Diagnostic significance of dual immunocytochemical staining of p53/cytokeratin20 on liquid-based urine cytology to detect urothelial carcinoma

INTRODUCTION

Urothelial carcinoma (UC) is the fifth most common cancer worldwide and is most prevalent among men.^[1,2] Although UC has a good prognosis, it is characterized by frequent recurrence and progression to a high grade that continuous follow-up is required.^[1,3,4] Cystoscopy and urine cytology are the most widely used methods for the diagnosis and follow-up of UC.^[1,2,5,6] Cystoscopy with surgical biopsy is the gold standard for the diagnosis of UC; however, the invasiveness and high cost of this method limit its use during routine patient follow-up.^[7,8] In addition, although cystoscopy is sensitive for detecting papillary lesions, UC in situ is often missed.^[5]

Urine cytology, on the other hand, is a noninvasive, inexpensive, and highly specific method for screening and monitoring UC.^[5,7,9] Although urine cytology has high specificity (99%), it frequently shows low sensitivity (15.8%–84.6%) and interobserver variability, particularly for detecting low-grade UC (LGUC).^{[5,6,8,10,11]} Another issue in the cytologic diagnosis of UC is the category of “atypical urothelial cells” (AUCs). As reactive cellular changes caused by inflammation, viral cytopathy, and therapy-induced changes, for example, are often diagnosed as atypical cells, ruling out UC can be difficult.^[3] A high incidence of AUCs is an important problem in urine cytology, and the proportion of AUCs is >20% according to large-scale studies.^[3,12] AUCs category can be overused as “wastebasket” that the proportion of AUCs needs to be reduced to increase the clinical significance of urine cytology.^[12] The development of an ancillary test capable of improving the efficacy of urine cytology would be of value.

Several ancillary tests have been investigated and developed to improve the accuracy of urine cytology. Although immunocytochemistry (ICC) of urine samples is rarely used, the method shows promise for improving diagnostic accuracy when used as an adjunct to urine cytology.^[5,13] Several recent studies evaluated the usefulness of ICC using biomarkers such as cytokeratin 20 (CK20), p53, CD44, and hTERT.^[3,14-16] Among them, immunohistochemical staining for p53 and CK20 is a widely used and promising method for the diagnosis of UC using surgical biopsy samples.

CK20 is a cytoplasmic protein and is considered a useful marker for detecting neoplastic changes in urothelial cells. In the normal urothelium, CK20 is expressed only in umbrella cells, whereas it is expressed in the entire epithelial layer in UC.^[17,18] In urine cytology, atypical cells expressing CK20 may indicate neoplastic changes rather than reactive changes.^[13] p53 is the product of the tumor suppressor gene TP53, and nuclear overexpression of p53 indicates the presence of TP53 gene mutation. TP53 mutation is involved in several stages of UC development. The expression of p53 characteristically shows a diffuse and strong nuclear pattern in neoplastic changes, whereas its expression is characterized by a focal and weak pattern in reactive changes.^[19,20] Strong positivity for p53 is a useful marker of UC in urine cytology.

Most previous studies investigating the efficacy of ancillary tests were performed using urine specimens obtained separately from urine cytology. These studies analyzed protein expression associated with molecular changes in the urine without a morphological analysis based on urine cytology. Previous studies analyzed the expression of CK20 using archival slides of urine cytology.^[6,11,21] To the best of our knowledge, there are no studies analyzing the expression of p53 using urine cytology.

The aim of the present study was to evaluate the expression of p53 and CK20 using dual immunocytochemical staining of liquid-based cytology (LBC) slides. The usefulness of these proteins for improving the diagnostic accuracy of urine cytology for UC was analyzed.

MATERIALS AND METHODS

RESULTS

A total of 61 urine samples diagnosed as UC by biopsy were examined, including 17 LGUC and 44 HGUC cases. There were 49 men and 12 women (ratio 4.1:1) with a median age of 72 years (range, 47–95 years). LBC specimens from the urine of 61 patients with UC were diagnosed as negative (n = 19), atypical (n = 20), suspicious for HGUC (n = 16), LGUN (n = 2), and HGUC (n = 4). The sensitivity of urine cytology for UC was 36%.

Dual staining for p53 and cytokeratin 20

The 61 destained urine cytology samples from patients with UC were evaluated by dual immunostaining for p53 and CK20. Ten cases confirmed as benign disease were also stained as negative controls. The results of immunostaining identified 38 (62.3%) p53-positive and 7 (11.5%) CK20-positive UC samples. The ten benign disease specimens demonstrated no reactivity for p53 or CK20. p53 expression showed a diffuse and strong nuclear staining pattern in scattered or clustered atypical cells [Figure 1]. CK20 was normally expressed in superficial urothelial cells with abundant cytoplasm and low nuclear–to-cytoplasmic ratio [Figure 2a and b]. CK20 expression detected in distinct superficial urothelial cells was excluded from positivity for UC. Of seven CK20-positive cases, six coexpressed p53 [Figure 3]. Only one case of LGUC showed cytoplasmic expression of CK20 without p53 expression [Figure 2c and d].

Close

Figure 1:

p53 expression in urine cytology. p53 staining in atypical scattered single cells (a and b) and clusters of atypical cells (c and d) showing strong nuclear expression of p53 (b and d) (×400).

Export to PPT

Close

Figure 2:

CK20 expression in urine cytology. Superficial urothelial cells are reactive for CK20 (red cytoplasmic stain) (a and b). A papillary cluster of urothelial cells in a patient with low-grade urothelial carcinoma (c) shows diffuse positivity for CK20 (d) (×400).

Export to PPT

Close

Figure 3:

Dual immunostaining for p53 and CK20 in urine cytology. Atypical urothelial cells are positive for p53 (brown nuclear stain) and CK20 (red cytoplasmic stain) (a and b) (×400).

Export to PPT

DISCUSSION

Urine cytology is widely used as a screening test for UC because it is inexpensive and easy. Urine cytology is also a noninvasive method and is useful for repeated and continuous surveillance of patients with UC. However, diagnosis by urine cytology is subjective, and it shows high specificity but low sensitivity.^[11] In particular, LGUC is difficult to diagnose by urine cytology. Previous studies reported a wide range of sensitivity values for urine cytology (15.8%–84.6%),^[11] and a meta-analysis reported an overall sensitivity of 42%.^[23] In the present study, the sensitivity of urine cytology was 36.1% for all cases, 29.4% for LGUC, and 38.6% for HGUC. Although the sensitivity was slightly higher for HGUC, urine cytology showed low sensitivity for both LGUC and HGUC.

Several adjunctive tests were developed to improve the low sensitivity of urine cytology. Six ancillary tests (UroVysion, uCyt+, and tests for NMP22 and BTA) are approved by the FDA and are commercially available. Among these, UroVysion and uCyt + are the most widely used, and these methods are based on urine sedimentation cells.^[7,24] However, these tests are costly, time-consuming, and require skilled technicians.^[5,7] An inexpensive, easy to perform, highly sensitive ancillary test to detect and monitor UC would be highly desirable.

Currently, ICC is rarely used for the diagnosis of UC using cytology.^[5] ICC has the advantage that it can be related to the morphology of cells in urine cytology.^[7] In the present study, dual immunostaining for p53 and CK20 was performed on archived urine cytology specimens to evaluate the diagnostic usefulness of adjunct ICC. Of 61 samples from patients with UC, 38 (62.3%) showed positive p53 expression. The combination of p53 ICC and urine cytology showed higher sensitivity (68.9%) than urine cytology alone (36.1%). The combination of urine cytology with p53 improved the sensitivity in both LUGC (from 29.4% to 47.1%) and HGUC (from 38.6% to 77.3%). Arville et al.^[3] reported the results of triple staining (CK20, p53, and CD44) using cell block samples of urine, and 67% and 79% of samples were p53 and CK20 positive, respectively. Two studies performing CK20 immunostaining of PAP slides of urine cytology showed 70.4% sensitivity for UC detection.^[11,21] Srivasta et al.^[21] reported that CK20 did not improve the detection of UC. Unlike the results of previous studies, we detected CK20 expression in 8.2% (5 of 61) of samples in this study. However, despite the low frequency of CK20 expression, the sensitivity of combining urine cytology with p53 and CK20 was >50% in LGUC. The diagnosis of LGUC by urine cytology has limitations, and the diagnostic value of cyto-morphological changes is still debatable.^[2,25] The combination of urine cytology with adjunct ICC can improve the accuracy of UC diagnosis.

Limitation of urine cytology is the high frequency of the diagnosis of atypical cells, which is reported at 2–23% in the literature.^[3] In these cases, the design of treatment is difficult for clinicians. As atypical cells are present in a variety of benign diseases, it is important to identify atypical cells associated with UC to reduce the rate of unnecessary cystoscopy. In the present study, 13 (65%) of 20 cases of atypia in urine cytology showed positivity for p53 or CK20. For patients diagnosed as atypia by urine cytology, additional staining for p53 and CK20 may be useful to identify patients with a high likelihood of UC, thereby facilitating treatment decisions by clinicians.

The application of immunochemical staining can be used as an adjunct to cytology-based diagnosis. However, extracellular materials for cell blocks or additional slides for immunochemistry are not always prepared.^[26] Furthermore, findings based on additional slides prepared for ICC can be inconsistent with those of the original urine cytology slides, and the amount of cells is likely to be insufficient. In the present study, ICC staining for p53 and CK20 was performed using destained archival slides of LBC. The main advantage of this method is that it allows simultaneous analysis of cytologic changes on PAP slides and the exclusion of nonspecific staining in normal cells.^[13]

In the present study, a small number of negative cases were enrolled because histologic confirmation was available only a few cases with negative cytology. It could limit the evaluation of specificity and predictive value of the ICC test. Urine cytology is one of screening test approaching urologic disorder. Nonetheless of this limitation, this study showed dual p53/CK20 ICC increased a diagnostic sensitivity of cytologic evaluation.

CONCLUSIONS

Combining urine cytology with p53 and CK20 ICC improved the diagnostic accuracy for both LGUC and HGUC. In addition, ancillary ICC staining for p53 and CK20 allowed discriminating between UC and benign diseases in patients diagnosed with atypia in urine cytology.