[CANCER RESEARCH 37, 2800-2805, August 1977]

Summary

Six years of experience with the cytological diagnosis ofcarcinoma of the urinary tract are reviewed. This includes 2years of participation in the National Bladder Cancer Project. With increasing experience of the cytopathologist andcytotechnologist, attention to detail of specimen collectionand preparation, the increasing use of bladder washing andcystoscopic urine specimen, and the interest and cooperation of the urologist, the false-positive rate of diagnosis oftransitional cell carcinoma of the urinary tract has decreased.

Papillomas and low-grade papillary, noninvasive transitional cell carcinomas have proven impossible to diagnoseby cytology alone but ‘maybe diagnosed rarely from cellblock material prepared from the bladder washings. A residual of cases of chronic cystitis, benign prostatic hyperplasia, renal calculi, chemotherapy effects, and cancer-associated atypias provide the cytopathologist with diagnosticproblems, since they remain nearly inseparable from transitional cell carcinoma on the basis of cytology alone.

Cytology has proven sensitive to the detection of recurrent and persistent transitional cell carcinoma followingtreatment with chemotherapy, radiation, on surgery. In suchcases the abnormal area of epithelium in the bladder may beextremely difficult for the urologist to find by biopsy. Heshould not be lulled by a normal-appearing bladder andnegative biopsies in these cases into believing that thecytology report is incorrect.

There are many factors that affect the efficiency, accuracy, and specificity of the cytological diagnosis of cancerof the entire urinary tract and urinary bladder: collectionand preparation of the cytological specimen, expertise andinterest of the cytotechnologist and cytopathologist, dm1-clan interest in obtaining good specimens and providingaccurate clinical data, and even participation in a specialstudy of the National Bladder Cancer Project. This reportdocuments the interaction of these factors on the performance of 1 participating cytology laboratory in Clinical Collaborative Group A of the National Bladder Cancer Project.

Materials and Methods

All of the urinary tract cytology specimens submitted to

our laboratory at the Medical College of Virginia betweenJanuary 1, 1970, and December 31, 1975, were reviewedand correlated with clinical and histological data as cornpletely as possible. This period includes 2 years of participation in Clinical Collaborative Group A, 1974 and 1975.During these 2 years specific protocols for the collection ofurine specimens for cytological study were in force. Theactual degree of compliance with these protocols cannot bedetermined. Voided urine samples were to be collected asfresh morning specimens and placed directly in an equalvolume of 50% ethyl alcohol before being transported tothe cytology laboratory. Over weekends and holidays theseprefixed specimens were to be stored in the cytology laboratony refrigerator, but in some cases they were undoubtedlyleft at room temperature in various areas of the hospital.

Urine was collected at the beginning of cystoscopy, andbladder washings were then obtained. Both specimenswere collected without fixation and transported immediately to the cytology laboratory for preparation of membrane filters. Collection of the bladder washing specimenwas performed according to the protocol of Clinical Collaborative Group A, using 50.0 ml of 0.9% NaCI solution andirrigating the bladder 3 to 5 times with this same fluid. In thelaboratory 4.0 ml of well-mixed cystoscopic urine specimenwere used to make 1 filter preparation, and 1 filter preparation was made for each 25.0 ml of bladder washing. Unclearspecimens were diluted with 0.9% NaCI solution until nearlyclean before filtering an aliquot. If the specimen was to bediluted, it was necessary for the cytotechnologist to exencise judgment concerning the volume of specimen thatcould be filtered to make a good cytological preparationwithout overloading the filter and/or distorting the cells.Two filter preparations were made of each voided or cystoscopic urine submitted to the laboratory. The Nucleoporemembrane filter (General Electric Co. , Pleasanton, Calif.)was used (7). The remaining sample of bladder washingsnot used to make filter preparations was centrifuged andprepared as a cell block by the method of Harris et aI. (7).

Also included in this study are a small number of ureteraland renal pelvis washings prepared in the same manner asbladder washings and ureteral and renal pelvis brushingsprepared as direct smears from the brush and as filterpreparations by agitating the brush in 4.0 ml of 0.9% NaCIsolution.

Results

, Presented at the National Bladder Cancer Conference, November 28 toDecember 1, 1976, Miami Beach, Fla. This work was supported in part byResearchGrant CA 15492from the NationalCancerInstitute through theNational Bladder Cancer Project.

2 Presenter. To whom requests for reprints should be addressed, at Box

115 MCv, Richmond, va. 23298.

The results of this review are summarized in Tables 1 to 4.The cytological diagnosis of suspected on definite cardnoma refers to individual specimens. The number of histological diagnoses aI@orefers to individual specimens and

CANCER RESEARCH VOL. 372800

Current Practice of Urinary Bladder Cytology1

WilliamJ. FrabIe,2LoisPaxson,Jo A. Barksdale,and WarrenW. Koontz,Jr.Division of Surgical and Cytopathology (W. J. F., L. P., J. A. B.J and Department of UrologyfW. W. K.], Virginia Commonwealth University, Health SciencesDivision, Medical College of Virginia, Richmond, Virginia 23298

on May 20, 2018. © 1977 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Cytologystudy of patients with caricer of the urmary tract, primary ormetastaticYrPatients

withknowncancerCytology

positive or suspected can

cer, individualsubmittedspecimensCytology

diagnosisconfirmedhistologi

callyFalse

positive/suspi

cious, individual submitted

specimensFalsenegativeTotal

specimens

submitted1970

19711972197319741975

Total12

312224485119

42 (3)―39 (6)37 (15)59 (28)101(57)10

28343045897

(36%)12 (28%)4 (10%)7 (18%)

10 (16%)12(11%)0

03113255

548576711739699

3488

Correlation of type of cytological specimen withfalse-positivediagnosespositive

versusPositivecytologyforcanci-

ConType of specimen noma firmedFalsepositiveCystoscopic

or voided ur- 191 15338(19%)meBladderwashings

93 7914(15%)Ureteror pelviswashings 16 88 (50%)

Number of false-positivediagnosesrelated to differurinary tract processesent

benignCysto

scopic orUreterorBenignurinary tract voided ur-Bladderpelvisprocesses

mewashingswashingsInfection1283Benign

prostatic hyper- 101plasiaCalculus

54Cancer-associated61changesChemotherapy

1Radiation2Normal

bladder 5 3

Cytology of Bladder Cancer

Table1

a Number in parentheses, number of bladder, ureteral, on renal pelvis washing speci

mensof the total positive or suspectedcytological diagnosesof cancer.

Table 2 Table 3

not to individual patients. Thus, of the 12 patients withbladder cancer seen in 1970, there were 19 suspected orpositive cytological diagnoses of cancer on individual specimens, 10 of which were confirmed by histological study.Until 1975 the number of suspected or positive cytologiesaveraged slightly over 1/case, but with active participationin the bladder cancer study the number rose to 2/case.Cytological specimens obtained by washing the bladder,ureter, or renal pelvis account for one-half of the abnormalcytology in 1974 and 1975. The application of this techniquealso correlated with improvement in the false-positive rateof diagnosis. When the results of studying cystoscopicurine samples and bladder washings submitted simultaneously from the same patient were compared, the cytologicaldiagnosis was the same for both types of material in 68% ofthe cases with bladder cancer. In the 32% of cases with adifference in diagnosis between the 2 sources of material,the bladder washing consistently indicated that a moresevere lesion was present. In a small number of cases wherevoided urine taken within 1 week of the cystoscopic specimen and bladder washing could be compared for similarityor difference in cytological diagnosis, the voided specimenwas significantly less diagnostic in 65% of the cases.

Table 2 compares 3 types of specimens in relation to thediagnosis of carcinoma, primary or metastatic, to the unnary tract, and the rate of false-positive diagnoses for eachtype of specimen. The false-positive rate is similar for bothcystoscopic and voided urine samples in comparison withbladder washings. The most difficult type of specimen tointerpret was a ureteral on renal pelvis washing. The propontion of false-negative diagnoses in this series was about the

same in bladder washings and cystoscopic or voided urines.

Certain urinary tract lesions are responsible for the majority of false-positive cytology findings from urine specimens.This is documented in Table 3. Inflammatory atypias wereequally difficult to interpret in bladder washings and voidedor cystoscopic unines. Benign prostatic hyperplasia andcancer-associated changes were interpreted as carcinomamost often in voided oncystoscopic unines. In the cases withcancer-associated changes, bladder washings were notsubmitted, so it cannot be assumed that those cases wouldhave been reported as negative. This curious phenomenonof malignant-appearing cells in patients with known cancerthat does not involve the epithelial surface from which thecytological specimen has been obtained was described indetail by Nieburgs (13). Patients with widespread metastaticcancer could have microscopic metastases in the urinarytract to account for the presence of tumor cells in the urine.In 1 case in this series, the positive urine cytology occurredin a patient with primary lung cancer without evidence ofdistant metastases.

Examples of false-positive cytology are illustrated in Figs.1 through 4, and comparison may be made between themand a typical cytological membrane preparation with thediagnosis of transitional cell carcinoma (Figs. 5 and 6).In all of the false-positive cases, there is a similarity Innuclear structure and cell arrangement. The nuclear detail

2801AUGUST 1977

on May 20, 2018. © 1977 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Correlation of positive cytological diagnosis and tissuediagnosis—______Cytological

diagnosisTransi

Squa- Adeno-Mixedtionalmouscell card- carci Pros Meta

cell car card- nomanomataticstaticcinomanoma(blad- (blad carci carci

Histological diagnosis(bladder)(bladder) der)der)nomanomaTransitional

cellcarcinoma732(bladder)Squamous

cell carcinoma4134(bladder)Adenocarcinoma

(bladder)2Mixedcarcinoma(bladder)132Prostatic

carcinoma917Metastaticcarcinoma1115Transitional

cellcarcinoma3(pelvisor ureter)

w. j. Frable et a!.

Table 4

is not usually as well preserved in the false-positive cases.The differences between cystoscopic unines and bladderwashings are slight. Quantitatively, there are usually fewerabnormal cells in the false-positive cases, but if membranefilters are used these quantitative differences disappear ordepend upon the amount of specimen used in preparing thespecimen for study.

This laboratory has reported its abnormal cytology in asspecific a diagnostic terminology as possible, using bothancillary clinical information when available and the cellblock material from bladder washing specimens. Evenwhen cytology allowed only a diagnosis of suspected carcinoma, an effort was made to determine the histologicaltype. Comparing only the positive diagnoses of carcinoma,primary or metastatic to the urinary tract, with the comesponding histology, Table 4 reveals the degree of specificityof cytology in this laboratory. Thus, 112 of 150 (74%) of thepositive samples were cytologically and histologically identidal. This is a slightly better result than the specificity ofcervical vaginal cytology in the same laboratory. The degreeof specificity in this particular investigation is undoubtedlyinfluenced by the care taken in procuring the specimen andthe simultaneous comparison of cystoscopic urine, bladderwashing, and cell block material when making a cytologicaldiagnosis. In essence, cancer can be diagnosed from urinespecimens, but the type of tumor can usually not be established by cytology alone.

Discussion

Obviously, the combination of factors previously outlinedresults in improved diagnostic accuracy of urinary cytology.Probably most important is improved preservation of celluIan detail through collection of both cystoscopic urine andbladder washings and their prompt preparation for cytological examination. Urine is a very inhospitable fluid for cells,but the basis for the deleterious effect is not known (10).Kern and Bales (8) attempted unsuccessfully to correlatecell size with urine tonidity, urine pH, and the number of inflammatory cells. Quick removal of the fluid portion of urineand prompt fixation of the freshly desquamated cells are the

main reasons for improved preservation. This must be thereason that bladder washings provide the most satisfactoryspecimen for cytological study.

Even when excellent specimens are available, the numberof false-positive diagnoses reported varies from less than1°/oto more than 20°/o(2, 5, 14, 15, 17, 19). Most series showa range of 5 to 15% false-positive cases, although the authons seldom indicate how the rate is calculated (i.e., as apercentage of all cancer diagnoses or a percentage of allspecimens examined). The latter method gives very impressive but unrealistically low rates. The causes of false-positive diagnoses in the reported series are the same as thosethat plagued the present study, namely, benign prostatichyperplasia, infections of the urinary tract, and calculi (2, 5,17,18).

False-negative diagnoses are also a problem, and usefulinformation in this regard is difficult to evaluate from theliterature. Long-term follow-up of a large series of negativecases is required. Series with low false-positive rates (highspecificity of diagnosis) have high false-negative rates (poorsensitivity to the detection of cancer) (17). If papillomas andGrade I noninvasive transitional cell carcinomas comprise asignificant part of the series, the rate of false-negative diagnoses will be high, up to 66% of papillomas and 30% ofGrade I carcinomas in 1 series (2). Esposti et a!. (3, 4)reported in 2 separate series false-negative cytology withGrade I tumors of 45 and 25%. Bladder washings in thepresent series allowed recognition of 2 of 3 well-differentiated noninvasive transitional cell carcinomas, but 2cases of papilloma were not recognized by us. Harris et al.(6) did not recognize 1 case of papilloma in a large series ofbladder washings from symptomatic patients. It is of interest from Kern's cell measurements that there is only a 5.0-

@maverage difference in cell and nuclear diameter betweenGrades I and II versus Grades Ill and IV transitional cellcarcinoma (9). False-negative cytology in low-grade lesionsis therefore not unexpected.

Umbrella cells are seen in papillomas, low-grade carcinomas, and fragments of bladder epithelium dislodged fromthe normal bladder during washing. The loss of that cellarrangement in some cases of cystitis and calculi probablyaccounts for false-positive diagnoses (Figs. 3 and 4).

2802 CANCERRESEARCHVOL. 37

on May 20, 2018. © 1977 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Cytology of Bladder Cancer

False-positive diagnoses, although higher here than inother areas of cytology, are probably not important. A positive cytological diagnosis of bladder cancer does not leaddirectly to cystectomy or other radical therapy. Patients inthis series were symptomatic, rnost of them with at leasthematunia. Some type of urinary tract disease was expected. False-negative cytology in such a situation can bemore damaging. As Levy and Jerusalem (11) have determined, false-negative cytology should be low even in thescreening situation in the totally asymptomatic, non-highrisk population. In fact, it should be so low (2/1000screened in his series) that routine screening from such apopulation is not indicated (11). It may not be indicatedbecause the false-positive rate in such a population couldbe high.

Although not part of the bladder cancer project, ureterand pelvic tumors were included in the analysis of cytologybecause they make up part of the differential diagnosis.These cases proved difficult, with a false-positive error rateof 5O%. Pelvic or ureteral calculi produced clusters of cellsthat the authors could not distinguish from well-diffenentiated transitional cell carcinoma. In 1 preliminary report,Bibbo et al. (1) did not have any false-positive diagnoses ofcancer from 20 uretenal brushing specimens. There were 8cancers in that series, all diagnosed from the cytology or, in1 case, from the histology of the cell block material (1). Fivecases had infections of the kidney and renal pelvis, and 4had calculi, so that there was ample opportunity for error(1).

A final and important part of the cytology studies withinthe Bladder Cancer Project is the follow-up of patientstreated for bladder cancer. The literature indicates somesuccess with the prediction of persistent and/or recurrenttumor following radiation, chemotherapy, or surgery. Recumrence may be detected well in advance of clinical orhistological evidence of tumor (3, 15, 16). Diagnosis depends on finding anaplastic tumor cells without the appearance of radiation or chernothenapeutic effects on even normal transitional epithelial cells (12). Esposti et al. (3) followed 182 bladder cancer patients treated by X-ray withdetection of 27 out of 29 nonresponders, 52 of 57 patientswith response followed by recurrence, and a 2.5% falsepositive error. There was an indeterminate group of 24cases in this series, 13 with cytology positive for recurrenceand 3 other patients suspected of having recurrence. Persistent tumor (on recurrence?) was detected from cytological studies between 3 and 24 months prior to clinical onhistological proof (3). From a smaller series reported byOrell (14), the data are similar.

In the authors' series there are 17 cases for analysisfollowing therapy by radiation, surgery (partiai cystectomy,fulguration), chemotherapy, or a combination of these.There are 2 negative reports, both from recurrent papillarywell-differentiated transitional cell carcinoma. Two othercases of this tumor were detected, along with 6 cases ofinvasive transitional cell carcinoma, 1 invasive squamouscell carcinoma, and 2 invasive mixed carcinomas of transitional and squamous type. These recurrences have beenidentified cytologically on an average of 4 months prior toclinical or histological detection. One case had a positivecytology 16 months before the recurrence was found. Cys

toscopy was frequently normal, or the changes in the bladden mucosa were minimal and nonspecific.

Particularly deceptive in follow-up cytology is chernotherapy with tniethylenethiophosphoram ide , which may destroylow-grade papillary tumors, leaving a normal-appearing unnary bladder. Tumor cells may still exfoliate from such abladder along with bizarre-appearing cells that representbenign but polyploid transitional cells. Recent experimentalwork by Murphy et@ shows that the effects of this drug onnormal urotheliurn are very transient. Histologically, thebladder returns to normal rapidly, after an initial short penod of cell swelling. Malignant-appearing cells are not exfoliated from the normal bladder following instillation oftniethylenethiophosphoramide (thio-tepa).3 In a patient withbladder cancer treated with this drug, there is a tendency toregard abnormal cells as indicating a drug effect, probablybecause this relationship has been established for otherchemotherapeutic agents instilled into the bladder (10).That is probably not true with triethylenethiophosphoramide. When abnormal cells are seen following treatmentwith this drug, they correlate with recurrent or persistenttumor, even if the bladder looks essentially normal.

Acknowledgments

This study and report could not have been realized without the cooperation of the entire faculty and housestaff of the Department of Urology.

References

1. Bibbo, M., Gill, W. B. Harris, M. J., Lu, C. T., Thomson, S., and Wied, G.L. Retrograde Brushing as a Diagnostic Procedure of Ureteral, RenalPelvic and Renal Calyceal Lesions. A Preliminary Report. Acta Cytol. , 18:137-141, 1974.

2. de voogt, H. J., and Wielenga, G. Clinical Aspects of Urinary Cytology.Acta Cytol., 16: 349-351 , 1972.

3. Esposti, P., Moberger, G. , and zajicek, J. The Cytological Diagnosis ofTransitional Cell Tumors of the Urinary Bladder and Its Histologic Basis.A Study 562 Cases of Urinary Tract Disorders Including 170 Untreatedand 182 Irradiated Bladder Tumors. Acta Cytol., 14: 145-155, 1970.

4. Esposti,P.,andZajicek,J. GradingofTransitionalCellNeoplasmsof theUrinary Bladder from Smears of Bladder Washings. Acta Cytol. , 16: 529-537, 1972.

5. Graham,R.TheCytologicDiagnosisof Cancer,Ed.3, pp. 332.Philadelphia:W. B.SaundersCo..1972.

6. Harris, M. J., Schwinn, C. P., Morrow, J. W., Gray, R. L., and Browell, B.M. Exfoliative Cytology of the Urinary Bladder Irrigation Specimen. ActaCytol, 15: 385—399,1971.

7. Keebler, C. M., Reagan, and Wied, G. L. Compendium on Cytopreparatory Techniques. Tutorials of Cytology, Vol. 3, pp. 26-34, 41-45. Chicago: University of Chicago, 1974.

8. Kern, W. H., and Bales, C. E. Quantitative Studies of Urine Cytology. Am.J.dIm,Pathol.51:225-228,1969.

9. Kern,W. H., Bales,C. E., andWebster,W. W. CytologicalEvaluationofTransitional Cell Carcinoma of the Bladder. J. Urol., 100: 616—622,1968.

10. Koss, L. G. Diagnostic Cytology and Its Histopathologic Basis, Ed. 2, pp.404-449. Philadelphia: J. B. Lippincott, Co., 1968.

11. Levy, E., and Jerusalem, K. Experience with Cytologic Examination ofUrines with the Assistance of Multiple Parallel Methods with and withoutClinical Indication of Tumor. Acta Cytol., 17: 121-124, 1973.

12. Loveless,K. J. The Effectsof Radiationupon Cytologyof Benign andMalignant Bladder Epithelia. Acta Cytol., 17: 355-360, 1973.

13. Nieburgs, H. E. Diagnostic Cell Pathology in Tissues and Smears, pp.158-175.NewYork: Grune& Stratton,1967.

14. Orell, S. R. Transitional Cell Epithelioma of the Bladder Correlation ofCytologic and Histologic Diagnosis. Scand. J. Urol. Nephrol., 3: 93-98, 1969.

3 W. M. Murphy, M. S. Soloway, and C. J. Lin. Morphological Effects of

Thio-tepa on Urothelium, submitted for publication to theJournal of Urology.

AUGUST 1977 2803

on May 20, 2018. © 1977 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

w. j. Frable et a!.

15. Park, C., Britsch, C., ljson, A. C., and Veenema, R. J. Reliability ofPositive Exfoliative Cytology of Urine in Urinary Tract Malignancy. J.Urol.,102:91-92,1969.

16. Reschborn-Kjennerud, S., and Hoeg, K. The Value of Urine Cytology inthe Diagnosis of Recurrent Bladder Tumors. A Preliminary Report. ActaCytol.,16:269-272,1972.

17. Sarnacki, C. T., McCormack, L. J., Kiser, W. J., Hazard, J. B., Mc

.@ “@*i

, 4

C

I

Laughlin, T. C., and Belovich, 0. J. Urinary Cytology and the ClinicalDiagnosis of Urinary Tract Malignancy. A Clinico-pathologic Study of1400 Patients. J. Urol., 106: 761-764, 1971.

18. Theologidis,A. D., Jamerson,R. M., and Scott, A. The ReliabilityofUrinary Cytology. Brit. J. Urol., 43: 598-602, 1971.

19. Wiggishoft,C. C., and McDonald,J. H. UrinaryExfoliativeCytologyinthe Diagnosisof BladderTumors.ActaCytor, 16: 139-141, 1972.

a

I

“4

$1

•4

1'

I

1 2

Fig. 1. Bladder washing with papillary clusters of atypical transitional cells reported as indicating low-grade papillary transitional cell carcinoma. Patienthas only benign prostatic hyperplasia. Papanicolaou, x 600.

Fig. 2. Bladder washing with severely atypical papillary clusters of transitional cells reported as carcinoma. Patient with metastatic lymphoepitheliomabeing treated by chemotherapy. Cells could represent cancer-associated atypia or chemotherapeutic effect on transitional epithelium. Papanicolaou, x 600.

Fig. 3. Cystoscopy urine from a patient with hematuria. Papillary clusters of atypical transitional cells suggesting a low-grade papillary transitional cellcarcinoma. Multiple biopsies indicate only chronic follicular cystitis. Papanicolaou, x 600.

Fig. 4. Bladder washing. Patient with hematuria. Papillary clusters of malignant-appearing transitional cells indicating papillary transitional cell carcinoma. Bladder mucosa appears red and inflamed. Multiple biopsies reveal only severe chronic cystitis. Papanicolaou, x 600.

Fig. 5. Bladder washing. Malignant-appearing cells from papillary transitional cell carcinoma, moderately well differentiated, noninvasive: for comparisonwith Figs. 1 through 4. Papanicolaou, x 600.

Fig. 6. Same case as Fig. 5. Papanicolaou, x 600.

2804 CANCER RESEARCH VOL. 37

on May 20, 2018. © 1977 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Cytology of Bladder Cancer

C

4@@

I,

4@

C

.@*@

V

ti :.@,

a

1'

@4

y

) : :11

qj'@

C,.

6.

•1

a

S

I

5.&4AUGUST 1977 2805

p.

on May 20, 2018. © 1977 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

1977;37:2800-2805. Cancer Res   William J. Frable, Lois Paxson, Jo A. Barksdale, et al.   Current Practice of Urinary Bladder Cytology

  Updated version

  http://cancerres.aacrjournals.org/content/37/8_Part_2/2800

Access the most recent version of this article at:

   

   

   

  E-mail alerts related to this article or journal.Sign up to receive free email-alerts

  Subscriptions

Reprints and

  .pubs@aacr.orgDepartment at

To order reprints of this article or to subscribe to the journal, contact the AACR Publications

  Permissions

  Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)

.http://cancerres.aacrjournals.org/content/37/8_Part_2/2800To request permission to re-use all or part of this article, use this link

on May 20, 2018. © 1977 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from