Risk stratification for upper tract urinary carcinoma.

Upper tract urinary carcinoma (UTUC) is a heterogeneous group of rare tumors. The aim of this article is to critically review current therapeutic strategies and to propose a change in the risk-stratification of the disease. A non-systematic review of the literature was performed using the Medline database with the search terms: "upper tract urothelial carcinoma" together with "prognostic factor", "risk stratification", "risk factor", "recurrence", "predictive tool", "nomograms" and "treatment". Preoperative risk factors can be viewed as patient-related risk factors (gender, age, ethnicity, body mass index, smoking status, or genetic factors), or tumor-related risk factors (stage, grade, size, architecture, multifocality, ureteric obstruction). Several biomarkers, available either in blood, urine, or the tumor itself have also been proposed. However, many of these prognostic factors lack accuracy and validation in predicting oncological outcomes, despite their use in predictive tools. After risk stratification, kidney-sparing strategies should be considered (endoscopic management and segmental ureterectomy) and could benefit from new diagnostic tools and technical improvements in in situ adjuvant endocavitary instillations. Radical nephroureterectomy remains the first choice therapy for high-grade disease and will probably be associated with other treatments in the future (lymphadenectomy, perioperative chemotherapy, or immunotherapy). In view of the major recent improvements in UTUC treatment strategies, a new classification should be proposed, including low-, intermediate-, high- and very high-risk disease. Subgroup analysis of good quality trials and better understanding of UTUC risk factors will help validate this new approach toward more personalized medicine. 2020 Translational Andrology and Urology. All rights reserved.

Introduction

Upper tract urinary carcinoma (UTUC) is a rare type of urothelial cancer, with an estimated annual incidence of 1–2 cases per 100,000 inhabitants (1). It accounts for 5–10% of all urothelial carcinomas (2), and while it shares similar features with urothelial carcinoma of the bladder, it represents a specific disease with specific management and recommendations.

Radical nephroureterectomy (RNU) is still considered to be the standard of care for most localized UTUCs in various recommendations. However, RNU carries a significant risk of morbidity and impaired renal function. Various strategies have emerged and are constantly developing to manage UTUC conservatively, including endoscopic or segmental resection, sometimes associated with perioperative (neo-)adjuvant treatment, whether localized or systemic. These strategies were initially limited to imperative indications (e.g., anatomic or functional solitary kidney), but the development of ureteroscopy over the past few decades has enhanced its role in the management of UTUC (3). These kidney-sparing strategies (KSS) appear to be a reasonable and safe option in well-selected patients even with a normal and functional contralateral kidney. Survival outcomes are similar after KSS or RNU in low-risk patients according to an EAU guideline panel systematic review (4).

As in bladder cancer, patient selection and risk stratification appears to be of utmost importance in UTUC. However, while bladder cancer has many validated per-operative risk stratification tools and objective criteria, UTUC patients remain poorly selected. Due to the low incidence of UTUC, the majority of studies consist of small, single institution, retrospective cohorts resulting in the lack of level 1 evidence for high-grade recommendations.

The aim of this brief review is to provide an overview of current risk stratification tools for UTUC patients and to offer an insight into how these patients can be better stratified in the light of new management options.

Materials and methods

Literature review

A non-systematic literature search was performed using the PubMed/Medline databases. Articles published after 2000 were screened using a combination of the following terms: “upper tract urothelial carcinoma” together with “prognostic factor”, “risk stratification”, “risk factor”, “recurrence”, “predictive tool”, “nomograms” and “treatment”. The recommendations of the main urological societies were also screened and additional articles were selected by cross-referencing the bibliography of previously selected articles.

Preoperative predictive factors for UTUC

Patient-related factors

Although the incidence of UTUC is higher in men than in women, female gender was thought to be associated with a worse oncologic outcome (5). However, after adjustment for other prognostic factors, no significant difference was found in cancer-specific (CSS) and overall survival (OS) between the two genders (5,6).

The same goes for advanced age. While many epidemiological studies suggested that advanced age was an independent factor for tumor recurrence and shorter CSS (7,8), as well as for invasive tumor patterns, large multi-institutional studies showed that advanced age was not associated with worse survival when adjusted for performance status (9).

The relationship between ethnicity and UTUC outcomes is also debatable. African-Americans were found to have a shorter survival than other ethnic groups (10), but this finding could be explained by worse access to care for African-Americans in the USA. Another study comparing Japanese, European, and US Caucasian patients did not find any differences in survival (11).

Altogether, the differences observed regarding gender, age, or ethnicity may reflect change in tumor biology and aggressiveness, and also heterogeneity in access to health care, they should in our opinion not be considered in treatment decision algorithms.

On the other hand, smoking is a well-established risk factor for UTUC. Long-term smokers (more than 20 cigarettes a day for more than 20 years) are at greater risk of more advanced disease stage or experiencing recurrence after RNU (12). Patients with specific factors for UTUC, such as Lynch syndrome, or long-term exposure to aristolochic acid or phenacetin should also be considered at higher risk of recurrence (13).

Tumor-related factors

The most accurate independent factors for outcome in UTUC remain stage and grade, which are evaluated by cytology, biopsy of the tumor, and imaging techniques. High-grade cytology is a significant predictor of more advanced disease, non-organ confined disease, loco-regional and distant recurrence (14). Cytology obtained during voiding or cystoscopy suffers from low sensitivity for low-risk disease (15), but has very good specificity if obtained in situ (selective ureteral cytology), notably for high-grade disease, exceeding 90% (16). Ureteroscopic biopsy has good accuracy for tumor grade, ranging from 69‒91% (17), but low performance for tumor stage. However, bioptic grading may be used to predict tumor stage: 68‒100% of grade 1 tumors on biopsy are non-muscle invasive tumors, while 62‒100% of grade 3 tumors on biopsy are ≥pT2 tumors (18). Pre-operative T-staging therefore relies mainly on imaging. The positive predictive value of computed tomography (CT) to detect muscle invasive tumors is high, despite a risk of under-staging tumors. Wang et al. (19) recently demonstrated that tumor stage is a significant prognostic factor for CSS and all-cause mortality in UTUC patients, and tumor stage and grade should therefore be used in decision treatment algorithms.

Ureteroscopy and imaging also provide important information about localization, size, and multifocality of the tumor. The impact of tumor location on oncologic outcomes is still under debate. Some studies have suggested that the rate of organ-confined tumors is the same irrespective of whether the tumor lies in the ureter or the renal pelvis (20), while others show that ureteral location is associated with a worse outcome (21). Tumors >3 cm in size were associated with a worse outcome in patients treated by RNU in retrospective studies (22). Similarly, tumor multifocality has been shown to affect CSS after RNU (23), and some authors have suggested that RNU is a more reasonable treatment option for patients with multifocal disease (24). This also accounts for the difficulty of selective treatment (endoscopic or extirpative) when the tumor is very large or multifocal.

Ureteroscopy can also help identify tumor architecture. Sessile architecture has been shown to be associated with a worse oncologic outcome (non-organ confined disease, high-grade disease, cancer recurrence) than papillary architecture (25). Most studies show that preoperative hydronephrosis is also associated with more advanced disease [more advanced T stage, more frequent lymphovascular invasion, and high-grade disease (26-28)].

Biomarkers

Aside from pathologic and radiologic factors like stage, grade, size, and multifocality, genetic markers can help predict the prognosis of UTUC. For example, microsatellite instability has been identified in many cancers, especially those associated with Lynch syndrome, and is associated with a better outcome in T2–T3N0 UTUC patients (29). Low methylation of the VIM promoter predicts worse CSS in UTUC patients (30), while FGFR3 mutations are frequently associated with lower stage and better survival of UTUC patients (31).

Several tissues biomarkers have also been investigated. Ki-67 expression was shown to be associated with adverse pathologic features in 475 patients treated by RNU (32). Similar findings were found for PI3K and Cyclin D (two mTOR biomarkers) (33). PD1 and PD-L1 expression have also been proposed as prognostic factors in UTUC (34,35). In a retrospective cohort of more than 400 patients undergoing extirpative surgery, PD-1 expression was associated with worse CSS and OS. In contrast, PD-L1 expression was predictive of a more favorable outcome.

Blood and urine biomarkers have also been investigated as prognostic markers in UTUC. Pre-operative C-reactive protein level was shown to be associated with a poor prognosis (36). Pre-operative neutrophil: lymphocyte ratio was also significantly associated with worse pathologic features and higher CSS (37). Other easily available markers such as platelet count, fibrinogen, albumin, or hemoglobin have been investigated. Most of these biomarkers however, whether derived from the tumor itself, from blood, or from urine, still lack validation in large studies.

Predictive tools

Based on these identified prognostic factors, and despite the challenge of their accurate characterization based on routine examination through imaging, endoscopy, or biopsy, multi-institutional research groups have developed per-operative models to predict high-grade, muscle-invasive or non-organ confined disease.

Several models focus on the prediction of muscle-invasive or non-organ confined disease. In 2010, Brien et al. first proposed the combination of hydronephrosis, high-grade disease on biopsy, and positive urinary cytology to predict advanced UTUC in a small retrospective cohort (38). If no criteria were present, the negative predictive value for muscle-invasive or non-organ confined disease reached 100%. However, when all three criteria were met, the positive predictive value for muscle-invasive or non-organ confined disease was 89% and 73%, respectively. Margulis et al. also proposed a model to predict non-organ confined disease with an accuracy of 77%, associating tumor grade, architecture, and location (39). In 2012, Favaretto et al. combined preoperative hydronephrosis, tumor grade, and location to predict muscle-invasive disease or locally advanced disease with an accuracy of 70% (40). A cohort of 683 Chinese patients helped to develop a model predicting non-organ confined disease with an accuracy of 79% based on gender, tumor grade, architecture, and multifocality (41). Finally, another model to predict muscle-invasive or non-organ confined disease was developed in 2015, including gender, locally advanced stage on pre-operative workup, and positive cytology with an accuracy of 65.3% and 67.2% for muscle-invasive and non-organ confined disease respectively (42). These models should be used with care in clinical practice and decision making as none have benefited from external validation.

Treatment options for localized disease

KSS

Endoscopic management and segmental resection

Ureteroscopic management of UTUC has emerged as an alternative to RNU for localized disease. In a single center, retrospective cohort with 52 months follow-up, 112 patients with tumors for which complete endoscopic eradication was achievable were treated with flexible ureteroscopic laser photoablation (43); the progression-free survival rate was 70%. The only predictive factor for disease progression was the presence of high-grade tumor (75% vs. 52% after 52 months). Tumor size (≤1 vs. >1 cm) or multifocality were not independent prognostic factors. Several other authors have reported encouraging results for endoscopic management of UTUC compared to RNU (44) and a meta-analysis comparing endoscopic treatment and RNU demonstrated equivalence between these treatment strategies in selected favorable cases (45). Usually, a biopsy is performed during an endoscopic procedure and treatment is initiated immediately (laser vaporization) as a second step of the procedure regardless of biopsy findings. Thus, one of the key issues of ureteroscopic management is that the tumor is treated based on a subjective decision (size, aspect, multifocality, and the endourologist’s own experience) without knowledge of the tumor grade. The use of endoscopic confocal laser endomicroscopy could help identify which are the best candidates for this strategy. Confocal laser endomicroscopy (Cellvizio®) provides microscopic images of tissues in vivo using a low-energy laser light source and can be used in real-time during ureteroscopy. In a small series, Breda et al. (46) were able to identify the correct grading using confocal laser endomicroscopy during endoscopic ablation of UTUC in 13/14 patients. This could allow real-time characterization of the lesion to identify patients with low-grade tumors suitable for conservative management and could be considered as “endoscopic frozen section analysis”.

Segmental ureterectomy describes the removal of the part of the ureter where the UTUC is localized, by either ureteroureterostomy or ureteroneocystostomy and allows for concomitant lymphadenectomy. Retrospective studies suggest that segmental ureterectomy has comparable long-term outcomes to RNU (47), with 5-year CSS around 85% when indicated carefully (4). Laparoscopic and robot-assisted ureterectomy have helped to decrease per-operative morbidity. A double approach with intra-operative ureteroscopy can also help to identify the limits of the tumor while performing robotic ureterectomy. A good tip is the use of near-infrared light mode (FireFly mode), which can be activated without any intravenous indocyanine green injection to better identify the ureteroscope and the tumor margins (see ). Intravenous indocyanine green injection could also help to enhance local visualization of ureteral vascularization.

Figure 1

Intraoperative view during segmental resection in regular light mode (top image) and near-infrared light (Firefly) mode (bottom image). The ureteroscope is inserted up to the margin of the ureteral tumor. When the Firefly mode is activated, the light of the ureteroscope is easily visible in green.

In situ endocavitary instillations

There is a strong rationale for endocavitary instillation of topical (neo-)adjuvant treatments, either antegradely or retrogradely. The outcomes associated with adjuvant endocavitary instillations have been poorly assessed in the literature (44). Like urothelial carcinoma of the bladder, the topical agents used are commonly mitomycin C and BCG. The major drawback is the reduced exposure time of these topical agents to the tumor site. To overcome this drawback, a novel hydrogel polymer with reverse thermal gelation properties (liquid at cold temperatures and solid at body temperature) might promote high-dose delivery of mitomycin C into the upper urinary tract. Several pre-clinical studies have already been published and an ongoing phase 3 trial (Olympus Study) (48) is underway to assess the safety and efficacy of such instillations in patients with low-grade UTUC.

Radical surgery

RNU with bladder-cuff excision is the standard treatment for high-grade UTUC (2). Bladder-cuff excision is mandatory due to the high recurrence rate in the ureteral stump (16–58%). Like segmental resections, RNU can be done through a minimally invasive approach. Laparoscopy can help diminish the postoperative burden and has been shown to be non-inferior to the open approach (49). The value of lymphadenectomy during RNU is still unclear but allows better staging to help post-operative adjuvant treatment. The anatomic landmarks and extent depend on the tumor localization and no strong recommendations can currently be made. A single post-operative intravesical mitomycin instillation should always be considered since its effect on reducing intravesical recurrence after RNU has been demonstrated (50).

Systemic treatment

It is very likely that systemic chemotherapy and immunotherapy will be part of the therapeutic armamentarium for very high-risk localized disease in the future, even for non-metastatic UTUC patients. The most interesting results presented recently are those from POUT Trial, a phase III randomized study conducted in the UK since 2012 (51). A total of 261 patients who underwent RNU for pT2–T4N0–3 M0 UTUC were randomized to four cycles of adjuvant chemotherapy or observation. Most of the patients were pT3 (approximately 65%) and pN0 (91%). The authors found a significant difference in disease-free survival at 2 years in favor of chemotherapy (51% vs. 70%), despite a high rate of grade ≥3 toxicity (62%). The trial was terminated early because of efficacy favoring the chemotherapy arm and paved the way for the development of a new trial (POUT2) that will randomize chemotherapy vs. a combination of chemotherapy and immunotherapy in the perioperative setting.

Can we improve the current risk stratification strategy?

The treatment strategy for localized disease without evidence of metastasis should be based on careful risk stratification. According to the most recent EAU guidelines, as tumor stage is difficult to identify clinically in UTUC, it is useful to “risk stratify” UTUC between low- and high-risk tumors to identify those patients who are more suitable for KSS rather than radical extirpative surgery. KSS should always be considered in cases of absolute indication, but also for localized disease amenable to complete excision (42). Patients should also be informed about the risk of more aggressive treatment, recurrence, or progression, and accept strict surveillance protocols. KSS for low-risk UTUC reduces the morbidity associated with radical surgery, without compromising oncologic outcomes or kidney function, as stated in a systematic review from the EAU guidelines panel. In low-risk cancers, it is the preferred approach with survival being similar after KSS vs. RNU. This option should therefore be discussed in all low-risk cases, irrespective of the status of the contralateral kidney.

International guidelines recommend risk stratification for many urologic cancers, such as prostate cancer or non-muscle invasive bladder cancer. The concept of low-risk and high-risk tumors has been developed by urological societies such as the European Association of Urology (EAU) (2) or the International Consultation on Urological Disease (ICUD) (15,52). A low-risk tumor is defined as unifocal disease, low-grade status on cytology or biopsy, and no invasive aspect on imaging. Historically, the maximum tumor size for low-risk disease was 1 cm, even though no data supported a poorer prognosis for tumors >1 cm; this was due to technical difficulties in conservatively managing tumors >1 cm. However, this limit has now been changed to 2 cm, since larger tumors should not be an exclusion criterion for endoscopic management or segmental resection. High-risk tumors include any high-grade status (on cytology or biopsy), lesions >2 cm, presenting with hydronephrosis or multifocality. A history of radical cystectomy or variant histology are also criteria for high-risk tumors.

In our opinion, elective indications for endoscopic management or segmental resection should be reserved for localized low-grade disease amenable to complete resection. Multifocality or large papillary tumors should not be excluded from elective indications.

The current criteria used to stratify patients remain limited. Molecular characterization will probably help to improve risk stratification in the future, but most molecular characterization studies have focused on RNU specimens and their predictive value on biopsies needs validation. Furthermore, most studies suffer from their retrospective design and small sample size. Therefore, further collaborative efforts are needed to validate predictive models and to identify and validate new biomarkers from endoscopic biopsies. Without these efforts, recommendations on the grade for risk stratification will remain poor. In the meantime, we strongly believe that risk stratification needs improvement in the light of new treatment options. The current EAU classification proposal considers only two categories () (2). This is not enough in the light of the treatment options currently available or under development. For example, low-risk tumors should certainly be treated endoscopically, usually when the disease is a small low-grade unifocal papillary tumor without any significant risk factor. An intermediate-risk group could be low-grade tumors of higher volume, which would still be accessible for endoscopic treatment or segmental resection, and would possibly benefit from in situ adjuvant instillation to prevent local recurrence. High-grade tumors or invasive disease should remain classified as high-risk disease requiring RNU. In our opinion, lymphadenectomy should be performed when possible, since its prognostic and therapeutic value has been suggested, mostly for muscle-invasive or locally advanced disease (53). A subgroup of these high-risk patients might be cured by surgery only, while another subgroup of very high-risk patients will benefit from perioperative systemic treatment. Subgroup analyses of ongoing trials will certainly help to better stratify the high-risk patients who could avoid the morbidity of systemic perioperative treatments in the same way that low-risk or intermediate-risk patients should avoid the morbidity of RNU. summarizes this new proposal and could be the basis for a new risk-adapted approach depending on the localization of the tumor as suggested in .

Table 1

Risk stratification according to the European Association of Urology

Low-risk upper tract urinary carcinoma (UTUC) (all criteria must be present)

Unifocal disease

Tumour size <2 cm

Low-grade cytology

Low-grade biopsy

No invasive aspect on CT-urography

High-risk UTUC (any criteria may be present)

Multifocal disease

Tumour size >2 cm

High-grade cytology

High-grade biopsy

Hydronephrosis

Previous radical cystectomy for bladder cancer

Variant histology

Table 2

Proposed new improved stratification in four groups

Low-risk upper tract urinary carcinoma (UTUC)

Unifocal disease

Low-grade cytology

No invasive aspect on CT-urography

Intermediate-risk UTUC

Multifocal disease if endoscopic or segmental treatment possible

Low-grade cytology

No invasive aspect on CT-urography

High-risk UTUC

High-grade cytology or biopsy

Carcinoma in situ or variant histology

Sessile lesions

Invasive aspect on CT-urography

Large, bulky disease

Very high-risk subgroup

Locally advanced or non-organ confined disease

Figure 2

Proposed new risk-adapted approach.

This new classification has not yet been validated and certain treatment options are optional. However, the timeline of evidence-based medicine is not always similar to the timeline of clinical daily practice. With all new possibilities on the table, we are convinced that some colleagues are already thinking out of the box and are moving forward in the direction of personalized medicine for UTUC management.

Conclusions

Current international guidelines recommend a risk-adapted approach to UTUC management. New tools and treatment options are being developed to help stratify our patients better and to offer them more options. A new stratification is therefore needed to integrate them when they have been fully validated.

The article’s supplementary files as