Reassessment of P53 Immunohistochemistry Thresholds in Invasive High Grade Bladder Cancer Shows a Better Correlation with TP53 and FGFR3 Mutations - Beyond the Abstract

p53 immunohistochemistry has been used for decades in the evaluation of solid organ malignancies. It is widely employed in gynecologic pathology in the assessment of predominantly endometrial but also ovarian carcinomas. Both these tumor types exhibit a dual pathway model of carcinogenesis, with p53 loss or overexpression commonly seen in more aggressive cancers. p53 immunostaining has now evolved into a surrogate marker for the TP53 mutation status of these carcinomas. Urothelial carcinoma shares a dual pathway model of carcinogenesis with these gynecologic malignancies such that FGFR3 mutations are over-represented in the low-grade urothelial pathway with TP53 mutations seen in the high-grade urothelial pathway. Intuitively, it would seem that p53 immunoexpression should follow a similar pattern in urothelial carcinoma. Historically, p53 expression in urothelial carcinoma has been assessed using a variety of cut-offs, most commonly a 10% level of tumor nuclear expression. However, this assessment does not consider the "null" expression pattern which is known to associate with TP53 mutations.

We have previously proposed alternative p53 thresholds in high grade urothelial bladder cancer that correlated with oncologic outcome.1 In this schema, p53 expression was assessed as "abnormal" when there was a complete absence of nuclear expression (with appropriate internal controls) or when >50% of tumor nuclei exhibited strong, diffuse p53 expression. When p53 staining was seen in 1-49% of nuclei, this was termed "wild type" pattern.

In this manuscript, we utilized this alternative scoring method in comparison with a traditional 10% cut off to assess concordance with TP53 and FGFR3 mutation status in a cohort of high grade, invasive urothelial carcinomas of the bladder (pT1= 63, pT2=45, pT3=159, pT4=77). All 344 cases were included on tissue microarrays and stained using p53 antibody (D0-7 clone, Ventana Medical Systems, Tucson, Arizona, USA) and scored as described above. Fifty cases were selected for next-generation sequencing (NGS) to determine the TP53 and FGFR3 status (Illumina NextSeq 550 platform). One-hundred and thirty-eight additional cases were separately assessed for FGFR3 status using SNaPshot polymerase chain reaction methodology.

We identified 202 cases with abnormal p53 staining and 142 with wild type pattern in comparison with 254 that had >10% expression and 90 <10% expression. Fifteen cases had TP53 mutations by next-generation sequencing (NGS) analysis and all 15 exhibited "abnormal" p53 expression using our novel scoring schema with only 12 cases demonstrating >10% expression using the traditional scoring method. Using both NGS and SNaPshot analysis, we found 27 tumours with FGFR3 mutations, 19 of which had "wild type" p53 expression versus only 6 which would have been deemed p53 negative using the 10% assessment threshold.

Receiver operating characteristic curve analysis demonstrated that a threshold of 45% tumor nuclear staining was the optimal cut-off for predicting TP53 mutations (sensitivity 100%, specificity 81%, area under the curve 0.904).

This contemporary p53 scoring scheme demonstrates a significant correlation with both TP53 and FGFR3 mutation status (p<0.0001 and p=0.002 respectively) unlike the widely used 10% cut off (p>0.05). The sensitivity and negative predictive value for TP53 mutations are also significantly improved (100% and 100%) when compared with the older system (80% and 88%).

These findings support our updated and biologically congruent scoring system which also correlates with oncologic outcome.

Written by: Michelle R. Downes MBBChBAO, MRCSI, MD, FRCPC, Division Head, Anatomic Pathology, Sunnybrook Health Sciences Centre, Genitourinary pathologist and Director of Research, Laboratory Medicine and Molecular Diagnostics, Sunnybrook Health Sciences Centre, Associate Professor, Laboratory Medicine and Pathobiology, University of Toronto, Toronto, Ontario, Canada


  1. Hodgson, Anjelica, Bin Xu, and Michelle R. Downes. "p53 immunohistochemistry in high‐grade urothelial carcinoma of the bladder is prognostically significant." Histopathology 71, no. 2 (2017): 296-304.
Read the Abstract