Is Diagnosis of Low-Grade Urothelial Carcinoma Possible in Urine Cytology?

BACKGROUND: Urine cytology is a useful modality, primarily for the diagnosis and follow-up surveillance of high-grade urothelial carcinoma (HGUC). Its utility in diagnosing low-grade urothelial carcinoma (LGUC) remains controversial because of low reported sensitivity compared to cystoscopy. AIM: To study the cytomorphology of LGUC in voided urine samples and analyze its utility in diagnosis.
MATERIALS AND METHODS: This is a retrospective study of one year, including 48 voided urine samples in cases which were confirmed as LGUC on subsequent histology. Urine cytology smears of these cases, originally stained with Papanicolaou stain were reviewed, critically analyzed and the specific cytomorphologic and cystoscopic findings were documented.
RESULTS: On review 18 samples were re-categorized as LGUC which included 10 samples initially diagnosed as Negative for HGUC, 2 as Atypical Urothelial Cells - Not Otherwise Specified (AUC-NOS) and 6 as Suspicious for Carcinoma. In addition, another 3 samples with initial diagnosis of LGUC remained as LGUC on review. Thus, a total of 21 LGUC samples were identified after the review. 26 (54%) samples with a diagnosis of negative for HGUC remained negative even after review, as the tumor cells were not identified either due to sampling error or unrecognizable morphology. One (2%) samples of AUC-NOS remained the same on review due to very scant atypical cells. In 21 LGUC samples, cytology showed a dual population of benign differentiated urothelial cells and small urothelial cells with subtle nuclear atypia such as irregular and thickened nuclear membrane with increased nuclear cytoplasmic ratio. In 12 false negative LGUC samples, the diagnostic cells were camouflaged by their subtle nuclear atypia coupled with an overwhelming background of differentiated benign urothelial cells as both appeared almost similar in morphology. Papillary fragments were identified only in 2 samples.
CONCLUSIONS: Diagnosis of LGUC on cytology is challenging and depends on the presence of diagnostic cells, pick up of diagnostic cells on screening and accurate interpretation. Special attention to papillary fragments and aforementioned nuclear atypia should be paid as tumor cells may resemble normal urothelial cells and can be easily missed. Copyright:

INTRODUCTION

Urine cytology is a simple, non-invasive and cost-effective diagnostic modality. Its role in the detection of high-grade urothelial carcinoma (HGUC) is well established with a reported sensitivity of 50% to 85% depending upon the type of urine sample collected.[12] The utility of urine cytology in diagnosing low-grade urothelial carcinoma (LGUC) remains controversial with low sensitivity of 10% to 43.6% as reported in the literature.[12] LGUC progresses to HGUC in only 5% to 25% of cases while a considerable proportion (48%-71%) of LGUC cases are known to recur.[3] The detection of recurrent tumor needs cystoscopic evaluation and pathological examination by urine cytology and biopsy. The diagnosis of LGUC on urine cytology poses great challenge mainly because the tumor cells show subtle nuclear changes resembling normal urothelial cells. Thus, urine cytology is not considered as a good diagnostic modality for the detection of LGUC. In majority of the cases of recurrent LGUC, cystoscopic examination can identify papillary growth in the bladder. However, in cases of LGUC where no growth is identified on cystoscopy, urine cytology may prove useful provided specific cytomorphological features for LGUC are defined and identified.

The role of urine cytology in HGUC has been assessed by several studies after the establishment of 2004 World Health Organization (WHO)/International Society of Urological Pathology (ISUP) classification scheme for urothelial tumors. Till date, only few studies have evaluated the cytomorphological features of LGUC, however these studies were focused primarily on instrumented urine samples.[456789] In the present study, the specific cytomorphological features of LGUC in fresh voided urine samples have been evaluated aiming to improve the detection rate of LGUC in voided urine samples.

MATERIAL AND METHODS

Patient selection

This was a retrospective study in which all the patients with urinary tract biopsy specimens diagnosed with LGUC during a period of one-year from June 2017 to May 2018 were included. The pathology database was searched for urinary bladder biopsy specimens which were diagnosed with LGUC during this period. All the corresponding clinical data including cystoscopy and urine cytology within six months of index biopsy were obtained from the electronic medical records. 48 voided urine samples (44 cases) were obtained with subsequent diagnosis of LGUC on biopsy. Cases diagnosed as HGUC on biopsy were excluded from the study.

Cytomorphological evaluation

Two Papanicolaou stained smears (conventional direct smears or cytocencentrifuged smears) prepared from each of the urine samples were reviewed by two independent observers.

The following seven cytomorphological features were evaluated and documented for each case:

Overall cellularity

Tumor cells architecture - groups/papillary clusters or single cells

Nuclear-cytoplasmic (N:C) ratio

Chromatin pattern

Nuclear membrane contour

Thickened nuclear membrane

Nuclear hyperchromasia.

Subjective assessment of cellularity was done as low or high based on the total number of cells in a smear regardless of the cell type (umbrella or non-umbrella cells). A minimum of 10 cells per smear were considered adequate. Three-dimensional papillary clusters were defined as cluster of urothelial cells with papillary architecture and significant nuclear overlapping. Small cell groups and presence of single tumor cells further provided an evidence of discohesion indicating malignancy. N:C ratio was assessed as <0.5 or >0.5. Nuclear chromatin pattern was described as finely granular, coarse or clumped. Contour of the nuclear membrane was assessed as regular or irregular. Any thickening of nuclear membrane and nuclear hyperchromasia in tumor cells was noted. Concurrent reassessment of discrepant cases led to the final consensus and accurate results.

For surgical specimens, papillary neoplasm of low malignant potential (PUNLMP) or LGUC were diagnosed according to the WHO/ISUP classification of urinary bladder cancers, 2016.[10]

Statistics

Statistical analysis was performed using SPSS software version 20. Frequency analysis of all the parameters/variables was performed to identify the cytomorphological features which independently or in combination predicted the diagnosis of LGUC.

RESULTS

A total of 48 fresh voided urine samples (44 patients) subsequently diagnosed as LGUC on biopsy were reviewed. None of the patients developed HGUC during the follow-up period. The study group comprised of 38 males and 6 females with a mean age of 59 years (range 39-79 years). Of these 48 samples, concomitant cystoscopy was performed in 46 samples. Out of these 46 samples, 45 (98%) showed a papillary growth with a size of ≤1 cm in 30 samples and >1 cm in 15 samples. Cystoscopy did not show any growth or abnormality in one samples.

Before review of the total 48 urine samples, only 3 had an initial diagnosis of LGUC on cytology while 6 were diagnosed with Atypical Urothelial Cells, Suspicious for Carcinoma (sensitivity 18.7%). 36 samples were diagnosed as negative for HGUC and 3 as AUC-NOS (Atypical Urothelial cells, Not Otherwise Specified). After review, 10 out of 36 samples initially categorized as Negative for HGUC, 2 out of 3 cases initially categorized as AUC-NOS and 6 samples initially categorized as AUC-Suspicious of Carcinoma (Atypical Urothelial Cells, Suspicious for Carcinoma) were re-categorized as LGUC (total 18) on the basis of above defined cytomorphological features. In addition, another 3 samples of initially diagnosed LGUC remained LGUC on review. Thus, a total of 21 LGUC samples could be identified on review.

From the 36 samples with Negative for HGUC, 26 (54%) remained as negative even after review. One samples of AUC-NOS remained as same on review.

Of the 45 samples which showed growth on cystoscopy, 35 samples were negative on urine cytology initially, 3 samples were diagnosed as AUS-NOS, 4 samples as AUC-Suspicious for Carcinoma and 3 samples as LGUC. After review, 10 samples with initial diagnosis of Negative for HGUC, 3 samples with AUS-NOS, 4 with AUC-Suspicious of carcinoma were re-categorized as LGUC. One samples which did not show any growth or abnormality on cystoscopy was diagnosed as Negative for HGUC in urine cytology initially. After review, the diagnosis of Negative for HGUC in urine cytology remained same for this samples [Table 1].

Table 1

Distribution of samples (n=48) in different diagnostic categories based on review of cytologic smears with corresponding cystoscopy findings

Diagnostic categories	Before review	After review	Cystoscopy
			Growth	No abnormality	Not done
Negative for HGUC	36	26	35	01	-
AUC-NOS	03	01	03	-	-
AUC-Suspicious for Carcinoma	06	00	04	-	02
LGUC	03	21	03	-	-
Total	48	48	45	01	02

Thus, a total of 21 out of 48 samples were diagnosed as LGUC after review (18 re-categorized on review and 3 samples with initial and review diagnosis of LGUC on cytology) owing to the presence of specific cytomorphological features such as increased N:C ratio, irregular nuclear membrane, thickening of nuclear membrane without any appreciable nuclear hyperchromasia. In this study, the sensitivity for the diagnosis of LGUC in urine cytology increased from 18.7% to 43.7% after review.

Out of 21 samples diagnosed as LGUC on urine cytology after review, the overall cellularity was low in 11 (52.3%) samples. In majority of the samples (76.2%), tumor cells were present as singly scattered cells while only 2 samples (9.5%) showed the classical three-dimensional clusters with papillary architecture. The N:C ratio was consistently high in 20 out of 21 samples (95.3%). The nuclear chromatin pattern was variable with majority of samples (62%) showing finely granular nuclear chromatin. 85% of the samples showed thickened nuclear membrane without any appreciable nuclear hyperchromasia. Nuclear membrane irregularity of variable degree was noted among majority (85.7%) of the samples [Table 2, Figures 1 and 2].

Table 2

Prevalence of different cytomorphological features among samples diagnosed with LGUC (n=21) on urine cytology

Cytomorphological feature	Number of cases
Cellularity
Low	11 (52.3%)
High	10 (47.7%)
Tumor cell architecture
Cell groups	03 (14.3%)
Papillary clusters	02 (9.5%)
Single cells	16 (76.2%)
Nuclear-cytoplasmic (N:C) ratio
<0.5	01 (4.7%)
>0.5	20 (95.3%)
Chromatin pattern
Finely granular	13 (62%)
Coarse	07 (33.3%)
Clumped	01 (4.7%)
Nuclear membrane contour
Smooth and regular	03 (14.3%)
Irregular	18 (85.7%)
Thickened nuclear membrane
Yes	18 (85.7%)
No	03 (14.3%)
Nuclear hyperchromasia
Present	03 (14.3%)
Absent	18 (85.7%)

Figure 1

(a) Urine cytology smear showing LGUC fragments with three-dimensional papillary configuration (small arrow) (×400 Papanicolaou stain) (b) Superficial urothelial cells (long arrow) and LGUC cells (small arrow) with increased N:C ratio and thick irregular nuclear membrane (×400 Papanicolaou stain) (c) Superficial urothelial cells (long arrow), intermediate urothelial cells (arrow head) and LGUC cells (small arrow) (×100 Papanicolaou stain) (d) LGUC cells (small arrow) show increased N: C ratio, thick and irregular nuclear membrane (×400 Papanicolaou stain)

Figure 2

(a) Urine cytology shows LGUC cells (small arrow) exhibiting increased N:C ratio, irregular thickened nuclear membrane and finely granular chromatin (×100 Papanicolaou stain) (b and c) LGUC cells from same case in different fields showing high N:C ratio, thickened irregular nuclear membrane and finely granular chromatin (×400 Papanicolaou stain). (d) Histopathology section from the same sample showing fused, branching papillae with enlarged nuclei exhibiting mild nuclear enlargement and minimal loss of polarity (H and E × 100)

DISCUSSION

The role of urine cytology as a diagnostic tool for the detection of LGUC in voided urine samples remains controversial. The reported sensitivity of urine cytology in the diagnosis of LGUC is low (10%-43.6%).[12] In this study, the sensitivity of cytologic diagnosis of LGUC in voided urine samples was 44%. Low cellularity in voided urine samples posed a great diagnostic challenge. In the present study, 26 samples (54%) which remained Negative for HGUC even after review either showed low cellularity or unrecognizable morphology of the tumor cells simulating normal urothelial cells. One sample (2%) remained AUC-NOS due to very scant atypical cells (only two cells). Low cellularity in the samples can be attributed to errors in sampling technique and increased cellular adhesion of the tumor.[4]

Only a few studies have made an attempt to suggest the cytological features which can predict the diagnosis of LGUC in voided urine samples but most of these features overlap with benign reactive conditions.[111213] Murata et al.[11] reported two cell patterns in LGUC in voided urine samples, “isolated single cell pattern” with low N:C ratio and “cluster pattern” with piled up cellular clusters with high N:C ratio, latter pattern being the predominant pattern. Both patterns showed hyperchromatic nuclei with uneven contour. In 2015, Chung et al.[14] described loss of polarity in papillaroid clusters with irregular contours, absence of cytoplasmic vacuolation and columnar cells, presence of nuclear hyperchromasia as key features of LGUC in Sure Path liquid based voided urine cytology. Several other studies have also described the nuclear enlargement, hyperchromasia, membrane irregularity and increased cellular/nuclear size as the frequent features of LGUC.[612] Contrary to these studies, nuclear hyperchromasia was not a diagnostic feature of LGUC in our study.

In this study, we evaluated seven cytomorphological features for each smear; the overall cellularity, tumor cells architecture in groups/papillary clusters or single cells, nuclear-cytoplasmic (N:C) ratio, chromatin pattern, nuclear membrane contour, thickened nuclear membrane and nuclear hyperchromasia. In the present study, most of the smears showed isolated single tumor cells while papillary fragments were observed only in 2 samples. The isolated single tumor cells may be mistaken for benign intermediate urothelial cells in the absence of any papillary fragments leading to a false negative diagnosis. The presence of low-grade nuclear atypia was not instantly picked by the screener as the cells resembled normal urothelial cells in low power objective (10×). In a background of overwhelming benign differentiated urothelial cells (without any atypia), fewer tumor cells with mild nuclear atypia pointing towards the diagnosis of LGUC could easily be overlooked.

On review, the N:C ratio in tumor cells was found to be consistently high in 20 out of 21 samples (95.3%) of LGUC. In 18 samples, the tumor cells could not be picked up on the initial screening, probably because the nuclei were pale (not hyperchromatic) and non-captivating at 10× objective and were missed during screening. As a matter of fact, nuclear hyperchromasia in tumor cells constitutes the most important feature which strikes during the microscopic screening. High N:C ratio with hyperchromasia enhances the pickup rate (as in HGUC) whereas only increased N:C ratio without any appreciable hyperchromasia can lead to overlook the diagnostic cells during screening.

In addition, none of these samples did show papillary fragments. One of the reasons for missing LGUC initially (before review) was due to lesser number of atypical cells in a background of numerous normal urothelial cells. Thus, a close and critical examination under the high-power objective (40×) is necessary to analyze the diagnostic features of LGUC.

Papillary fragments although considered as a diagnostic feature of LGUC were rarely seen in urine cytology. Increased N:C ratio, slight nuclear membrane irregularity and thickened nuclear membrane were found to be the most frequent features in LGUC. Chromatin pattern was variable in this study with majority of LGUC cells showing fine nuclear chromatin. Nuclear hyperchromasia, one of the features essential for the diagnosis of HGUC, was not a predominant feature in tumor cells in our study.[15]

Features which favored a low-grade malignant pathology over benign reactive conditions include:

Increased N:C ratio

Subtle nuclear membrane irregularity

Thickened nuclear membrane

Papillary cell fragments with overlapping nuclei

Absence of nuclear hyperchromasia.

Though both HGUC and LGUC can show single cells with high N:C ratio, the presence of moderate to severe hyperchromasia and marked irregularity of the nuclear membrane with coarse/clumped chromatin can lead to a final diagnosis of HGUC.

In this study, cystoscopy had a sensitivity of 98% in the detection of LGUC. However, a false negative urine cytology may extend the scope of cystoscopy during the follow-up of LGUC. The improved detection of LGUC in urine cytology has better clinical application of immediate cystoscopic examination. 25% increase in sensitivity for the detection of LGUC in voided urine sample was noted after re-evaluating the specific cytomorphological features diagnostic of LGUC.

A major drawback of this study was the exclusion of cases without subsequent biopsy confirmation of LGUC. Urine samples with features predictive of LGUC but without concurrent biopsy should have been closely followed-up prospectively to see the utility of these features in the diagnosis of LGUC.

In short, the diagnosis of LGUC in voided urine cytology is challenging and depends essentially on the presence of diagnostic cells in the sample, right pick up on screening and accurate interpretation. Tumor cells with mild nuclear atypia coupled with distraction by benign urothelial cells in the background can lead to the under-diagnosis of LGUC. Therefore, special attention should be paid to the aforementioned nuclear atypia, especially under the high-power objective to avoid false negative diagnosis.

Further studies with a large number of cases and longer duration of follow-up period are required to validate the findings of the present study. Although the main focus of urine cytology is to detect HGUC, its utility for the detection of LGUC cannot be underestimated, especially in a setting of long-term follow-up requirement of treated cases of urothelial carcinoma.

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.