Development of renal cell carcinoma (RCC) diagnostics and impact on prognosis

Authors


Abstract

Objective

  • To evaluate imaging methods and prognoses between small renal cell carcinomas (RCCs) and larger tumours according to the era of diagnostics.

Patients and Methods

  • In all, 784 consecutive patients diagnosed with RCC between 1964 and 1997 at the Pirkanmaa Hospital District in Finland were included.
  • Patients were divided into two groups: tumours of ≤3.0 and >3.0 cm in diameter.
  • Prognosis was analysed according to the era of diagnostics: (i) pre-computed tomography (CT) and pre-ultrasound (US), (ii) US era and (iii) CT era.

Results

  • Small tumours became more common: in the pre-CT and pre-US era, only 4.4% of tumours were small; however, in the CT era 16% were small tumours.
  • More diagnostic methods were used in studying small tumours.
  • CT proved to be the most reliable method, although it was actually better at diagnosing large tumours.
  • Relapses occurred less frequently among patients with small tumours; more than half of the tumours that developed distant metastases (16.0%) already evinced them at the time of diagnosis. There were no relapses after 14 years of follow-up among small tumours, whereas large tumours relapsed within that time. RCC was the cause of death in 14.9% of patients with small tumours vs 50.7% with large tumours.
  • The best prognosis was among patients with small tumours diagnosed with CT.

Conclusion

  • Among patients with small tumours, prognosis has improved along with better diagnostics, although some showed relapse during a surveillance period of up to 14 years.

Introduction

In RCC, a threshold size of 3 cm has long been considered significant [1]. Beyond this size, there is a sharp increase in the incidence of high stage and grade, multifocality, primary metastases and relapses [2, 3]. While small tumours represented only 5.3% of findings in the pre-CT and pre-ultrasound (US) era, this percentage increased almost five-fold after their general adoption [4]. Today, 17.3% of all RCCs and 43.4% of all stage I tumours are discovered at sizes of <3 cm [5, 6].

In IVU, a finding of parenchymal, especially central calcification is suggestive of a malignancy [7]. In angiography, a typical pattern involves a hypervascular mass with irregular vessels [8]. With US, most small RCCs are hyperechoic and present as homogeneous solid masses, although isoechoic, hypoechoic and cystic lesions are seen as well. The hypoechoic rim of the capsule is a characteristic feature; protrusion from the kidney is seen in 71% of all cases [9, 10]. A common presentation of small RCC in CT is a noncalcified homogeneous lesion with a baseline attenuation of >20 Hounsfield units (HU), enhancing by at least 10 HU after i.v. contrast [11, 12].

Occult RCC is identified in 0.3% of patients referred for abdominal CT [13] and in 0.04% referred for US [14]. Anywhere from 10–70% of small tumours are symptomatic: dorsolumbar pain, haematuria, flank mass, hypertension, weight loss, leukocytosis and even paraneoplastic cachexia can be observed. However, in tumours of >5 cm, symptoms are more likely [15-19]. Among incidental tumours, 27.5% are small; of symptomatic tumours, only 2–9.5% are small [18, 20].

In most studies, new imaging methods have changed the presentation of RCC. Between 1973 and 1999 the proportion of patients with tumours of >10 cm decreased from 55% to 26% whereas those with tumours of <5 cm increased from 0% to 29% [21]. Likewise, the proportion of small tumours increased from 10.2% to 62.7% between 1981 and 2006 [22]. However, this change has not been confirmed in all reports [23].

In the context of increasing diagnoses of small tumours, we evaluated changes and trends associated with their diagnostic methods and prognoses according to the imaging-method era and symptoms, in comparison to larger tumours.

Patients and Methods

RCC cases from the Pirkanmaa Hospital District (Tampere University Hospital and four secondary hospitals) diagnosed between 1964 and 1997 were collected from the Finnish Cancer registry. Wilms, uroepithelial and benign tumours were excluded, as well as lymphomas. Similarly, patients were excluded if their treatment was conducted only partially in our hospital district or if their medical records were destroyed 20 years after death. In addition, those patients without any imaging studies and patients without pathological or surgical confirmation of tumour sizes were excluded. Originally, 970 patients were identified; of which 784 were eventually included (demographics are presented in Table 1). Patients were divided into three subgroups according to the era of diagnostics. The threshold point was the year when at least 25% of patients were imaged with US and again with CT: pre-CT and pre-US era (1964–1979), US era (1980–1988) and CT era (1989–1997).

Table 1. Patient and tumour demographics according to study group
VariableGroup A, n (%)Group B, n (%)
  1. Group A, tumours ≤3.0 cm in diameter; Group B, tumours >3.0 cm in diameter.
Gender:  
Female41 (43.6) 
Male53 (56.4)385 (55.8)
Age, years:  
<2002 (0.3)
20–291 (1.1)2 (0.3)
30–394 (4.2)24 (3.5)
40–498 (8.5)69 (10.0)
50–5919 (20.2)166 (24.0)
60–6929 (30.9)228 (33.1)
70–7928 (29.8)164 (23.2)
≥805 (5.3)35 (5.1)
Decade of diagnosis:  
1960s1 (1.1)17 (2.5)
1970s6 (6.4)134 (19.4)
1980s26 (27.7)253 (36.7)
1990s61 (64.9)286 (41.4)
Era of diagnostics:  
Pre-CT and pre-US7 (7.4)151 (21.9)
US24 (25.5)222 (32.2)
CT63 (67.0)317 (45.9)
Operation:  
Nephrectomy76 (80.9)661 (95.8)
Partial resection15 (15.9)10 (1.4)
Not done3 (3.2)19 (2.8)
Stage:  
I78 (83.0)219 (31.7)
II0104 (15.1)
III6 (6.4)150 (21.7)
IV10 (10.6)217 (31.4)
Primary metastases:  
Lymph node3 (3.2)104 (15.1)
Distant metastases8 (8.5)153 (22.2)
Re-evaluated grade (n = 25 and 184):  
I01 (0.5)
II3 (24.0)13 (7.1)
III13 (52.0)94 (51.1)
IV6 (24.0)76 (41.3)

Information on symptoms, diagnostic delays, radiological investigations, surgeries, histology, tumour size, TNM stage and grade, recurrence, last date of surveillance or death and cause of death, were collected from the original medical records. The indication for the imaging studies was recorded depending on the question formulated by the clinician. Tumours were stratified according to pathological or operative measurements. Of all tumours, 50% were assessed pathologically, 84% radiologically and 75% by operative size measurement. Tumour grades diagnosed between 1985 and 1995 were re-evaluated by one uropathologist (P.M.K.) according to Fuhrman's grading. The Pirkanmaa Hospital District Ethical Committee and the National Authority for Medicolegal Affairs gave permission to collect information and to re-evaluate the tissue samples, respectively. Follow-up continued until death or August 2007.

Overall survival was analysed using the life-table method: the observed survival rates were compared with rates based on year, gender- and age-specific survival tables for the entire Finnish population. Calculations were based on the individual life expectancies of the target population for the target years. The relative survival of the reference population would be 1.0. If the survival curve remains below this, there is excess mortality. The 95% CIs of relative survival are calculated using Hakulinen's standard error formula [24, 25]. For survival rates statistical differences were analysed by t-test, other statistical differences were analysed by chi-squared or Wilcoxon rank test using a two-tailed interpretation. A P < 0.05 was considered to indicate statistical significance. The rating highly statistically significant (HSS) is given with a P < 0.000001.

Results

In this study population, 94 (12.0%) tumours were ≤3.0 cm in diameter (group A) and 690 (88.0%) were larger (group B), with no differences in gender or age. Small tumours had been diagnosed more recently (Table 1). Asymptomatic tumours were found more often in group A (Table 2). Of all symptomatic cases small tumours were found in 9.1%, of all asymptomatic in 29.2%. Group A had fewer symptoms; differences in the incidence of haematuria, abdominal mass and elevated erythrocyte sedimentation rate were seen (Table 2). Symptomatic patients in group B contacted health care later than symptomatic patients in group A: the median time was 12.5 days vs 1 day (P < 0.05).

Table 2. Indication, diagnostic studies needed and symptoms in the study groups
 Group A, n (%)Group B, n (%)P
  1. Group A, tumours ≤3.0 cm in diameter; Group B, tumours >3.0 cm in diameter; ESR: erythrocyte sedimentation rate.
Indication for imaging studies:  <0.001
Suspicion of RCC16 (17.0)246 (35.7)
Abdominal complaint27 (28.7)162 (23.5)
General health examination7 (7.4)15 (2.2)
Other44 (46.8)267 (38.7)
Diagnostic studies:  <0.05 < 0.01
12 (2.1)14 (2.0)
220 (21.3)204 (29.6)
338 (40.4)316 (45.8)
429 (30.9)135 (19.6)
54 (4.3)20 (2.9)
≥61 (1.1)1 (0.1)
Kidney biopsy9 (9.6)22 (3.2)
Symptom number:  HSS
033 (35.1)80 (11.6)
140 (42.6)299 (43.3)
215 (16.0)158 (22.9)
35 (5.3)106 (14.2)
4037 (5.4)
51 (1.0)8 (1.2)
≥602 (0.3)
Symptoms:   
Haematuria16 (17.0)241 (34.9)<0.001
Abdominal mass1 (1.1)64 (9.3)<0.05
Elevated ESR11 (11.7)179 (25.9)<0.01

The indication for imaging studies was suspicion of RCC more often in small tumours than in large ones (Table 2). The median time to the first imaging study after contact with health care was 1 day in group A and 7 days in group B (P < 0.01). In group A more diagnostic studies and kidney biopsies were needed (Table 2) and in more recent years more imaging studies have been performed (HSS) resulting in longer operation delays. The median time from the first diagnostic imaging study to operation was 57 days for small tumours vs 27 days for large tumours (HSS). Preoperative angioembolisation was done in 10.0% of cases in group B. There was no difference in the percentage of operated patients, but partial nephrectomy was conducted more often in group A (Table 1; HSS).

The imaging studies undertaken are shown in Fig. 1a,b; results are shown in Table 3. CT came in to general use earlier for studying small tumours than large ones; IVU became rarer along with this change. CT, US and IVU proved superior in large tumours when compared with the study results of small tumours. Altogether 37.5% of IVUs missed malignancy in small tumours. Altogether, a preoperative malignant imaging result was obtained in 93.6% of patients in group A and in 96.7% in group B.

Figure 1.

Percentages of different imaging studies according to study year in groups A (a) and B (b).

Table 3. Results with different imaging methods in the study groups
Imaging methodStudy groupDoneNormalAbnormalIndefinite malignantDefinite malignantP
n%n%n%n%n%
  1. Group A, tumours ≤3.0 cm in diameter; Group B, tumours >3.0 cm in diameter; n.s., non-significant; n.a., not available.
IVUA2425.5312.5625.0937.5625.0<0.001
B31545.7123.83511.17925.118960.0
AngiographyA6569.123.134.657.75548.6n.s.
B48970.961.291.8132.746194.3
CavographyA0000000000n.a.
B111.6436.419.1218.2436.4
USA8489.456.056.03136.94351.2<0.001
B19972.340.8193.87915.839779.6
CTA6164.912.323.32642.63252.5<0.001
B21831.620.920.9219.619388.5
MRIA33.20000133.3266.7n.a.
B121.7000018.31191.7

A re-evaluated grade was defined in 26.6% in both groups. Large tumours were higher grade (P < 0.05, Table 1) and higher stage (HSS). They evinced also more often primary lymph node involvement (P < 0.05) and distant metastases (P < 0.01, Table 1). Small tumours had better prognosis than large ones (Fig. 2, Table 4); differences were seen in the amount of relapses (18.1% vs 55.7%), local relapses (2.1% vs 3.2%) and distant metastases (16.0% vs 52.5%, HSS). Again, RCC was the cause of death more often in patients with large tumours (50.7% vs 14.9%, HSS).

Figure 2.

Overall survival according to study groups.

Table 4. Relative overall survival (OS) with 95% CI among all patients and according to study group and diagnostic era
 Group AGroup BP
OS, %95% CIOS, %95% CI
  1. Group A, tumours ≤3.0 cm in diameter; Group B, tumours >3.0 cm in diameter.
All, 5-year8382–855958–59HSS
All, 10-year6866–714544–46HSS
All, 20-year6748–853027–33<0.001
Before US and CT, 5-year6321–955351–55n.s.
Before US and CT, 10-year5110–924037–44n.s.
US, 5-year6555–755756–59<0.01
US, 10-year173–704442–46<0.001
CT, 5-year9291–946261–63HSS
CT, 10-year9188–945048–51<0.001

According to imaging-method era there were survival differences: the best prognosis was seen in patients diagnosed with small tumours in the CT era (Fig. 3, Table 4). No differences were seen between group A patients diagnosed in the first two eras; in group B this difference was statistically significant (P < 0.001) after 5 years, but not after 10 years. When comparing the US era with the CT era, there were marked differences in survival among both study groups (P < 0.001). Between the two study groups there was no difference in the first era, while in the US era group A survived better after 5 years, but the situation was the opposite after 10 years. In the last era small tumours had better survival. When the background factors concerning this survival improvement were considered, we found that the proportion of patients with primarily metastatic disease stayed the same in group B, but decreased among group A being 29%, 19% and 2% in the respective diagnostic eras (P < 0.01). Cancer-specific survival showed a trend of improvement, but it did not reach statistical significance because of the few cases: the cause of death was RCC in 43%, 30% and 15%, respectively. There was no difference in disease-free survival between the diagnostic eras. The proportion of asymptomatic patients increased (28.9%, 26.9% and 39.3%, respectively), but this did not reach statistical significance.

Figure 3.

Overall survival according to diagnostic era and study group.

Discussion

The diagnostics of renal masses have undergone a marked change since the introduction of CT and US. During the study period, the use of IVU and angiography was diminishing while the use of US and CT became more common (Fig. 1a,b), these were also the most common combination of imaging methods. Among British urologists in 1983 only 16% used CT and 93% recommended US [26]. In Finland in 1998 IVU, renography and cytology/biopsy examinations were no longer used while these were still in common use in other Scandinavian countries [27]. After the introduction of helical CT in 1989, more reliable measurements were achieved and the method came in to general use. In 1998, CT machines were equipped with multiple row detector arrays increasing dramatically volume coverage speed. Nowadays, the diagnostic algorithm is to perform abdominal US and then 2/3 phase helical CT with 5 mm collimation to minimise partial-volume artefacts [28]. Kidneys are scanned early during the corticomedullary and the tubular nephrogram phase [29]. Staging is recommended to be done by abdominal and chest CT and bone scintigraphy; cerebral CT is performed only when there is suspicion of brain metastases. If tumour is local, but there is a suspicion of a thrombus after primary imaging, MRI is recommended [28]. The performance of 18F-fluorodeoxyglucose positron-emission tomography (18F-FDG PET) CT in the detection of primary disease is limited because of renal excretion of FDG, resulting in high frequency of false-negative results. However, this method has a good performance in staging and the diagnosis of metastases; the major advantage is the detection of occult lymph node or bone metastases and differential diagnosis of tumour and coagulative thrombus [30].

The indication for imaging studies was often other than a suspicion of RCC. However, almost two thirds of small tumours were symptomatic, even if in 83% of cases, RCC was not suspected. RCC is notorious for variable and unspecific symptoms and even in two thirds of large tumour this malignancy was not primarily suspected. Both local and systemic symptoms were recorded up to five symptoms per patient reflecting the long study period; in more recent studies most tumours are incidental [20]. Only one patient with a small tumour and 14 with large tumours were studied by one method. Over 70% of group A were studied with three of four methods, while large tumours needed one method less. The diagnosis remained indefinite despite all methods in 6.4% of small tumours and in 3.3% of large, this mainly due to cystic tumours. The diagnosis of small tumours was more difficult than that of large tumours; more imaging studies and preoperative biopsies were required, this resulting in longer delays to surgery than for large tumours. The number of studies needed may also reflect the increasing number of methods available. While CT was the most reliable method (Table 3), 4.9% of small tumours were nonetheless missed. Usually, most tumours that are not detected by US can be identified by CT [31], but also CT is known to fail in tumours measuring <1 cm in diameter [32]. CT should detect all surgically verified tumours of >15 mm [33]. MRI has been shown to be slightly superior to CT: sensitivity is the same, but MRI is better in differential diagnosis [34]. As MRI offers no clear advantage, it is most useful in patients with iodine allergy or renal failure [35].

Here angiography yielded better accuracy than US due to its use in ‘road mapping’ for the surgeon. Currently, angiography is indicated only for interventions and in the cure of complications of nephron-sparing surgery [8]. US missed 12.0% of small tumours and <5.0% of large ones. The main limitations of US are known to emerge in cases of small isoechoic intraparenchymal tumours causing no deformity, and polar tumours with extrarenal growth that may be obscured by bowel gas [36, 37]. When comparing US with CT, it detects only 26% of CT-confirmed renal masses of <1 cm, but 85% of lesions of >3 cm [38]. US should detect all surgically verified tumours of >25 mm [33]. IVU failed in as many as 37.5% of small tumours, as unless the tumour is large enough to effect a change in renal contour or distortion of the collecting system, it can be missed [39]. IVU detects only 10% of CT-confirmed renal masses of <1 cm in diameter, but 85% of masses of >3 cm [38]. In another study sensitivity in detecting small RCCs was very similar: 67% for IVU, 74% for angiography and 94% for CT, but only 79% for US [40].

The use of fine-needle aspiration was not frequent. Only 40% of aspirations are reported to yield diagnostic malignant cells [41]. For this reason in 2005, 43% of urologists in the UK never used biopsy and 23% used it only for selected patients [42]. However, there is growing literature after 2001 that currently the success rate is >90% and insufficient material is noted only in 5% [43]. This is achieved by using helical CT-fluoroscopy guided core biopsy [44].

Small tumours became more common during the study period. These were of lower stage and grade evincing fewer nodal and distant metastases yielding better prognosis than large tumours, as also reported elsewhere [2, 3]. Most small tumours with metastases were already primarily disseminated, while few patients developed metastases during the follow-up up to 14 years after diagnosis, but not after that. Larger tumours did not achieve plateau (Fig. 2). In another series, 7.0% of small tumours were primarily metastasised; 7.0% developed recurrence at ≤5 years and 16% at ≤10 years [45], but as long a follow-up as ours has not been previously reported. The prevalence of concurrent metastases was in one large collected series 2.5% [46]. Even 6 mm primary tumours have been reported to have metastasised [45], but the risk of synchronous metastases increases by 22% for each 1 cm increase in tumour size [47]. These few disseminated small RCCs are very aggressive from the outset, but most tumours will grow slowly without relapse [46]. Cancer-related death was noted in 14.9% of patients with small tumours vs 50.7% of patients with larger tumours. Even if the difference in cancer-related deaths between diagnostic eras did not reach statistical significance, it is of clinical relevance that in the first diagnostic era patients died almost three times more often because of RCC than patients in the CT era. Incidental tumours found most often by CT have a good prognosis as these are usually operable.

During the present study period there was not an effective systemic oncological therapy for metastatic RCC [48]. In the study population there were no major perioperative complications to explain differences in survival (data not shown). Along with the change in diagnostics resulting in increasing numbers of incidental findings, patients with small tumours diagnosed in the CT era had the best survival. This is explained by the small proportion of primarily metastatic disease, as we did not find any statistically significant differences in the percentage of asymptomatic tumours or in disease-free and cancer-specific survival. Also, among large tumours, those diagnosed in the most recent era had a better prognosis than tumours diagnosed earlier. Small tumours achieved a plateau in survival after 14 years of follow-up, which was not seen among larger tumours. With multiple imaging methods there was no difference in the reliability of achieving preoperative diagnosis of RCC of any size. As more small tumours have been diagnosed incidentally with developed imaging methods, the overall prognosis of RCC has improved along with this change; clearly some of this is due to improvements in the diagnostics.

Acknowledgements

Funding for this research was received from Tampere University Hospital/Research Center/Elna Savolainen Fond.

Conflict of Interest

None declared.

Abbreviations
FDG

fluorodeoxyglucose

HSS

highly statistically significant

US

ultrasonography/ultrasound

Ancillary