- Top of page
- SUBJECTS AND METHODS
The number of small and incidentally detected RCCs is increasing; some of these tumours are amenable to nephron-sparing surgery, with 5 and 10-year survival rates comparable to those achieved with radical nephrectomy [1,2]. Whether removed by nephrectomy or nephron-sparing surgery, the early detection of small, organ-confined RCCs should improve the long-term prognosis of the disease, as it is insensitive to both chemotherapy and radiation. This makes RCC a logical candidate for a screening programme, provided that a diagnostic tool of sufficient sensitivity and specificity is available.
Several studies have reported the detection of renal masses by abdominal ultrasonography (US) [3–6]. In all of these studies the prevalence of RCC was higher than that of other pathological (benign) findings. The high sensitivity of US for RCC depends on tumour size; it is 96% for tumours of> 3 cm in diameter and 79% for those < 3 cm . Thus it is apparent that US facilitates the detection of RCC, but the practicability and efficacy of such systematic screening has not been studied before. The German Ministry of Health funded the present pilot study to investigate screening by US for RCC in two cities over a 2-year period.
SUBJECTS AND METHODS
- Top of page
- SUBJECTS AND METHODS
A screening programme using US for RCC was established in Mainz (180 000 inhabitants) and Wuppertal (370 000 inhabitants). Participants were screened by 55 GPs, 79 internists and 19 urologists in private practice, all experienced in and equipped for renal US. Overall, 68% of all eligible physicians participated in the programme. Eligibility was determined by documented qualifications for abdominal US and an active medical practice. Involvement in the already established German screening programmes was an additional prerequisite.
In cooperation with different health insurers, an infrastructure was developed to inform physicians and the eligible study population. All citizens of both cities could attend the screening free of charge. Inclusion criteria were: age ≥ 40 years (according to the inclusion criteria of the already established German screening programmes), no symptoms of possible renal origin (e.g. haematuria, flank pain) and no known renal disease. Informed consent was obtained in writing from all subjects.
There were three modes of recruitment of subjects for the RCC screening programme, i.e. exclusively for the RCC screening, during participation in the other established German screening programmes (for skin, colon, breast, cervical and prostate cancer) and during an office visit for other than renal complaints. The reason for participation was recorded. Based on data available in German tumour registries, on the reported incidence of RCC  and on a previous retrospective abdominal screening study , it was calculated that 10 000 screening participants would be necessary to detect enough RCCs for statistical analysis.
The screening was in two consecutive phases of 13 months each; the first started in December 1996 and the second in January 1998. The second phase offered follow-up US to the population screened in the first phase. Reference US at one of the two university urology departments was recommended but was not obligatory for every individual with an equivocal renal mass on initial screening US.
Masses discovered on screening were described either as equivocal or positive; the criterion for equivocal findings was a solid renal mass suspicious of RCC, and that for a positive finding a solid mass typical of RCC. The criterion for a negative result was no evidence of tumour. Other solid tumours indistinguishable from RCC by imaging studies, e.g. TCC, oncocytoma, angiomyoma, leiomyoma and lymphangioma, were naturally included. Masses consistent with angiomyolipomas on US and/or CT were excluded.
To compare cases of RCC in the screened population to those from an unscreened population, the clinical data on all patients presenting to the Departments of Urology in Mainz and Wuppertal with a renal mass were registered and analysed. The TNM system of tumour classification (UICC 1992) was used. The results were assessed using descriptive statistical analyses, the epidemiological measures of interest being the sensitivity, specificity and positive predictive value (PPV).
- Top of page
- SUBJECTS AND METHODS
In all, 9959 volunteers (49% men and 51% women, mean age 61 years, range 40–94; 55% aged 55–70) participated in the first screening phase. The participation rate in the second phase was 79% (7851) of all participants in the first phase. Of all screened individuals, 49% (4763) came to the screening examination exclusively because of the new programme; the rest were informed about the programme while at the physician's office for other reasons: 15% (1441) came for other screening programmes offered by the German Ministry of Health, while 38% (3755) presented with other than urological symptoms.
Table 1 lists the data from the first screening phase, including US and CT findings, surgical procedures, histology, tumour stage and size of detected tumours. RCC was confirmed on histology in nine cases, including six men and three women (mean age 61.6 years, range 40–85). Of these, three were aged 40–49 years and two 60–69 and 70–79, respectively (Table 2). Two other tumours had benign histology (nos 10 and 11) while pre-existing medical comorbidities made surgical exploration too risky in two further patients (nos 12 and 13, aged 60 and 61 years). Of the nine cases with RCC, six were declared positive and three equivocal for tumour on initial screening US. All but two (nos 2 and 6) positive cases underwent immediate CT.
Table 1. Detected renal tumours on screening US during the first phase of screening
|screen US||reference US||CT||stage||size, cm|
| 1||equiv||equiv|| +ve||Neph||RCC||pT2, N0, M0||6.0|
| 2|| +ve|| +ve|| +ve||Neph||RCC||pT2, N0, M0||7.7|
| 3|| +ve|| || +ve||Neph||RCC||pT1, N0, M0||2.5|
| 4|| +ve|| || +ve||Neph||RCC||pT3b, N1, M1||7.0|
| 5|| +ve|| || +ve||Neph||RCC||pT3b, N2, M1||13.0|
| 6|| +ve|| +ve|| +ve||Neph||RCC||pT2, N0, M0||9.0|
| 7|| ||equiv|| +ve||NSS||RCC||pT2, Nx, Mx||2.6|
| 8|| +ve|| || +ve||Neph||RCC||pT3b, N0, M0||7.5|
| 9||equiv|| || +ve||Neph||RCC||pT2, N0, M0||6.5|
|10||equiv|| || +ve||Neph||oncocytoma|| || |
|11||equiv|| || +ve||NSS||leiomyoma|| || |
|12||equiv|| +ve|| +ve||None|| – || ||2|
|13||equiv|| +ve|| +ve||None|| – || ||2.5|
Table 2. The number of cases of RCC detected in different age groups
|Age group, years||RCC, n||Screening population, n|
Two additional cases of RCC were detected in patients with no evidence of tumour in the first phase. One participant moved away from the study location; a small (2.5 cm) centrally located renal tumour (T1N0M0) was detected incidentally in the course of abdominal US for other reasons 13 months after initial screening. In the second, both screening investigations were negative, but a third abdominal US 6 months later for other than urological symptoms revealed a renal mass of 3.8 cm in diameter (pT1N0M0). From these data, the sensitivity of US screening for detecting RCC as assessed at the 1-year of follow-up for 79% of the original cohort was 82%. The specificity was 98% in the first phase and 99% in the second phase.
In all, 481 patients (38% women and 62% men) were admitted to both hospitals with renal tumours during the same period. The screened patients comprised 2% of these, while 34% were detected secondary to symptoms and 64% incidentally; 78% of the incidental findings were made by US. A radical nephrectomy was performed in 358 (74%) and nephron-sparing surgery (NSS) in 124 (26%). RCC was found on histology in 415 (86%) patients. Table 3 shows the tumour stage of the 415 RCC and distinguishes between incidental and symptomatic tumours in comparison to the screened population. The mean (range) tumour size of RCC was 5.0 (1–15) cm in the incidental group, 6.2 (1.2–16) cm in the symptomatic group and 6.9 (2.5–13) cm in the screened group. The tumour size of the screened cases did not differ significantly from either of the other two groups.
Table 3. Tumour stages of all 415 surgical cases of RCC from 1997 to 1998 from both study centres
|Stage||Cases, n (%)|
|Screen detected||Incidental finding||Symptomatic|
|pT1N0M0||1||110 (40.7)|| 27 (20.0)|
|pT2N0M0||5||117 (43.3)|| 64 (47.4)|
|pT3N0M0||1|| 28 (10.4)|| 36 (26.7)|
|pT4N0M0||0|| 0|| 0|
|pT1–2N+M0||0|| 1 (0.4)|| 0|
|pT3–4N+M0||0|| 3 (1.1)|| 2 (1.5)|
|pTxNxM1||2|| 10 (3.7)|| 6 (4.4)|
|Total||9||271 (100)||135 (100)|
Thirteen (0.1%) results were classified as positive on initial screening US. Of the 12 patients eligible for further diagnosis, six were confirmed to have RCC. The PPV of a positive screening finding was thus 50%. In all, 175 (1.8%) results were classified as equivocal on initial screening US; a tumour was confirmed in seven of 171 patients eligible for further diagnosis. Confirmation was by histology in five (three RCC, two benign lesions) and by imaging alone in two inoperable cases. In these latter two, CT was highly suspicious for RCC, with tumours of 2 and 2.5 cm, but with no enlarged regional lymph nodes or evidence of metastases. The PPV of an equivocal screening finding was thus 4.1% for a solid renal tumour, excluding angiomyolipoma, and 1.8% for RCC. The PPV for both equivocal or positive screening results was 7.1% for a solid renal tumour and 4.9% for RCC. In the second screening phase, there were no positive findings; the findings were classified as equivocal in 64 of 7851 patients (0.8%), of whom none were confirmed to have a renal tumour by reference US, CT or MRI.
Positive or equivocal findings on screening US in the first phase were falsely positive on subsequent imaging studies in 93% of cases. Overall, 48% were deemed negative after reference US alone; 62% of those referred directly for reference US before further imaging were reclassified as negative. CT was used in 68 cases after 9959 underwent renal US in the first phase; 40 of these followed a reference US, while 28 were direct referrals from the screening physician. Of the latter 28, four had been classified as positive on US and 24 equivocal. The false-positive rate in the second phase was 100%; 32 CT scans were taken after 7851 renal US, most (23) with no reference US, although all were only classified as equivocal on screening US.
Other positive findings of less significance (renal pathology) were angiomyolipoma in nine, hydronephrosis in 13, kidney stones in 214, and a renal anomaly (small kidney, dysplasia, aplasia) in 40 (total 276). None required further therapy; all detected angiomyolipomas were too small to warrant surgery. Minor findings such as renal cysts, duplex systems and renal parenchymal scars were documented in 1264 (13%) cases in the first phase and in 1016 (13%) in the second; none of these required treatment.
- Top of page
- SUBJECTS AND METHODS
With the advances in the diagnostic capabilities of US and CT, the number of incidentally detected small RCCs has increased markedly [9–11]. Abdominal US is particularly important in this respect, not only because it is not invasive, of low cost and easy to apply, but also because of its widespread use. Many RCCs have been found during US for other diseases (liver, gall bladder, pancreas, etc.) or during annual health checks in which US is routinely used. Several abdominal US screening studies [3–6] have shown the prevalence of RCC to be substantially greater than other benign findings, but these studies were all retrospective investigations on routine hospital populations or healthy adults, and all included complete abdominal US. A prospective study to investigate the use of renal US alone for RCC screening has not been described previously.
The early detection of organ-confined RCC would improve prognosis and long-term survival . For small peripherally located tumours the option of NSS could be offered . A key question addressed in the current study was the willingness of physicians and patients to participate in screening; the participation rate was 68% for all physicians and 90% for urologists. Almost half (48%) of all screened patients participated exclusively in the new programme, indicating a high acceptance rate. Currently, ≈ 14% of the eligible male and 34% of the female population comply with the existing screening programmes offered in Germany . Compared with a DRE for detecting prostate and rectal cancer, and the cervical smear for detecting cervical cancer, renal US is less invasive and disturbing to the patient. This assumption is confirmed by a 79% return rate for the second screening phase of renal US.
In all, 13 renal tumours were detected, of which nine were RCC on final histology, giving a prevalence of 9/10 000 for the first phase of this screening study. Based on older German cancer registry data and abdominal US screening studies [3,8], we expected to find about three cases of RCC for every 10 000 participants in the given age group. Whether the difference of observed and expected prevalence can be explained by the limitations of the databases from which the primary estimates were derived, or by our small study cohort, by (self-) selection of the population, or by actual early detection by systematic screening as opposed to incidental or symptom-guided detection, cannot be decided with certainty. Further follow-up of the cohort with the detection of new cases (or lack thereof) will provide values on the incidence of RCC in the screened population. The current values, in contrast, are representative of the prevalence in the first screening phase. The trends were similar in abdominal screening studies with a comparable prevalence of RCC [6,15].
In contradiction to our expectations, the proportion of higher stage and larger tumours was greater in the screened than in the incidentally detected group. As a result, fewer tumours were amenable to NSS. Early detection by screening did not lead to a shift towards lower tumour stages, one of the key goals of screening. One explanation may again be the small sample size of the study cohort. Another explanation may be that detection at an early stage by screening requires screening at regular, tumour-specific intervals to depict the true incidence of newly developing tumours. These tumours will be detected at early stages if the intervals of screening are matched with the speed of tumour growth, and if tumours do not metastasize early at small volumes. For RCC, there is a positive correlation between tumour size and the rate of lymph node and distant metastases . However, in the first screening we were not evaluating tumour incidence but rather tumour prevalence. Even if detecting prevalent tumours by systematic screening is by definition ‘earlier’ than detecting tumours by symptoms in the same patients (as they had no symptoms at the time of screening), it does not necessarily mean that ‘earlier’ detection is also at an ‘earlier’ stage. PSA screening of prostate cancer resulted in detecting prostate cancer at earlier stages and in parallel a more frequent diagnosis of all stages, which led to considerable debate about the value of such a screening programme [17,18].
For a screening programme to be useful, it must lead to a net benefit in patient survival. Six cases with low stage (T1 and T2) tumours in the current cohort (Table 1) are likely to have benefited from screening by detecting an organ-confined tumour. Conversely, the two patients with metastases presumably have not benefited from the programme as there is no effective treatment for metastatic disease. Furthermore, the two patients with RCC on CT who were not surgical candidates because of poor overall health have suffered psychological distress secondary to the screening; they know of their potentially fatal disease but are untreatable.
The sensitivity of US for detecting renal tumours is reported to be high [3,7], but it depends on tumour location and size . The sensitivity in the current study was 82%, as judged at the 1-year follow-up, within which two cases of RCC were detected after previous negative US (interval cases). This is based on the assumption that these tumours were present but overlooked at the time of screening (false-negatives). Another possibility is that they were present but too small to detect at the time of screening. A third possibility is that these tumours truly developed after the last negative screening (true incidence). In the first case a 2.5-cm centrally located tumour almost iso-echogenic to normal renal parenchyma was detected 13 months after US screening. Although considered a false-negative on screening US, the time elapsed between US, the tumour size and the estimated speed of growth of this grade 1 tumour could justify its classification as a newly developed tumour . In contrast, the second case was most likely missed on US during the second phase of screening, if not also in the first. This peripheral tumour was 3.8 cm and detected 6 months after the second US.
Other pathological findings occurred at a low rate of 12% in both phases; this is much less than reported in other studies [3–6], which may be related to the use of complete abdominal US in those, as opposed to renal US only in the present study.
A further concern was the induction of a sequence of costly imaging studies in cases of equivocal findings. Based on the 17810 renal US results in both phases of screening, 100 CT scans were taken. Although it was not obligatory in the study protocol, reference US was efficient for further evaluating equivocal findings; half of equivocal lesions were deemed negative after reference US, and made further imaging studies superfluous. Our experience thus dictates that reference US should be used in these cases before more costly imaging studies.
Cost-effectiveness is a critical issue for evaluating any screening programme. Assessing the cost-effectiveness of US as a screening tool for RCC in the current study population is difficult, because there is no information about both the cost of treatment of metastatic RCC and the prevalence of incurable disease. As for the costs of the screening process alone, the low incidence of RCC compared with other screened cancers (e.g. breast) [21,22] raises doubts about the economic benefit of screening a population of ≥ 40 years old. However, the cost of further evaluations generated by the initial screening, when compared with mammography , are low if reference US is used routinely in cases with equivocal findings. The widespread availability and low cost (12 DM, or US $6 per investigation) of renal US in Germany favour its use as a potential screening tool for RCC. However, whether the screening benefit will justify the cost of screening for the whole population aged> 40 years can only be answered in a comparative prospective long-term screening study of a screened and unscreened population, although the cost factors differ greatly in other countries. In the USA, for example, renal US is reserved for the radiologist and is thus much less accessible. At the same time, the examination is many times more costly (US$≈ 295) and thus there less costly screening approaches for RCC, e.g. molecular tests, might be more promising.