Original research  
Enumerating pelvic recurrence following radical cystectomy for  
bladder cancer: A Canadian multi-institutional study  
1
2
3
4
Libni J. Eapen, MD; Edward Jones, MD; Wassim Kassouf, MD; Carole Lambert, MD;  
1
5
6
7
Scott C. Morgan, MD; Madeleine Moussa, MD; Robert Nam, MD; Matthew Parliament, MD;  
7
8
5
3
6
Laurie Russell, MD; Fred Saad, MD; D. Robert Siemens, MD; Luis Souhami, MD; Ewa Szumacher, MD;  
2
1
6
1
1
Scott Tyldesley, MD; Yan Xu, MD; Ingrid Zbieranowski, MD; Rodney H. Breau, MD; Eric Belanger, MD;  
2
7
9
9
10  
Peter Black, MD; Eric Estey, MD; Julie Bowan, MD; Bishwajit Bora, MD; Michael Brundage, MD;  
1
1
11  
11  
12  
12  
Peter Chung, MD; Neil Fleshner, MD; Andrew Evans, MD; Glenn Bauman, MD; Jonathan Izawa, MD;  
5
3
Chris Davidson, MD; Fadi Brimo, MD  
1
2
3
4
The Ottawa Hospital Cancer Centre, University of Ottawa, Ottawa, ON, Canada; BC Cancer Agency, BC, Canada; McGill University Health Centre, Montreal, QC, Canada; Notre-Dame Hospital, Montreal,  
5
6
7
8
QC, Canada; Queen’s University, Kinston, ON, Canada; Odette Cancer Centre, Sunnybrook Health Sciences Centre, Toronto, ON, Canada; Cross Cancer Institute, Edmonton, AB, Canada; Université de  
Montréal, Montreal, QC, Canada; Sudbury Cancer Centre, Sudbury, ON, Canada; Kingston Cancer Centre, Kingston, ON, Canada; Princess Margaret Hospital, Toronto, ON, Canada; London Cancer  
Centre, London, ON, Canada  
9
10  
11  
12  
Cite as: Can Urol Assoc J 2016;10(3-4):90-4. http://dx.doi.org/10.5489/cuaj.3456  
disease and the systemic micrometastatic tumour burden.  
Unfortunately, cure rates have remained static for over two  
decades, reflecting the lack of more effective systemic treat-  
ment and a continuing problem with pelvic tumour eradica-  
See related article on page 95.  
3
tion. We report the results of a Canadian academic centre  
Abstract  
survey designed to enumerate the proportion of patients  
developing pelvic tumour recurrence following contempo-  
rary cystectomy.  
Introduction: We aimed to enumerate the rate of pelvic recurrence  
following radical cystectomy at university-affiliated hospitals in  
Canada.  
Methods  
Methods: Canadian, university-affiliated hospitals were invited to  
participate. They were asked to identify the first 10 consecutive  
patients undergoing radical cystectomy starting January 1, 2005,  
who had urothelial carcinoma stages pT3/T4 N0-2 M0. The first 10  
consecutive cases starting January 1, 2005 who met these criteria  
were the patients submitted by that institution with information  
regarding tumour stage, age, number of nodes removed, and last  
known clinical status in regard to recurrence and patterns of failure.  
Results: Of the 111 patients, 80% had pT3 and 20% pT4 disease,  
with 62% being node-negative, 14% pN1, and 27% pN2; 57%  
had 10 or more nodes removed. Cumulative incidence of pelvic  
relapse was 40% among the entire group  
In 2011, 17 university-affiliated cancer clinics across Canada  
were invited to participate in this effort. Specifically, each  
centre was asked to identify a trio of investigators comprised  
of a pathologist, urologist, and radiation oncologist. Following  
ethics approval, the pathologist generated a chronologically  
accurate list of consecutive patients undergoing cystectomy  
for bladder cancer starting January 1, 2005. To be eligible  
for the study, the following criteria had to be met: 1) primary  
urothelial carcinoma with or without any admixed other  
histologies (no other primary histologies such as squamous,  
adenocarcinoma, or small cell were included); 2) no known  
hematogenous metastases or nodal involvement above the  
iliac bifurcation; 3) surgery performed with curative intent  
and no gross residual disease; 4) neoadjuvant or adjuvant  
chemotherapy was permitted; 5) only pT3/T4 N0-2 stages.  
The first 10 consecutive cases meeting these criteria, regard-  
less of whether the subsequent clinical outcome was fully  
known or not, constituted the required patient sample from  
each institution. It was the expectation that this method  
of case selection from each hospital’s surgical pathology  
chronological record would minimize case selection bias.  
Conclusions: This review demonstrates a high rate of pelvic tumour  
recurrence following radical cystectomy for pT3/T4 urothelial cancer.  
Introduction  
The current gold standard management for locally advanced  
urothelial bladder cancer consists of neoadjuvant chemo-  
therapy, followed by radical cystectomy with pelvic node  
dissection and neo-bladder reconstruction, as appropri-  
1
,2  
ate. This combination addresses the loco-regional pelvic  
9
0
CUAJ • March-April 2016 • Volume 10, Issues 3-4  
2016 Canadian Urological Association  
©
Pꢀꢁvꢂꢃ ꢄꢀꢃuꢄꢄꢀꢅꢃꢀ ꢆftꢀꢄ ꢄꢆdꢂꢃꢆꢁ ꢃyꢇtꢀꢃtomy foꢄ bꢁꢆddꢀꢄ ꢃꢆꢅꢃꢀꢄ  
Abstracted clinical data included birthdate, date of radical  
cystectomy, pathological stage, number of removed lymph  
nodes recorded in the pathology report, date of last clinical  
contact, and last known clinical status. Defining any hema-  
togenous spread or nodal disease above or at L5 as distant  
metastases (DM) and any pelvic soft tissue or nodal disease  
below L5 as pelvic recurrence (PR), the eventual clinical  
status was categorized as: 1) no tumour recurrence; 2) pelvic  
recurrence only (with date); 3) distant metastases only (with  
date); 4) distant metastases and pelvic recurrence regardless  
of whether identified simultaneously or not; and 5) pelvic  
tumour status unknown. The anonymized individual patient  
data sheets were collated centrally for compilation and anal-  
ysis. Arithmetic proportions of patients developing pelvic  
recurrence with or without distant metastases were tabulated  
and cumulative incidence calculations were performed to  
compute the risk of developing pelvic relapse.  
(pT3a=39, pT3b=41, pT3NOS=1) and pT4 in 30 (pT4a=27,  
pT4b=2, pT4NOS=1). N stage was N0 in 62 patients, N1 in  
14, N2 in 27, and Nx in 8. The median number of nodes dis-  
sected was 11 and ranged from 0–67. With regard to nodes,  
26.6% had more than 15 nodes resected; 21.1% had 11–15  
resected; 26.6% had 6–10 resected; 22.9% had 5 or less  
resected. Median time to last followup was 15 months, and  
ranged from 0–87 months. PR occurred in 34.2% of patients;  
51.4% had no PR. Pelvic status was unknown in 14.4%, but  
these were considered as having no PR. Of the 38 patients  
with PR, 25 (65.8%) developed DM and 13 (34.2%) had  
only PR. Fig. 1 shows the 40% pelvic recurrence rate when  
calculated as cumulative incidence.  
Discussion  
Starting with Whitmores’ emphasis on the nodal dissection  
4
component of a radical cystectomy in 1962 to examina-  
tion of planned perioperative adjuvant radiotherapy in the  
Results  
7
0s and 80s, Skinners’ compilation of the large, single-  
Eleven university-affiliated cancer clinics (Vancouver,  
Edmonton, London, Kingston, Ottawa, Princess Margaret,  
Sunnybrook, McGill, CHUM, Saint John, and Sudbury) from  
five provinces participated.  
institutional testament to comprehensive, meticulous surgi-  
cal technique through the 90s, and calibrated further by  
5
Herrs’ elucidation of the importance of resected node counts  
6
in 2003, these strategies have addressed the all-important  
Ten centres from five provinces provided data on 111  
patients (Table 1). Age ranged from 43.9–92.7 years, (mean/  
median 69.6/70.9 years). T stage was pT3 in 81 patients  
matter of pelvic tumour eradication in bladder cancer. In a  
2007 comprehensive review, Cagiannos et al tabulated con-  
temporary post-cystectomy pelvic recurrence rates ranging  
from 3.929%. These numbers have significantly underes-  
7
timated the actual occurrence of loco-regional failures due  
to: 1) dilution by reporting overall rather than per T stage  
recurrence rates; 2) calculation of simple proportions rather  
than cumulative incidence risk; 3) requirement of biopsy  
confirmation; 4) discounting pelvic failure if accompanied  
or preceded by systemic metastases; and 5) relative inatten-  
tion to pelvic imaging following the development of distant  
disease. These detriments to accurate risk determination are  
compounded by the known insensitivity of pelvic imaging  
in the context of subcentimetric tumour deposits.  
Table 1. Patient characteristics and recurrence status  
Characteristics  
N (%)  
Age  
<65  
65  
Unknown  
31ꢀ(27.9)  
79ꢀ(71.2)  
1ꢀ(0.9)  
Nodeꢀstatus  
pN0  
pN1  
pN2  
Unknown  
62ꢀ(55.9)  
14ꢀ(12.6)  
27ꢀ(24.3)  
8ꢀ(7.2)  
Two experiences published in 2005 and 2012 define  
the scope of this ongoing pelvic control issue. The SWOG  
trial of neoadjuvant chemotherapy featured a 32% biopsy-  
Pathologicꢀstageꢀatꢀcystectomy  
pT3a  
pT3b  
pT3NOS  
pT4a  
pT4b  
39ꢀ(35.1)  
41ꢀ(37.0)  
1ꢀ(0.9)  
27ꢀ(24.3)  
2ꢀ(1.8)  
8
proven crude simple pelvic relapse rate in T3/T4 patients.  
The MRC trialists’ report of the long-term results of the  
seminal European neoadjuvant chemotherapy trial yield-  
pT4NOS  
1ꢀ(0.9)  
9
ed a 49% loco-regional relapse rate. While documenting  
Nodesꢀremoved  
<10  
10  
Unknown  
52ꢀ(46.9)  
57ꢀ(51.4)  
2ꢀ(1.8)  
the high intrapelvic recurrence rates, both these trials also  
demonstrate that although overall survival is improved by  
5%, neoadjuvant chemotherapy does not have a discernible  
Recurrenceꢀstatusꢀatꢀfollowup  
impact on pelvic tumour control.  
None  
36ꢀ(32.4)  
13ꢀ(11.7)  
22ꢀ(19.8)  
25ꢀ(22.5)  
15ꢀ(13.5)  
The 40% pelvic relapse rate in this cross-Canada  
study is in concert with the experiences detailed above.  
Unfortunately, it is likely that even this high risk is an under-  
estimate, given that we used the conservative assumption  
Pelvicꢀonly  
Distantꢀonly  
Pelvicꢀandꢀdistantꢀ  
Unknown  
CUAJ • March-April 2016 • Volume 10, Issues 3-4  
91  
eꢆpꢀꢅ ꢀt ꢆꢁ.  
Fig. 2. Cumulative incidence of pelvic recurrence following cystectomy of 110  
patients; subgroups defined by nodal status at time of surgery.  
Month  
0
12  
24  
36  
At risk (events)  
110 (25)  
46 (10)  
27 (3)  
19 (0)  
demic institutions. This is in contrast to the general literature,  
wherein pelvic control is not the prime focus of published  
reports.  
Weaknesses of this data set are the relatively small num-  
ber of cases audited at each centre (10), the limited rigour  
in determining pelvic tumour status in those patients for  
whom it was deemed “unknown” from the readily available  
documentation, and omission of surgical margin status. The  
latter two are a reflection of the practicalities of a multicentre  
survey.  
Whereas enumeration defines the magnitude of the  
problem, understanding causation requires unraveling the  
interplay between the culprit clinical, surgical, and patho-  
logical factors. In this Canadian patient cohort, node posi-  
tivity predicted for higher PR (Fig. 2), while the number of  
lymph nodes resected appeared to not materially influence  
Fig. 1. Cumulative incidence of pelvic recurrence following cystectomy of 110  
patients; number of patients at risk and events occurring per 12-month period  
at bottom of graph.  
that patients in whom pelvic tumour status was unknown  
were free of tumour.  
The strength of our data is two-fold. Firstly, the data is  
a random snapshot of outcomes. By virtue of the patient  
retrieval strategy used, the selection biases that can occur  
when measuring outcomes in patients selected for partici-  
pation in a clinical trial or abstracting data from cultivated  
institutional patient data sets is reduced. Secondly, and likely  
most importantly, the sole and simple object of this exer-  
cise was to enumerate pelvic relapse in unselected patients  
undergoing contemporary surgery at widely dispersed aca-  
Fig. 3. Cumulative incidence of pelvic recurrence following cystectomy of 110  
patients; subgroups defined by number of nodes removed.  
Fig. 4. Cumulative incidence of pelvic recurrence following cystectomy of 110  
patients; subgroups defined by pathological tumour stage.  
9
2
CUAJ • March-April 2016 • Volume 10, Issues 3-4  
Pꢀꢁvꢂꢃ ꢄꢀꢃuꢄꢄꢀꢅꢃꢀ ꢆftꢀꢄ ꢄꢆdꢂꢃꢆꢁ ꢃyꢇtꢀꢃtomy foꢄ bꢁꢆddꢀꢄ ꢃꢆꢅꢃꢀꢄ  
PR (Figs. 3 and 4). This latter observation is at odds with  
what currently comprises a cardinal aspect of radical cyst-  
ectomy, whereby still more extensive nodal clearance is  
rent tumour in the pelvis can and do experience significant  
morbidity that is poorly palliated, whether or not distant  
disease coexists.  
1
0,11  
being advocated and investigated.  
Lymph node counts,  
This study was undertaken because of the clinical percep-  
tion that PR remains a significant problem following curative  
surgery in pT3/T4 urothelial bladder cancer, an issue that is  
not readily discernible in the uro-oncological literature. The  
multi-institutional data presented corroborates this mater-  
ially high risk of loco-regional failure following contempor-  
ary radical cystectomy in advanced-stage presentation and  
emphasizes the need to address the issue definitively. Pelvic  
control is a necessary, but insufficient requirement for cure.  
Giving it the focus it warrants will enable uro-oncologists to  
elucidate the optimal, nuanced amalgam of surgery, radio-  
therapy, and systemic therapies that maximizes the cure  
potential for the individual bladder cancer patient.  
be they absolute or proportional, are vulnerable to varia-  
tions in retrieval from the resected specimen. Additionally,  
the relatively small number of patients and the possibility  
that patients with node-positive disease are more likely to  
manifest metastatic disease, thus reducing the clinical prob-  
ability of identifying PR, likely explain this anomalous result.  
Our data must not be misconstrued as minimizing the  
importance of adequate nodal dissection. Christodouleas  
has validated, extramurally, the robustness of assigning  
post-cystectomy patients to low-, intermediate- and high-  
risk categories on the basis of pathological stage, numbers  
12  
of resected lymph nodes, and margin status. The soundness  
1
3
15  
of this model has been further corroborated by Froehner  
In North America, the NRG clinical trials group has  
1
4
and Ku This stratification of PR risk accounts for tumour-  
related parameters. Other factors can be considered as being  
patient- or treatment-related. Given that, by definition, all  
patients undergoing curative radical cystectomy had to be  
sufficiently well-suited in terms of medical condition and  
performance status to undergo the surgery, it is unlikely that  
other patient variables have contributed to the probability  
of PR. Treatment factors include the use of neo or adjuvant  
chemotherapy, thoroughness of the operation, and case vol-  
ume issues that speak to the surgical team’s experience. As  
noted previously, randomized trial data show that adminis-  
tering chemotherapy to these patients does not reduce PR  
rates. While nodal status is unequivocally a tumour par-  
ameter, resected nodal counts definitely and margin status  
possibly also reflect surgical rigor. That all these patients  
underwent surgery at academic hospitals with cancer clinic  
affiliations speaks to the likely sufficient patient volumes and  
requisite surgical expertise.  
opened a randomized trial examining adjuvant radiother-  
apy and similar studies are being launched in France, Asia,  
and the U.K. This is the currency of clinical investigation  
into the treatment of all other solid malignancies that has  
yielded ever-improving local, regional, and systemic control,  
a scenario that can be realistically anticipated for bladder  
cancer as well.  
Competing interests: Dr. Eapen has received grants/honoraria from Abbott and AstraZeneca; and  
has participated in numerous clinical trials. Dr. Kassouf has received grants/honoraria from Amgen,  
Astellas, and Janssen; and is also the recipient of a Research Scholar Award from the FRSQ. Dr.  
Lambert has received grants/honoraria from Ferring and has participated in clinical trials for Ferring  
and Janssen. Dr. Morgan has been an Advisory Board member for Accuray, Bayer, Janssen, and  
Sanoꢀ; has received grants/honoraria from Abbvie and Astellas; and has participated in clinical trials  
for Bayer and Janssen. Dr. Siemens has participated in clinical trials for Amgen, Astellas, Ferring,  
and Janssen. Dr. Tyldesley has received grants/honoraria from Amgen, Bayer, and Janssen. Dr.  
Black has been an Advisory Board member for Abbvie, Amgen, Astellas, Biocancell, Cubist, Janssen,  
Novartis, and Sitka; has been on Speaker Bureaus for Abbvie, Janssen, Ferring, Novartis, and Red  
Leaf Medical; has received grants/honoraria from Pendopharm; has participated in clinical trials for  
Amgen, Astellas, Ferring, Janssen, and Roche; and has received research funding from GenomeDx,  
iProgen, Lilly, and New B Innovation. Dr. Bowen has been an Advisory Board member for Astellas  
and Janssen; and has received grants/honoraria from AstraZeneca. Dr. Evans has been an Advisory  
Board member for Omnyx Digital Pathology. Dr. Bauman has received grants/honoraria from Sanoꢀ  
and has participated in clinical trials for Sanoꢀ. Dr. Izawa has received grants/honoraria from Abbott,  
AstraZeneca, Astellas, Janssen, Sanoꢀ, and Pꢀzer. Dr. Chung has received grants/honoraria from  
Sanoꢀ and has participated in clinical trials for Abbvie. The remaining authors declare no competing  
In the final analysis, this high rate of PR cannot be attrib-  
uted to any unfortunate case mix or singular compromise  
of accepted bounds of contemporary surgical management.  
Rather, it is almost certainly a reflection of the true rates of  
pelvic failure revealed when the singular focus of the study  
is the number of pelvic relapses.  
The approach of discounting pelvic recurrence if it  
coincides with or succeeds distant metastases speaks to a  
perspective that relegates loco-regional failure to the status  
of being an unfortunate, but clinically unimportant event.  
This derives from the fact that the majority of patients with  
pelvic relapse also develop distant disease. However, it  
remains the clinical reality that: 1) essentially, no patient  
with pelvic failure can be salvaged, rendering the magnitude  
of the pelvic relapse problem an absolute ceiling on surgical  
curability; 2) in the 15% minority of patients who develop  
isolated pelvic relapse, the clinical outcome is determined  
by local and not distant disease; and 3) patients with recur-  
nancial or personal interests.  
This paper has been peer-reviewed.  
CUAJ • March-April 2016 • Volume 10, Issues 3-4  
93  
eꢆpꢀꢅ ꢀt ꢆꢁ.  
9
1
1
1
.
International Collaboration of Trialists. International phase III trial assessing neoadjuvant cisplatin, metho-  
trexate, and vinblastine chemotherapy for muscle-invasive bladder cancer: Long-term results of the BA06  
References  
3
0894 trial. J Clin Oncol 2011;29:2171-7. http://dx.doi.org/10.1200/JCO.2010.32.3139  
1
2
.
.
NCCN Guidelines version 2.2015. https://www.nccn.org/store/login/login.aspx? ReturnURL=http://  
www.nccn.org/professionals/physician_gls/PDF/bladder.pdf  
Stenzl A, Cowan NC, De Santis M, et al. Treatment of muscle-invasive and metastatic bladder  
cancer: Update of the EAU guidelines. Eur Urol 2011;59:1009-18. http://dx.doi.org/10.1016/j.  
eururo.2011.03.023  
0. Zehnder P, Studer UE, Skinner EC, et al. Super-extended vs. extended pelvic lymph node dissection in  
patients undergoing radical cystectomy for bladder cancer: A comparative study. J Urol 2011;186:1261-8.  
http://dx.doi.org/10.1016/j.juro.2011.06.004  
1. Zlotta AR. Limited, extended, super-extended, mega-extended pelvic lymph node dissection at the time of  
radical cystectomy: What should we perform? Eur Urol 2012;61:243-4. http://dx.doi.org/10.1016/j.  
eururo.2011.11.006  
2. Christodouleas JP, Baumann BC, He J, et al. Optimizing bladder cancer loco-regional failure risk stratiꢀcation  
after radical cystectomy using SWOG 8710. Cancer 2014;120:1272-80. http://dx.doi.org/10.1002/  
cncr.28544  
3. Froehner M, Novotny V, Wirth MP, et al. External validation of a model to predict loco-regional failure after  
radical cystectomy. Cancer 2014;120:3584. http://dx.doi.org/10.1002/cncr.28876  
4. Ku JH, Kim M, Jeong CW, et al. Risk prediction models of loco-regional failure after radical cystectomy  
for urothelial carcinoma: External validation in a cohort of Korean patients. Int J Radiat Oncol Biol Phys  
3
.
Grossman HB, Natale RB, Tangen CM, et al. Neoadjuvant chemotherapy plus cystectomy compared with  
cystectomy alone for locally advanced bladder cancer. N Engl J Med 2003;349:859-66. http://dx.doi.  
org/10.1056/NEJMoa022148  
4
5
6
.
.
.
Whitmore WFJ, Marshall VF. Radical total cystectomy for cancer of the bladder: 230 consecutive cases  
ve years later. J Urol 1962;87:853.  
1
1
Stein JP, Lieskovsky G, Cote R, et al. Radical cystectomy in the treatment of invasive bladder cancer:  
Long-term results in 1054 patients. J Clin Oncol 2001;19:666-75.  
Herr HW, Bochner BH, Dalbagni G, et al. Impact of the number of lymph nodes retrieved on outcome in  
patients with muscle-invasive bladder cancer. J Urol 2002;167:1295-8. http://dx.doi.org/10.1016/  
S0022-5347(05)65284-6  
2
014; 89:1032-7. http://dx.doi.org/10.1016/j.ijrobp.2014.04.049  
5. NRG Oncology GU 001. https://www.nrgoncology.org/Clinical-Trials/NRG-GU001. Accessed March 21,  
016.  
1
7
.
.
Cagiannos I, Morash C. Surveillance strategies after deꢀnitive therapy of invasive bladder cancer. Can  
Urol Assoc J 2009;3:S237-42.  
Herr HW, Faulkner JR, Grossman HB, et al. Surgical factors inꢁuence bladder cancer outcomes: A coopera-  
tive group report. J Clin Oncol 2004;22: 2781-9. http://dx.doi.org/10.1200/JCO.2004.11.024  
2
8
Correspondence: Dr. Libni Eapen, The Ottawa Hospital Cancer Centre, Ottawa, ON, Canada;  
leapen@toh.on.ca  
9
4
CUAJ • March-April 2016 • Volume 10, Issues 3-4