Upper-tract urothelial carcinoma : diagnostics and prognostic factors

Abstract

Upper-tract urothelial carcinoma (UTUC) is rare, constituting 5-10% of all urothelial cancers. Urothelial carcinoma of the bladder is a common disease and is more studied than UTUC. The incidence of UTUC is rising, with people living longer and, especially, surviving bladder cancer to a greater extent than before; therefore, these individuals are at risk of developing UTUC, as the two diseases are closely linked. As the majority of patients presenting with UTUC are ≥65 years old and have risk factors for renal impairment, as well as for complications from general anaesthesia, large surgical procedures and chemotherapy, it is crucial to make as good a risk assessment as possible prior to deciding with the patient what treatment to undertake. Survival is poor for invasive UTUC, despite radical treatment, but if low-risk UTUC can be identified, that minority of patients can have equal survival if offered kidney-sparing treatment (KSS), generally URS laser ablation.

In Paper I, forty-five patients who underwent URS prior to radical nephroureterectomy (RNU) were included, and 43 were included in the final analysis. Samples were analysed, comparing the agreement of grade and ploidy in endoscopic biopsies and in barbotage samples from the renal pelvis and from the ureter (fluid collected from the bladder after instrumentation of the upper tract) with those in RNU specimens. Almost half of the tumours (20/43) were grade 3 (high-grade) cancer. Thirteen of the low-grade cancers were classified as grade 2, and 10 were classified as grade 1. The overall agreement of grade was 94%, with cytology being equal to the histology of biopsies in identifying cancer. Ureteral-barbotage specimens were more accurate than the other samples in 4/16 cases of UTUC located in the ureter.

The same patient cohort that was analysed in Paper I was included in Paper II, where the tumour characteristics associated with tumour stage were found to be tumour grade, DNA ploidy and cell proliferation (S-phase fraction, SPF). Five years had passed since inclusion and URS of the last patient; thus, the follow-up time was long in the calculation of survival. Additionally, the risk of death from UTUC in relation to SPF was calculated. An increased risk of death from UTUC with increasing SPF was found. Ploidy was useful for strengthening the assessment of grade 1 (diploid) and grade 3 (aneuploid) but not grade 2 UTUC.

Paper III was a larger study of cancer-specific survival (CSS) and the prognostic role of the different tumour characteristics studied in Paper II associated with survival and with invasive tumour stage. The cohort was extended to include 99 in the final analysis. SPF and stage were confirmed in multiple Cox analysis to be independent prognostic markers. The area under the ROC curve indicated that SPF was a good predictor for both the invasive stage and death from UTUC. CSS stratified by stage, grade and ploidy confirmed the large difference in survival between superficial and invasive stages, between different grades (when using the WHO 1999 classification) and between diploid and aneuploid UTUC. The risk of death from UTUC increased by 17% for every percent increase in SPF.

In summary, with a thorough and systematic work-up, the individual patient ́s risk can be estimated after URS so that those with high-risk UTUC can be treated with RNU, whereas patients with low-risk UTUC can consider KSS, i.e., URS laser ablation or segmental ureterectomy. Samples taken at URS are reliable for grading and the determination of ploidy and SPF. Tumour stage and SPF were found to be independent prognostic markers. The tumour characteristics that proved useful for “indirect staging” of UTUC were grade, ploidy and SPF, as these were independently associated with tumour stage. SPF was shown to strengthen risk stratification and can be analysed in barbotage taken at URS. We also showed that the WHO 1999 classification system was more informative than the two-tiered WHO 2004 classification and better predicted CSS.