Abstract
As a rare disease, the management of upper tract urothelial carcinoma (UTUC) continues to evolve. While radical nephroureterectomy remains the gold standard, there is a growing desire to pursue nephron-sparing approaches through endoscopic management, particularly for low-risk disease or in the setting of imperative indications. A particular challenge for those undertaking endoscopic management is appropriate grading and staging of tumors, and thereby appropriate patient selection. In this review we will cover the current state of diagnostics for UTUC as well as highlight the challenges in accurate diagnosis. Additionally, we will focus on emerging technologies to aid in optimizing diagnostic accuracy in UTUC. This will include discussion of narrow band imaging, photodynamic diagnosis, optical coherence tomography, and confocal laser endomicroscopy.
Introduction
As a rare disease, there exists a paucity of high-level evidence on the diagnosis and management of upper tract urothelial carcinoma (UTUC). 1 The gold standard for management of UTUC remains radical nephroureterectomy (RNU), including excision of the ipsilateral bladder cuff. However, indications for a nephron-sparing approach exist, and represent a diagnostic and therapeutic challenge for urologists. Development of European Association of Urology (EAU) guidelines, most recently updated in 2017, provide a basis to establish standards of care in management of UTUC, including the role of nephron-sparing approaches. 2 To succeed at a nephron-sparing approach, patients must be appropriately selected with accurate staging. However accurate grading and staging of UTUC remains difficult and without out any major advances in the past two decades, thereby leading to inevitable shortcomings in diagnostic accuracy. This is reflected in the need to risk stratify patients into low-risk versus high risk as a surrogate for accurate identification of clinical stage. 2 Needless to say, poor patient selection for an organ-sparing approach may lead to inferior outcomes. Herein we review current indications and outcomes for endoscopic management of UTUC and discuss emerging technology to improve diagnostic accuracy and subsequent patient selection.
Indications and outcomes endoscopic management
UTUC represents only 5–10% of urothelial tumors. 1 Risk factors for UTUC mirror those of bladder urothelial carcinoma, classically tobacco exposure and exposure to aromatic amines. 1 Internationally, however, exposure to aristolochic acid has resulted in a dramatic rise in cases of UTUC among certain locations, particularly in the east. 3 Found as a byproduct of Aristolochia fangchi and Aristolochia clematis, aristolochic acid has been implicated in Balkan nephropathy as well as nephropathy resulting from the use of Chinese herbs. It results in p53 mutations, and increases the risk of urothelial carcinoma as well as end-stage nephropathy. 4 Such changes likely represent a field change effect, and further the impetus towards endoscopic management, when feasible, due to an increased risk of recurrence in the contralateral urothelium as well as urinary bladder.
Hereditary syndromes, most notably hereditary nonpolyposis colorectal cancer, carry an increased risk of UTUC.5,6 In fact, urothelial carcinoma represents the third most common malignancy after colon and endometrial cancer in this syndrome. 5 These patients carry a marked increase of UTUC, with a high risk of recurrence, and mandate attempts at nephron-sparing approaches when technically feasible due to the high rate of recurrence as well as bilateral disease.
Ideally, elective (nonimperative) endoscopic management is restricted to smaller, low-grade tumors. Indeed, a single institution review of patients undergoing endoscopic management found the 5-year cancer-specific survival (CSS) of 96% for low-grade tumors compared with 39% for high-grade tumors, 7 and multiple studies have confirmed tumor grade as a driving predictor patient outcome.8–10 The EAU guidelines separate UTUC by risk (Table 1), mirroring changes made to guidelines for primary bladder carcinoma. 2 Among patients without imperative indications, criteria for endoscopic management include solitary, low-grade lesions without radiographic evidence of invasion. In contrast, imperative or relative indications for a nephron-sparing approach include the presence of an anatomic or functional solitary kidney, bilateral disease, comorbidities disqualifying from RNU, chronic kidney disease, disease states representing a high risk of contralateral recurrence (i.e. Balkan nephropathy and Lynch syndrome), or patient refusal to undergo RNU. As will be discussed, accurately identifying high and low-risk patients remains a challenge.
Risk groups for UTUC. 2
UC, urothelial carcinoma; URS, ureteroscopy; UTUC, upper tract urothelial carcinoma.
As in bladder primary, UTUC has a high rate of recurrence but may show favorable survival for endoscopic management in appropriately selected patients. 11 Currently no high-level evidence exists comparing RNU with endoscopic management of UTUC. Yakoubi and colleagues performed a meta-analysis comparing RNU with endoscopic management for UTUC. 12 In total, eight retrospective studies met criteria and resulted in pooled data of 1002 patients. Importantly, they found no significant difference in overall survival and CSS between patients undergoing endoscopic management and RNU. However, with retrospective, low-level data, the true outcomes remain poorly defined. Patients undergoing endoscopic management tend to be a highly-selected, biased population and therefore a clear comparison is difficult to draw. What is at least apparent is that well selected patients maintain good survival outcomes following endoscopic management. Our ability to appropriately select those patients needs improvement
The optimal endoscopic approach for UTUC tumors remains debated. Recently, Cutress and colleagues reported their systematic review of ureteroscopic and percutaneous management of UTUC, including 885 papers, of which 56 met criteria for analysis. 13 They found a disease-specific survival stratified by tumor grade, with 98% DSS for low-grade tumors whether treated by ureteroscopy (URS) or a percutaneous approach, which decreased among high-grade disease. There was a striking difference in recurrence rates within the upper tract between the URS and percutaneous resection, with URS showing a 52% recurrence compared with 23% for percutaneous resection among low-grade disease. Among G3 tumors, recurrence rose to 76% for URS and 40% for percutaneous tumor resection. This suggests that a more through resection and ablation of disease with larger instruments via a percutaneous approach results in a higher rate of initial tumor clearance. When feasible, a percutaneous resection may offer the greatest benefit to achieve a recurrence-free state.
Of particular concern, may be those patients with UTUC presenting following a primary bladder urothelial carcinoma (UC). In a review of 90 patients with UTUC following primary bladder tumor, Krambeck and colleagues found a 5 year CSS of only 71.2%, and that the risk of death from UC was associated with UTUC grade, stage, imperative indication for organ preservation as well as a history of cystectomy. 14 This is reflected in the EAU guidelines for management of UTUC, where upper tract disease after prior radical cystectomy for urothelial carcinoma is a criteria for high-risk disease. 2 Since RNU following cystectomy carries increased risk and morbidity, there is often a desire to manage such patients endoscopically. Better stratification would be use at clinical counseling and patient selection for radical surgery versus endoscopic control.
Whether by percutaneous or ureteroscopic approach, studies suggest approximately 20–30% of patients undergoing endoscopic management will progress to RNU.11,13,15 Patients electing for nephron-sparing approaches, therefore, require vigorous follow up to monitor for recurrence. With high recurrence rates associated with endoscopic management of UTUC, diligent follow up is required for patients electing a nephron-sparing approach. Currently, EAU guidelines indicate that, following an organ-sparing approach, patients should be followed up to 5 years or beyond. 2 This should include computed tomography (CT) urogram at 3 months, 6 months, and then yearly, as well as cystoscopy with URS and in situ cytology at 3 months, 6 months, and then every 6 months until 2 years then yearly. Such diligent follow up is prudent, as indeed Thompson and colleagues, in their series of 84 patients undergoing endoscopic management with 4.6 years of follow up, found that 55% of patients recurred within the upper tract and 45% recurred within the bladder, while 33% of patients progressed to RNU. 15 Addressing this issue, Villa and colleagues, in a novel study, examined the role of early repeat URS at 6 weeks and the association of tumor present at second look with outcomes. 16 They found that the presence of tumor 6 weeks later predicted poor subsequent recurrence-free rates. One wonders if this truly represents early recurrence, or instead an incomplete identification and resection of tumor at the initial endoscopy, and therefore incomplete treatment. To that extent, improved diagnostics are sorely needed in order to improve outcomes among those undertaking the long road of endoscopic management.
Numerous studies have found an association between diagnostic URS and an increased risk of intravesical tumor recurrence following radical nephroureterectomy, suggesting that manipulation of UTUC promotes tumor seeding.17–22 In their study of patients undergoing RNU, Lee and colleagues found an increased rate of bladder recurrence among patients who underwent URS prior to RNU (median 5 days later) compared with patients who underwent RNU without previous diagnostic URS. 19 Luo and colleagues found, in their review of 115 patients who underwent URS prior to RNU and 281 patients who proceeded directly to RNU, that URS was associated with an increased risk of bladder cancer recurrence. 20 A 2017 systematic review and meta-analysis that included 2382 patients who underwent RNU for UTUC, among whom 765 and a diagnostic URS, found that previous URS significantly increased the rate of intravesical tumor recurrence following RNU [hazard ratio (HR) 1.56, p < 0.0001]. 22 It follows that the high rate of bladder recurrence associated with URS prior to RNU would carry into those undergoing endoscopic management. This supports the need for rigorous surveillance of patients electing for endoscopic management, mandates patients counseling on the risks of bladder recurrence and the need for subsequent treatment.
Current state and emerging technology in diagnosis of UTUC
Any discussion of organ preservation in the setting for UTUC begins with diagnosis. As outcomes of organ-sparing approaches will depend most significantly on accurate diagnosis and staging, many of the advances of treatment are, in fact, improvements in diagnostics. Suspicion of UTUC most commonly arises from the presence of gross or microscopic hematuria, or the presence of concerning findings on imaging obtained for other indications. The EAU guidelines for UTUC recommend a baseline evaluation with contrast-enhanced imaging using CT urography if possible (magnetic resonance urography is an option when CT is contraindicated), cystoscopy to evaluate for bladder lesions, and urine cytology (Grade A recommendation). 2 Whereas, they previously recommended diagnostic URS, the most recent update has amended that statement to include ‘Use of diagnostic ureteroscopy and biopsy in cases where additional information will impact treatment decisions’ and is therefore best reserved for patients with diagnostic uncertainty or considering an organ-sparing approach. 2 Despite improved treatment guidelines, available diagnostic tools have not significantly evolved over the last few decades. Ultimately, accurate diagnosis and subsequent patient selection for an organ-sparing approach remains a challenge. Given the previous discussion, accurate staging is paramount to predict outcomes of endoscopic management, and any improvements in diagnostics may significantly improve patient outcomes.
Diagnostic ureteroscopy
The last few decades have seen a rapid development in both surgeon skill as well as the technical capabilities of URS. Beginning in the 1980s and 1990s the growing use of ureteroscopes prompted research into the feasibility and success of endoscopic biopsy and management of selected cases of UTUC.23–30 Despite attempts to optimize biopsies, nondiagnostic and understaged results are frequent and little improvement has been made. Kleinmann and colleagues found that flat wire baskets are superior to 3 F cup biopsy for diagnosis. 31 Malm and colleagues identified that the use of both barbotage cytology as well as biopsy may improve successful diagnosis. 32 Multiple studies have found that tumor grade at the time of ureteroscopic biopsy predicts adverse pathology at the time of RNU.33–35 Indeed findings by Grasso and colleagues indicate that grade at the time of URS may be the single most important predictive factor of overall survival and CSS, regardless of the ultimate treatment approach. 35 Based on this, it is generally recommended that endoscopic management of UTUC is typically reserved to low-grade tumors for nonimperative indications. 2
Nonetheless, accurate diagnosis and staging remains poor. According to the series by Yamany and colleagues, up to 25% of patients undergoing URS for UTUC will have missed lesions, and up to 50% may have missed carcinoma in-situ (CIS). 36 In a review of 77 patients undergoing RNU, Straub and colleagues found that even combining URS biopsy and cytology results, the sensitivity for diagnosing high-grade disease was 84% and that 15% of high-grade tumors were misdiagnosed as low-grade preoperatively. 37 Currently, various nomograms for predicting advanced disease are available based on imaging and ureteroscopic variables.38,39 Brien and colleagues found that combining the presence of preoperative hydronephrosis along with high-grade ureteroscopic biopsy as well as positive cytology predicted muscle-invasive and nonorgan-confined disease, while the absence of all three variables had 100% negative predictive value in ruling out advanced disease. 38 Favaretto and colleagues found, in their series of 324 patients treated with RNU, that only ureteroscopically high-grade tumors, as well as the presence of local invasion on imaging, predicted tumor stage pT2 or higher or nonorgan-confined disease. 39 These are useful tools for clinical decision-making, but nonetheless remain imperfect. This reinforces the need for close observation with short interval follow up for surveillance. However, it also highlights the dramatic need for improved diagnostic tools in order to better stratify patients.
Novel diagnostic tools
Research continues to focus on improving diagnosis in order to better stratify patient risk and, therefore, more accurately define appropriate treatment options. However, the majority of prognostication available involves information available at the time of surgical pathology for RNU, with little available for the appropriate triage of patients to radical surgery or organ-sparing approaches. Recently, however, there as been a push of new studies to bridge this gap. Improving diagnostic URS through better visualization as well as imaging allowing ‘real-time’ pathology’ shows promise. New avenues into narrow band imaging (NBI), photodynamic diagnosis (PDD), optical coherence tomography (OCT), and confocal laser endomicroscopy (CLE) offer hope of advancing diagnostics of UTUC (Table 2).
Novel diagnostic tools for UTUC.
CLE, confocal laser endomicroscopy; NBI, narrow band imaging; OCT, optical coherence tomography; PDD, photodynamic diagnosis; UTUC, upper tract urothelial carcinoma.
Narrow band imaging
NBI has been well described for UC of the bladder, but there is little published experience in the kidney. The fundamental goal of NBI is to improve visualization, and there for identification, of disease. To date, there is one report of NBI used for diagnosis in UTUC, and the series included 27 patients of whom 14 had known and 13 had suspected UTUC. 40 The kidney was inspected with standard white light followed by NBI, and the authors found subjective improvement visualization. Objectively there were five additional tumors identified in four patients, along with wider extent of three tumors identified when compared with white light. Improved visualization of upper tract tumors offers a clear and significant benefit; one cannot treat what one cannot see. Adding to this is the relative familiarity of NBI to practicing urologists as well as the need for minimal additional equipment aside from scopes equipped for NBI. This makes the expansion of NBI from the bladder to the upper tract a logical evolution.
Photodynamic imaging
In contrast with NBI, photodynamic diagnosis involves the use a fluorochrome agent, typically, 5-aminolaevulinic acid (5-ALA) and hexyl aminolevulinate (HAL). While well described in the bladder, use in the upper tract is only recently being explored.41–44 In a series of 106 patients with UTUC, Kata and colleagues reported their experience of patients undergoing URS for suspected or diagnosed UTUC. 42 They identified 48 tumors, of which 95.8% were seen with PDD compared with 47.9% for white light (p < 0.0001). Additionally, PDD found more CIS and dysplasia compared with white light (93.75% versus 18.75%; p = 0.0006), as well as more lesions of UC (96.9% versus 62.5%; p = 0.007). PDD provided greater sensitivity for UTUC (95.8 versus 53.5) with no difference in specificity (96.6 versus 95.2). While not yet widely used for diagnosis and management of UTUC, PDD offers a compelling new tool for diagnosis. By improving tumor localization, we may be better able to select appropriate patients for endoscopic management and decrease recurrence rates for those undergoing a more conservative approach. However the ability to improve detection of CIS might be the greatest benefit of PDD, as in a study of 300 patients undergoing RNU for UTUC, it was found that the presence of CIS is significantly under detected on preoperative ureteroscopic biopsy. 45 As a high-grade disease with increased risk of recurrence, failure to identify the presence of CIS within the upper tract will lead to greater treatment failures of patients undergoing endoscopic management. Furthermore, the improved ability of PDD to identify CIS may allow better selection of candidates for intracavitary therapy, especially as the primary evidence in favor of upper tract Bacillus Calmette-Guerin (BCG) is the presence of CIS with less evidence to support its use in Ta and T1 disease. 46
However, challenges to PDD exist within the upper tract. Achieving successful placement of the fluorochrome with poor dwell time in the upper tract may be difficult. It is also notable that side effects may occur with administration of the agent. A recent systematic review of PDD in UTUC found relatively high rate of minor adverse events (25.8%) associated with the use of 5ALA, with the most common being hypotension followed by facial rash and abnormal liver function tests. 47 Fortunately there were no long-term sequelae. In addition, assessment of tumors along the ureter may be altered as the technology functions best when tumors are viewed head on. Tangential views may result in an increase in fluorescence and greater false positives. Nonetheless, improved tumor visualization, particularly of CIS, may improve diagnosis and patient selection for appropriate treatment.
Optical coherence tomography
OCT is an existing medical technology, currently utilized for imaging the retina, that has recently been used to examine UTUC.48,49 Analogous to ultrasound, OCT receives and processes backscattered light (as opposed to sound waves) to build a three-dimensional image of tissues. As such, it can assess beyond the mucosal surface to offer impressions of not only tumor grade, but stage. By visualizing the distortion of tissue layers, an assessment of tumor depth may be assessed. One drawback, however, is depth of imaging due to decreased light scatter with deeper tissues thereby limiting assessment to approximately 2 mm of depth. This may be adequate to assess a tumor as non-invasive versus invasive, but more challenging to separate T1 and T2 lesions. An assessment of tumor grade is made according to the decrease in light intensity (mm−1) measured by the attenuation coefficient (µoct), with a higher coefficient corresponding to higher grade tumors.
In an initial pilot study, Bus and colleagues report on an initial series of eight patients undergoing URS for suspected UTUC who eventually progressed to RNU. 49 They found that OCT staging correlated with histopathology in seven of eight cases, and represented a promising technology for UTUC. In a 2016 follow-up study to assess the diagnostic accuracy of OCT for UTUC, Bus and colleagues then evaluated 26 patients who underwent OCT prior to RNU in order to assess the sensitivity, specificity, negative and positive predictive values. They correlated OCT stage/grade with biopsy stage/grade as well as surgical pathology as the reference standard. They found that that OCT staging accurately reflected the surgical pathology in 83% of cases, whereas biopsy results corresponded in only 49% of cases. The tumor grade correlated with surgical pathology in 88% of patients according to OCT, as opposed to 79% for biopsy. Overall, OCT staging showed excellent sensitivity (100%), specificity (92%), positive predictive value (PPV) (92%) and negative predictive value (NPV) (100%), and significantly outperformed biopsy with regards to staging. This results in clinically useful information that may better select appropriate patients for endoscopic management.
Confocal laser endomicroscopy
One novel new diagnostic tool is CLE. The technology utilizes the intravenous administration of fluorescent dye (fluorescein). Under excitation from the CLE probe, the fluorescein-stained tissue emits light that is filtered through a pinhole filter and received on a photo detector. This results in tissue resolution similar to standard histology. 50 Since the fluorescein does not cross cell membranes, nuclear features are not evaluated. Initial studies found CLE to be well tolerated and feasible, and allowed visualization of pathologic features similar to histologic examination. 51 In a recent publication Breda and colleagues described their use of real-time CLE for 14 patients undergoing URS with biopsy, and report correlation with biopsy histopathology as well as pathologic correlation with RNU or segmental ureterectomy specimens when available (6 patients). 52 Correlation between CLE and biopsy was 100% (7/7 patients) for low-grade tumors, 83% (5/6 patients) for high-grade tumors and 100% (1/1 patients) for CIS. Among the six patients who progressed to RNU or segmental ureterectomy, there was 100% correlation between the CLE prediction of tumor grade and the final pathologic specimen (5 high grade and 1 low grade). While extremely promising, this small series of essentially six patients with true pathologic correlation requires additional study. With the frequency of nondiagnostic or poorly diagnostic specimens obtained at URS, the ability to obtain a real-time pathologic interpretation may dramatically improve the diagnostic accuracy for UTUC. In a large single institution series of 184 patients who underwent diagnostic URS prior to RNU, Wang and colleagues found that a surprising 96% of patients with grade 1 UTUC were found to have higher grade tumors at final pathology. 53 This series included patients from 1995 to 2008, and therefore improvements in technique have most likely improved the accuracy of diagnosis. However, the point remains: diagnostic URS is at risk to under stage UTUC. Since the primary driver of a successful organ-preserving approach appears to be appropriate patient selection, the ability to more accurately characterize tumors is paramount. CLE offers a potential step forward for real-time assessment of tumor grade and stratification towards radical surgery or an organ-sparing approach.
Although the addition of NBI, PDD, OCT and CLE each offer promise of improved diagnostics, their true role and utility relative to each other remain incomplete. Current studies show improvement of disease identification, and a correlation with pathologic staging. However, comparative accuracy of tumor identification between NBI and PDD have yet to be reported in UTUC. Similarly, the comparative accuracy of ‘live pathology’ in OCT and CLE has not been demonstrated. Finally, it remains to be seen if the use of these novel technologies will ultimately improve oncologic outcomes. Nonetheless, they offer promise to a field that has been stagnant in new developments.
Conclusion
Despite increased focus and collaboration on the diagnosis and management of UTUC, advances in the past 20 years have been limited. Endoscopic management of UTUC depends on accurate diagnosis for appropriate patient selection. Rates of recurrence and progression are high, and existing techniques for biopsy significantly under sample existing disease. However new research into better techniques of endoscopic assessment of upper tract tumor yield promise. NBI, PDD, CLE and OCT offer new ways to visualize tumors and improve the accuracy of grading and staging. While additional research is needed to recognize their impact and potential, we appear to be on the cusp of truly new tools for use in UTUC.
Footnotes
Funding
This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.
Conflict of interest statement
The authors declare that there is no conflict of interest.
