[1731] Use of Multispectral Microscopy To Distinguish Reactive Urothelium from Neoplastic

CM Gilbert, AV Parwani. University of Pittsburgh Medical Center, Pittsburgh, PA

Background: The interpretation of urothelial atypia in a setting of chronic inflammation and reactive changes can prove quite difficult. Ancillary studies such as CK20, P53, and CD44 have been shown to be of great value in these instances. The aim of this study is to evaluate a triple-immunostain with the assistance of multi-spectral microscopy in order to, in a single slide, evaluate their simultaneous expression and location in urothelium, and, at the same time, diminish the risk of loosing the area of interest in further recut sections.
Design: 53 bladder biopsies with previous diagnosis of benign/reactive, dysplastic, carcinoma in situ or carcinoma (approximately 39 cases spanning 1998-2006) were prepared using a triple-immunostain cocktail consisting of CK20, P53 and CD44. Three control stains were used for the purpose of creating a spectral library for the Nuance CRI Flex microscopy system (spectral range of 420 to 720 nm). The bright-field mode was used for analysis. All specimens were captured and analyzed with the custom built spectral template and analyzed using the provided Nuance software version 2.71. CK20 was given a score of 1 if it involved the upper 1/3, 2 for 2/3, and 3 for full thickness staining. P53 was scored as 0 (neg), 1+ (weak), 2+ (moderate) and 3+ (heavy). CD44 was not scored as it proved difficult to interpret in many cases.
Results: The results demonstrated that it was possible to separate the multiple stains co-localized in the biopsies, including the counterstains. Separation of the stains demonstrated a correlation of p53 and CK20 dual expression in biopsies diagnosed as carcinoma. Low or undetectable levels of expression were seen in biopsies later diagnosed as reactive or benign. Using this method, for benign or reactive cases, CK20 was restricted to the upper 1/3 in 17/23 biopsies (74%) and p53 was weak or negative in 11/23 biopsies (55%). No benign or reactive biopsies stained heavily (3+) for p53. For cases of carcinoma in situ or invasive disease CK20 was expressed in the upper 2/3 or full thickness in 19/26 (73%) and p53 was moderate or heavy in 19/26 cases (73%). Dysplastic biopsies appeared to resemble benign reactive biopsies with only 2/6 (30%) expressing moderate or heavy p53 and CK20 restricted to the upper 1/3 in 66% (4/6) of the biopsies.
Conclusions: The combination of multispectral microscopy and multiple immunostain cocktails form a powerful and useful tool for the interpretation of small biopsies with faint or difficult to interpret staining and for cases with limited material such as small needle core biopsies.
Category: Techniques

Wednesday, March 11, 2009 1:00 PM

Poster Session VI # 226, Wednesday Afternoon