Does a p53 "Wild-type" Immunophenotype Exclude a Diagnosis of Endometrial Serous Carcinoma?

An aberrant p53 immunophenotype may be identified in several histotypes of endometrial carcinoma, and is accordingly recognized to lack diagnostic specificity in and of itself. However, based on the high frequency with which p53 aberrations have historically been identified in endometrial serous carcinoma, a mutation-type immunophenotype is considered to be highly sensitive for the histotype. Using an illustrative case study and a review of the literature, we explore a relatively routine diagnostic question: whether the negative predictive value of a wild-type p53 immunophenotype for serous carcinoma is absolute, that is, whether a p53-wild type immunophenotype is absolutely incompatible with a diagnosis of serous carcinoma. The case is an advanced stage endometrial carcinoma that was reproducibly classified by pathologists from 3 institutions as serous carcinoma based on its morphologic features. By immunohistochemistry, the tumor was p53-wild type (DO-7 clone), diffusely positive for p16 (block positivity), and showed retained expression of PTEN, MSH2, MSH6, MLH1, and PMS2. Next generation sequencing showed that there indeed was an underlying mutation in TP53 (D393fs78, R213). The tumor was microsatellite stable, had a low mutational burden (4 mutations per MB), and displayed no mutations in the exonuclease domain of DNA polymerase epsilon (POLE) gene. Other genomic alterations included RB1 mutation (R46fs*19), amplifications in MYST3 and CRKL, and ARID1A deletion (splice site 5125-94_5138del108). A review of the recent literature identified 5 studies in which a total of 259 cases of serous carcinoma were whole-exome sequenced. The average TP53 mutational rate in endometrial serous carcinoma was only 75% (range, 60 to 88). A total of 12 (33%) of 36 immunohistochemical studies reported a p53-aberrant rate of <80% in endometrial serous carcinoma. We discuss in detail several potential explanations that may underlie the scenario of serous carcinoma-like morphology combined with p53-wild-type immunophenotype, including analytic limitations, a nonserous histotype displaying morphologic mimicry of serous carcinoma, and true biological phenomena (including the possibility of a TP53-independent pathway of endometrial serous carcinogenesis). Ultimately, our central thematic question is provisionally answered in the negative. At present, the available data would not support a categorical conclusion that a p53 alteration is a necessary and obligate component in the genesis and/or diagnosis of endometrial serous carcinoma. On the basis of their collective experience, the authors proffer some recommendations on the use of p53 immunohistochemistry in the histotyping of endometrial carcinomas.