Changing prognosis of renal cell carcinoma: a single-centre experience over 25 years

Introduction

Carcinoma of the kidney accounts for approximately 2% of all cancers in the United Kingdom (UK), with on average 6700 new cases diagnosed each year. ¹ The incidence is increasing by approximately 2% each year, in part because of the increased use of imaging modalities. ² Despite this increase in early detection, about one-third of patients who present with localised kidney cancer develop metastases, and more than half will die from the disease. Renal cancer, therefore, remains a significant cause of mortality and morbidity.

From 1975 to 2003 incidence rates in men and women have increased by 69% and 100%, respectively, in the UK. This may be partly because of the increasing average age of the population, with the majority of the increase seen in men over 65, and women over 55 years old. ³ Another contributing factor is the increasing use of diagnostic imaging, with a subsequent rise in the incidental finding of a renal mass. There has been a 73% rise in the use of ultrasound scanning (USS) and computerised tomography (CT) in the United States (US) from 1986 to 1994. ⁴ However, this has also demonstrated a significant rise in asymptomatic metastatic renal cell carcinoma, so the improved survival of kidney cancer may not be wholly accountable to this.

Materials (patients) and methods

The inclusion criteria for the subsequent data analysis were all nephrectomies (radical (open/laparoscopic) or partial (open/laparoscopic) with documented cancer involvement reported by the pathology department (Oxford Radcliffe Hospitals NHS Trust) from January 1983 to December 2007. Pathological data from these cases were assessed following review by specialist uropathologists within the Cancer Research Uro-Oncology Database (CRUD©) system ⁵ to extract pathological tissue diagnosis, tumour T-stage (all tumours reclassified according to 2002 criteria using pathology report; however, only tumours after 2001 were prospectively assessed using new classification), Fuhrman grade, and maximum tumour diameter (cm) as measured at histological cut-up.

Each tumour was related to available clinical data (age, sex and date of operation) and patient survival data from the Oxford Cancer Intelligence Unit ⁶ (OCIU) within the CRUD© system. Survival data from the OCIU originate from death certificates, with deaths directly caused by kidney cancer being taken as those classified as cause of death category 1a on the death certificate. The death status of patients was classified by alive/dead, date of death, cause of death (kidney cancer/other).

Statistical analysis and graphs were generated using GraphPad Prism v4.03. The log rank test was used to perform a univariate analysis for survival and the survival curves were estimated by Kaplan-Meier method. P values were considered statistically significant when p < 0.05.

Results

A total of 750 cancer nephrectomies were performed at the John Radcliffe or Churchill Hospitals from January 1983 to May 2007. Of these, 664 (89%) were for renal cell carcinoma (RCC), with the remaining 11% mainly transitional cell carcinoma (TCC). The one-, five- and 10-year survival of patients post-nephrectomy for RCC was 89%, 73% and 63%, respectively.

Renal cancer survival — trends over time

To assess how prognosis from kidney cancer has changed over time, all RCC nephrectomies were divided into six cohorts of four inclusive years based on their date of operation. Figure 1 shows a clear change in kidney cancer- specific survival over the past 25 years (log rank test p < 0.0001), with one-year survival improving from 74% (1983–1986) to 95% (2003–2007), and 10-year survival changing from 17% (1983–1986) to 70% (1995–1998) (see Table 1).

OPEN IN VIEWER

Figure 1.

Kaplan Meier Survival Curve (death from kidney cancer) of all cancer nephrectomies for RCC in Oxford 1983 to 2007, split into four-year cohorts.

OPEN IN VIEWER

Table 1.

One-, five- and 10-year survival patients undergoing nephrectomy for RCC, in Oxford 1983 to 2007, split in four-year cohorts.

	Year of nephrectomy
	1983–1986	1987–1990	1991–1994	1995–1998	1999–2002	2003–2007
One-year survival	74%	76%	84%	89%	90%	95%
Five-year survival	42%	58%	69%	75%	73%	89%
10-year survival	17%	47%	59%	70%	N/A	N/A

RCC: renal cell carcinoma.

Number of nephrectomies

The number of RCC nephrectomies performed each year has increased by a factor of 10 over the past 25 years and has a significant association with five-year survival from kidney cancer (Chi squared p < 0.0001).

Fuhrman grade

Nuclear Fuhrman grade shows clear univariate survival significance (log rank test p < 0.0001), with Fuhrman grade 4 tumours having the shortest survival. Mean grade per year has not changed significantly over this time; however, the distribution of low-/high-grade tumours has changed significantly, with high-grade tumours increasing and low-grade tumours decreasing (p = 0.009, Table 2). Survival per Fuhrman grade category (1983–2001 vs 2002–2007) is not significantly altered for grade 3 or 4 tumours, but is significantly better for Fuhrman grade 1 and 2 tumours.

OPEN IN VIEWER

Table 2.

Change in low-grade/high-grade RCCs in Oxford 1983–2007. Change in grade: p = 0.009 (Chi squared).

	% 1983–2001	% 2002–2007
Fuhrman grade 1 or 2 (low grade)	68	55
Fuhrman grade 3 or 4 (high grade)	32	45

RCC: renal cell carcinoma.

Tumour pT-stage

Tumour pT-stage (as defined by the World Health Organization (WHO) classification) is an important univariate prognostic factor (log rank test p < 0.0001). Analysis of the prognostic significance of stage after 2002 demonstrates consistent five-year survival for pT3 and pT4 tumours, but a significantly improved survival in pT2 tumours from 85% to 97% (p = 0.033), and a significantly improved survival in pT1 tumours from 93% to 100% when compared to 1983–2001 (see Table 3) (p = 0.030). The distribution of pT-stage has not changed over time.

OPEN IN VIEWER

Table 3.

Change in distribution of pT1 and pT2 RCC five-year survival in Oxford 1983–2007.

Change in pT1 stage survival: p = 0.031 (Log-rank (Mantel-Cox) test)

Change in pT2 stage survival: p = 0.033 (Log-rank (Mantel-Cox) test)

	Five-year survival 1983–2001	Five-year survival 2002–2007
pT1 Tumours	93%	100%
pT2 Tumours	85%	97%

RCC: renal cell carcinoma.

Tumour size — trends over time

Mean tumour size has not changed over the past 25 years. Further analysis of tumour size was performed by assessing the percentage of cases per centimetre. The size of tumour was subcategorised into three similarly sized cohorts (A: <6 cm, B: 6–8 cm and C: >8 cm) tumours. Figure 2 shows a five-fold increase in the number of size A tumours, with a corresponding decrease in size B tumours. Of note is that there is no change in size C tumours over the past 25 years. This subcategorisation of maximum tumour diameter is a significant univariate prognostic marker (log rank test p < 0.0001), with size C tumours having the poorest survival (see Figure 3).

OPEN IN VIEWER

Figure 2.

Percentage distribution of maximum tumour diameter per four-year cohort according to size subcategory (A–C) of all cancer nephrectomies for RCC in Oxford, 1999 to 2007.

OPEN IN VIEWER

Figure 3.

Kaplan Meier Survival Curve (death from kidney cancer) of all cancer nephrectomies for RCC in Oxford, 2002 to 2007, split into three cohorts of size subcategory (A–C).

Discussion and conclusions

Five-year cancer-specific survival for RCC in Oxford in 2007 was 73%. This compares to national survival rates in the UK of 42%. ¹ This was certainly not the case 20 years ago, when the five-year survival in Oxford was 43%, i.e. the same as the national rates now. The trend towards improved survival over the years has been progressive, with gradually increasing survival seen for every four-year cohort analysed. The causes of this significant change are likely to be multifactorial in our centre.

The number of nephrectomies performed in our institution has risen by a factor of 10, which correlates to the improving survival. In our centre from 2002 onwards the introduction of subspecialisation led to two surgeons rather than seven performing all nephrectomies. There has also been increased specialisation through the multidisciplinary team over this time. It is proposed that high throughput cancer centres with experienced urological surgeons performing large numbers of operations is beneficial to patient outcome ⁷ and this has resulted in changes in provision of cancer services within the UK in recent years. It is not clear whether this factor alone may have resulted in the dramatic increase in RCC survival demonstrated, although it may be a contributing factor.

Fuhrman grade is a recognised prognostic marker in RCC. This study showed an improved survival in grade 1 and 2 RCC tumours in the past 20 years. Associated with this there has been a significant increase in the proportion of G3 and G4 tumours at the expense of G1 and G2 tumours. This suggests a change in the tumour biology of these tumours towards higher grade; however, improved pathological reporting due to subspecialisation could also contribute to these findings. Despite this the survival rates for G1 and G2 tumours has improved significantly. A further contributing factor may be improved surgical outcomes.

Tumour T-stage is a known prognostic marker in RCC. This study showed a significant increase in the percentage of tumours <6 cm with a concomitant reduction in the percentage of tumours measuring 6–8 cm. There was no change in the percentage of tumours >8 cm. This may relate to increased detection and the trend of progressively improving survival, as the five-year survival for size A tumours is 95% whilst for size B tumours it is 85%. However, this change would appear to account for only a 10% increase in five-year survival and does not completely explain the changes documented.

The consistent proportion of size C tumours undergoing nephrectomy is interesting and seems to go against the changes seen in tumours ≤8 cm. This cohort of large tumours could represent an alternative tumour biology that grows rapidly, avoiding early detection and curative nephrectomy. These patients may highlight a cohort to include in future research trials.

A compounding factor with the T-staging during this study was a change in the technique used to dissect the macroscopic specimen for pathology examination in order to determine the TNM classification since 2002. The inclusion of more blocks of the renal sinus into pathological assessment increases the proportion of tumours where vascular invasion of the renal sinus veins is recognised and has meant that a significant proportion of tumours which would have been staged as T2 prior to 2002 are now classified as T3b. ⁸ This could explain some of the demonstrated improved survival for pT1 and pT2 tumours as a population of small but vascular invasive tumours are now removed from this group (see Table 4).

OPEN IN VIEWER

Table 4.

Log rank survival analysis of comparison between three size cohorts subcategories (A–C).

			p value	HR	95% CI
Tumour size A (<6 cm)	vs	Tumour size B (6–8 cm)	0.0230	0.58	0.32–0.92
Tumour size B (6–8 cm)	vs	Tumour size C (>8 cm)	0.0080	0.59	0.40–0.87
Tumour size C (>8 cm)	vs	Tumour size A (<6 cm)	<0.0001	0.32	0.22–0.45

HR: hazard ratio; CI: confidence interval.

With the increased quality and use of USS and CT for abdominal imaging, a greater number of incidental (pre-symptomatic) tumours are being detected. ⁴ The true significance of these tumours is currently a hotly debated subject in urology. It is difficult to know what the true clinical course of these tumours would be if left undetected and whether the removal of all of them is warranted. However, the increased use of imaging could explain the shift in tumour sizes to the greater numbers of smaller (<6 cm) tumours seen in this study.

It is also not clear cut whether the increased detection of these renal tumours is solely due to improved diagnostic imaging or whether there is a natural pattern of increasing incidence of RCC from another source. The increase in detection rates has also included locally advanced and metastatic renal tumours, which does suggest that the increased detection of localised tumours is not the sole reason for the observed improvement in patient survival.

The five-year disease-specific survival rate in patients with pT1 renal carcinoma was 95% and in those with stage pT2 disease, 88%. Patients with pT3 renal carcinoma have a five-year survival rate of 59%, and those with pT4 disease, 20%. This appears to be comparable with other contemporary series where five-year disease-specific survival was 95% in pT1a, 93% in pT1b, 86% in pT2a, 70% in pT2b, 65% in pT3a, 55% in pT3b, 18% in pT3c, and 27% in pT4. ⁹