The modern management of upper urinary tract urothelial cancer: tumour diagnosis, grading and staging


Francis X. Keeley, Bristol Urological Institute, Southmead Hospital, Bristol, BS10 5NB, UK.


Three of the four mini-reviews this month deal with the topic of upper tract urothelial cancer, intending to cover completely the issues of diagnosis, staging and grading, as well as surgical treatment, both open and laparoscopic. In this way the reader should be able to have at hand most of the relevant important information about the subject, all in one issue.


upper urinary tract




multiphasic multidetector-row CT.


TCC of the upper urinary tract (UUT-TCC) is relatively uncommon, accounting for 2–5% of all urothelial tumours [1]. Times to diagnosis of renal and ureteric TCC for the UK compare unfavourably with all other urological cancers, with a median of 48 days for UUT-TCC compared to 27 days for bladder TCC [2]. This delay might be related to the relative difficulty in identifying small UUT lesions using a combination of currently available endoscopic, imaging and urinary investigations. However, earlier detection of UUT-TCC could be expected to lead to a significantly better outlook for patients, with 5-year survival rates for Ta tumours of >80%, vs 15–30% for T3 tumours.

The standard treatment of UUT-TCC, previously consisting of nephroureterectomy alone, has widened, with nephron-sparing surgery for intrarenal TCC in cases of solitary kidneys, bilateral disease or renal dysfunction. In addition, low-grade, low-stage lesions might fall within elective indications for minimally invasive endoscopic treatment strategies for patients with two normally functioning kidneys. Accurate staging and grading methods are therefore imperative in such situations, and the development of small-diameter actively deflectable ureteroscopes, together with biopsy devices, allow for a precise assessment of such tumours, thus reducing the risk of clinical under-grading and -staging. The need for rigorous surveillance, based on the risk of recurrence or local progression, means that endoscopic treatment is often labour-intensive and expensive. Despite these limitations, endoscopic treatment in well-selected patients is appropriate, and its precise role is currently being defined as experience accumulates.


Patients with UUT-TCC often present with simple general abnormalities of urinary testing, e.g. microscopic or dipstick-detected haematuria. Standard assessment includes cystoscopic visualization of the bladder mucosa, imaging of the UUTs, and urinary cytology. In only a minority of patients will these investigations lead to a firm diagnosis of UUT-TCC. For many patients, further assessment will be required in the form of retrograde contrast imaging, selective ureteric urinary sampling and the reference standard, UUT endoscopy and biopsy.

Regardless of the management planned for UUT-TCC, accurate diagnosis is essential before starting treatment. Just as cystoscopy and bladder biopsy allows for visual and pathological diagnosis of bladder TCC, endoscopic inspection and biopsy are essential to ensure an accurate diagnosis of UUT-TCC. Flexible ureteroscopes are now produced with distal tip dimensions of 6.8 F (e.g. DUR-8, Gyrus-ACMI) and active deflection of 180° both up and down (UPF P3, Olympus; Flex-X, Storz); such instruments, when combined with semi-rigid ureteroscopes for inspecting the distal ureter, allow for a reliable assessment of the entire collecting system [3]. This inspection should be carried out in a standard fashion, as outlined in Table 1[4], taking particular care to visualize the urothelium before passing guidewires, to avoid unnecessary biopsy of traumatized areas.

Table 1.  Diagnostic technique; adapted from Bagley [4]
2Gentle insertion of retrograde catheter to distal ureter only
Ureteric drip specimen for cytology (aided by i.v. diuretic bolus)
3Retrograde ureteropyelography
4Aspiration cytology
5Semi-rigid ureteroscopic inspection of ureter before guidewire insertion
6Guidewire insertion only to position already inspected
7Flexible ureteroscopic inspection of proximal ureter and intrarenal collecting system
8With ureteroscope positioned in area of interest:
Aspirate; saline wash; biopsy with basket or cup forceps; repeat saline wash; treat visible  tumour with Ho:YAG and/or Nd:YAG laser or fulguration; repeat saline wash
9All specimens delivered fresh to cytology for cytospin and cell block


The sensitivity of urinary cytology for UUT-TCC as a whole has been shown to be poor in many studies, with detection rates as low as 29%[5]. However, whilst the changes associated with low-grade disease are very subtle [6], cytological detection rates for G3 tumours or carcinoma in situ approach 100%[7] and are sometimes the only abnormality detected at initial assessment. Xu et al.[8] showed how the sensitivity of cytology for low-grade disease can be improved by additional staining for uroplakin.

More modern developments in urinary testing such as ImmunoCyt, Fibrin Degradation Product Test (AuraTek FDP, PerImmune Inc., Rockville, MD, USA) and Bladder Tumour Antigen Test (BTA, Bard UK Ltd., UK) have been shown to improve on the sensitivity and specificity of urinary cytology in some small studies, but have not been widely accepted [5,7].


Endoscopic sampling combined with cytopathological techniques can permit a positive diagnosis and accurate grading of UUT-TCC [4]. Whilst simple inspection of UUT lesions has been shown to accurately predict low- and high-grade TCC in 71% and 80% of cases, respectively [9], biopsy of the suspect lesion within the UUT leads to a formal diagnosis, and information on stage and grade. Such information is becoming more relevant with modern minimally invasive approaches to management. The extent to which endoscopic biopsy correlates with eventual pathological stage and grade has been assessed in several studies. Keeley et al.[10] assessed the accuracy of ureteroscopic biopsy in 51 cases, and the correlation between biopsy grade and surgical stage. In 90% of cases graded low or high on biopsy, the surgical specimen grade was identical; 87% of low-grade biopsies were eventually staged pTa or pT1, whilst 67% of high-grade biopsies were from surgical T2-3 tumours.

Guarnizo et al.[11] showed that the biopsy accurately predicted the eventual grade in 78% of lesions but in all the remaining 22% the eventual nephroureterectomy specimen was upgraded. If tumour invaded the lamina propria on biopsy it was never found to be over-staged, whilst if tumour was graded Ta on biopsy there was a 45% rate of understaging of the final surgical specimen.

Skolarikos et al.[12] showed that nine of 32 G2 biopsies were upgraded to G3 on final surgical specimen, but combining urinary cytology results with biopsy grade improved the sensitivity and specificity for high-grade disease from 43% to 55%, and 23% to 85%, respectively. Shiraishi et al.[13] showed the correlation between grade and stage in the UUT, with 85% of G2 or G3 tumours showing muscle invasion.

These findings lend weight to the suggestion that the grade of the tumour appears to correlate with stage more closely in UUT-TCC than in bladder TCC [10,14]. Some investigators have even suggested that tumour grade is a more important prognostic factor than pathological stage [14].

In reviewing the Bristol experience of 30 nephroureterectomies biopsied before surgery, we found that biopsy grade corresponded accurately with nephroureterectomy grade in 75% of cases, whilst 90% of the remaining cases were under-graded by biopsy [15]; 68% of grade 1 or 2 ureteroscopic specimens were from low-stage tumours (pT0, Ta or T1), whereas 75% of grade 3 biopsies came from tumours staged ≥pT2. However, some concern should be reserved for lesions graded G2 on biopsy, in which six of 18 had a final surgical stage of ≥pT2. Adding a cytological analysis of ureteric-sampled urinary aspirates to this subset might improve stage prediction. In cases of biopsies graded as G2 with frankly malignant urine, we found a 60% rate of stage ≥pT2 for surgical specimens, whilst this rate was only 17% if the urine was negative.

Thus, the clinical staging of UUT-TCC is currently limited. The small calibre of ureteroscopic biopsy instruments commonly results in samples that are inadequate for conventional histopathological staging. Our current practice therefore is to send all endoscopic biopsies to the cytology laboratory, relying on the grade of the lesion to guide further treatment. Percutaneous access permits deeper biopsies using larger instruments, but aggressive tissue sampling runs a higher risk of perforation, severe bleeding and tumour implantation [16].

In conclusion, we recommend direct ureteroscopic inspection, sampling of ureteric urine, and ureteroscopic biopsies, with all samples sent for cytopathological examination so as to maximize the diagnostic yield. We emphasize the importance of frankly malignant cells on cytology, especially in combination with grade 2 or 3 TCC on biopsy, as a predictor of high-grade, high-stage disease [12,15].


Many departments still currently use one imaging method in evaluating haematuria. For solid renal masses of <3 cm, urinary tract ultrasonography (US) has been shown to be more sensitive than IVU [17]. However, US is notoriously unreliable for identifying small UUT-TCCs. Khadra et al[18] investigated the relative efficacy of renal US and IVU in detecting UUT tumours in 1930 patients presenting with haematuria. The authors advocated the use of a combination of both methods in all patients, in view of significant false-negative rates for both tests if used in isolation (42% for US, 27% for IVU). They also found no subsequent development of UUT-TCC in any patient who had a negative US and IVU. However, in all, only two UUT urothelial tumours were found, and so it is difficult to draw firm conclusions. More recent studies suggested that the use of IVU can be safely reserved for specific groups of patients, which might aid in resource allocation. Edwards et al.[19], in a retrospective study of >4000 patients, reserved IVU for those with an abnormal US or where there was persistent dipstick haematuria after normal first-line investigations. Overall, 13 cases of UUT-TCC were diagnosed (incidence 0.32%) of which three were missed by US and found on IVU. No UUT tumour was identified in an individual aged <50 years. The main drawback is that the study did not report the long-term outcome of the apparently negative cohort.

The ability of multiphase spiral CT to detect renal pelvic and ureteric anatomy has also been compared with IVU, with intrarenal and mid-ureteric results favouring CT [20]. In one study, of 86 patients with a negative IVU, CT detected eight small renal parenchymal neoplasms [21]. Similar studies lead to the suggestion that CT should be used routinely in the initial assessment of haematuria [22]. Certainly, IVU suffers by comparison in obese patients or in those with collecting systems obscured by bowel contents, and often necessitates retrograde studies under anaesthetic. MR urography, using gadolinium-labelled contrast media, is yet to find a role in the initial diagnosis of UUT-TCC, with several reports highlighting a significant incidence of false-positive filling defects [23,24] and one reported of a missed PUJ TCC [24].

Three-dimensional reconstructed CT is useful in some urological situations for understanding the spatial relationship between structures. However, no extra information is obtained in addition to cross-sectional imaging, and therefore the diagnostic sensitivity for UUT-TCC is unlikely to be improved. More advanced still is the development of virtual ureterorenoscopy. The ability to accurately survey the entire urothelium without having to resort to anaesthesia and potentially morbid invasive procedures is obviously attractive, especially in the realm of surveillance. Takebayashi et al.[25,26] reported sensitivities of >80% for ureteric and renal lesions, but this application remains limited due to difficulties with spatial resolution and the inability to detect subtle changes of mucosal appearance suggestive of carcinoma in situ.


Imaging of the UUT uncommonly provides accurate staging information, with several studies showing MRI to be insensitive in local staging [27]. The use of CT is standard in the assessment of the local and distant spread of UUT-TCC. Extensive local invasion on CT usually corresponds with histological findings and carries a grim prognosis, but standard CT has been shown to be relatively insensitive for lower stages [28]. The same study also showed a 40% over-staging rate, which might have been improved in many cases by ignoring accompanying hydronephrosis. In the future, multiphasic multidetector row CT (MDCT) could improve the staging of the disease. This technique results in faster data acquisition and higher resolution than standard helical CT, in addition with a slice thickness down to 1 mm. The entire scan can be done in a single breath-hold and radiation doses for MDCT are roughly 1.5 times that of a full series IVU. Fritz et al.[29] recently reported a detection rate of 100% and an accurate stage prediction in 88% of cases.

Endoluminal US using 6 F catheters introduced retrogradely might have some promise in assessing invasion. The earliest report of the use of endoluminal US in staging human UUT-TCC was by Goldberg et al. in 1991 [30]. Grotas and Grasso [31] subsequently developed three-dimensional reconstructed images to further refine the technique. However, although other authors have also reported success [32], to date no large study has confirmed the accuracy of this technology.


Many genetic factors have been studied in relation to the diagnosis and prognosis of UUT-TCC, but almost all have very limited value. There seems to be a relationship between over expression of the tumour suppressor gene p53 and the presence of UUT-TCC. One study showed p53 over-expression in all 28 biopsies positive for TCC, with no false-negatives but eight false-positives [33]. There was no correlation between degree of over-expression and grade. Expression of p63, essential for normal urothelial differentiation, was shown in nearly all TCC but no cases of RCC, suggesting that it might be useful in differentiating these two tumour types [34]. p27 expression decreases with increasing tumour stage but was independent of tumour grade; the immunoreactivity of its ubiquitin ligase subunit, Skp2, increased with grade and stage, and was associated with angio-invasion [35]. The finding of an increased prevalence of UUT-TCC in patients with hereditary nonpolyposis colorectal cancer syndrome has lead to the discovery of significantly increased rates of microsatellite instability in UUT-TCC, unlike in bladder cancer where it is a rare event [36]. Future work in these areas might provide clinicians with further tools to help diagnose UUT-TCC and aid the selection of treatments, prognostic prediction and adjuvant therapy.


For many patients the modern diagnosis of UUT-TCC has changed significantly since the days when radical renal surgery was undertaken with only an equivocal IVU to guide the decision. Full assessment of suspected lesions takes advantage of advances in cytological, radiological and endoscopic techniques.

Only by providing accurate tumour staging and grading can clinicians decide which patients should be offered more conservative therapies. Endoscopic treatment is now possible for many patients with UUT-TCC and it might be the preferred option in patients unfit for major surgery, with bilateral disease or a solitary kidney. Low-grade, low-stage disease also appears to be particularly well suited to this approach. In other patients, especially those with bulky or high-grade lesions, it might be suboptimal therapy that should only be undertaken with the patient’s full understanding and consent. Cases where ureteric biopsies are graded as G2 in combination with positive urinary cytology should raise particular concern, even if taken from apparently low-volume lesions.


None declared.