Author's Accepted Manuscript Reassessment of prognostic heterogeneity in pT3 renal pelvic urothelial carcinoma: Analysis in terms of three proposed pT3 subclassification systems Jinsung Park, Tomonori Habuchi, Youichi Arai, Chikara Ohyama, Takamitsu Inoue, Shingo Hatakeyama, Seong Soo Jeon, Ghee Young Kwon, Cheol Kwak, Kyung Chul Moon, Choung-Soo Kim, Hanjong Ahn PII: DOI: Reference:
S0022-5347(14)03314-X 10.1016/j.juro.2014.04.001 JURO 11379
To appear in: The Journal of Urology Accepted Date: 4 April 2014 Please cite this article as: Park J, Habuchi T, Arai Y, Ohyama C, Inoue T, Hatakeyama S, Jeon SS, Kwon GY, Kwak C, Moon KC, Kim CS, Ahn H, Reassessment of prognostic heterogeneity in pT3 renal pelvic urothelial carcinoma: Analysis in terms of three proposed pT3 subclassification systems, The Journal of Urology® (2014), doi: 10.1016/j.juro.2014.04.001. DISCLAIMER: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our subscribers we are providing this early version of the article. The paper will be copy edited and typeset, and proof will be reviewed before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to The Journal pertain. All press releases and the articles they feature are under strict embargo until uncorrected proof of the article becomes available online. We will provide journalists and editors with full-text copies of the articles in question prior to the embargo date so that stories can be adequately researched and written. The standard embargo time is 12:01 AM ET on that date.
1
ACCEPTED MANUSCRIPT
Reassessment of prognostic heterogeneity in pT3 renal pelvic urothelial carcinoma:
RI PT
Analysis in terms of three proposed pT3 subclassification systems
Jinsung Park1, Tomonori Habuchi2, Youichi Arai3, Chikara Ohyama4, Takamitsu Inoue2, Shingo Hatakeyama4, Seong Soo Jeon5, Ghee Young Kwon6, Cheol Kwak7, Kyung Chul
SC
Moon8, Choung-Soo Kim9, and Hanjong Ahn9
Department of Urology, Eulji University Hospital, Daejeon, Korea
2
Department of Urology, Akita University Graduate School of Medicine, Akita, Japan
3
Department of Urology, Tohoku University School of Medicine, Sendai, Japan
4
Department of Urology, Hirosaki University Graduate School of Medicine, Hirosaki, Japan
5
Department of Urology and Pathology6, Samsung Medical Center, Sungkyunkwan
TE D
M AN U
1
University School of Medicine, Seoul, Korea 7
Department of Urology and Pathology8, Seoul National University College of Medicine,
9
EP
Seoul, Korea.
Department of Urology, Asan Medical Center, University of Ulsan College of Medicine,
AC C
Seoul, Korea
Running title: Prognostic heterogeneity of pT3 renal pelvic urothelial carcinoma
Conflict of interest disclosures: The authors made no disclosures.
Source of funding: None
1
2
ACCEPTED MANUSCRIPT
Acknowledgements: We thank Dr. Seung-Sik Hwang who works at the Department of Social and Preventive Medicine, Inha University School of Medicine for his excellent
RI PT
supervision in the statistical analysis.
SC
IRB No.: S2012-0912-000224
Correspondence: Hanjong Ahn, MD, PhD.
M AN U
Department of Urology, Asan Medical Center, University of Ulsan College of Medicine, 388-1 Pungnap 2 Dong, Songpa-gu, Seoul, 138-736, Korea Fax: 82-2-477-8928 Tel: 82-2-3010-3733
TE D
e-mail:
[email protected] Key words: prognosis, renal pelvicalyceal system, pT3 subclassification, urothelial
AC C
EP
carcinoma, upper urinary tract
2
3
ACCEPTED MANUSCRIPT ABSTRACT
Purpose: We determined whether the three pT3 subclassification systems reported by the Asan, Cornell and Nagoya groups provide an accurate estimation of patient prognosis and
RI PT
which one is most predictive of the heterogeneous oncological outcomes in pT3 renal pelvic urothelial carcinoma (RPUC).
Materials and Methods: From a Korea-Japan multi-institutional, retrospective database, 250
SC
pT3 RPUCs treated with radical nephroureterectomy (RNU) were assigned into the three pT3 subcategories according to tumor location and depth of parenchymal invasion after
M AN U
pathologic re-evaluation. Recurrence-free (RFS) and cancer-specific survival (CSS) were assessed according to the three pT3 subclassifications. Predictive accuracy (PA) for survival in four different models (a baseline model and three models including pT3 subclassifications) was quantified and increments in PA for each model were compared.
TE D
Results: In the baseline multivariate Cox regression model, nodal metastasis and high grade were significant for survival. In the multivariate analyses including pT3 subclassifications, the three subclassifications remained significantly associated with survival rates. Among the
EP
three pT3 subclassification systems, the Cornell subclassification had the highest PA for discriminating the heterogeneous prognosis of pT3 RPUC, followed by the Nagoya
AC C
subclassification. Compared with the baseline model, adding the Cornell subclassification significantly increased the PA for RFS (from 0.687 to 0.742, p=0.029) and CSS (from 0.713 to 0.758, p=0.047).
Conclusions: The criteria of microscopic versus macroscopic parenchymal invasion and/or peripelvic fat invasion provide the most accurate differential classification of prognostic heterogeneity of pT3 RPUC. Further studies should be performed to determine the need to modify the current pT3 RPUC staging system.
3
4
ACCEPTED MANUSCRIPT INTRODUCTION
Among the prognostic parameters in upper tract urothelial carcinoma (UTUC), the prognostic value of the location of UTUC remains controversial. The European Association of Urology
RI PT
(EAU) guidelines have been inconsistent regarding the prognostic impact of tumour location.1, 2 In the 2011 EAU guidelines, which were based on three large sample size studies from Memorial Sloan-Kettering Cancer Center,3 the Surveillance Epidemiology and End Results registry4 and ten international multi-centre studies,5 initial tumour location was not
SC
accepted as a prognostic factor.1 However, the 2013 EAU guidelines based on French6, 7 and
M AN U
nine international multi-centre studies8 state that tumour location in UTUC is a prognostic factor, i.e. ureteral and multifocal tumours have a worse prognosis than renal pelvic tumours.2 While few prior studies examined the possible reasons of inter-institutional discrepancies regarding prognostic impact of tumour location in UTUC, we have demonstrated that
TE D
location-dependent survival difference exists only in patients with pT3 UTUC, and conflicting institutional results may be due to differences in the proportion of parenchymal versus peripelvic fat invasion in pT3 renal pelvic urothelial carcinoma (RPUC).9, 10 In line
EP
with the prognostic significance of tumour location in pT3 UTUC,9-13 several studies suggest new subclassification systems for pT3 RPUC according to tumour location and depth of
AC C
parenchymal invasion.14, 15 The purpose of the current study was to determine whether the proposed subclassifications9, 14, 15 for pT3 RPUC provide an accurate estimation of patient prognosis and which of the proposed subclassifications is most predictive of the heterogeneous oncological outcomes in pT3 RPUC on the basis of a Korea-Japan multiinstitutional collaborative database.
4
5
ACCEPTED MANUSCRIPT PATIENTS AND METHODS Study design and patient population
The present study was a retrospective, multi-centre study that included six academic
RI PT
institutions in Korea and Japan. The study protocol was approved by the Institutional Review Board of each institute. Inclusion criteria were patients who were diagnosed with pT3 RPUC following radical nephroureterectomy (RNU) between 1990 and 2010. Exclusion criteria
SC
were kidney sparing or palliative surgery, histologic variants of urothelial carcinoma such as micropapillary, squamous, sarcomatoid and small cell carcinoma, patients receiving
M AN U
neoadjuvant chemotherapy, and those with distant metastasis or concomitant invasive bladder cancer. Medical files were reviewed, and a digitised database was generated. Before the analysis, the database was frozen and the final dataset was produced. The study population comprised 250 patients with pT3 RPUC. The database included the
TE D
following parameters: age at diagnosis, sex, date of surgery, history of bladder tumour (no, previous, synchronous), pathologic stage, grade, carcinoma-in-situ, lymphovascular invasion (LVI), resection margin, adjuvant chemotherapy or radiation therapy, tumour location,
EP
follow-up, and oncologic outcomes including recurrence and cause of death. Regional lymph node (LN) dissection was performed in patients with enlarged LNs on preoperative imaging
AC C
or with suspicious findings intraoperatively. Adjuvant chemotherapy and radiation therapy were administered to 59 (23.6%) and 15 (6.0%) patients, respectively, according to the clinicians’ decision. After RNU, patients were generally followed every 3–4 mo for the first year, every 6 mo from the second through the fifth year, and thereafter. Follow-up consisted of history, physical examination, blood laboratory tests, urinary cytology, chest radiography, cystoscopy for bladder recurrence, abdominopelvic CT scans and elective bone scans.
Pathology evaluation 5
6
ACCEPTED MANUSCRIPT
All surgical specimens were processed according to standardised procedures, and all slides were thoroughly re-reviewed according to a predefined pathologic-consensus protocol by dedicated genitourinary pathologists at each institute. As described in our prior study, LVI
RI PT
was defined as the presence of tumour cells in the luminal space lined by endothelial cells in hematoxylin-eosin stained sections.16 Pathological staging was based on the 7th AJCC/UICC TNM staging system and tumours were graded according to the 2004 WHO classification.17 Tumour location and depth of parenchymal invasion were determined according to the
SC
three different pT3 subclassifications, i.e., from the studies of Park et al. for Asan,9 Shariat et
M AN U
al. for Cornell,14 and Sassa et al. for Nagoya.15 The Asan subclassification9 does not consider the depth of parenchymal invasion and is based on the tumour location only, i.e., tumours invading the renal parenchyma (Asan-T3a, designated as ‘AT3a’) versus tumours invading peripelvic fat (AT3b). The other two subclassifications differentiate among tumours invading the renal parenchyma according to a depth of invasion of 5 mm from the basement membrane
TE D
(Cornell) and corticomedullary junction (Nagoya). Thus in the Cornell subclassification,14 pT3 RPUC was further classified into infiltration of the renal parenchyma on a microscopic
EP
(< 5 mm) level (Cornell-T3a, designated as ‘CT3a’) and infiltration of the renal parenchyma visible on gross inspection of the resection specimen and/or invasion of peripelvic fat (CT3b).
AC C
Meanwhile, the Nagoya subclassification15 differentiates extension only into the renal medulla (Nagoya-T3a, designated as ‘NT3a’) from extension into the renal cortex and/or peripelvic fat invasion (NT3b).
Outcomes and Statistical Analysis Recurrence-free survival (RFS) and cancer-specific survival (CSS) rates were analyzed. Disease recurrence was defined as local failure in the tumour bed or regional LNs, or distant metastasis. Bladder recurrence following RNU was separately analysed. Cause of death was 6
7
ACCEPTED MANUSCRIPT
determined by the treating surgeons or by chart review corroborated by death certificates. Fisher’s exact test was used to compare frequencies of categorical variables. Student’s t-test was used to compare means of continuous variables. Univariate and multivariate analyses
RI PT
were performed using the Kaplan-Meier method with the log rank test and Cox-proportional hazards regression model. Estimates of predictive accuracy (PA) for RFS and CSS in four different models (a baseline model and models including the three pT3 subclassifications) were quantified by Harrell’s concordance index.18 Increments in PA for each model were
SC
compared. Internal validation with 500 bootstrapping samples was performed.19 In addition,
M AN U
to rule out the potential confounding by LN metastasis and adjuvant treatment, the same analyses were performed in 220 pT3 pN0/pNx patients and 179 pT3 patients who did not receive either adjuvant chemotherapy or radiation. All tests were two-sided, with p < 0.05 considered significant. Statistical analyses were performed using Stata/SE software, version
AC C
EP
TE D
12.1 (Stata Corporation, College Station, Texas).
7
8
ACCEPTED MANUSCRIPT RESULTS Patient characteristics and clinicopathologic associations
Baseline characteristics for the entire patient cohort and their associations with the three pT3
RI PT
subclassifications are shown in Table 1. Median patient age was 66 years (range: 32−93 years). Among the patients included, 30 (12.0%) had LN metastasis, and 81 (32.4%) did not undergo LN dissection. Overall, 122 patients (48.8%, AT3a) had tumours invading the renal
SC
parenchyma and 128 (51.2%, AT3b) had tumours invading peripelvic fat. Among 122 AT3a tumours, 57 and 22 cases were changed into CT3b and NT3b tumours according to the depth
M AN U
of parenchymal invasion, respectively. In the subclassification, patients with pT3b tumours were more likely to undergo adjuvant chemotherapy than those with pT3a tumours (p