Prognostic Factors of Overall Survival in Upper Urinary Tract Carcinoma: A Systematic Review and Meta-analysis

Accepted Manuscript Title: Prognostic Factors of Overall Survival in Upper Urinary Tract Carcinoma: a Systematic Review and Meta-Analysis Author: Fausto Petrelli, Mohamed Ismail Yasser Hussein, Ivano Vavassori, Sandro Barni PII: DOI: Reference:

S0090-4295(16)30496-4 http://dx.doi.org/doi: 10.1016/j.urology.2016.07.036 URL 19954

To appear in:

Urology

Received date: Accepted date:

24-5-2016 29-7-2016

Please cite this article as: Fausto Petrelli, Mohamed Ismail Yasser Hussein, Ivano Vavassori, Sandro Barni, Prognostic Factors of Overall Survival in Upper Urinary Tract Carcinoma: a Systematic Review and Meta-Analysis, Urology (2016), http://dx.doi.org/doi: 10.1016/j.urology.2016.07.036. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof 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.

1

Prognostic factors of overall survival in upper urinary tract carcinoma: A systematic review and meta-analysis Fausto Petrelli MD1, Mohamed Ismail Yasser Hussein MD2, Ivano Vavassori MD2 and Sandro Barni MD1 Affiliations: 1

Oncology Unit, Oncology Department, ASST Bergamo Ovest, Piazzale Ospedale 1,

24047 Treviglio (BG), Italy 2

Urology Unit, Surgical Department, ASST Bergamo Ovest, Piazzale Ospedale 1,

24047 Treviglio (BG), Italy Word count: 3033 Abstract count: 98 Keywords: upper urinary tract; urothelial carcinoma; prognostic factors; overall survival; meta-analysis Corresponding author: Dr. Fausto Petrelli, Azienda Ospedaliera Treviglio, 24047 Treviglio (BG), Italy, [email protected]; +390363424420; fax +390363424380.

Abstract Several prognostic factors that influence overall survival (OS) after radical nephroureterectomy (RNU) for upper urinary tract urothelial carcinoma (UUTUC) have been described. We have performed a systematic review of the literature and metaanalysis. The clinico-pathological factors associated with an increased risk of death

Page 1 of 20

2

were: age, multifocality , lympho-vascular invasion, pT3–4 stage, pT2 vs < pT2 stage, node-positive disease, tumour grade, tumour size, and positive surgical margins. This meta-analysis of published data identified 9 significant predictors of mortality that should be systematically assessed to propose a risk-based strategy for the choice of adjuvant therapy and surveillance after RNU.

1. Introduction Urothelial tumors of the upper urinary tract (Upper Urinary Tract Urothelial Carcinomas—UUTUCs) are relatively rare, accounting for approximately 10% of all renal tumors and only 5% of all urothelial tumors of the urinary tract. Transitional cell carcinoma (TCC) represents more than 95% of UUTUCs. Traditional prognostic factors such as cancer stage and tumor grade, common to other solid tumors have been demonstrated to be prognostic factors for cancer-specific survival in a series of patients treated at MD Anderson Cancer Centre between 1984 and 2004 [1]. In this period, lasting approximately 2 decades, Brown et al. showed that there was no improvement in the outcome, leading to the reconsideration of a potential shift from a traditional surgical-based paradigm, toward a more aggressive approach including multimodal therapy to improve the cure rate. The aim of the treatment and follow-up of UUTUC is the radical removal of all gross diseases and the early detection of any local and/or distant relapse within or outside the urothelium. Standard prognostic factors such as the TNM stage are classically used to predict prognosis. In addition, nomograms have also been useful in predicting cancer-specific survival. In 2014, Seisen et al. developed a specific nomogram that was able to predict disease-related outcome in a multicenter series of 2223 chemotherapy naive patients with resected

Page 2 of 20

3

UUTUC [2]. The 6 variables included in the final model were age, tumor grade, tumor location, pT stage, architecture, and lympho-vascular invasion. The aim of treatment in urothelial cancers is to improve overall survival (OS), and to achieve a definitive cure. In addition, this end point includes any noncancer related deaths. Furthermore, OS is the primary end point of systemic adjuvant therapies in the oncological settings. Owing to the poor prognoses in patients with more advanced disease stages with long-term OS <20–30%, increasing the prognostic accuracy in any resected (nonmetastatic) patient, considered as the best candidate for adjuvant therapy, should be attempted. Therefore, our purpose was to evaluate the significant predictors of survival after radical nephroureterectomy (RNU) for UUTUC by a systematic review of the literature and a meta-analysis of the published data. 2. Materials and methods This systematic review adheres to the guidelines provided by the Preferred Reporting Items for Systematic Reviews and Meta-Analyses report (PRISMA Statement) [3]. 2.1 Search strategy Computerized bibliographic search of the PubMed, Embase, SCOPUS, Web of Science, LILACS, CINAHL, and Cochrane databases from inception to April 2016 was performed. The search terms used were (upper tract OR upper urinary tract OR renal pelvis OR ureter) AND (cancer OR carcinoma) AND (hazard ratio) AND (multivariate OR multivariable OR cox) AND (overall survival), without any language restriction. 2.2 Inclusion criteria

Page 3 of 20

4

The inclusion criteria for this systematic review and meta-analysis were publications that considered patients diagnosed with UUTUC and aimed to determine the independent clinicopathological predictors of death, using multivariate logistic regression analyses. Case reports and meeting abstracts were excluded. Studies were selected according to the following specific inclusion criteria: (i) large studies with more than 100 patients who had been surgically treated for UUTUC; (ii) studies that defined events as deaths of any cause (cancer or noncancer-related), whichever came first; (iii) studies that included patients with muscle-invasive (>pT2) UUTUC; (iv) studies that provided hazard ratios (HRs) from multivariate logistic regression analyses, with their corresponding 95% confidence intervals (CIs). Finally, if 2 or more studies had reported results from overlapping case series (i.e., by the same institution), we selected the one with the more recent publication date. 2.3 Data extraction and outcome measures Data from all included studies were independently extracted by two authors (F. P. and Y. U.) and subsequently crosschecked to ensure their accuracy. Discrepancies in data extraction were resolved by consulting a third author (S. B.). We recorded the country and the number of patients included, along with the rates of radical surgery and chemotherapy (adjuvant or neoadjuvant), and the median follow-up period for assessing the risk of death after previous UUTUC. In addition, the type of study and median OS (if reported) were extracted. The primary outcomes were findings reported in at least 3 papers concerning the adjusted prognostic effect of clinical and pathological factors on OS. We used the Newcastle–Ottawa Scale (NOS) for assessing the risk of bias. This scale assesses the likelihood of bias in 3 domains [4, 5]: (1) selection of the study groups; (2) comparability of the groups; and (3) ascertainment of exposure and outcome. Studies with scores ≥7 were considered to

Page 4 of 20

5

have low risks of bias, those with scores of 4–6 were considered to have moderate risks of bias, and those with scores <4 were considered to have high risks of bias. 2.4 Statistical analyses Because all studies were prospective nonrandomized cohort studies, observational studies and/or retrospective series that reported adjusted HRs from multivariate analyses were preferred, and given the identified clinical heterogeneity, random effect models were adopted for each analysis. A meta-analysis of each potential risk factor for death was performed using the predictor effect, which was assessed using the adjusted HRs that were available along with their corresponding 95% CIs. The cumulative effects of the factors of interest were evaluated using the inverse variance method. Statistical heterogeneity was assessed using both the Tau2 and the I2 statistics, which describe the percentage of total variation across studies caused by heterogeneity rather than by chance, which resulted in the use of the random-effects model based on the DerSimonian method for estimating the Tau value [6]. We suggested that a median follow-up period of 3 years was adequate. The extracted data were aggregated into a meta-analysis using the RevMan 5.3 software (Cochrane Collaboration, Copenhagen, Denmark).

3. Results In accordance with all the previously described inclusion criteria, 20 articles among the 443 publications retrieved, which were published between 2007 and 2016, were selected [7–26]. The PRISMA flow-chart depicting the procedure for the systematic literature search and selection of studies is shown in Figure 1. The number of participants in each selected study ranged 100–1029 (mean, 400; median, 304).

Page 5 of 20

6

Overall, a total of 7816 patients were analysed in the meta-analysis. Among the included studies 11 were Eastern, 3 were European, and 4 were American/Canadian, respectively. Two publications included international multicenter series. The median follow-up period was 47.5 months (range, 19.8–70.8 months; Table 1). The median follow-up time was not reported in 2 studies. The series were mainly retrospective in nature (16/20). The median NOS scale classification was 7. Approximately all of the patients underwent RNU.

3.1 Age Multivariate analyses conducted in 14 studies identified older age as a poor prognostic factor. The conflicting results concerning the predictive value of age translated to inter-study heterogeneity of observed outcomes according to an I2 statistic of 78% (p < 0.001). A meta-analysis of available HRs revealed that age (HR, 1.04; 95% CI, 1.02–1.06; p < 0.001) was a weak but significant predictor of mortality. 3.2 pT3–4 versus pT2) at diagnosis was identified as a predictor of increased mortality in 12 studies. The data were homogeneous (P for heterogeneity, 0.08, I2 = 39%) and the cumulative analyses of the available HRs demonstrated that more invasive pT stage (HR, 2.62; 95% CI, 2.17–3.16; p < 0.001) was a significant predictor of mortality (Fig. 2). 3.3 Node-positive disease (pN+ UUTUC) Lymph node involvement was reported as a risk factor for death in 9 studies. The predictive value of pN+ stage for the risk of mortality was moderately heterogeneous.

Page 6 of 20

7

In fact, inter-study heterogeneity of outcomes was reflected by an I2 statistic of 51% (p = 0.04). A meta-analysis of the available HRs revealed that pN+ stage (HR, 2.16; 95% CI, 1.75–2.66; p < 0.001) was a significant predictor of mortality (Supplementary Fig. 1). 3.4 Tumor grade Tumor grade was reported as a prognostic factor for OS in 13 studies with moderate heterogeneity (I2 = 57%; p = 0.006). A meta-analysis of the available HRs revealed that tumor grade (HR, 1.7; 95% CI, 1.38–2.09; p < 0.001) was a significant predictor of mortality (Supplementary Fig. 2). 3.5 Lympho-vascular invasion (LVI+ UUTUC) Eight studies reported that LVI+ was included as a covariate in the multivariate analysis for OS. The results concerning the predictive value of LVI+ for the risk of death were controversial. This translated to a high inter-study heterogeneity of observed outcomes according to an I2 statistic of 67% (p = 0.003). However, a metaanalysis of the available HRs revealed that LVI+ (HR, 1.55; 95% CI, 1.18–2.06; p = 0.002) was a significant predictor of overall mortality in patients with UUTUC (Supplementary Fig. 3). 3.6 Tumor multifocality Tumor multifocality as a risk factor for mortality was reported in 3 studies. Inter-study heterogeneity of outcomes was absent according to an I2 statistic of 0% (p = 0.83). A meta-analysis of available HRs revealed that tumor multifocality (HR, 1.6; 95% CI, 1.21–2.1; p < 0.001) was a significant predictor of mortality. 3.7 Positive resection margins (R+)

Page 7 of 20

8

Nonradical resection (R+ surgery) was reported as a predictor of mortality in 5 studies. Overall, the heterogeneity was moderate (I2 = 51%, p = 0.09). A metaanalysis of all the studies demonstrated that positive resection margins are risk factors of death (HR, 1.75; 95% CI, 1.26–2.43; p < 0.001). 3.8 Tumor size Larger median tumor size (3–3.5 cm) was presented as a predictor of mortality in 3 studies. The data were homogeneous, and a cumulative analysis of available HRs revealed that tumor size (HR, 2.54; 95% CI, 2–3.23; p < 0.001) was a significant predictor of mortality. 3.9 pT2 versus pTa-is-1 stages In 8 studies, muscle-invasive UUTUCs were found to be associated with an increased risk of death compared to superficial disease (pTa, pTis, pT1). The data had no heterogeneity, and a cumulative analysis of the available HRs demonstrated that the pT2 stage (HR, 1.72; 95% CI, 1.36–2.17; p < 0.001) was a significant predictor of mortality. Visual inspection of the funnel plots revealed no obvious publication bias for the predictors studied. 3.10 Other factors associated with survival (in 1 paper only) Tumor necrosis, histology variants, tumor architecture (papillary vs. other), sex, blood transfusions, bladder recurrence or metachronous bladder tumor, IMP3 and KPNA2, LDH, site (ureteral vs. renal pelvis), pain and weight loss, progression to renal failure, inflammation parameters, adjuvant CT, renal function, ratio of positive nodes to

Page 8 of 20

9

nodes removed, performance status, and lymphocyte/monocyte ratio were found significantly associated with poorer OS only in 1 study each. 3.11 Other covariates not associated with OS Among the covariates significantly associated with reduced survival, 4 were not associated with OS in multivariate analysis: Hb level, laparoscopic RNU, smoking status and carcinoma in situ. After exclusion of studies that included patients treated with (neo)adjuvant chemotherapy, pT3-4, grade, size and multifocality remained the only variables significantly associated with adverse prognosis. 4. Discussion In the present systematic review and meta-analysis, we identified 8 strong predictors of OS, supporting the concept that the biological and anatomical characteristics of the disease burden are of paramount importance in predicting prognosis and guiding treatment and follow-up. In fact, we identified advanced T and N stages (pT3–4 and/or node-positive tumors), high tumor grade, larger tumor size/multifocality, positive LVI, and nonradical resection to be independent prognostic factors of survival in 7816 UUTUC patients treated with nephroureterectomy. Adjuvant chemotherapy, albeit used in a minority of the patients analysed, was not a positive prognostic factor, suggesting that systemic treatment is inadequate in abrogating the bad prognoses associated with tumors that are more aggressive. All the prognostic factors validated in the present study are currently included in the pathology reports of surgical specimens and in the guidelines of the College of American Pathologists [27]. Not all patients underwent neoadjuvant therapy; therefore, the pathologic variables were not influenced by the effect of chemotherapy. In addition,

Page 9 of 20

10

approximately all subjects underwent standard resection treatment, suggesting optimal staging. To the best of our knowledge, the present study is the first metaanalysis that systematically evaluated any variable that was independently associated with survival in UUTUC. Therefore, the results are of paramount importance for any decision-making in the oncological setting. In a large series by Rink et al., among the 2494 patients treated with RNU without neoadjuvant therapy, approximately 25% showed relapse, and 80% of the patients who showed relapse died within 2 years from the detection of recurrence [28]. These data confirm that after recurrence, UUTUC is incurable in almost all cases. These results, coupled with the results of the present study, stress the importance of using the identified predictors in the selection of appropriate neoadjuvant treatments for large (>3 cm) tumors and for macroscopically node-positive cases; as well as to expand the indications for adjuvant therapy based on the biology of the disease (high grade, LVI+, and multifocal/pT3–4 tumors). Finally, more intensive follow-up strategies should be implemented in high-risk patients. Neoadjuvant more than adjuvant chemotherapy is the standard of care in muscleinvasive bladder cancer. Similar data does not exist for UUTUC and case series on preoperative treatment are limited [29]. In a systematic review and meta-analysis of the adjuvant and neoadjuvant chemotherapy for UUTUC, a significant OS and disease-free survival benefit was conferred by adjuvant chemotherapy, limited by the retrospective nature of the studies and their relatively small sample sizes [30]. According to the NCCN and European Association of Urology, adjuvant chemotherapy should instead be considered in the treatment of stage pT2–4 and/or pN+ disease, even though the strength of recommendation is low [31, 32].

Page 10 of 20

11

Grade, lympho-vascular invasion, and multifocality were identified as prognostic factors in the present meta-analysis and the architecture (papillary or sessile) and site (ureteral vs. pelvic) of the tumor were not identified as prognostic factors. In a meta-analysis of 17 studies performed by Ku et al., lympho-vascular invasion was previously demonstrated as a predictor of mortality [33]. Divergent histology, although described as a poor prognostic factor in some papers, was not significantly associated with poor outcome [34]. Other authors showed that in addition to multifocality, the ureteral site was an independent prognostic factor for cancer-specific survival but not for OS in a meta-analysis of 17 studies [35]. Surprisingly, no laboratory variables (e.g., anemia or inflammation parameters) or performance status were found to be significantly associated at all neither in more than 1 series with OS nor with cancer specific-survival, as found, for example, in bladder cancer. The reasons could be the lack of sufficient data to prove statistically significant power and/or a different biology [36]. Finally, follow-ups with cytological urine analysis to discover intravesical/contralateral and distant recurrences are imperative in the initial years after diagnosis, in particular, in high-risk disease. After radical RNU, local recurrence was rare, with metachronous bladder recurrence rates ranging 15–50%. The bladder should be surveyed routinely by using the risk factors for intravesical recurrence that were recently evaluated in a systematic review and meta-analysis. Some of these factors (R+ surgery, pT stage, and multifocality) are the same predictors of survival identified in the present meta-analysis and should be systematically assessed to propose adjuvant intravesical instillation and even systemic chemotherapy other than cystoscopic surveillance after RNU [37]. Detection of distant recurrences is also warranted in high-risk disease. The 2015 UUTUC European Association of Urology guidelines suggest a follow-up period of at least 5 years, with both cystoscopy/urinary Page 11 of 20

12

cytology to be performed at 3 months and yearly thereafter, and CT urography every 6 months over a period of 2 years and yearly thereafter [32]. However, a more intensive surveillance of distant recurrences (nodes, liver, or lung) might be warranted in patients with higher stage and/or node-positive diseases, as well as those with high-grade or multi-focal disease, because they have a higher likelihood of distant recurrence compared to local recurrence. The present systematic review and meta-analysis had some limitations. First, data were mainly retrospective, with an overall moderate–high heterogeneity in most analyses. Secondly, some studies have relatively short follow-up periods, likely to fail in capturing most of the survival events. Thirdly, some variability in the evaluation of pathological variables such as lympho-vascular invasion could have been taken into account. However, the present analysis validated the knowledge on the prognostic factors for UUTUCs, even though mainly in a retrospective manner. Some of these variables could eventually be adopted by clinicians in clinical practice. The present study included 20 publications, encompassing approximately 8000 patients, and represented a more comprehensive analysis of literature on UUTUCs, thereby providing to both urologists and oncologists the criteria above described for selecting the patients at higher risk of death and may be candidates for further therapy (adjuvant chemotherapy?) or more intense follow-up. In addition, the results highlighted the importance of reporting an exhaustive pathological analysis of surgical specimens by pathologists. Overall survival, instead of cancer-specific survival, was chosen to observe the effect on risk of death as it represents the main endpoint of cure in all cancer studies. Other intermediate endpoints such as disease recurrence have not been considered because the topic of other similar revisions are poorly reported in our series [37]. We recognize however that, due to low number of chemotherapy exposure (0–30%) and high number of surgeries (100%), risk of death Page 12 of 20

13

due to anticancer treatment complications could be very low (or absent) and similar among series, but risk of death due to other compelling causes of mortality (cardiovascular disease, smoking status) could be high, and partially influence the probability of all causes of death. In conclusion, owing to the lack of specific biomarkers in clinical practice, which could predict early recurrence and improve survival, classical prognostic factors based on standard TNM staging system and histological features are still useful for predicting the outcome. While awaiting the validation of all prognostic and predictive markers that are clinically available [38] in randomized prospective trials, we identified significant patient-specific (age), tumor-specific (i.e., node-positive disease, tumor grade, tumor size, multifocality, pT2–4 stage, and positive LVI), and treatmentspecific (e.g., nonradical resection surgery) predictors of OS that should be systematically assessed in order to be able to propose a risk-adapted approach for surveillance and postoperative treatment, with the aim of improving the cure rates of these rare urological tumors. References

1. Brown GA, Busby JE, Wood CG et al. Nephroureterectomy for treating upper urinary tract transitional cell carcinoma: Time to change the treatment paradigm? BJU Int. 2006 Dec;98(6):1176-80. 2. Seisen T, Colin P, Hupertan V et al. Postoperative nomogram to predict cancer-specific survival after radical nephroureterectomy in patients with localised and/or locally advanced upper tract urothelial carcinoma without metastasis. BJU Int. 2014 Nov;114(5):733-40.

Page 13 of 20

14

3. Liberati A, Altman DG, Tetzlaff J, et al. The PRISMA statement for reporting systematic reviews and meta-analyses of studies that evaluate health care interventions: explanation and elaboration. PLoS Med 2009;6:e1000100. 4. Deeks JJ, Dinnes J, D’Amico R , et al. Evaluating non-randomized intervention studies. Health Technol Assess 2003;7, iii–x, 1–173. 5. http://www.ohri.ca/programs/clinical_epidemiology/oxford.asp last accessed 29th December 2015. 6. Higgins JPT, Thompson SG, Deeks JJ et al. Measuring inconsistency in metaanalyses. BMJ 2003;327:557-60. 7. Hellenthal NJ, Shariat SF, Margulis V et al. Adjuvant chemotherapy for high risk upper tract urothelial carcinoma: results from the Upper Tract Urothelial Carcinoma Collaboration. J Urol. 2009 Sep;182(3):900-6. 8. Makise N, Morikawa T, Kawai T et al. Squamous differentiation and prognosis in upper urinary tract urothelial carcinoma. Int J Clin Exp Pathol. 2015 Jun 1;8(6):7203-9. 9. Lee KS, Kim KH, Yoon YE et al. Impact of adjuvant chemotherapy in patients with upper tract urothelial carcinoma and lymphovascular invasion after radical nephroureterectomy. Korean J Urol. 2015 Jan;56(1):41-7. 10. Tai YS, Chen CH, Huang CY et al. The effect of tumor location on oncologic outcomes in patients with upper urinary tract urothelial carcinoma stratified by pathologic stage. Urol Oncol. 2016 Jan;34(1):4.e19-25. 11. Zhang XK, Zhang ZL, Lu X et al. Prognostic Significance of Preoperative Serum Lactate Dehydrogenase in Upper Urinary Tract Urothelial Carcinoma. Clin Genitourin Cancer. 2016 Jan 22.

Page 14 of 20

15

12. Chung SD, Huang KH, Lai MK et al. CKD as

a risk factor for bladder

recurrence after nephroureterectomy for upper urinary tract urothelial carcinoma. Am J Kidney Dis. 2007 Nov;50(5):743-53. 13. Dragicevic D, Djokic M, Pekmezovic T et al. Survival of patients with transitional cell carcinoma of the ureter and renal pelvis in Balkan endemic nephropathy and non-endemic areas of Serbia. BJU Int. 2007 Jun;99(6):135762. 14. Fairey AS, Kassouf W, Estey E et al. Comparison of oncological outcomes for open and laparoscopic radical nephroureterectomy: results from the Canadian Upper Tract Collaboration. BJU Int. 2013 Oct;112(6):791-7. 15. Grasso M, Fishman AI, Cohen J et al. Ureteroscopic and extirpative treatment of upper urinary tract urothelial carcinoma: a 15-year comprehensive review of 160 consecutive patients. BJU Int. 2012 Dec;110(11):1618-26. 16. Kang M, Kim HS, Jeong CW et al. Conditional Survival and Associated Prognostic Factors in Patients with Upper Tract Urothelial Carcinoma after Radical Nephroureterectomy: A Retrospective Study at a Single Institution. Cancer Res Treat. 2016 Apr;48(2):621-31. 17. Kim HS, Jeong CW, Kwak C et al. Can body mass index predict survival outcomes in patients treated with radical nephroureterectomy for upper-tract urothelial carcinoma? Int Urol Nephrol. 2015 Aug;47(8):1311-20. 18. Lee DJ, Xylinas E, Rieken M et al. Insulin-like growth factor messenger RNAbinding protein 3 expression helps prognostication in patients with upper tract urothelial carcinoma. Eur Urol. 2014 Aug;66(2):379-85. 19. Mason RJ, Kassouf W, Bell DG et al. The contemporary role of lymph node dissection during nephroureterectomy in the management of upper urinary

Page 15 of 20

16

tract

urothelial

carcinoma:

the

Canadian

experience.

Urology.

2012

Apr;79(4):840-5. 20. Rink M, Soave A, Dahlem R et al. Impact of Perioperative Allogenic Blood Transfusion on Survival After Radical Nephroureterectomy for Upper Tract Urothelial Carcinoma. Clin Genitourin Cancer. 2016 Feb;14(1):96-104. 21. Shi B, Su B, Fang D et al. High expression of KPNA2 defines poor prognosis in patients with upper tract urothelial carcinoma treated with radical nephroureterectomy. BMC Cancer. 2015 May 9;15:380. 22. Yan S, Liu L, Wei Q, Liao H, Song T, Lei J, Yang L, Yuan Z, Jiang Y, Fu G, Li Y, Cao D. Impact of tumor size on prognosis of upper urinary tract urothelial carcinoma after radical nephroureterectomy: a multi-institutional analysis of 795 cases. BJU Int. 2016 Mar 2. 23. Sung HH, Jeon HG, Jeong BC et al. Clinical significance of prognosis using the neutrophil-lymphocyte ratio and erythrocyte sedimentation rate in patients undergoing radical nephroureterectomy for upper urinary tract urothelial carcinoma. BJU Int. 2015 Apr;115(4):587-94. 24. Yeh HC, Jan HC, Wu WJ et al. Concurrent Preoperative Presence of Hydronephrosis and Flank Pain Independently Predicts Worse Outcome of Upper Tract Urothelial Carcinoma. PLoS One. 2015 Oct 15;10(10):e0139624. 25. Inman BA, Tran VT, Fradet Y et al. Carcinoma of the upper urinary tract: predictors of survival and competing causes of mortality. Cancer. 2009 Jul 1;115(13):2853-62. 26. Hutterer GC, Sobolev N, Ehrlich GC et al. Pretreatment lymphocyte-monocyte ratio as a potential prognostic factor in a cohort of patients with upper tract urothelial carcinoma. J Clin Pathol. 2015 May;68(5):351-5.

Page 16 of 20

17

27. Amin MB, Srigley JR, Grignon DJ et al. Updated protocol for the examination of specimens from patients with carcinoma of the urinary bladder, ureter, and renal pelvis. Arch Pathol Lab Med. 2003 Oct;127(10):1263-79. 28. Rink M, Sjoberg D, Comploj E et al. Risk of cancer-specific mortality following recurrence after radical nephroureterectomy. Ann Surg Oncol. 2012 Dec;19(13):4337-44. 29. Matin SF, Margulis V, Kamat A et al. Incidence of downstaging and complete remission after neoadjuvant chemotherapy for high-risk upper tract transitional cell carcinoma. Cancer. 2010 Jul 1;116(13):3127-34. 30. Leow JJ, Martin-Doyle W, Fay AP et al. A systematic review and metaanalysis of adjuvant and neoadjuvant chemotherapy for upper tract urothelial carcinoma. Eur Urol. 2014 Sep;66(3):529-41. 31. http://www.nccn.org/professionals/physician_gls/pdf/bladder.pdf (v 1.2016) last accessed 2016 April 18th 32. Rouprêt M, Babjuk M, Compérat E et al. European Association of Urology Guidelines on Upper Urinary Tract Urothelial Cell Carcinoma: 2015 Update. Eur Urol. 2015 Nov;68(5):868-79. 33. Ku JH, Byun SS, Jeong H et al. Lymphovascular invasion as a prognostic factor in the upper urinary tract urothelial carcinoma: a systematic review and meta-analysis. Eur J Cancer. 2013 Aug;49(12):2665-80. 34. Kim SP, Frank I, Cheville JC, et al. The impact of squamous and glandular differentiation on survival after radical cystectomy for urothelial carcinoma. J Urol 2012;188:405–9. 35. Wu Y, Dong Q, Liu L et al. The impact of tumor location and multifocality on prognosis for patients with upper tract urothelial carcinoma: a meta-analysis. Sci Rep. 2014 Sep 15;4:6361. Page 17 of 20

18

36. Mathieu R, Bensalah K, Lucca I, Mbeutcha A, Rouprêt M, Shariat SF. Upper urinary tract disease: what we know today and unmet needs. Transl Androl Urol. 2015 Jun;4(3):261-72. 37. Seisen T, Granger B, Colin P et al. A Systematic Review and Meta-analysis of Clinicopathologic Factors Linked to Intravesical Recurrence After Radical Nephroureterectomy to Treat Upper Tract Urothelial Carcinoma. Eur Urol. 2015 Jun;67(6):1122-33. 38. Matin SF, Shariat SF, Milowsky MI et al. Highlights from the first symposium on upper tract urothelial carcinoma. Urol Oncol. 2014 Apr;32(3):309-16. Supplementary Fig.1 Forrest plots of meta-analyses of pN+ stage as predictor of overall survival Supplementary Fig.2 Forrest plots of meta-analyses of histological grade as predictor of overall survival Supplementary Fig.3 Forrest plots of meta-analyses of lymphovascular invasion stage as predictor of overall surviva

Fig.1 Overview of trials search and selection. Fig. 2 Forrest plots of meta-analyses of pT3-4 stage as predictor of overall survival

Page 18 of 20

1

Table 1: characteristics of included studies Author/year [reference]

Type of study

Country

N° pts

Surgery (RNU %)

Lee KS/2015 [9] Tai/2016 [10] Makise/2015 [8] Zhang/2016 [11] Shibing/2016 [22]

Retrospecti ve Prospective Prospective

Korea

344

Taiwan Japan

Retrospecti ve Retrospecti ve

Sung/2015 [23]

Rink/2016 [20] Grasso/2012 [15] Lee DJ/2014 [18] Kang/2016 [16] Shi/2015 [21] Fairey/2013 [14] Mason/2012

Median follow up (months) 53.9

Median OS (months)

Covariates associated with OS

NOS Scale

100

Neoadj/ Adj CT (%) 0/18.6

31% at 5y

Age, pT3-4, pN+, LVI+,

8

503 140

100 100

0/1.6 0/30

52 53

NR NR

pT, age Age, grade, LVI+

8 7

China

100

100

0/0

45.8

37

7

China

795

100

0/25.4

32

52.6% at 5y

Retrospecti ve

Korea

410

100

0/22.2

40.2

50.4-90.7% at 5y*

Retrospecti ve Prospective series

German

285

100

0/16.8

30

42% ay 5y

US

160

100

0/0

NR

59% at 5y

Retrospecti ve Retrospecti ve Retrospecti ve Retrospetiv e Retrospetiv

Internatio nal Korea

622

100

1.9/10.3

27

330

100

0/25.8

58,7

15.6-64.8% at 5y° NR

pT, metacronous bladder ca, tumor necrosis, LDH pT, site, pN+, grade, variant histology, margin +, size, tumor architecture pT3-4, age, pN+, tumor architecture, inflammation risk pT3-4, pN+, LVI+, adj CT, blood transfusion Age, grade, bladder recurrence, progression to renal failure Age, pN+, pT3-4, multicentric disease, adj CT, IMP3+ Age, site

China

176

100

0/0

36

7

Canada

849

100

0/0

26,4

Sex, pT2-4 stage, size, high KPNA2 Age, pT3-4, grade

Canada

102

100

0/0

19,8

LNR, site, age, multicentric

6

77.3-88-2% at 5y^ 67-68% at 5y# 22.3-66.1%

7 8 7 6 7 8

6

Page 19 of 20

2

[19] Yeh/2015 [24] Kim/2015 [17] Chung/2007 [12] Dragicevic/200 7 [13] Inman/2012 [25]

Hutterer/2015 [26]

e Retrospecti ve Retrospecti ve Retrospecti ve Prospective

Taiwan

9 472

at 5y** 64.2-75.5% at 5y°° 58.8-76.3% at 5y^^ NR

disease, pT3 Age, renal function, pT3-4, pN+, Age, pT2, pT3-4, grade, LVI+, multicentric disease, margin + pT2-4

100

0/18

48,6

Korea

445

100

0/23.8

50,3

Taiwan

150

100

0/0

47,5

Serbia

114

100

0/0

67

Sex, size, pT3-4

7

70,8

0-90% at 5y## NR

Retrospecti ve

Canada

168

1/5

Retrospecti ve

Austria

182

Retrospecti ve

Internatio nal

542

80 (also: 4 RN, 1 ureterectomy, 4 TUR,11 distal ureterectomy) 100 (in selected cases segmental ureterectomy) 100

Age, pain/weight loss, grade

8

NR

NR

NR

Age, pT2-4, lymphocyte– monocyte ratio

6

0/22

26

7 8 7

26 (no adj Age, PS, pT3-4 or pN+, grade 7 CT) & 24 (with adj CT) pT, pathological stage; LVI+, limphovascular invasion present; OS, overall survival; pts, patients; RNU, radical nephroureterectomy; NR, not reported; LDH, lactate dehydrogenase; pN+, pathological node positive disease; adj CT, adjuvant chemotherapy; IMP3+, Insulin-like Growth Factor Messenger RNA-binding Protein 3 positive disease; PS, performance status; LNR, ratio of positive nodes removed to nodes removed; *, according to inflammation risk class; °, according to expression or not of IMP3; ^, according to high or low expression of KPNA2; #, according to open or laparoscopic surgery; **, according to nodal status; °°, according to hydronephrosis status; ^, according to weight status; ##, according to stage;

Hellenthal/2009 [7]

Page 20 of 20