A

ibd

ls

tr3(fl9

nE

K

1

mpbt

9

1d

Original articleA prognostic index (bladder prognostic index) for bilharzial-related

invasive bladder cancer

Hussein Khaled, M.D.a,*, Omar El Hattab, M.D.b, Doaa Abdel Moneim, M.D.a,Hatem Aboul Kassem, M.D.c, Ahmed Morsi, M.D.c, Ghada Sherif, M.D.b,

Tarek Darwish, M.D.a, Rabab Gaafar, M.D.a

a Department of Medical Oncology, National Cancer Institute, Cairo University, Cairo, Egyptb Department of Cancer Epidemiology and Biostatistics, National Cancer Institute, Cairo University, Cairo, Egypt

c Department of Surgical Oncology, National Cancer Institute, Cairo University, Cairo, Egypt

Received 20 September 2004; received in revised form 11 March 2005; accepted 14 March 2005

bstract

Purpose: Bladder cancer is still the most common solid tumor among adult males in Egypt because of the prevalence of bilharzialnfestation, especially in the countryside. In this prospective study, we have recorded the prognostic factors for 180 patients with invasiveladder cancer for whom standard radical cystectomy had been performed to develop a prognostic index (bladder prognostic index) thatefines high risk patients who are more vulnerable to disease relapse after surgery and who may benefit from additional therapy.Patients and Methods: The study was performed between January 1997 and December 1999, in which 180 patients with histopatho-

ogically proved invasive bladder cancer associated with bilharziasis underwent radical cystectomy or anterior pelvic exenteration. Afterurgery, patients were regularly followed for a minimum of 2 years.

Results: Our patients included 141 males and 39 females. Squamous cell carcinoma was the most common type (53.3%), and most ofhe tumors were grade II (61.1%). A total of 173 patients had their tumors operable, while 7 were inoperable. We had 5 (2.8%) operativeelated mortalities. At 5 years postoperatively, free and overall survival rates for the whole group of patients were 31.44% � 5.9% and2.5% � 6.8%, respectively. Tumor pathologic stage, grade, and nodal affection were the only significant factors with impact on survivalP � 0.008, 0.051, and 0.004, respectively). These 3 prognostic indexes were used to design a model to predict an individual patient’s riskactor for recurrence. Patients were then assigned to one of the 4 risk groups according to the score achieved in this prognostic index (0 �ow risk, 1 � intermediate risk, and 2 or 3 � higher risk). These 4 risk groups had distinctly different rates of disease-free survival, i.e.,1.7%, 53%, 13%, and 7% for low, intermediate, and higher risk groups, respectively.Conclusion: Although this prognostic index appears to be of a significant clinical relevance, it needs to be more validated on a larger

umber of patients, and it could be a surrogate variable for biologic factors responsible for the heterogeneity of bladder cancer. © 2005lsevier Inc. All rights reserved.

Urologic Oncology: Seminars and Original Investigations 23 (2005) 254–260

eywords: Bladder cancer; Bilharziasis; Prognostic index; Egypt

ldpcmi7Hf

. Introduction

Carcinoma of the urinary bladder is still the most com-on solid tumor among adult males in Egypt [1]. The

rocess of bladder carcinogenesis is related to the associatedilharzial infestation, where Egypt has a high prevalence ofhis parasite, especially in the countryside. Bilharzial-re-

* Corresponding author. Tel.: �2012-215-1040; fax: �202-368-711.

fE-mail address: Khaled@internetegypt.com (H. Khaled).

078-1439/05/$ – see front matter © 2005 Elsevier Inc. All rights reserved.oi:10.1016/j.urolonc.2005.04.005

ated bladder cancer has a clinicopathologic pattern thatiffers from that seen in Europe and North America [2]. Theeak age of individuals diagnosed with bilharzial bladderancer in Egypt is younger (50 � 5 years), with a higherale-female ratio (5:1), and most tumors present bulky

nvasive lesions with deep muscle infiltration in more than5% of cases, and are mainly of the squamous cell variety.owever, a recent trend toward a relative increase in the

requency of transitional cell variety has been reported [3].Radical cystectomy is the standard treatment of choice

or invasive bladder cancer in Egypt. It achieves the best

lohc4hnp

atottbca

2

bhsNtgaputu

2

rvctedwtptsoticppot

bAetafwpTf

ptpstmhdnbw

2

y5fvtlwdDcppi

3

3

fimyat

l

255H. Khaled et al. / Urologic Oncology: Seminars and Original Investigations 23 (2005) 254–260

ocal control of invasive bladder cancer whether associatedr not with schistosomiasis. The results of this operationave been reported in a number of large series, with quiteonsistent results of 5-year survival rates between 35% and5%. Different reports clearly showed that tumor stage,istologic grade, and the nodal status are independent prog-ostic factors after adjusting for all other relevant clinico-athologic characteristics [4].

Efforts investigating the value of adding neoadjuvant ordjuvant radiotherapy and/or chemotherapy to improvereatment outcome and increase the low survival rates arengoing [5]. Therefore, the need for developing a prognos-ic model to define high risk groups that are more vulnerableo disease relapse after radical cystectomy and, thus, mayenefit from additional preoperative therapy is highly indi-ated, and may explain the controversial data of such ther-py.

. Patients and methods

This prospective study was open for the patient’s entryetween January 1997 and December 1999. Patients withistopathologically proved invasive bladder cancer pre-ented to the Surgical Oncology Department at the Cairoational Cancer Institute were eligible for entry. Preopera-

ive evaluation studies included routine laboratory investi-ations, renal and liver function assessment, chest x-ray,bdominal ultrasound, bone scan, and intravenous urogra-hy. Preoperative evaluation of involvement of the prostaticrethra was reviewed during cystoscopy. In all patients, theumor had invaded into the muscle layer. Radiologic eval-ation for metastasis was negative in all patients.

.1. Surgical procedure

A total of 180 patients underwent exploration for en blocadical cystectomy or anterior pelvic exenteration, with pel-ic lymph node dissection and urinary diversion. Radialystectomy performed in men involved en bloc removal ofhe urinary bladder with its perivesical fat, peritoneal cov-ring, lower ureters, prostate, seminal vesicles, pelvic vaseferens, and proximal urethra. The pelvic lymph nodesere also removed, including the distal common iliac, in-

ernal, and external iliac lymph nodes. In this procedure, thelane of dissection followed the pelvic wall starting fromhe aortic bifurcation down to the inguinal ligament, withkeletonization of the iliac vessels, and complete clearancef the obturator fossa saving the obturator nerve. The ure-hra was removed only in cases proved to have tumornvolvement of the prostatic urethra. In women, radicalystectomy was in effect an anterior pelvic exenterationrocedure involving removal of the urinary bladder with itseritoneal covering, entire urethra, uterus, fallopian tubes,varies, anterior vaginal wall with pelvic cellular tissue, and

he aforementioned lymph nodes. T

Patients were staged both clinically and pathologicallyy the American Joint Committee-International Uniongainst Cancer System. All cystectomy specimens were

xamined using the same pathologic protocol. Tissue sec-ions were obtained from the tumor, bladder wall, ureters,nd regional lymph node. In males, tissues were also studiedrom the seminal resides and prostate. In females, sectionsere obtained from the ovaries and uterus as well. Theresence or absence of bilharzial eggs was also evaluated.he World Health Organization Classification was adopted

or histopathologic typing and grading of tumors.The clinicopathologic features evaluated for potential

rognostic importance included age, sex, bilharzial history,umor pathologic subtype and grade, pathologic stage, andelvic lymph node affection by disease. After the primaryurgery, regular follow-up was performed bimonthly duringhe first year, every 3 months during the second year, and 6onths thereafter. During each visit, clinical assessment,

epatic, and renal function profiles were performed. Ab-ominopelvic ultrasound, x-ray of the chest, and intrave-ous urography were performed routinely on an annualasis. Isotopic bone scan and computerized tomographyere also performed according to the clinical need.

.2. Statistical methods

The minimum follow-up for surviving patients was 2ears, with a median observation time of 21 months. The-year disease-free and overall survival rates were per-ormed using the Kaplan-Meier method. Disease-free sur-ival was defined as the period from the date of surgery tohe date of detection of the first disease recurrence, eitherocal or systemic, death or to the date of last evaluation;hile overall survival was defined as the interval from theate of diagnosis to the date of death or last evaluation.eath from any other cause than disease-related causes was

onsidered censored. The log-rank test was used for com-arison between survival curves. The Cox multivariate pro-ortional hazards analysis was applied to determine thendependent risk factors for disease recurrence.

. Results

.1. Patient characteristics

Descriptive statistics representing the clinicopathologiceatures of the patients included in the study are illustratedn Table 1. There were 141 males and 39 females with aale-female ratio of 3.6:1. There were 29 patients who were

ounger than 45 years, while 99 and 52 patients were in thege groups 45�60 years and older than 60 years, respec-ively.

Squamous cell carcinoma was the most common patho-ogic type, and it accounted for 53.3% (96) of patients.

ransitional cell carcinoma accounted for 38.9% (70) of

pcpi

pdecpnpgBsp

wpc6g5rdaol(

TD

* Presence of bilharzias ova in the pathology samples.

Fig. 1. Disease-free survival am

256 H. Khaled et al. / Urologic Oncology: Seminars and Original Investigations 23 (2005) 254–260

atients, adenocarcinoma 5% (9), and undifferentiated car-inoma 2.7% (5). Tumors were pathologic grade II in 110atients (61.1%), while grades III and I tumors were presentn 50 (27.8%) and 20 (11.1%) patients, respectively.

Radical cystectomy was performed in males, anteriorelvic exenteration was performed in females, and urinaryiversion was possible in 173 patients, while 7 were inop-rable because of fixed posterolateral extension. Neurovas-ular bundles were preserved in 28% of the explored maleatients (40 of 141). Pathologic stages P2–P3 were diag-osed in 153 of 173 patients who could be assessed forathologic tumor staging, while histologic evidence of re-ional lymph node involvement was seen in 28 (16.6%).ilharzial eggs were seen in the majority of the excised

pecimens (79.7%). However, the relative frequency of theirresence was higher among squamous tumors.

We had 5 (2.8%) operative mortalities (defined as deathithin 30 days of surgery), including 2 patients who died ofulmonary embolism, 2 of liver failure, and 1 as a result ofardiogenic shock secondary to myocardial infarction. The-year disease-free and overall survival rates for the wholeroup of 180 patients included in the study was 31.44% �.9% and 32.5% � 6.8%, respectively (Fig. 1). Diseaseecurrence, including local, distant, or combined local andistant, occurred in 43 patients. Sites of disease recurrencefter radical cystectomy, occurring alone or combined withther sites, included bone metastases in 15 (43.9%) patients,ocal recurrence in 13 (30.2%), liver metastases in 716.3%), pulmonary involvement in 6 (14%), soft tissue in

able 1escriptive statistics

Number of patients (%)

Age (yrs)�45 29 (16.1)45–60 99 (55)�60 52 (28.9)

SexMale 141 (78.33)Female 39 (21.67)

HistopathologySquamous cell 96 (53.33)Transitional cell 70 (38.89)Adenocarcinoma 9 (5)Undifferentiated 5 (2.78)

GradeI 20 (11.11)II 110 (61.11)III 50 (27.78)

StageP1 2 (1.2)P2 10 (5.8)P3a 58 (33.5)P3b 83 (48)P4a 14 (8)P4b 6 (3.5)

Lymph nodesNegative 141 (83.4)Positive 28 (16.6)NA 11

Bilharziasis*Negative 33 (20.3)Positive 129 (79.7)NA 18

Abbreviation: NA � nonapplicable.

ong the whole group.

5(

aitad

3

epUpo0

nrswlf�

TR

A

S

H

G

S

L

B

* Presence of bilharzias ova in the pathology samples.

TT

AA

S

H

G

S

L

B

* Presence of bilharzias ova in the pathology samples.

257H. Khaled et al. / Urologic Oncology: Seminars and Original Investigations 23 (2005) 254–260

(11.6%), and para-aortic lymph node metastases in 37%).

Meanwhile, 86 (47.8%) patients were alive at evaluation,nd the remaining 94 (52.2%) were dead. Causes of deathncluded disease recurrence in 43 patients, inoperability ofhe tumor in 7, operative mortality in 5, intractable ascitesnd liver failure in 11, and renal failure in 6. The cause ofeath was unknown in 20 patients.

.2. Independent risk factors

Tumor recurrence and the estimates of the 5-years dis-ase-free and overall survival in relation to the tumor andatient characteristics are summarized in Tables 2 and 3.nivariate analysis of the data indicated that tumor grade,athologic stage, lymph node involvement, and the presencef bilharziasis had a significant impact on survival (P �.009, 0.0003, 0.0009, and 0.038, respectively).

The patient and tumor characteristics, which had a sig-ificant impact on survival, were then evaluated by the Coxegression model (Table 4). With multivariate analysis, theignificance of bilharzias affection in relation to survivalas not sustained. With such analysis, the tumor’s patho-

ogic stage, grade, and nodal disease remained as the onlyactors with a statistically significant impact on survival (P

0.008, 0.051, and 0.004, respectively).

tors

(%) P Value Overall survival (%) P Value

— 32.5 � 6.8 —

0.14 35.8 0.2735.724.5

0.16 27.7 0.2548.4

39.40.16 24.3 0.37

40.038.13

0.002 73.5 0.00928.6

0.0003 56.7 0.000312.3

0.0003 38.2 0.000911.3

0.055 30.02 0.03840.8

able 2ecurrence in relation to prognostic factors

Recurrencenumber of patients (%)

P Value

ge (yrs)�45 7 (25.93) 0.16945–60 27 (31.40)�60 9 (16.98)

exMale 33 (25.58) 0.859Female 10 (27.03)

istopathologySquamous cell 20 (25.97)Transitional cell 18 (27.69) 0.931Adenocarcinoma 2 (22.22)Undifferentiated 3 (20.00)

radeI 0 (00.00) 0.006II, III 43 (29.25)

tageP1-3a 9 (12.86) 0.001P3b-P4 34 (36.56)

ymph nodesNegative 30 (21.90) 0.014Positive 12 (44.44)

ilharziasis*Negative 13 (40.63) 0.025Positive 25 (21.19)

able 3he 5-year disease-free survival and overall survival in relation to prognostic fac

Number of patients Disease-free survival

ll patients 180 31.44 � 5.9ge (yrs)�45 29 31.645–60 99 36.8�60 52 23.3

exMale 141 25.9Female 39 49.8

istopathologySquamous cell 96 39.4Transitional cell 70 22.7Adenocarcinoma 9 40.0Undifferentiated 5 36.9

radeI 20 73.0II, III 160 27.4

tageP1-3a 70 54.9P3b-P4 103 11.93

ymph nodesNegative 141 38.2Positive 28 6.4

ilharziasis*Negative 33 30.2Positive 129 39.97

3

iatamrtwptrrfrFd

as

4

bshoTcTl�ltb

TM

LBGSP

258 H. Khaled et al. / Urologic Oncology: Seminars and Original Investigations 23 (2005) 254–260

.3. Bladder Prognostic Index (BPI)

Thus, the 3 pretreatment characteristics that remainedndependently significant in relation to disease recurrencend survival were pathologic stage (P1�3a vs. P3b�4),umor grade (I vs. II and III), and nodal involvement (neg-tive vs. positive). These 3 factors were used to design aodel to predict an individual patient’s risk of death. The

elative risk of death could be characterized by calculatinghe number of risk factors present at diagnosis. Risk groupsere defined by comparing the relative risk of death inatients with each possible number of presenting risk fac-ors (0, 1, 2, or 3) and combining categories with similarelative risk. Thus, patients were assigned to one of the 3isk groups on the basis of their number of presenting riskactors: 0, low risk; 1, intermediate risk; and 2 or 3, highisk. The survival curves for the 4 risk groups are shown inig. 2 and Table 5. These 4 risk groups had distinctlyifferent rates of disease-free survival: 91.7%, 53%, 13%,

able 4ultivariate analysis results (proportional hazards analysis) of disease-fre

Regressionestimate

Standardof error

ymph node 0.80 0.28ilharziasis* 0.46 0.27rade 1.17 0.60tage 0.75 0.28athology 0.12 0.13

* Presence of bilharzias ova in the pathology samples.

Fig. 2. Disease-free survival in relation to the prognostic g

nd 7% for low, intermediate, and higher risk groups, re-pectively.

. Discussion

Treatment protocols for invasive bladder cancer areased on clinical, radiologic, endoscopic, and pathologictaging of the disease. Radical cystectomy can result in aigher rate of cure, as reported in all series. The reportedverall survival ranges between 48% and 53% [6,7].hese data are related to cases with transitional cellarcinoma, which did not receive any adjuvant therapy.he overall survival rate after cystectomy and pelvic

ymphadenectomy in the current series was 32.5%6.8%. This rate is lower than that reported in the

iterature, and could be explained by the large number ofhe high-grade and advanced stage of primary tumorsecause more than 90% of the patients had tumors more

al

95% Confidence interval P Value

1.28 3.89 0.0040.37 1.07 0.0850.99 10.5 0.0511.22 3.67 0.0080.68 1.15 0.367

e surviv

roup. (Color version of figure is available online.)

tgsradt

mswtc[ibs[sP1tthcba

mc5a4lsvsp

tlltitaa

Ttadtmnts

nkintpFpirbhb

R

[

[

TT

0123

259H. Khaled et al. / Urologic Oncology: Seminars and Original Investigations 23 (2005) 254–260

han or equal to P3a, with 88.9% more than or equal torade II. This result indicates that despite the extent ofurgery, additional modalities are needed to improve theesults, particularly in high stage, high-grade tumors. Itlso emphasizes the need to develop a prognostic index toefine such patients who will need more aggressiveherapy.

The impact of tumor grade on survival was confirmed byany studies [4,6,7,10]. In the present study, the overall

urvival rate of patients with grade I tumors was 73.5%,hile that of those with grade II and III tumors combined

ogether was 28.6%. These results are of statistical signifi-ance and go well with the results reported in the literature8]. The reported long-term survival rates for patients hav-ng tumors with superficial muscle invasion (P2) rangeetween 65% and 76%. With further stage progression, theurvival expectancy was markedly decreased. Frazier et al.9] reported a 5-year survival of 39% for patients withtages P3 and more. Pagano et al. [10] stratified cases into3a, P3b, and P4; they reported a 5-year survival of 50%,5%, and 21%, respectively. The same investigators noticedhat patients with stage P4 with contiguous involvement ofhe prostate had a 5-year survival of only 6%. Our resultsad the same conclusion and have indicated that tumorsonfined to the muscle wall of the bladder (P2 and P3a) hadetter survival than extending through bladder wall (P3bnd P4), being 56.7% and 12.3%, respectively.

The reported incidence of regional lymph nodes involve-ent varied between 14% and 27% [11]. This incidence

orrelates with the P-stage of the disease [12]. The reported-year survival after radical cystectomy and pelvic lymph-denectomy in patients with positive nodes ranged between% and 29% [13]. In the present study, the incidence ofymph node involvement was 16.6%, and the disease-freeurvival was 6.4% for patients with positive nodes. Multi-ariate analysis of our results proved that tumor grade,tage, and lymph node involvement remain to be the inde-endent factors for prognosis in bilharzial bladder cancer.

Lerner et al. [14] and Vieweg et al. [15] reported thathe survival advantage of radical cystectomy with pelvicymph node dissection was more evident in patients withow stage primary tumors. Smith et al. [16] emphasizedhe importance of the number of involved nodes, with anncreased risk of recurrence and death from cancer ifheir number is 6 or more. Therefore, Skinner [17] hasdvocated that an extended lymphadenectomy up to the

able 5he 5-year disease-free survival in relation to prediction groups

Groups Number of patients Disease-free survival (%) P Value

12 91.6765 53.01 0.00172 12.9931 6.95

ortic bifurcation is necessary. However, Wishnow and

enney [18] suggested that a restricted lymphadenec-omy up to the bifurcation of the common iliac artery isdequate in patients who have no evidence of nodalisease grossly during cystectomy. In the present study,he level of lymph node dissection was to the mid com-on iliac artery. In our view, the presence of positive

odes above the common iliac bifurcation indicates thathe disease is no more locoregional, and extended dis-ection alone would not be sufficient for disease control.

Although many predictive models for different malig-ant tumors were previously reported [19,20], to ournowledge, this study is the first to report on a prognosticndex for invasive bladder cancer, both bilharzial andon-bilharzial related. However, the prognostic value ofhis index needs to be validated in a larger number ofatients and in different centers treating such patients.inally, it is important to recognize that the clinicalrognostic features included in the bladder prognosticndex are, most probably in part, surrogate variables thateflect the biologic heterogeneity of bladder cancer. Theiologic factors responsible for this heterogeneity mayelp in future refinements of the current index that isased solely on clinical factors.

eferences

[1] National Cancer Institute. Statistical Report for NCI, 2001. CairoUniversity. Available at: http://www.nci.edu.eg/english/statistics.htm. Accessed August 2004.

[2] El Bolkainy MN, editor. Cancer of urinary tract. In: Topographicpathology of cancer. Cairo (Egypt): NCI, Cairo University, 1998. p.59–63.

[3] Gad El Mawla N, El Bolkainy MN, Khaled HM. Bladder cancer inAfrica: Update. Semin Oncol 2001;28:174–8.

[4] Ghoneim MA, El Mekresh MM, El Baz MA, et al. Radical cystec-tomy for carcinoma of the bladder: Critical evaluation of the result in1026 cases. J Urol 1997;158:393–9.

[5] Khaled HM, Zaghloul MS, Ghoneim M, et al. Gemcitabine andcisplatin as neoadjuvant chemotherapy for invasive bladder cancer:Effect on bladder preservation and survival For the Egyptian BladderCancer cooperative Group. National Cancer Institute. Proceedings of39th ASCO Annual Meeting. JCO 2003;22:411.

[6] Ghoneim MA, Ashamallah AG, El-Hammady S, et al. Cystectomyfor carcinoma of the bilharzial bladder:138 cases 5 years later. Br JUrol 1979;51:541–4.

[7] McCarron JP, Marshall VF. The survival of patients with bladdertumor treated by surgery: Comparative results of an old and a recentseries. J Urol 1979;22:322–4.

[8] Zaghloul MS. Distant metastasis from bilharzial bladder cancer. Can-cer 1996;77:743–9.

[9] Frazier HA, Robertson JE, Dodge RK, et al. The value of pathologicfactors in predicting cancer-specific survival among patients treatedwith radical cystectomy for transitional cell carcinoma of the bladder.Cancer 1993;71:3993–4001.

10] Pagano F, Bassi P, Caletti TP, et al. Results of contemporary radicalcystectomy for invasive bladder cancer: A clinicopathologic studywith an emphasis on the inadequacy of the tumor, nodes and metas-tases classification. J Urol 1991;145:45–50.

11] Wishnow KI, Johnson DE, Rom JY, et al. Incidence, extent and

location of unsuspected pelvic lymph node metastasis in patients

[

[

[

[

[

[

[

[

[

260 H. Khaled et al. / Urologic Oncology: Seminars and Original Investigations 23 (2005) 254–260

undergoing radical cystectomy for bladder cancer. J Urol1987;137:408–10.

12] Zincke H, Patterson DE, Utz DC, et al. Pelvic lymphadenectomy andradical cystectomy for transitional cell carcinoma of the bladder withpelvic nodal disease. Br J Urol 1985;57:156–9.

13] Zungri E, Chechile G, Sarroca J, et al. Value of radical cystectomy inbladder cancer with pelvic node involvement. Eur Urol 1988;14:123–6.

14] Lerner SP, Skinner E, Skinner DG. Radical cystectomy in regionallyadvanced bladder cancer. Urol Clin North Am 1992;19:713–23.

15] Vieweg J, Whitmore WF Jr, Herr HW, et al. The role of pelviclymphadenectomy and radical cystectomy for lymph node positivebladder cancer. The Memorial Sloan-Kettering Cancer Center expe-

rience. Cancer 1994;73:3020–8.

16] Smith JA, Whitmore WF Jr. Regional lymph node metastasis frombladder cancer. J Urol 1981;126:591–3.

17] Skinner DG. Management of invasive bladder cancer. A meticu-lous pelvic node dissection can make a difference. J Urol 1982;128:34 – 6.

18] Wishnow KI, Tenney DM. The results of radical cystectomyfor invasive bladder cancer. Urol Clin North Am 1991;18:529 –37.

19] Goldhirsch A, Wood WC, Gelber RD, et al. Meeting highlights:Updated international expert consensus on the primary therapy ofearly breast cancer. J Clin Oncol 2003;21:3357–65.

20] The Non-Hodgkin’s Lymphoma Prognostic Factors Project. A pre-dictive model for aggressive non-Hodgkin’s lymphoma. N Engl

J Med 1993;329:987–94.