LAUSR

I have read with great interest the manuscript on undifferentiated carcinoma of the pancreas with osteoclast-like giant cells (UCOGC) by Hrudka and colleagues, published in the last issue of this Journal [1]. Histologically, UCOGS are heterogeneous, hypercellular tumors composed of three types of cells: (1) histiocytes, which are mononuclear and usually located adjacent to necrotichemorrhagic foci, typically present in this tumor type, (2) neoplastic cells, which are pleomorphic and markedly atypical, and (3) multinucleated giant cells, which show a very large cytoplasm and can have phagocytic activity, resembling true osteoclast [2]. Genetically, UCOGC are pancreatic tumors very similar to conventional ductal adenocarcinoma, sharing somatic mutations in the most common genetic drivers of conventional pancreatic ductal adenocarcinoma (PDAC) [3]. However, compared to PDAC, UCOGC show an important clinical difference: indeed, if they are “pure” (i.e., not associated with a PDAC component), their prognosis is clearly better than PDAC and even than UCOCG with an associated PDAC component [3, 4]. In their paper, Hrudka and colleagues, following an already well-demarcated line of research [5], investigated the expression of programmed death ligand (PD-L1) in UCOGC. PD-L1 is a molecule that is frequently overexpressed on the membrane of tumor cells of several types of solid malignancies. It is able to suppress anti-cancer immunity through the association with programmed death-1 (PD-1), a molecule that is expressed above all on cytotoxic lymphocytes. Hrudka and colleagues demonstrated the presence of a significant enrichment in the expression of programmed death ligand (PD-L1) in UCOGC, compared with PDAC, and that the prognosis in PD-L1-positive tumors is poor, compared with PD-L1 negative cases. Furthermore, a subset of patients with PD-L1negative UCOGC showed a surprisingly long survival. Such findings are certainly of interest and are in line with already published data. Indeed, a previous manuscript on this topic has well demonstrated the negative prognostic role of PD-L1 expression in UCOGC [5]. Notably, for the PD-L1 immunohistochemical analysis, the first paper that studied this topic used the clone E1L3N (Cell Signaling, Leiden, the Netherlands) [5], but in the following confirmatory manuscript, a different clone, 22C3 (pharmDx, Agilent, Santa Clara, CA, USA) has been used [1]. The fact that very similar results have been obtained with different clones represents an important confirmation of the reliability of the findings. The recent histological, molecular, and immunological insights on UCOGC can be summarized as follows: (i) histology: “pure” UCOGC have a better prognosis than UCOGC with an associated PDAC component [3, 4]; (ii) molecular profile: UCOGC share with PDAC the most common genetic alterations, thus representing a real PDAC variant [3]; (iii) immunology: UCOGC are reach in tumor-associated macrophages, and those expressing PD-L1 have a poor prognosis [1, 5]; (iv) biology: epithelial-to-mesenchymal transition is not a central process in the oncogenesis of this tumor time, differently from undifferentiated carcinoma of the pancreas without osteoclast-like giant cells [7]. These recent findings shed light on the biology of UCOGC; however, currently, there are not direct consequences for new specific therapeutic approaches against this tumor type. There is only a preliminary report of a patient with a pancreatic UCOGC with a lung metastasis that showed a good response to anti-PD-1 immunotherapy, with the possibility of curatively resection after therapy [6]. This calls for new clinical trials on the topic: such efforts should urgently explore more in depth the possibility of immunotherapy in * Claudio Luchini [email protected]