Plasmacytoid dendritic cells in keratoacanthoma and squamous cell carcinoma: A blinded study of CD123 as a diagnostic marker

Abstract

Background: Histopathologic distinction between keratoacanthoma (KA) and squamous cell carcinoma (SCC) is challenging. We surmised that a discriminatory immunostain would be clinically meaningful. Previous investigators have found CD123-positive plasmacytoid dendritic cells (PDCs) are more prominent in KA than SCC. We sought to determine if CD123 immunostaining might have value as a diagnostic test for distinguishing KA from SCC.Methods: We used blinded, semi-automated image analysis to compare CD123 expression in 66 KAs and 63 SCCs in a tissue microarray.Results: PDCs were present in both KA and SCC. Mean PDC frequency was higher in KA than SCC (14.2 vs 11.2 mean cells/0.0945 square mm) but the difference was not statistically significant (P = 0.1240). There was no significant difference in mean PDC cluster frequency, mean intratumoral PDC frequency, or the percentage of PDCs as proportion of the total mononuclear inflammatory cell infiltrate between KA and SCC. Conclusion: CD123 immunostaining is not a clinically useful test for distinguishing KA from SCC. K E Y W O R D S CD123, keratoacanthoma, plasmacytoid dendritic cells, Standards for Reporting of Diagnostic Accuracy Studies (STARD), squamous cell carcinoma

The Role of Plasmacytoid Dendritic Cells in Cancers

“…The subtyping of pDCs is also of great significance in SSC. BDCA2 is a specific marker of human pDCs, but high BDCA2 is expressed by immature pDCs, while pDCs expressing CD123 have higher maturity and ability to secrete cytokines ( 60 , 111 ). In head and neck SSC, the Poropatich group identified a subgroup pDCs expressing high levels of OX40 in the tumor microenvironment, which is conducive to anti-tumor immunity by increasing the expression levels of local IL-12 and IFN-α and enhancing the interaction between cDC and CD8 + T cell via OX40/OX40L-signaling axis ( 112 ).…”

Differentiating keratoacanthoma from squamous cell carcinoma—In quest of the holy grail

“…In a recent issue of the Journal of Cutaneous Pathology , Fraga et al also evaluated the number and distribution of PDCs using CD123, in a larger cohort (66 KA and 63 SCC) . While Abbas et al evaluated the entire lesion, Fraga et al opted to evaluate one 200× image per tumor on tissue microarrays in a blinded manner.…”

“…The following features may have contributed variably to the difference in the observed results of the two studies: (a) type of tissue evaluated: complete sections and relative bias due to availability of at least some histopathologic information vs single 200× image per sample and lack of any histopathologic information; (b) sample size: 20 each vs at least 60 each; (c) use of different markers: BDCA2 vs CD123; (d) definition of PDC cluster size: not defined vs three cells; and (e) Fraga et al only included cases where the diagnosis of KA was concordant clinically and histopathologically, while Abbas et al did not specify. As it can be sometimes impossible to distinguish KA from SCC, it is possible that some of the samples in both studies may not have been classified correctly.…”

“…Another limitation of both studies is that neither specified the stage of KA being studied. Because KA regression may be associated with PDCs, further studies using well‐defined criteria and potentially evaluating larger area of the lesions are needed to reconcile this difference and to critically evaluate PDCs as a marker for differentiating KA from SCC …”

“…Because KA regression may be associated with PDCs, further studies using well-defined criteria and potentially evaluating larger area of the lesions are needed to reconcile this difference and to critically evaluate PDCs as a marker for differentiating KA from SCC. 20,21 Thus, it is extremely challenging to confidently diagnose KA and…”