Renal cell carcinoma: stage, grade and histology migration over the last 15 years in a large Australian surgical series

Authors


Associate Professor Manish I. Patel, University of Sydney, Urological Cancer Surgeon, Westmead and Sydney Adventist Hospital, Sydney, Australia.
e-mail: mpatel@med.usyd.edu.au

Abstract

Study Type – Therapy (case series)
Level of Evidence 4

What’s known on the subject? and What does the study add?

Recent studies suggest that the incidence of RCC is increasing among early-stage, incidentally discovered tumours, presumably detected because of the increased use of soft tissue imaging for evaluating a wide variety of abdominal and pelvic symptoms. This study did not show migration towards earlier stage and smaller RCCs.

OBJECTIVE

• To verify whether migration towards earlier-stage renal cell carcinoma (RCC), which has been observed in the USA over the last decade, also applies to Australia.

PATIENTS AND METHODS

• Between January 1993 and December 2007, 547 nephrectomies performed in public and private hospitals in western Sydney were analysed from a retrospectively collected database.

• Data were divided into three consecutive time groups.

• Tumour-node-metastasis (TNM) stage as well as patient demographics, size, grade and histology of tumours and proportion of benign tumours were also assessed.

RESULTS

• In all, 499 nephrectomies were performed for RCC. The median age was 62 years, with a male:female ratio of 1.9 : 1. Similarly sized tumours were identified in each time group [group 1 (1993–1997), 54.8 mm; group 2 (1998–2002), 52.0 mm; group 3 (2003–2007), 52.2 mm, P= 0.6].

• Pathological stage II disease decreased from 18.1 to 11.1%, but stage III disease showed an increase from 13.9 to 21.5% over that time period (P= 0.02).

• The proportion of stage I and stage IV disease has remained relatively the same. There has been a statistically significant upward histological migration for the papillary subtype from 1.3 to 10.2% (P= 0.01).

• There has also been an increasing representation of Fuhrman grade III tumours over time, from 17.6 to 30.8%, and a decreasing proportion of Fuhrman grade I tumours from 16.2 to 7.1% (P= 0.03).

• There was a decrease in the incidence of benign tumours originally thought to be malignant on preoperative investigations, from 10% in group 1 to 4% in group 3 (P= 0.03).

CONCLUSION

• The recent US phenomenon of migration towards earlier-stage, smaller RCCs as well as increased representation of benign tumours was not observed in the present study. The results of the present study, however, show an upward histological migration for papillary RCCs and an increasing representation of more aggressive Fuhrman grade III tumours.

INTRODUCTION

Renal cell carcinoma accounts for 3% of cancers in adults [1] and is the most common malignant tumour of the kidney [2]. According to the Cancer Institute of New South Wales (NSW), Australia, there are approximately 950 new cases each year in NSW. It is responsible for 320 deaths each year in NSW.

Recent studies suggest that the incidence of RCC is increasing among early-stage and also later-stage tumours [3]. A large American study has shown, however, that the biggest increase is among early-stage, incidentally discovered tumours, presumably detected because of the increased use of soft tissue imaging for evaluating a wide variety of abdominal and pelvic symptoms [4]. Other studies have also shown an increase in the detection of benign renal tumours [5].

The present study aims to examine whether there has been a migration by tumour size and stage for RCC according to the American Joint Committee on Cancer (AJCC) staging system in a major Australian referral centre over the last 15 years, and the magnitude, if any, of this stage migration.

Tumour-node-metastasis stage has been identified as one of the most reliable and accurate predictors of survival for patients with RCC [2,6]. Fuhrman nuclear grade (based on nuclear size and shape and the prominence of nucleoli) and histological subtype are also recognized prognostic factors in RCC. Higher grades correlate with the biological aggressiveness of the tumour and increased metastatic potential. There is, however, more debate about histological subtype [2,7]. There are three main histological subtypes: conventional clear-cell carcinomas, papillary carcinomas and chromophobe carcinomas. Other rarer subtypes include collecting duct carcinomas and unclassified RCCs. These other prognostic factors are also analysed in the present study for trend and magnitude.

PATIENTS AND METHODS

All nephrectomies, including partial nephrectomies, performed for suspected RCC between January 1993 and December 2007 by six urology consultants at Westmead Public Hospital, Westmead Private Hospital and Sydney Adventist Hospital in Sydney were retrospectively incorporated into a database.

The recorded data included patient demographics, TNM stage, size on preoperative CT, operative details and final pathology.

The data were divided into three equal consecutive time groups: group 1, from January 1993 to December 1997; group 2, from January 1998 to December 2002; and group 3, from January 2003 to December 2007. Pathological stage was classified according to the 2002 AJCC staging system (sixth edition) [8]. Cases diagnosed before the implementation of the sixth edition were restaged using the sixth edition criteria.

Results were analysed using SPSS, version 16. Significance of trend in stage distribution over time was assessed with the Pearson chi-squared test for trend. Significance of trends in Fuhrman nuclear grade, histological subtype, sex distribution, tumour size and benign lesions was also assessed. Descriptive statistics and T-test for equality of means were used to analyse age. The Mann – Whitney test for significance of trend was used to analyse pathological tumour size. No adjustment was made for the multiple comparisons done in the present study.

Results reported in the present study were in accordance with guidelines set out by the Sydney West Area Health Service Human Research Ethics Committee.

RESULTS

Between 1993 and 2007 a total of 547 patients from the three centres mentioned earlier fitted our case definition. Stage information was only missing for three patients. All remaining patients had complete pathological information for review.

In the present study, we also looked at the proportion of lesions originally believed to be malignant on preoperative investigation and that were ultimately benign on final histology. All imaging was performed using CT. Table 1 shows the proportion of these benign lesions in each time group. Interestingly, there has been a statistically significant decrease in the proportion of benign lesions between groups 1 and 3, from 10 to 4% (P= 0.03).

Table 1.  Benign vs malignant* renal lesions by time group [N (%)]
GroupsBenignMalignant
  • *

    Malignant lesions include RCCs and other malignant lesions originally thought to be RCCs.

1 9 (10) 81 (90)
215 (8)167 (92)
3 11 (4)264 (96)
Total35 (6)512 (94)

In all, 499 patients operated on were diagnosed as having RCC on final pathology. There were 75, 163 and 261 patients with RCC in groups 1, 2 and 3 respectively. The median (range) age was 62 (17–100) years. There were 328 men and 171 women. The male:female ratio was 1.9 : 1. The median age and sex ratio confirm that this is a typical population of patients with RCC [9]. No difference in sex distribution between time periods was observed (P= 0.09). There was, however, a significant difference in age between earlier and later groups (P= 0.004). The mean (95% CI) age of patients in group 1 was 59 (56–62) years, compared with 63 (62–65) years for group 2.

As can be seen from Table 2, over the 15-year period, a mean of 66.6% of RCCs were stage I, 12.9% were stage II, 20.1% were stage III, and 0.4% were stage IV. Figure 1 shows a decrease in stage II disease from 18.1 to 11.1% but an increase in pathological stage III disease from 13.9 to 21.5% over that time period (P= 0.02). The proportion of stage I and stage IV disease remained relatively the same.

Table 2.  RCC stage – percentage distribution by time period [N (%)]
GroupsStage IStage IIStage IIIStage IV
  1. Group 1, 1993–1997; group 2, 1998–2002; group 3, 2003–2007.

1 48 (66.7)13 (18.1)10 (13.9)1 (1.4)
2108 (65.9)22 (13.4)34 (20.7)0 (0)
3175 (67)29 (11.1)56 (21.5)1 (0.4)
Total331 (66.6)64 (12.9)100 (20.1)2 (0.4)
Figure 1.

RCC pathological stage distribution by diagnosis year group.

With regard to Fuhrman nuclear grade, one of the other prognostic indicators, it can be seen from the results in Table 3 that most RCCs are classified as grade II. Figure 2 shows that there is an increasing representation of Fuhrman grade III tumours over time, from 17.6 to 30.8%, and a decreasing proportion of Fuhrman grade I tumours from 16.2 to 7.1%, with a significant test for trend for these tumour grades (P= 0.03).

Table 3.  RCC Fuhrman grade – percentage distribution by time period [N (%)]
GroupsGrade IGrade IIGrade IIIGrade IV
112 (16.2)44 (59.5)13 (17.6) 5 (6.8)
222 (13.9)93 (58.9)32 (20.3) 11 (7.0)
318 (7.1)136 (53.8)78 (30.8)21 (8.3)
Total52 (10.7)273 (56.3)123 (25.4)37 (7.6)
Figure 2.

RCC Fuhrman grade distribution by diagnosis year group.

Analysis of histological subtypes shows that most are conventional clear-cell carcinomas (Table 4). Figure 3 shows that there is a trend towards less conventional RCCs, but this trend is not statistically significant (P= 0.09). There has, however, been a statistically significant upward histological migration for the papillary subtype, from 1.3 to 13.0% (P= 0.01), as compared with earlier years.

Table 4.  RCC histological subtype*– percentage distribution by time period [N (%)]
GroupsConventionalPapillaryChromophobe
  • *

    Unclassified and collecting duct tumours are not included.

1 64 (85.3) 1 (1.3) 6 (8.0)
2132 (81.0)16 (9.8) 11 (6.7)
3195 (74.7)34 (13.0)24 (9.2)
Total391 (78.4)51 (10.2)41 (8.2)
Figure 3.

RCC histological distribution by diagnosis year group.

In analysing pathological size, we have identified similarly sized tumours in each time group. The mean pathological sizes for groups 1, 2 and 3 were 54.8, 52.0 and 52.2 mm, respectively (P= 0.6). Considering size in more detail (Table 5), particularly stage I tumours, there was no significant decrease in size over time (P= 0.6). This contrasted with the experience in the USA where the size of stage I tumours decreased significantly from a mean of 41 mm in 1993 to 36 mm in 2004 [10]. This has prognostic significance because the threshold between stage Ia and stage Ib is 4 cm. This is also the arbitrary limit, although recently challenged, for elective nephron-sparing surgery.

Table 5.  Mean pathological size over time for T1 RCCs: USA vs Australia
Time groupsPathological size T1 tumours, mm
USAAustralia
15 years ago4142
Today3641

DISCUSSION

The experience of the Australian sample population described in the present study is not similar to that seen in the USA. In a large retrospective American study of patients between 1993 and 2004, researchers found an increasing representation of stage I tumours (defined as a tumour ≤7 cm less that is confined to the kidney) from 51 to 60%, and decreasing proportions of stage II and III disease [4]. Pathological stage II disease decreased from 14.5% in 1993 to 11.6% in 2004 (P < 0.001), whereas there was a larger decline in pathological stage III disease, from 22.8% in 1993 to 16.3% in 2004 (P < 0.001). Pathological stage IV disease remained relatively constant throughout the 12-year period at about 12% (P > 0.25), a finding similar to that in the present study (Table 6). This evidence suggests that all stages, except stage IV, are migrating to earlier stages over time and that a greater proportion of newly diagnosed patients with RCC in the USA are presenting with stage I disease than was the case in previous years.

Table 6.  Comparison of RCC stage distribution: USA [4] vs the present study
Time groupsStage I (%)Stage II (%)Stage III (%)Stage IV (%)
USAPresent studyUSAPresent studyUSAPresent studyUSAPresent study
15 years ago516714.518.122.813.9121.4
Today6067 11.6 11.116.321.5120.4

This is an important observation because the TNM classification has been identified as one of the most reliable and accurate predictors of survival for patients with RCC. Previous studies have shown that the 5-year cancer-specific survival rates by the 2002 tumour classification were 97, 87, 71, 53, 44, 37 and 20% in patients with pT1a, pT1b, pT2, pT3a, pT3b, pT3c and pT4 RCC, respectively [11].

In the USA it is possible that improvements in survival have not been observed because of the lack of change in stage IV RCC over time. Although there has been no stage migration in the sample population in the present study for stage I cancer, we have been operating on a higher proportion of stage I cancer than the USA and have done so consistently over the last 15 years. Table 6 shows that 67% of patients with RCC operated on in Australia have stage I disease. In the USA, however, 15 years ago 51% of patients with RCC were at stage I and this has increased to about 60% in 2004. We have also seen much less stage IV disease in the Australian population in the present study. Therefore, a higher proportion of our patients will have improved cancer-specific survival, will require less rigorous postoperative surveillance and will be eligible for nephron-sparing surgery. There is also room for further improvement with the increasing use of, and advances in, abdominal imaging.

On the down-side, there has been over time a migration towards more aggressive Fuhrman grade III tumours. The cause is not clear but it could represent changing patterns of scoring by pathologists, as has been observed with Gleason scoring for prostate cancer, rather than a true worsening of tumour aggressiveness. This hypothesis could be tested by central rescoring of Fuhrman grades of all RCCs from all three time periods. Alternatively, this observed increase in more aggressive Fuhrman grade III tumours could explain why, despite increased detection and treatment, mortality from kidney cancer appears to be increasing.

The present study also examined histological subtype and has shown a significant migration towards more papillary RCCs and a trend towards less conventional RCCs. The cause of the change in histology is not obvious, but, if real, it would confer a survival advantage in our population group. Histology is clearly a prognostic indicator in RCC and several studies have shown significant differences in outcome. The study by Cheville et al. [12] showed that patients with clear-cell RCC had a poorer prognosis than patients with papillary and chromophobe RCC (P < 0.001). The 5-year cancer-specific survival rate for clear-cell RCC was only 68.9% compared with 87.4 and 86.7% for papillary and chromophobe RCC, respectively.

It is also interesting to note that the proportion of lesions originally thought to be malignant on preoperative investigation and that turned out to be benign on final histopathology has decreased over time. To a degree these data run counter to the USA data, which indicate benign tumour resections in the range of 25%. We hypothesize that this could be due to a combination of improved imaging modalities and increasing use of active surveillance, where only small tumours that are growing are treated, as well as the availability of minimally invasive ablation techniques such as radiofrequency ablation and the greater use of sub-specialized uroradiology imaging services. Our institution performed radiofrequency ablation on 60 patients with small renal masses during the last time period (group 3), and percutaneous biopsy at the time confirmed benign histology in only five patients. It is therefore unlikely that adding this cohort to the series in the present study would substantially alter the percentage of benign tumours diagnosed. It is conceivable, therefore, that in Australia the diagnosis and treatment of small renal tumours is not as aggressive as in the USA. However, if this hypothesis is correct, it raises the question as to why this has not been observed in surgical series in the USA.

The profound stage migration in the USA is probably due to the increased and widespread use of cross-sectional imaging. In Australia, however, we have not experienced the same stage migration towards earlier-stage tumours as in the USA, possibly due to lower availability or clinical dependence on abdominal imaging assessment, reflecting the different pattern of clinical management. However, a large institutional series has shown a lack of survival benefit despite a stage migration over the last 15 years [13]. Despite more aggressive treatment and earlier detection, overall survival rates are not improving as was seen in the analogous situation with prostate cancer in the PSA era. The possibility exists that environmental or lifestyle co-factors are leading to more virulent strains of RCC despite treatment of smaller lesions and those with indolent pathology [13]. Therefore the demographic trends seen in Australian could also be happening in the USA and elsewhere, in addition to the more aggressive treatment of the smaller and more benign tumours.

Interestingly, other studies have shown that incidental detection was not an independent prognostic factor of death, indicating that these tumours are of a similar biological nature as symptomatic RCCs, only diagnosed earlier. In fact, the improved cancer-specific survival of incidental cases has not, thus far, resulted in a decrease in overall mortality from kidney cancer, perhaps because of this lead-time bias [14]. Therefore, despite increased detection and treatment, mortality from kidney cancer appears to be rising in the USA [15]. Indeed, we have identified in the present study a significant trend towards more aggressive Fuhrman grade III tumours, which could explain this. However, other larger series have shown that symptoms on presentation are a poor prognostic sign [16].

Survival data were not collected in the present study but this would be an interesting analysis and could be carried out in subsequent studies to compare relative survival according to stage, grade and histology. The most important limitation of the present study is that our database captures only pathologically confirmed RCCs. Thus RCCs that have been ablated or observed are not included. The precision of the information in the present study depends on the quality of the tumour registry data entered. Errors in staging could also have occurred and no central case review for staging accuracy and pathological review was performed. Comparisons of variables are performed independently, although it is likely that these variables are correlated (e.g. stage, grade and subtype). It would therefore be interesting to examine their combined effects in a multivariable model. Numerous pathologists over the years were involved in examining specimens for this sample. There could have been a significant change in pathologists that might explain some of the findings of the present study.

Despite these limitations, the present study is the largest multicentre database analysis of kidney cancer stage, grade and histological migration performed in Australia and the southern hemisphere. The authors do acknowledge, however, that 499 nephrectomies for RCC over a period of 15 years do not represent a very large multicentre cohort, compared with the volumes found in large referral centres in the USA.

In conclusion, the present study did not show migration towards earlier-stage and smaller RCCs in Australia over the last 15 years. However, despite this, it was shown that a larger overall proportion of surgically treated RCCs in Australia are stage I than is the case in the USA. The other positive findings are decreasing representation of stage II tumours and increasing representation of stage III tumours. There has been a decrease in the proportion of surgical resections of benign lesions over the years, possibly due to the increasing use of dedicated uroradiology services for preoperative assessment of suspected tumours. There has also been a significant upward histological migration for the papillary subtype, but at the same time there has been a significant increase in more aggressive Fuhrman grade III tumours.

CONFLICT OF INTEREST

None declared.

Ancillary