Morphological distinction of UC from PC can often be a diagnostic challenge especially in poorly differentiated tumors. Additionally, serum PSA levels may be raised in urothelial cancers that infiltrate the prostate gland adding to the diagnostic dilemma. Immunohistochemistry is often used as a diagnostic tool to accurately distinguish between these two tumors and establish the final diagnosis . The difficulty in making an accurate distinction is further compounded when there is only limited tissue available, such as in needle biopsies, cell blocks and fine needle aspirations with only small foci of carcinoma, when additional sections may have to be ordered and there may not be adequate tissue remaining to perform multiple immunohistochemical stains on separate slides.
Although the diagnostic utility of p63 and P501S in distinguishing between primary PC and UC have been individually evaluated previously [3, 11, 15], thus far, there have been no previous studies evaluating these two markers together either as a cocktail or applied sequentially. Our results indicate that dual-color immunohistochemistry with p63 and P501S applied sequentially shows excellent specificity for distinguishing UC from PC. The differential localization (diffuse nuclear for p63 versus granular cytoplasmic for P501S) combined with the double chromogen reaction facilitate easy visualization (brown for p63 and red for P501S) and enable quick and easy interpretation of the markers all in one slide. Additionally, this technique can be easily performed in the laboratory and conserves tissue as it is performed on a single slide. Thus, this immunostain could be a potentially valuable tool to aid in the distinction between these two cancers in the presence of limited tissue. However, in order to accurately characterize this double sequential immunostain, further studies comparing its performance with immunohistochemistry performed using the same antibodies individually as well as comparing its performance in prostate needle biopsies versus radical prostatectomies are required.
In the current study, p63 positivity was seen in 119/132 (90%) of UC cases. None of the PCs were positive for p63. Of the 13 UC cases that were negative for p63, 3 cases had micropapillary features. Our results are comparable to Kunju et al  who also found diffuse nuclear p63 positivity in 92% of their UC cases using the same p63 monoclonal antibody. Chuang et al  using the same p63 antibody found p63 in 83% of their UC cases. Similar to our study, Kunju et al and Chuang et al did not observe nuclear p63 positivity in any of their PCs. Thus, p63 appears to be a useful marker in distinguishing between UC and PC due to its high specificity for UC.
We found granular perinuclear cytoplasmic P501S expression in 22/23 (96%) PC cases. There was no difference in P501S staining intensity across PC cases according to Gleason scores. None of the UCs in our study showed P501S positivity. Our results are similar to Kalos et al  who found prostein expression in 111/118 (94%) primary and metastatic prostate cancers. They also found prostein to have excellent specificity with no expression detected in 4,635 normal and malignant non-prostatic tissues. Chuang et al observed P501S positivity in all of their 38 PC cases. These authors also found P501S to have high specificity with only 2/35 (6%) high grade UC showing focal weak positivity. Other studies [16, 17] have also shown prostein expression to be a highly specific marker for identification of prostatic origin of tumors. The granular perinuclear cytoplasmic expression of prostein is an important feature in establishing the prostatic origin of tumors. A recent study by Lane et al  found moderate diffuse cytoplasmic P501S staining in 11% of urinary bladder adenocarcinomas.
The single P501S negative case in our study was diagnosed as a poorly differentiated prostatic adenocarcinoma (Gleason score 9). The focus of carcinoma seen on the immunostain slide of this case showed atrophic features and p63 negativity for basal cells. Although prostein expression was absent in this focus of carcinoma, adjacent areas of high-grade prostatic intraepithelial neoplasias (HGPIN) expressed prostein and also showed p63 reactivity in the basal cells. It has been previously described that the expression of AMACR, is absent or decreased in atrophic PCs . It is possible that prostein expression may similarly be decreased in prostate cancer with atrophic features. This could lead to a potential diagnostic pitfall especially in rare cases of prostate cancer when p63 is aberrantly expressed in a non-basal distribution . Due to the potential impact on clinical practice, further studies are required to validate our finding and determine if prostein immunoreactivity varies among specific morphological variants of PC.
Our findings indicate that p63/P501S dual immunostaining shows excellent specificity and good sensitivity in distinguishing urothelial from PC. A p63+/P501S- immunoprofile favors a diagnosis of UC with 90.2% sensitivity and 100% specificity, while a p63-/P501S+ profile establishes a diagnosis of PC with 95.7% sensitivity and 100% specificity. The p63-/P501S- profile was seen in 14 cases (9%), 13 of which were urothelial cancers. Caution should be exercised in interpreting a p63-/P501S- profile as lack of p63expression does not rule out UC. Kunju et al  using a panel comprising PSA, HMWCK, and p63 to distinguish between PC and UC in 26 diagnostically challenging cases found p63 positivity in 10/13 UCs. The 3 remaining p63 negative UCs were also negative for HMWCK, and they established the urothelial origin using CK 7 and CK 20 expression. In a study by Higgins et al  that included 321 bladder UCs (n = 238 high grade, n = 83 low grade) and 267 PCs, placental S100 (S100P) and GATA 3 emerged as markers associated with urothelial differentiation, and S100P was found to have higher sensitivity for UCs. They found p63 expression in 87% of UCs and only 0.4% of PCs. However, when the expression of S100P was also considered, 94.9% of all UCs expressed one or both markers, while none of their PCs expressed both p63 and S100P. Higgins et al concluded that the expression of S100P and p63 are partly complementary and when used in combination each marker may identify UC cases missed by the other. Some studies have shown high sensitivity for detection of UC with HMWCK (clone 34βE12)  and thrombomodulin . Immunohistochemistry using a triple antibody combination of p63, P501S, and S100P or HMWCK or thrombomodulin may increase the sensitivity of the p63/P501S immunostain for detection of UC, and needs to be confirmed in future studies.