A

d

fwo

iwapsm

tpwn

K

1

cmc[

1d

Original article

Long-term results of primary adenocarcinoma of the urinary bladder:A report on 192 patients

Mohamed S. Zaghloul, M.D.a,b,*, Akram Nouh, M.D.b,c, Mohamed Nazmy, M.D.a,b,Samy Ramzy, M.D.b,d, Ashraf S. Zaghloul, M.D.d, Mohamed Abou Sedira, M.D.d,

Ehab Khalil, M.D.a

a Department of Radiation Oncology, National Cancer Institute, Cairo, Egyptb Minia Oncology Center, Minia, Egypt

c Department of Pathology, National Cancer Institute, Cairo, Egyptd Department of Surgery, National Cancer Institute, Cairo, Egypt

Received 6 January 2005; received in revised form 23 May 2005; accepted 27 May 2005

bstract

Objectives: To evaluate the clinical presentation and treatment end results of primary adenocarcinoma of the urinary bladder, and toetermine the significant independent prognostic factors that determine this outcome.Patients and Methods: Of 3659 patients who underwent cystectomy, 192 had adenocarcinoma of the urinary bladder, with a relative

requency of 5.2%. Most of these patients (68.2%) presented in late stages (P3 � P4). The incidence of pelvic lymph nodes involvementas 25.5%. Mucinous adenocarcinoma was reported in 28 patients (14.6%), papillary in 20 (10.4%), signet ring in 14 (7.3%), while nottherwise specified was reported in 130 (67.7%) in the cystectomy specimens.Results: Mucinous and signet-ring histologic subtypes showed increased frequency of high stages and high grades, and more nodal

nvolvement than the papillary and not otherwise specified. All patients were treated with radical cystectomy and pelvic lymphadenectomyith (69 patients) or without (123) postoperative radiotherapy. The 5-year disease-free survival rate was 46 � 4% for all patients with

denocarcinoma. Postoperative radiotherapy improved the disease-free survival significantly. The 5-year disease-free survival rate for theostoperative radiotherapy group was 61 � 6% compared to 37 � 5% for the cystectomy alone group (P � 0.002). Local control rate wasignificantly improved from 53 � 7% for cystectomy alone to 96 � 3% for postoperative radiotherapy patients (P � 0.00001). Distantetastases were the leading cause of death in the postoperative radiotherapy group.Conclusions: Within the limitations provided by retrospective studies, it could be concluded that postoperative radiotherapy improved

he disease-free survival through its effect on local control. The disease-free survival independent prognostic variables were tumor stage,ostoperative radiotherapy, nodal involvement, and adenocarcinoma subclassification. These factors, except the adeno-subclassification,ere also found to determine the local control rate. On the other hand, the independent prognostic factors for distant metastasis were lymphodal involvement, stage, and adeno-subclassification. © 2006 Elsevier Inc. All rights reserved.

Urologic Oncology: Seminars and Original Investigations 24 (2006) 13–20

eywords: Urinary bladder neoplasm; Adenocarcinoma; Radical cystectomy; Adjuvant radiotherapy; Prognostic factors

wttfbtvpc

. Introduction

Adenocarcinoma of the urinary bladder is classified ac-ording to its origin into 3 categories: primary, urachal, andetastatic [1]. The reported incidence of primary adenocar-

inoma was 0.5% to 2% of all urinary bladder malignancy2]. However, it is more frequently encountered in areas

* Corresponding author. Tel.: �20-10-1720664; fax: �20-2-3644720.

tE-mail address: mszagh@yahoo.com (M.S. Zaghloul).

078-1439/06/$ – see front matter © 2006 Elsevier Inc. All rights reserved.oi:10.1016/j.urolonc.2005.05.027

here bilharziasis is endemic. This incidence ranged be-ween 5% and 11.4% [3–7]. Metaplastic changes of poten-ially unstable urothelium were considered as the causativeactor for the development of adenocarcinoma of the urinaryladder [8]. It was proposed that the metaplastic potential ofhe urothelium has 2 distinct patterns [9]. Progressive in-agination of hyperplastic epithelial buds into the laminaropria (Von Brunn nests) leads to the formation of cystitisystica. Subsequently, metaplasia of the urothelial lining of

hese cysts to columnar mucin-producing cells results in the

pltitea

riusnetmdClschs

ppamebnutbeta

2

jfItwaO(wsesih

tiwlwafimsftot

pwm615pevltp1iossnaltorm

2

ctrwmaowdfrma

14 M.S. Zaghloul et al. / Urologic Oncology: Seminars and Original Investigations 24 (2006) 13–20

roduction of cystitis glandular, which is a premalignantesion [10]. Moreover, cuboidal or columnar metaplasia ofhe surface epithelium can occur with no downward invag-nation. Chronic vesical irritation and infection were men-ioned as predisposing factors for these changes [8,9]. Thisffect may explain the higher incidence of these tumorsmong patients with bilharzial cystitis.

The pathologic characteristics, clinical behaviors, andesponse to different lines of treatment were not clearlydentified, mostly because of the rare incidence and lack ofniversal standardized treatment. The largest 2 publishederies reported on 185 patients with primary adenocarci-oma treated with radical cystectomy and pelvic lymphad-nectomy [6] and 142 patients treated with radical cystec-omy with or without postoperative radiotherapy [7]. Theajority (76%) of patients in the first series had early stage

isease (i.e., P1 and P2, Union Internationale Contre leancer 1997), while the majority (67.6%) of the second had

ate stage disease (i.e., P3 and P4). The 5-year disease-freeurvival rate for patients with early stage disease was ex-ellent, while patients with advanced stages (i.e., P3 and P4)ad a statistically significant lower survival rate in the 2eries.

Postoperative radiotherapy as an adjuvant therapyroved to increase the disease-free survival through im-rovement of the local control of patients with relativelydvanced stages P2b, P3, and P4a with transitional, squa-ous and adenocarcinoma [4,7,11]. Irradiation has shown

fficacy in a series [12] reporting on small numbers ofladder adenocarcinoma, and it is clearly effective in ade-ocarcinomas that occur at other sites (e.g., breast, cervix,terus, prostate, parotid, rectum). The aim of this study waso report on a large series of primary adenocarcinoma of theladder, its clinicopathologic characteristics, and treatmentnd results. All patients were treated with radical cystec-omy and bilateral pelvic lymphadenectomy with or withoutdjuvant postoperative radiotherapy.

. Patients and methods

A total of 3659 patients with bladder cancer were sub-ected to radical cystectomy and pelvic lymphadenectomyrom January 1994 to December 2003, at National Cancernstitute Cairo University and Minia Oncology Center (Oc-ober 1998 to December 2003). The files of the patientsere reviewed, and those who had the pathology of primary

denocarcinoma of the bladder were thoroughly examined.f the 3659 patients who underwent cystectomy, 192

5.2%) had adenocarcinoma. The pathologic specimensere reviewed by the same pathologist (A.N.) and subclas-

ified according to Grignon et al. [13]. Specimens withxtension to the prostate had to prove negative prostate-pecific antigen immunohistochemical stain to be includedn this study. They divided vesical adenocarcinoma into 5

istologic subtypes: (1) enteric (papillary), when the archi- w

ectural and cytologic features resembled those of the typ-cal colonic adenocarcinoma; (2) mucinous, when the tumoras characterized by single cells or nests of cells floating in

akes of extracellular mucin; (3) signet ring, when the tumoras composed of single signet-ring cells diffusely perme-

ting the tissues; (4) adenocarcinoma not otherwise speci-ed, when the pattern did not fit into any of these previouslyentioned categories; and (5) mixed, when the tumor

howed 2 or more patterns with no single pattern accountingor more than 75% of the materials. Detailed characteristics,reatment, and follow-up of these 192 patients were thor-ughly examined because they were the main concern ofhis report.

Radical cystectomy and pelvic lymphadenectomy wereerformed in 133 males, and anterior pelvic exenterationith pelvic lymphadenectomy was the treatment in 59 fe-ales according to El-Sebai [14]. Of the 192 patients, only

9 (35.9%) received postoperative radiotherapy using 6 or5 MV photon, according to the patient’s separation, giving0 Gy/5 weeks/25 fractions using 3 fields (1 anterior and 2osterolateral) technique. The target volume included thentire true pelvis and extending upward from the first sacralertebra to the lower border of both obturator foramina. Theateral borders of the radiotherapy fields were 1 cm lateral tohe pelvic brim. The anterior border included half the sym-hysis pubis, and the posterior border included the anterior/3 of the rectal wall [4,11]. Postoperative radiotherapy wasndicated in pathologic stages P2b, P3, and P4a. The choicef adjuvant postoperative radiotherapy depends entirely onurgeon belief. In the absence of definite evidence, someurgeons believe that adenocarcinoma is radioresistant andot expected to benefit from adding postoperative radiother-py, while others believe the opposite. Patients were regu-arly followed up and examined to detect signs and sites ofreatment failure, depending on clinical findings, radiologicr histopathologic evidence, whenever possible. Follow-upanged between 0.1 and 108 months, with a median of 49onths.

.1. Statistical analysis

Comparison between different percentages and frequen-ies were performed using the standard Student t-test, F-est, and chi-square test. The 5-year disease-free survivalates, local control rates, and distant metastasis-free ratesere measured using the Kaplan-Meier product limitethod [15]. The disease-free survival period was defined

s the period from the date of radical cystectomy to the datef occurrence of any relapse, either local or distant. Patientsho died of any other cause were considered failures ateath. The local control period was defined as the periodrom radical cystectomy to the date of detection of localecurrence. Any other event, either distant metastasis, ure-ia, or death caused by accidental disease was considered

s censored. The period of freedom from distant metastasis

as defined as that from the date of surgery to the detection

olctfmc

3

uclowcmcra

iwescobh

a4b

(wfitsPhuddig(

3

as7ca5fn1

TT

C

MMS

PP

G

N

T

radioth

15M.S. Zaghloul et al. / Urologic Oncology: Seminars and Original Investigations 24 (2006) 13–20

f the first distant metastasis or the date of evaluation. Theog-rank test was used for comparison between survivalurves [16]. The Cox multivariate analysis was performedo determine the independent prognostic variables that af-ected the disease-free survival, local control, and distantetastasis-free rates. Interactions among variables were

onsidered in determining the independent risk factors [17].

. Results

This is a retrospective analysis of 192 patients withrinary bladder adenocarcinoma. They underwent radicalystectomy or anterior pelvic exenteration, and pelvicymphadenectomy with (69 patients) or without (123) post-perative radiotherapy. The urinary diversion proceduresere ileal conduit (58 patients), rectal bladder (87), ileoce-

al continent diversion (16), and urethrocutaneous anasto-osis (31 with bad kidney function and/or fair general

onditions). The male-female ratio was 2.25:1. Their ageanged from 26 to 72 years old, with a mean of 50.4 � 10.9nd a median of 52 years.

Ten patients died postoperatively, within 30 days, lead-ng to a postoperative mortality rate of 5.2%. Three deathsere caused by liver cell failure, and 2 were caused by

sophageal varices as sequelae of the associated bilharzia-is. Another 4 patients died as a result of cardiopulmonaryomplications, while 1 because of peritoneal sepsis. A totalf 99 patients (51.6%) had category PT3 disease, followedy 54 with PT2 disease. Two thirds of the patients (n � 127)

able 1he clinicopathologic characteristics of patients with adenocarcinoma of t

ategory All patients(N � 192)

Mucinous(N � 28)

ean age � SD (yrs) 50.4 � 10.6 51.2 � 9.8ale-female ratio 2.25:1 8.3:1

tageP1 7 (4%) 1 (4%)P2 54 (28%) 6 (21%)

P2a 7P2b 47

3 99 (51%) 18 (64%)4 32 (17%) 3 (11%)P4a 25P4b 7

radeI 21 (11%) 0II 127 (66%) 15 (54%)III 44 (23%) 13 (16%)

odalNegative 143(74%) 19(68%)Positive 49 (26%) 9 (32%)

reatmentCystectomy alone 123 (64%) 19 (68%)PORT 69(36%) 9(32%)

Abbreviations: NOS � not otherwise specified; PORT � postoperative

ad grade II tumors, while 21 and 44 patients had grades I �

nd III, respectively. Pelvic nodal involvement was found in9 patients, giving a nodal involvement rate of 25.5% (Ta-le 1).

Mucinous adenocarcinoma was reported in 28 patients14.6%), signet ring in 14 (7.3%), papillary in 20 (10.4%),hile 130 (67.7%) had the pathology not otherwise speci-ed. The distribution of patients among the different his-

opathologic subtypes of bladder adenocarcinoma did nothow differences of statistical significance (P � 0.642).atients who received postoperative radiotherapy had aigher percentage (81%) of stages P3 and P4 than those whonderwent cystectomy alone (61%). The postoperative ra-iotherapy group also had more patients with grade IIIisease (40% vs. 16%) and a higher incidence of nodalnvolvement (45% vs. 15%) than the cystectomy aloneroup. All these differences were statistically significantTable 2).

.1. Disease-free survival

The 5-year disease-free survival rate of all 192 patients withdenocarcinoma was 46 � 4%. There was no difference inurvival rates between males and females (46 � 5% vs. 49 �%, respectively). The 5-year disease-free survival rate de-reased with the advancement of the level of tumor infiltrations all 7 patients with P1 survived free of disease, while the-year disease-free survival rate decreased steadily from 54%or those with P2 to 35% for those with stage P4 disease. It isoteworthy that patients with stage P2a (n � 7) enjoyed a00% 5-year disease-free survival rate, while those with P2b (n

ary bladder

net ring� 14)

Papillary(N � 20)

NOS(N � 130)

P Value

.9 � 11.2 50.6 � 9.7 49.9 � 114:1 1.9:1 0.021

4 (20%) 2 (2%)(14%) 6 (30%) 40 (31%)

0.007(79%) 6 (30%) 64 (49)1 (7%) 4 (20%) 2 (2%)

3 (15%) 18 (14%)(14%) 15 (75%) 95 (73%) 0.001(86%) 2 (10%) 17 (13%)

(57%) 17(85%) 99(76%)(43%) 3(15%) 31(24%) 0.239

(57%) 15 (75%) 81 (62%) 0.642(43%) 5(25%) 49(38%)

erapy; SD � standard deviation.

he urin

Sig(N

501:1

02

11

02

12

86

86

47) had a survival rate of 44 � 11%. The histologic grade

hsa5

f3mth5ssseip

3

etlsftwhcgwp7

slmda0mbln

3

1T(

3

(tTs

TP

C

MMS

G

N

d

TF

AS

S

G

N

TCPH

16 M.S. Zaghloul et al. / Urologic Oncology: Seminars and Original Investigations 24 (2006) 13–20

ighly affected the disease-free survival because patients withtage G3 had a dismal rate of 38 � 8%, while those with G1nd G2 had a disease-free survival rate of 46 � 11% and 49 �%, respectively.

Nodal involvement had a major negative effect on disease-ree survival, leading to a 5-year disease-free survival rate of1 � 8%, compared to 51 � 5% for those with no involve-ent. Patients with papillary and not otherwise specified sub-

ypes of the adenocarcinoma had a statistically significantigher 5-year disease-free survival rate (68 � 11% and 50 �%, respectively) than that for those with the mucinous andignet-ring histologic subtypes (36 � 10% and 7 � 7%, re-pectively) (Table 3). The difference among these disease-freeurvivals was statistically significant (P � 0.00001). Postop-rative radiotherapy improved the disease-free survival signif-cantly from 37 � 5% for cystectomy alone to 61 � 6% forostoperative radiotherapy (P � 0.002).

.2. Local control and distant metastasis-free rates

The 5-year local control rate was 70%, with no differ-nce between males and females (72% vs. 66%, respec-ively). The pathologic stage had some influence on theocal control rate, although it did not reach the level ofignificance (P � 0.184). The local control rate decreasedrom 100% for patients with stages P1 and P2a to 68% forhose with stages P2b�3�4a. A further decrease to 40%as noticed for patients having stage P4b. However, theistopathologic grade did not influence the 5-year localontrol rates because they were 62%, 71%, and 72% forrades I, II, and III, respectively. The local control ratesere not affected by the metastatic involvement of theelvic lymph nodes (69% for negative involvement and

able 2atients’ characteristics according to treatment groups

ategory Cystectomy(N � 96)

PORT(N � 46)

P Value

ean age � SD (yrs) 50.8 � 9.6 49.3 � 10.2 —ale-female ratio 1.7:1 2.3:1 —

tageP1 7 (6%) 0 0.006P2a 7 (6%) 0P2b 34 (27%) 13 (19%)P3 58 (47%) 41 (59%)P4a 10 (8%) 15 (22%)Pb4 7 (6%) 0

radeI 15 (12%) 5 (13%) 0.05II 92 (75%) 26 (57%)III 16 (13%) 14 (30%)

odal involvementNegative 105 (85%) 38 (55%) 0.0001Positive 18 (15%) 31 (45%)

Abbreviations: PORT � postoperative radiotherapy; SD � standardeviation.

4% for positive involvement). There was also no effect of r

tatistical significance of the pathologic subtypes on theocal control (Table 4). On the other hand, the distant

etastasis-free rate for the 192 patients was 74.7%. Theistant metastasis-free rates were affected by the stage,lthough they did not reach a level of significance (P �.099). The grade of the tumor affected the 5-year distantetastasis-free rates significantly (P � 0.035). Moreover,

oth the involvement of pelvic lymph nodes and the patho-ogic subtypes affected the distant metastasis-free rates sig-ificantly (P � 0.0001 for each) (Table 4).

.3. Radiation acute side effects

A total of 28 patients (41%) had grade 1 diarrhea, while9 had grade 2 (28%). Only 2 patients (3%) had grade 3.here were 41 patients (59%) who had tenesmus grade 1

early rectal reaction).

.4. Late side effects

Intestinal obstruction was experienced by 10 patients8%) who underwent cystectomy alone compared to 8 pa-ients (12%) who received postoperative radiotherapy.here were 4 patients in each group (3.3% and 5.8%, re-pectively, P � 0.4) who required surgical intervention.

able 3ive-year disease-free survival rates in different subgroups

No. ofpatients

5-Yeardisease-freesurvival � SE

P Value

ll patients 192 46 � 4 —exMale 133 46 � 5 0.608Female 59 49 � 7

tageP1 7 100P2a 7 100P2b 47 47 � 10 0.0002P3 99 40 � 6P4a 25 44 � 10P4b 7 0

radeI 21 46 � 11II 127 49 � 5 0.126III 44 38 � 8

odal involvementNegative 143 51 � 5 0.012Positive 49 31 � 8

reatment groupystectomy alone 123 37 � 5 0.002ORT 69 61 � 6istologic subtypeMucinous 28 36 � 10Signet ring 14 7 � 7Papillary 20 68 � 11 0.00001NOS 130 50 � 5

Abbreviations: NOS � not otherwise specified; PORT � postoperative

adiotherapy; SE � standard error.

Oa

ptaoc

3

taw(tgd

3

tdrtpdpnvd

4

t

TF

V

AG

S

G

L

S

radioth

TC

C

LDPUIIU

TMm

V

SPNAG

17M.S. Zaghloul et al. / Urologic Oncology: Seminars and Original Investigations 24 (2006) 13–20

ne patient in each group died, while the other 6 recoveredfter surgery.

Mild deterioration of kidney functions was detected in 9atients (7%) who underwent cystectomy alone comparedo 7 patients (10%) who received postoperative radiother-py. This difference was not statistically significant. Onlyne patient died of uremia 29 months after treatment withystectomy alone.

.5. Causes of treatment failure

Table 5 shows the different causes of failure for the 2herapeutic groups, cystectomy and postoperative radiother-py. The leading cause of failure in the cystectomy groupas local recurrence (32%), followed by distant metastasis

11%) and postoperative mortality (7%). On the other hand,he main cause of death in the postoperative radiotherapyroup was distant metastasis (29%), while local recurrenceecreased to 3% only.

able 4ive-year local control and distant metastasis-free rates in different patien

ariable Local control � SE (%) P

ll patients 70 � 4 —roupCystectomy 53 � 7 0PORT 96 � 3

tageP1 � 2a 100P2b � 3 � 4a 66 � 9 0P4b 50 � 25

radeI 62 � 11II 71 � 5 0III 72 � 9

ymph nodeNegative 69 � 5 0Positive 74 � 7

ubtypesPapillary 75 � 10Mucinous 55 � 15 0Signet ring 67 � 13NOS 72 � 5

Abbreviations: NOS � not otherwise specified; PORT � postoperative

able 5auses of failure in both therapeutic groups

ause Cystectomy PORT P ValueNumber (%) Number (%)

ocal recurrence 39* (32) 2 (3) 0.001istant metastasis 13* (11) 20 (29) 0.001ostoperative death 10 (8) — —remia 1 (1) 0 —

ntestinal obstruction 1 (1) 1 (1) —ntercurrent disease 2 (2) 1 (1) —nknown 4 (3) 3 (4) 0.92

Abbreviation: PORT � postoperative radiotherapy.

* Four patients had both local recurrence and distant metastasis.

.6. Multivariate analysis

Cox multivariate analysis [17] was performed to identifyhe independent working prognostic factors determining theisease-free survival. Tumor stage, adjuvant postoperativeadiotherapy, pelvic nodal involvement, and pathologic sub-ypes were the 4 independent prognostic variables. Theathologic grade was not a prognostic factor because it wasependent on other prognostic variables. Tumor stage andostoperative radiotherapy were the only independent prog-ostic factors for local control, while the stage, nodal in-olvement, and adeno-subtyping were independent factorsetermining distant metastasis-free rate (Table 6).

. Discussion

Primary adenocarcinomas are uncommon tumors ofhese urinary bladders. This category is a rare entity of

roups

Distant metastasis-free � SE (%) P Value

73 � 4 —

78 � 6 0.027164 � 6

10069 � 7 0.09950 � 35

77 � 1276 � 5 0.03557 � 11

81 � 5 0.000146 � 10

10063 � 16 0.000112 � 1175 � 5

erapy; SE � standard error.

able 6ultivariate analyses for the disease-free, local control, and distantetastasis-free survival

ariable Disease-freesurvival Pvalue

LocalcontrolP value

Distantmetastasis-freesurvival P value

tage 0.00001* 0.001* 0.013*ORT 0.00001* 0.00001* 0.761odal involvement 0.004* 0.099 0.008*deno subtype 0.023* 0.410 0.025*rade 0.890 0.150 0.157

Abbreviation: PORT � postoperative radiotherapy.

ts’ subg

Value

.00001

.184

.314

.959

.716

* Statistically significant.

bmedfrubpd

t1omvaehlasoctws5i

ctosarapspwbosescsttv

wta9

rph5ipt

etPfppt[

clittmdksncaeppanp0ltgct(r3t

tlagr�PpP

18 M.S. Zaghloul et al. / Urologic Oncology: Seminars and Original Investigations 24 (2006) 13–20

ladder malignancies [2,18]. However, it is the most com-only reported malignancy of the extrophic bladder [19],

nterocystoplasty [20]. In areas where bilharziasis are en-emic, the incidence of adenocarcinoma is higher, rangingrom 5% to 11.4% [4–7] It is believed that these tumorsesult from metaplastic changes of potentially unstablerothelium [8]. The urothelium in a bladder infested withilharziasis, that of extrophic bladder or after enterocysto-lasty, is more likely to be unstable and more liable toevelop metaplasia.

The classification by Grignon et al. [13] was adopted inhe present study. Mucinous adenocarcinoma was found in4.6%, papillary in 10.4%, signet ring in 7.3%, and nottherwise specified in 67.7% of cases. Signet ring anducinous adenocarcinoma had a higher percentage of ad-

anced stages (P3�P4), high grades (i.e., III) (P � 0.007nd 0.001, respectively) than that in papillary and not oth-rwise specified. Their nodal involvements were alsoigher, although these differences did not rank up to theevel of statistical significance (P � 0.239). Signet-ringdenocarcinoma, being frequently diagnosed in advancedtage, had a survival rate that was poor [21]. To the best ofur knowledge, the present study is the first that couldompare the 5-year disease-free survival of signet ring tohe other 3 categories (papillary, mucinous, and not other-ise specified). The signet ring had a 5-year disease-free

urvival of 7 � 7% compared to 68 � 11%, 36 � 10%, and0 � 5% for papillary, mucinous, and not otherwise spec-fied, respectively (P � 0.00001).

The clinicopathologic features of patients with adenocar-inoma shown in the present study did not differ much fromhat of the other Egyptian studies [3,4,10] of mixed pathol-gy bladder cancer. Mean age and male predominance wereimilar. The pathologic stage and grade distributions werelso similar to that reported in the same Egyptian studieseporting on all histologic subtypes: transitional, squamous,nd adenocarcinoma. Pelvic nodal involvement was re-orted in 25.5% of the patients, which coincided with theeries occurring from bilharzial endemic areas. However,atients with adenocarcinoma in the present study presentedith higher stages than that reported by El Mekresh et al. [6]ecause patients with P3 and P4 disease represented 68.2%f the whole population compared to 13% in the El Mekresheries. They reported 16% nodal involvement. This differ-nce in presentation and nodal involvement could be under-tood in view of the earlier presentation of patients at theirenter. This earlier presentation was also noticed in anothertudy [5] of 1026 patients with bladder cancer, including allumor cell types, treated at the same center. It was provedhat nodal involvement is correlated with the depth of in-asion of bladder cancer [5,6,22,23].

The 5-year disease-free survival rate for the 192 patientsith adenocarcinoma was 46 � 4%, which correlated with

he results of other large adenocarcinoma series. Ghoneim etl. [5] reported a 5-year overall survival rate of 46.4 �

.6%, while El Mekresh et al. [6] showed an overall survival a

ate of 55 � 4.3%. It is noteworthy that the patients in theresent study had more advanced stages, higher grades, andigher incidence of pelvic lymph node involvement. This-year disease-free survival kept the same level, after add-ng more patients and allowing longer follow-up as ourrevious report showed 46.6 � 5% disease-free survival forhe first 142 patients [7].

The pathologic stage affected markedly the 5-year dis-ase-free survival as the rate decreased from 100% in pa-ients with stage P1 disease to 36 � 9% in those with stage4 (P � 0.049). The 5-year disease-free survival was af-ected also by the histopathologic grade (P � 0.011) and theelvic nodal involvement (P � 0.019). These results werereviously confirmed in the adenocarcinoma [6,7], transi-ional cell carcinoma [23–25], and mixed pathology series4,5].

Although radical cystectomy remained, the treatment ofhoice for bladder cancer, including adenocarcinoma, yetocal recurrence accounted for the majority of the causes ofts treatment failure. Postoperative radiotherapy improvedhe local control rate in mixed pathology series that con-ained patients with adenocarcinoma [4,10]. This improve-ent of local control led to a considerable improvement in

isease-free survival. Although adenocarcinoma has beennown to be radioresistant, postoperative radiotherapy wasuccessful in decreasing local recurrence markedly in ade-ocarcinoma of the breast. The rarity of primary adenocar-inoma of urinary bladder made controlled randomized tri-ls extremely difficult, even in areas where bilharziasis isndemic. The relatively large number collected in theresent study, which included 36% of patients who receivedostoperative radiotherapy, allowed testing the value ofdjuvant postoperative radiotherapy in primary adenocarci-oma of the bladder. The group of patients who receivedostoperative radiotherapy had statistically significant (P �.006) higher stages (P3�P4), a higher incidence of pelvicymph nodes involvement (P � 0.0001) (Table 2). Despitehis distribution bias against the postoperative radiotherapyroup, its 5-year disease-free survival was higher than theystectomy alone group (61 � 6% vs. 37 � 5%, respec-ively). The statistical significance was as high as 0.002Table 4). Local failure was much lower in the postoperativeadiotherapy group than the cystectomy alone group (3% vs.2%) while distant metastasis was higher in the postopera-ive radiotherapy group (29% vs. 13%, respectively).

The other causes of failure were equally distributed be-ween the 2 groups (Table 4). This result was reflected inocal control rates because that of postoperative radiother-py had a much higher rate (96 � 3%) than in cystectomyroup (53 � 7%) (P � 0.00001). The distant metastasis-freeate for the postoperative radiotherapy group was lower (64

6%) than that for the cystectomy alone group (78 � 6%,� 0.027). Nevertheless, the retrospective nature of the

resent study warrants it as a level II evidence-based study.rospective randomized controlled studies, in bilharzial-

ssociated and nonassociated adenocarcinoma, are needed

tetemeamie

lfefagpecaf(twpscaoog

uceoOct(ogctcpcpsTZavtf

5

anaipitstsrt

R

[

[

[

[

[

[

[

[

19M.S. Zaghloul et al. / Urologic Oncology: Seminars and Original Investigations 24 (2006) 13–20

o emphasize these results. However, multivariate analysisliminated the effects of distribution bias and the interac-ions between the working variables. Of 5 variables foundffective in the univariate analysis, tumor grade proved byultivariate analysis to be not an independent factor. Its

ffect was dependent on the other factors. Tumor stage,djuvant postoperative radiotherapy, pelvic nodal involve-ent, and adenocarcinoma subtypes proved to be the only

ndependent risk factors for disease-free survival in differ-nt probability values (Table 5).

El Mekresh et al. [6] found that tumor stage, grade, andymph nodal involvement were the only 3 independent riskactors for patients treated with radical cystectomy. How-ver, El Sobky et al. [26] denied any prognostic significanceor tumor stage, grade, and deoxyribonucleic acid profile,nd restricted the prognostic significance to tumor angio-enesis tested by mean vascular density. Their study re-orted on only 55 patients, which probably did not containnough numbers in each subgroup to illustrate the signifi-ance of different variables. On the other hand, multivariatenalysis for local control proved that the only independentactors were postoperative radiotherapy and tumor stageTable 5), while that for distant metastasis-free rates showedhat the nodal involvement, tumor stage, and adeno-subtypeere the only independent prognostic factors. Adjuvantostoperative radiotherapy proved to be dependent on thetage, nodal involvement, and/or subtype. This result mayorrelate the lower distant metastasis-free rate in postoper-tive radiotherapy in univariate analysis to the selection biasf inclusion of more advanced stages and higher percentagef nodal involvement in the postoperative radiotherapyroup.

Postoperative radiotherapy in bladder cancer remainednpopular as a result of the expectation of late intestinalomplications. Warning results were published by Reisingert al. [27], as they reported 37% (15 of 40) intestinalbstruction in patients received postoperative radiotherapy.f these 15 patients, 9 required surgery, and 3 died. On the

ontrary, Zaghloul et al. [4] had 14 (18%) of the 78 patientsreated with conventional postoperative radiotherapy, and 45%) of the 75 patients treated with hyperfractionation post-perative radiotherapy had chronic enteritis of differentrades. Only 3 (4%) and 4 (5%) of patients treated withonventional and hyperfractionation, respectively, had fis-ula that necessitated surgery. The present study resultsoincided with that of Zaghloul et al. [4] because 12% of theatients who received postoperative radiotherapy hadhronic intestinal complications of different grades com-ared to 8% in the cystectomy alone group. No difference oftatistical significance was detected between the 2 groups.he difference between the present results and that ofaghloul et al. [4] on the one hand, and that of Reisinger etl. [27] on the other hand, was probably a result of the largerolume of radiation used in Reisinger et al. study, which ledo severer late bowel complications with similar disease-

ree survival, when compared stage by stage [28].

. Conclusions

Because adenocarcinoma of the urinary bladder was ex-mined more closely and thoroughly using a considerableumber of patients using both univariate and multivariatenalyses, it seemed to be similar to other types of musclenvasive bladder cancer (e.g., squamous cell carcinoma androbably transitional cell carcinoma) in behavior after rad-cal cystectomy. The use of adjuvant postoperative radio-herapy could improve the disease-free survival in this largeeries through improvement of local control. Distant metas-asis is still a major problem, especially in more aggressiveubtypes of adenocarcinoma (e.g., mucinous and signeting). This result suggests the strong need for adding effec-ive chemotherapy

eferences

[1] Wheeler JD, Hill WT. Adenocarcinoma involving the urinary blad-der. Cancer 1954;7:468–76.

[2] Thomas DC, Ward AM, Williams JL. A study of 52 cases of adeno-carcinoma of the bladder. Br J Urol 1971;43:4–15.

[3] El Bolkainy MN, Mokhtar NM, Ghoneim MA, et al. The impact ofschistosomiasis on the pathology of bladder carcinoma. Cancer 1981;48:2643–8.

[4] Zaghloul MS, Awwad HK, Omar S, et al. Postoperative radiotherapyof carcinoma in bilharzial bladder. Improved disease-free survivalthrough improving local control. Int J Radiat Oncol Biol Phys 1992;22:511–7.

[5] Ghoneim MA, El Mekresh MM, El Baz MA, et al. Radical cystec-tomy for carcinoma of the bladder: Critical evaluation of the resultsin 1026 cases. J Urol 1997;158:393–9.

[6] El Mekresh MM, El Baz MA, Abol Eniein H, et al. Primary adeno-carcinoma of the urinary bladder. A report of 185 cases. Br J Urol1998;82:206–12.

[7] Zaghloul MS, Abdel Aziz SA, Nouh A, et al. Primary adenocarci-noma of the urinary bladder. Risk factors and value of postoperativeradiotherapy J Natl Cancer Inst 2003;13:193–200.

[8] Allen TD, Henderson BW. Adenocarcinoma of the bladder. J Urol1965;93:50–6.

[9] Mostofi FK. Potentialities of bladder epithelium. J Urol 1954;71:705–14.

10] Kittridge WE, Coliett AJ, Morgan C. Adenocarcinoma of the bladderassociated with cystitis glandularis. A case report. J Urol 1964;91:145–50.

11] Zaghloul MS, Awwad HK, Soliman O, et al. Postoperative radiother-apy of carcinoma in bilharzial bladder using a three-fractions per dayregimen. Radiother Oncol 1986;6:257–65.

12] Johnson DE, Hogan JM, Ayala AG. Primary adenocarcinoma of theurinary bladder. South Med J 1972;65:527–30.

13] Grignon DJ, Ro JY, Ayala AG, et al. Primary adenocarcinoma of theurinary bladder. Cancer 1991;67:2165–72.

14] El-Sebai I. Cancer of the bilharzial bladder in Egypt. Kasr El-AiniJ Surg 1961;2:183–241.

15] Kaplan EL, Meier P. Non-parametric estimation from incompleteobservations. J Am Stat Assoc 1972;53:457–65.

16] Peto R, Pike MS, Armitage P, et al. Design and analysis of random-ized clinical trials requesting prolonged observation of each patient.II. Analysis and examples. Br J Cancer 1977;35:1–39.

17] Cox DR. Regression methods and life tables. J R Stat Soc 1972;34:

187–220.

[

[[

[

[

[

[

[

[

[

[

20 M.S. Zaghloul et al. / Urologic Oncology: Seminars and Original Investigations 24 (2006) 13–20

18] Bennet JK, Wheatly JK, Walton KN. Ten years experience withadenocarcinoma of the bladder. J Urol 1984;131:262–3.

19] Clup DA. The histology of extrophied bladder. J Urol 1964;91:538–48.20] Filmer RB, Spencer JR. Malignancies in bladder augmentations and

intestinal conduits. J Urol 1990;143:671–8.21] Holmang S, Borghede G, Johansson SL. Primary signet-ring cell

carcinoma of the bladder: A report on 10 cases. Scand J Urol Nephrol1997;31:145–8.

22] Poulsen AL, Horn T, Steven K. Radical cystectomy: Extending thelimits of pelvic lymph node dissection improves survival for patientswith bladder cancer confined to the bladder wall. J Urol 1998;160:2015–20.

23] Leissner J, Hohenfellner R, Thuroff JW, et al. Lymphadenectomy inpatients with transitional cell carcinoma of the urinary bladder; sig-

nificance for staging and prognosis. BJU Int 2000;85:817–23.

24] Stein JP, Lieskovsky G, Cote R, et al. Radical cystectomy in treat-ment of invasive bladder cancer: Long term results in 1054 patients.J Clin Oncol 2001;19:666–75.

25] Medersbacher S, Hochreiter W, Burkhard F, et al. Radical cystectomyfor bladder cancer today�A homogeneous series without neoadju-vant therapy. J Clin Oncol 2003;21:690–6.

26] El Sobky E, Gomha M, El Baz M, et al. Prognostic significance oftumour angiogenesis in schistosomiasis associated adenocarcinomaof the urinary bladder. Br J Urol 2002;89:126–32.

27] Reisinger S, Mohiuddin M, Mulholland S. Combined pre- and post-operative adjuvant radiation therapy for bladder cancer�A ten yearsexperience. Int J Radiat Oncol Biol Phys 1992:24:463–8.

28] Zaghloul MS. Radiation as adjunctive therapy to cystectomy forbladder cancer: Is there a difference for bilharzial association? Int J

Radiat Oncol Biol Phys 1994;28:783.