mercedes Gómez-Morales,
  • Inés Díaz-Cano,
  • Álvaro Andrades,
  • Mercedes Caba-Molina,
  • María Teresa Miranda-León,
  • Pedro Pablo Medina,
  • Joel Martín-Padron,
  • María Esther Fárez-Vidal

DOI
https://doi.org/10.1080/03009734.2019.1692101
Journal volume & issue
Vol. 125, no. 1
pp. 19 – 29

Abstract

Read online

Background: An antibody panel is needed to definitively differentiate between adenocarcinoma (AC) and squamous cell carcinoma (SCC) in order to meet more stringent requirements for the histologic classification of lung cancers. Staining of desmosomal plaque-related proteins may be useful in the diagnosis of lung SCC. Materials and methods: We compared the usefulness of six conventional (CK5/6, p40, p63, CK7, TTF1, and Napsin A) and three novel (PKP1, KRT15, and DSG3) markers to distinguish between lung SCC and AC in 85 small biopsy specimens (41 ACs and 44 SCCs). Correlations were examined between expression of the markers and patients’ histologic and clinical data. Results: The specificity for SCC of membrane staining for PKP1, KRT15, and DSG3 was 97.4%, 94.6%, and 100%, respectively, and it was 100% when the markers were used together and in combination with the conventional markers (AUCs of 0.7619 for Panel 1 SCC, 0.7375 for Panel 2 SCC, 0.8552 for Panel 1 AC, and 0.8088 for Panel 2 AC). In a stepwise multivariate logistic regression model, the combination of CK5/6, p63, and PKP1 in membrane was the optimal panel to differentiate between SCC and AC, with a percentage correct classification of 96.2% overall (94.6% of ACs and 97.6% of SCCs). PKP1 and DSG3 are related to the prognosis. Conclusions: PKP1, KRT15, and DSG3 are highly specific for SCC, but they were more useful to differentiate between SCC and AC when used together and in combination with conventional markers. PKP1 and DSG3 expressions may have prognostic value.

Keywords