Journal of the National Cancer Institute Advance Access originally published online on August 28, 2007
JNCI Journal of the National Cancer Institute 2007 99(17):1296-1303; doi:10.1093/jnci/djm101
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
© The Author 2007. Published by Oxford University Press.
ARTICLES |
Interval Cancers in Prostate Cancer Screening: Comparing 2- and 4-Year Screening Intervals in the European Randomized Study of Screening for Prostate Cancer, Gothenburg and Rotterdam
Affiliations of authors: Department of Urology, Erasmus Medical Centre, Rotterdam, The Netherlands (MJR, FHS); Department of Urology, Sahlgrenska University Hospital, Göteborg, Sweden (AG, JH)
Correspondence to: Monique J. Roobol, PhD, Department of Urology, Erasmus Medical Centre, Rm NH 224, PO Box 2040, 3000 CA Rotterdam, The Netherlands (e-mail: m.roobol{at}erasmusmc.nl).
 |
ABSTRACT |
|---|
 Top Abstract Context and CaveatsSubjects and MethodsResultsDiscussionFundingReferences |
|---|
Background: The incidence of prostate cancer has increased substantially since it became common practice to screen asymptomatic men for the disease. The European Randomized Study of Screening for Prostate Cancer (ERSPC) was initiated in 1993 to determine how prostate-specific antigen (PSA) screening affects prostate cancer mortality. Variations in the screening algorithm, such as the interval between screening rounds, likely influence the morbidity, mortality, and quality of life of the screened population.
Methods: We compared the number and characteristics of interval cancers, defined as those diagnosed during the screening interval but not detected by screening, in men in the screening arm of the ERSPC who were aged 55–65 years at the time of the first screening and were participating through two centers of the ERSPC: Gothenburg (2-year screening interval, n = 4202) and Rotterdam (4-year screening interval, n = 13301). All participants who were diagnosed with prostate cancer through December 31, 2005, but at most 10 years after the initial screening were ascertained by linkage with the national cancer registries. A potentially life-threatening (aggressive) interval cancer was defined as one with at least one of the following characteristics at diagnosis: stage M1 or N1, plasma PSA concentration greater than 20.0 ng/mL, or Gleason score greater than 7. We used Mantel Cox regression to assess differences between rates of interval cancers and aggressive interval cancers at the two centers. All statistical tests were two-sided.
Results: The 10-year cumulative incidence of all prostate cancers in Rotterdam versus Gothenburg was 1118 (8.41%) versus 552 (13.14%) (P<.001), the cumulative incidence of interval cancer was 57 (0.43%) versus 31 (0.74%) (P = .51), and the cumulative incidence of aggressive interval cancer was 15 (0.11%) versus 5 (0.12%) (P = .72).
Conclusion: The rate of interval cancer, especially aggressive interval cancer, was low in this study. The 2-year screening interval had higher detection rates than the 4-year interval but did not lead to lower rates of interval and aggressive interval prostate cancers.
Prior knowledge Prostate cancer incidence has risen substantially since screening men without symptoms of the disease has become common practice. Study design Prostate screening trial of men in Sweden (Gothenburg, 2-year interval) and in The Netherlands (Rotterdam, 4-year interval). Numbers of prostate cancers that were clinically diagnosed within the screening interval (interval cancers) in the two centers were compared. Contributions The cumulative incidence of prostate cancer for a period of 10 years was higher in Gothenburg than in Rotterdam, but the numbers of interval prostate cancers, including those that were potentially life threatening, were similar in the two centers. Implications Although more prostate cancers were detected overall in the center using the 2-year than the 4-year screening interval, the 2-year interval did not decrease the detection of interval prostate cancers. Limitations In Rotterdam, men were randomly assigned to the screening arm after consent, whereas in Gothenburg, men were assigned before consent. This difference in the randomization procedure may have led to selection bias in the Gothenburg center of the study and to a healthy screenee bias in the Rotterdam center of the study. In addition, response rate was slightly higher in Gothenburg than in Rotterdam. It is unknown whether these results can be generalized to other populations.
|
Since the introduction of serum prostate-specific antigen (PSA) testing in the late 1980s, prostate cancer incidence has risen substantially, mainly due to mass screening of clinically asymptomatic men (1). In 1984, 5.1% of all newly diagnosed prostate carcinomas were detected by PSA testing. By 1990 this percentage had already increased to 60.6% of prostate cancers that were diagnosed in the United States (2). Screening for prostate cancer remains a controversial issue. There is still a lack of scientific data that prove the effectiveness of PSA screening with respect to the prevention of deaths from prostate cancer and whether it will outweigh the loss of quality of life due to overdetection (3) and overtreatment and will justify the expenditure of a considerable part of the total budget of health care resources (4).
The European Randomized Study of Screening for Prostate Cancer (ERSPC) (5) was initiated with the intent to show or exclude an effect of screening on prostate cancer mortality. In addition, the ERSPC will allow a risk-to-benefit analysis, including parameters of quality of life and costs. Variations in the screening protocol, particularly in the interval between different screening rounds, are likely to influence the outcome of screening, whereas the length of the screening interval affects the detection rate, efficacy, and costs (4).
The rate of interval cancers (those clinically diagnosed within a screening interval) gives an indication of the sensitivity of the screening program and the appropriateness of the length of the screening interval. Different screening intervals are used within the ERSPC. In this study, we compared the number and characteristics of interval cancers in the Swedish center (Gothenburg, 2-year screening interval) and the Dutch center (Rotterdam, 4-year screening interval) of ERSPC. These two centers of ERSPC were chosen because the duration and completeness of available follow-up data are comparable.
|
Subjects and Methods |
|---|
|
TopAbstractContext and CaveatsSubjects and MethodsResultsDiscussionFundingReferences |
|---|
Characterization of the Study Population
The Dutch center (Rotterdam) of ERSPC started to randomly assign subjects on November 17, 1993, after a series of four pilot studies (6,7). A total of 42376 men aged 55–75 years were randomly assigned (21210 to the screening arm and 21166 to the control arm) through December 31, 1999. The Swedish center (Gothenburg) started inviting subjects for screening on January 2, 1995. A total of 9973 men were randomly assigned to the screening arm and 9973 to the control arm (aged 50–65 years). Legal requirements with respect to randomized trials were different in The Netherlands and Sweden. In The Netherlands, written informed consent was required before random assignment, whereas in Sweden, written informed consent was only necessary for those men who were in fact randomly assigned to the screening arm of the trial. The randomization at both centers was computer based. The screening algorithms used in both centers have been extensively described previously (8,9). All participants provided written informed consent, and the studies were approved by review boards in each country (Sweden: Göteborg University; The Netherlands: Erasmus Medical Center and a national board). The ERSPC trial is registered at the International Standard Randomized Controlled Trial Number Register under number SRCTN49127736 and date assigned December 20, 2005.
In general, according to the ERSPC, a prostate biopsy was indicated for men with a PSA level of 3.0 ng/mL or higher. However, for some men who were screened in Rotterdam at the beginning of the study, a plasma PSA concentration of 4.0 ng/mL or higher and/or an abnormal digital rectal exam (DRE) and/or transrectal ultrasound examination (TRUS) was required to indicate a biopsy. In Gothenburg, men were invited for subsequent biennial screening until age 70 years, and in Rotterdam the upper age limit for rescreening was 75 years. Both centers used the lateralized sextant prostate biopsies technique, as described by Eskew et al. (10). To achieve a similar age distribution between the two centers, only men who were 55–65 years of age at the time of first screening were included, and only data from those who responded to the first invitation have been analyzed. All participants who were diagnosed with prostate cancer through December 31, 2005, but not more than 10 years after initial screening were assessed by linkage with regional and national cancer registries.
Stage and Grade Classification
All cancers were classified according to the primary tumor size–lymph node status–distant metastasis classification of 1992 (11). Grading of the cancers was done using the Gleason grading system (12). This grading system has been used within the screening study since the second half of 1994 and is currently used in almost all pathologic laboratories. The pathologists who performed the assessment were blinded to study arm and to interval versus screen-detected cancers.
Definition of Interval Cancers
An interval cancer was defined as any prostate cancer that was diagnosed outside the screening study protocol within a screening interval. Interval cancers were identified by linkage to national cancer registries. The characteristics at the time of diagnosis were assessed using data from patient medical charts. Men with T1c prostate cancer and plasma PSA concentration less than 10.0 ng/mL were classified as M0, even though a bone scan was not performed.
A potentially life-threatening interval cancer or aggressive interval cancer was defined as an interval cancer that had at least one of the following characteristics at diagnosis: stage M1 or N1, plasma PSA concentration greater than 20.0 ng/mL, or a Gleason score greater than 7. Cancers with these characteristics are often incurable. Because these cancers could be considered as cancers that might have been detected with more favorable characteristics with a shorter screening interval, we choose to analyze this group separately. Prostate cancers that were diagnosed outside the screening study after a longer period than the screening interval (due to patient nonparticipation or old age) were not classified as interval cancers.
Statistical Analysis
A possible difference between the two centers in the rate of all prostate cancers, interval cancers, and aggressive interval cancers up to 10 years after initial screening was tested by Mantel Cox regression analysis (log-rank test in Kaplan–Meier analysis).
The rate of interval cancers was compared with the rate of prostate cancers detected in the control arm at each center. Follow-up time was measured from date of random assignment until 1) date of prostate cancer diagnosis; 2) date of death; or 3) date of last follow-up, which was defined as the date of last cross-matching with the cancer registry (and no finding of registration of prostate cancer). In addition, the clinical stage, Gleason score, and plasma PSA concentration of the interval cancers and screen-detected cancers at the time of diagnosis were compared between the two centers. Percentages of T1c screen-detected prostate cancers were compared after adjustment for the difference in biopsy indication at the beginning of the study in The Netherlands.
All statistical tests were two-sided. P values less than .05 were considered to be statistically significant.
|
Results |
|---|
|
TopAbstractContext and CaveatsSubjects and MethodsResultsDiscussionFundingReferences |
|---|
Cancer Detection Rates
After age selection, i.e., of men who were aged 55–65 years at the time of first screening, 13301 men from the screening arm in Rotterdam were included. During the study period (up to December 31, 2005), all of these men had two screening visits and some had three. A total of 1061 prostate cancers were detected at these screening visits (i.e., without the interval cancers) (cancer detection rate = 7.98%). Mean follow-up time was 7.16 years. In Gothenburg, 4202 men from the screening arm were included in the study. These men all had five screenings, and some had six. A total of 521 prostate cancers were detected, without the interval cancers (cancer detection rate = 12.40%). Details are indicated in the trial flow diagram (Fig. 1). Mean follow-up time was 7.38 years. During the same period, 317 men were diagnosed with prostate cancer in the control group (N = 13309) in Rotterdam (detection rate = 2.38%), whereas 402 prostate cancers were found in the control arm in Gothenburg (N = 5951; detection rate = 6.76%).
|
Identification of Interval Cancers
During follow-up, 57 patients in Rotterdam were diagnosed with interval cancer, of whom 15 were diagnosed with aggressive cancer; the corresponding detection rates were 0.43% and 0.11%, respectively. In Gothenburg, 31 patients were diagnosed with interval cancer, of whom five were diagnosed with aggressive cancer, resulting in detection rates of 0.74% and 0.12%, respectively. Only five patients with interval cancers in Rotterdam and none in Gothenburg had a plasma PSA concentration less than 1.0 ng/mL at their first screening visit. The ratios of interval cancer to all other prostate cancers that were detected in the screening and control groups (i.e., the rate of interval cancer divided by the rate of screen-detected prostate cancer and prostate cancer detected in the control arm, respectively) were calculated (Table 1). The differences in the rates of interval cancers, which were higher in the center with the shorter interval (0.74% in Gothenburg, using a 2-year interval, versus 0.43% in Rotterdam, using a 4-year interval), diminished when correcting for the rate of prostate cancers detected in the men randomly assigned to the control arm.
|
Interval Cancers Per Screening Interval
The number of aggressive interval cases, classified as such on the basis of chart review and the characteristics described above, were categorized per year after initial screening (Table 2). During the first screening interval in Rotterdam (i.e., 4 years), eight interval cancers were detected, among which five were considered as aggressive. The number of interval cancers as well as aggressive interval cancers increased over time in Rotterdam. In Gothenburg, eight interval cancers were detected, among which two were aggressive, that surfaced clinically during the first 2 years after initial screening. In Gothenburg, there was also an increase over time in the number of interval cancers but not in the number of aggressive interval cancers; four of five aggressive interval cancers were detected within the first 4 years of the study.
|
The clinical stage, Gleason score, and plasma PSA concentration of the interval cancers and screen-detected cancers at time of diagnosis were compared between the two centers (Table 3). In Rotterdam, five of the 57 interval cancers were stage T3–4, whereas the 31 interval cancers in Gothenburg included no stage T3–4 cancers. Gleason scores in the interval cancer and screen-detected cancers were similarly distributed in both centers. Gleason scores 6 or lower were the most predominant, representing approximately 70% of the interval cancers and 75% of the screen-detected cancers. The percentage of T1c cancers in the group of screen-detected cancers was considerably higher in Gothenburg (74.5% versus 50.5% in Rotterdam). Even after adjustment for those prostate cancers found in Rotterdam men by means of abnormal DRE and/or TRUS and men with a plasma PSA concentration less than 3.0 ng/mL, the percentage of T1c cancers in Rotterdam remained lower (536/[1061 – 136] = 58%).
|
The 10-year cumulative incidence of all prostate cancers found in Rotterdam versus Gothenburg was 1118 (8.40%) versus 552 (13.14%) (P<.001), the cumulative incidence of interval cancer was 57 (0.43%) versus 31 (0.74%) (P = .51), and the cumulative incidence of aggressive interval cancer was 15 (0.11%) versus 5 (0.12%) (P = .72). We estimated percentages of men free of all prostate cancers detected per time, men free of an interval prostate cancer per time, and men free of an aggressive prostate cancer per time using Kaplan–Meier analyses (Fig. 2). A statistically significant difference in cumulative incidence between the two centers was seen only when comparing the numbers of all prostate cancers detected (screen-detected plus interval prostate cancers). Cumulative incidences of both interval cancers and aggressive interval cancers separately showed no statistically significant differences when using a 2- or 4-year interval.
|
|
Discussion |
|---|
|
TopAbstractContext and CaveatsSubjects and MethodsResultsDiscussionFundingReferences |
|---|
This study reports the number and characteristics of interval cancers in two cohorts of men screened for prostate cancer—one at 2-year intervals and the other at 4-year intervals. The 10-year cumulative incidence of all prostate cancers in Rotterdam (4-year interval) versus Gothenburg (2-year interval) was 1118 (8.41%) versus 552 (13.14%) (P<.001), the cumulative incidence of interval cancer was 57 (0.43%) versus 31 (0.74%) (P = .51), and the cumulative incidence of aggressive interval cancer was 15 (0.11%) versus 5 (0.12%) (P = .72).
Both cohorts were part of the randomized populations of the screening studies. Screened men, nonparticipants, and control subjects were closely followed by regular cross-matching with national cancer registries, and data of men with prostate cancer have been regularly traced from patients' records, laboratory reports, etc. This design generates high-quality data. Furthermore, it is standard procedure at both centers for study databases to be checked regularly for data entry errors. All centers of the ERSPC are regularly reviewed by the ERSPC Quality Control Committee (13).
The study has some potential limitations. The differences in the randomization procedure at the two centers may lead to uncertainties because selection bias in noncompliant men in Gothenburg cannot be excluded (14). How this bias compares with the healthy screenee bias in the Rotterdam protocol is impossible to elucidate (15). Both biases include the possibility that men who were ill at the time of random assignment might have been suffering from unrecognized prostate cancer. In Rotterdam, where consent was obtained before random assignment, the response rate among men in the age category 55–65 years was approximately 55%, whereas in Gothenburg the response rate after random assignment in the same age category was slightly higher, 64%. Because these response rates are similar, the net result of these two different biases may be small. However, the combination of selection and healthy screenee bias remains and cannot be quantified. In addition, this study applies to Dutch and Swedish populations, and it may therefore not be generalizable to other populations and ethnic groups.
Our data show that the 10-year cumulative incidence of interval cancers was 0.43% with a 4-year screening interval and 0.74% with a 2-year screening interval, which correspond to 18% and 11% of the detection rate observed in the control groups (0.43/2.4 in Rotterdam and 0.74/6.8 in Gothenburg). These more comparable percentages (18% and 11%) suggest that the large difference in the 10-year cumulative incidence of interval cancers can be explained at least partly by the differences in background incidence in the two centers. The detection rates of interval cancers in this study are low compared with those in breast cancer screening, in which the ratio of interval cancer and expected cancer detection rate has been reported to be approximately 50% for a 2-year interval (16).
However, there is a large difference between the potential aggressiveness of interval breast cancers and interval prostate cancers. The term "interval cancer" usually refers to cancers with more aggressive features than those described by the stage distribution observed in this study (17). Many of the interval cancers detected in this study were of low stage and grade and thus locally confined and potentially curable—probably a result of opportunistic screening because a considerable proportion of the interval cancers were staged as T1c, and previous studies have shown that opportunistic screening is present (18,19). Aggressive prostate cancers with features matching the highly aggressive interval cancers that are commonly detected in breast cancer trials were uncommon, and their rate of detection was not different between the centers with 2- and 4-year screening intervals. Moreover, the rate of detecting aggressive interval cancer was only 2% and 5% of that in the control groups. The absolute numbers of aggressive interval cancer with 2- and 4-year intervals were also small, 5 and 15, respectively, indicating that PSA screening is usually effective in diagnosing prostate cancer at an earlier stage than these aggressive interval cancers, even with a 4-year interval. This observation is also supported by earlier reports showing that PSA screening results in stage migration and reduces the number of men with metastatic disease (14,20,21).
The observation that the incidence of interval cancers was higher in the center with the shorter screening interval seems counterintuitive. This relatively high rate of interval cancers in Gothenburg is due in large part to the detection of eight cancers within the first 2 years after random assignment. One explanation of this relatively high rate of interval cancers in Gothenburg might be that some men became aware of their elevated PSA concentrations and, despite negative biopsies, sought medical attention and were rebiopsied before the next screening visit 2 years later. Because only sextant biopsies were performed, some cancers might have been missed at the first screening and therefore surfaced as interval cancers. Although this retesting was also present in Rotterdam, it is known that in Rotterdam, opportunistic PSA testing in men randomly assigned to the screening arm rarely led to a prostate biopsy. Of 1982 men who had a PSA test, 62 men (3.1%) were actually biopsied (19).
Opportunistic screening in both countries was uncommon during the mid-1990s but has since increased, especially in Sweden. The percentage of men in both the control and screening group who had a PSA test in 1995 was 3%, whereas in 2005 it increased to approximately 25% (22). In Rotterdam, the percentages for the years 2001 and 2005 were 14.4% and 19.4%, respectively (23). In The Netherlands, this PSA testing in men randomly assigned to the control arm resulted in 10% effective contamination [i.e., a prostate biopsy (19)]. This opportunistic screening is probably the explanation for the relatively high incidence of interval cancer in Gothenburg. Many men in both locations now have annual PSA measurements at regular health checkups. Although the clinical stage distributions did not immediately show such an effect (the percentage of T1c interval cancers is somewhat higher in Rotterdam), correction for the number of prostate cancers detected in the control arm resulted in more comparable interval cancer/control group ratios in the two centers (Table 1).
Another possible explanation for the relatively high incidence of interval cancers in the Swedish center is the difference in biopsy indication between the two centers at the start of the trial. However, the number of additional cancers found indicated by an abnormal DRE and/or TRUS in men with low PSA levels is limited. The positive predictive value of an abnormal DRE and/or TRUS at a PSA level of 3.0 ng/mL and less or 4.0 ng/mL is 7.3% and 9.7%, respectively (24). Furthermore, similar numbers of interval cancers are found up to 4 years after applying a screening algorithm with and without a DRE- and/or TRUS-driven biopsy indication at low PSA levels (25). It is therefore unlikely that the difference in biopsy indication that existed only during a short period at the initial screening round had an effect on the number of interval cancers in Rotterdam.
In Rotterdam, the number of interval cancers increased after the first interval. This increase was shown to be due to missing a visit or refusing biopsy when indicated by the PSA level at repeat screening (eight patients). In the years after the third screening (i.e., 8 years after initial screening), 10 interval cancers surfaced among men who had refused biopsy at the third screening visit. Therefore, these interval cancers are not a result of a too long screening interval but more a result of noncompliance to the screening algorithm used.
The appropriateness of the length of the screening interval can be evaluated not only by the rate of interval cancers but also by the tumor characteristics of cancers detected at repeat screening. In Rotterdam, 15 of 433 (3.5%) of the cancers detected at repeat screening were greater than stage T2, and two patients (0.46%) had distant metastases (26). In Gothenburg, these numbers were two of 111 cancers detected (1.8%) and one M+ cancer in 111 cancers detected (0.9%), respectively (20).
Although numbers of interval cancers with stage greater than T2 were higher in Rotterdam (4-year interval) than in Gothenburg (2-year interval), they were low in an absolute sense, and the question from the clinical and public health points of view is whether a shorter screening interval is warranted. An additional screening visit of the 13301 men in Rotterdam would have resulted in more than 10500 additional PSA tests after 2 years (the loss in number of men available for repeat screening 2 years later is due to prostate cancer detected at initial screening, interval prostate cancers, and refusal to take a next screening round). Approximately 8000 additional PSA tests would have been done within the second 4-year interval (i.e., 6 years after the initial screening and again men would be lost for screening due to the reasons mentioned above). Using a PSA cutoff of 3.0 ng/mL as biopsy indication would result in approximately 3700 (0.20 x 18500) additional biopsy procedures. The two additional screening visits (2 and 6 years after the initial screening) could result in avoiding four interval cancers within the first screening interval (2–4 years after initial screening, see Table 2) and 23 interval cancers within the second 4-year interval (6–8 years after initial screening, see Table 2). Furthermore, an additional screening visit after 2 years would have advanced the diagnosis by 2 years of the 20 stage T3 cancers that were detected at repeat screening 4 years after the initial screening. In total, a possible gain of detecting 47 potentially curable prostate cancers at the cost of 18500 PSA tests and 3700 additional biopsy procedures would result.
Several other studies (27–29) have recently suggested that the screening interval should be related to the initial PSA level. Aus et al. (30), using data from the Swedish ERSPC center, found that after a median follow-up of 7.6 years and biennial screening PSA measurements, the cancer detection rate in men aged 50–66 years with an initial plasma PSA concentration of less than 1.0 ng/mL was 0.9%. Studies done in the Dutch ERSPC center showed that only 0.9% of men aged 55–74 years with an initial PSA value of less than 1.0 ng/mL progressed to having a PSA value of 3.0 ng/mL or higher (31). Roobol et al. (32), using data from three consecutive screening rounds in the Dutch ERSPC center, concluded that the screening interval of men aged 55–65 years with a plasma PSA concentration of 1.0 ng/mL could be as long as 8 years with a minimal risk of missing aggressive prostate cancer at a curable stage.
Our data on interval cancers seem to confirm these previous observations. Of the men with an interval cancer within 10 years after initial screening, only four of those in Rotterdam and none in Gothenburg had plasma PSA concentrations of 1.0 ng/mL or less at their first screening. However, recent data from Rotterdam also suggest that PSA progression to 3.0 ng/mL or greater occurs most frequently in men who initially have plasma PSA concentrations between 2.0 and 3.0 ng/mL (31). Improvements in the specificity of testing in this PSA concentration range are therefore desirable.
Even if a shorter screening interval might result in a more favorable stage distribution of detected cancers, it is obvious that the cumulative incidence of prostate cancers will increase with the number of screens performed. This increase would, however, result in many unnecessary repeated biopsies in the same men unless some form of risk stratification is applied. In this study, the cumulative incidence of screen-detected prostate cancer was greater than 50% higher in Gothenburg than in Rotterdam (12.4% versus 7.98%), in spite of similar age distribution and follow-up period. This difference may be the result of differences in the background incidence of prostate cancer in the two countries. In 2003, incidence rates in Sweden were 1.7 times higher than those in The Netherlands (22,23). World standard rates (WSR, cases per 100000 men) in Sweden and The Netherlands were 104.4 and 61.4, respectively. However, these differences were considerably less in the 1990s, the period during which the first three of the five screening visits took place in Sweden [WSR for prostate cancer in Sweden and The Netherlands in 1995 were 62.7 and 56.3, respectively (22,23)]. It is therefore more realistic to believe that each screening will again lead to prostate cancer diagnoses among some men from a large pool harboring small and often clinically insignificant disease and thus that more cancers were detected in Sweden because screening was more frequent.
The percentage of overdiagnosis in the Rotterdam center of ERSPC is described by Draisma et al. (33) and is estimated to be on the order of 27% for a single screening test, whereas for a screening program with a 4-year screening interval from age 55 to 67, the overdetection rate was 48% (range = 44%–55%). A comparative study using the Micro Simulation Screening Analysis model for the Swedish and Dutch data is ongoing. Steyerberg et al. (34), using a nomogram for the prediction of indolent prostate cancer that is adapted to a screening setting, estimated that 23% of screen-detected prostate cancers can be characterized as indolent. (A higher percentage of indolent cancers was found at repeat screening 4 years later [44% versus 23%; P<.001]). These percentages will probably not be lower with a shorter screening interval but rather are likely to be higher.
One should also consider that the 10-year cumulative incidence of 12.4% in Gothenburg was achieved with a 2-year screening interval. In most countries where PSA testing is advocated, annual testing is recommended. Provided that PSA testing is ultimately shown to reduce prostate cancer mortality, the future goal should be to evaluate and individualize the screening interval and the age when screening should be offered to obtain an optimal balance between a favorable stage distribution and risk of overdiagnosis. With the present results, it does not seem justified to recommend annual PSA testing except in men at high risk of prostate cancer, who may be identifiable at secondary screening using recently developed algorithms (35).
|
Funding |
|---|
|
TopAbstractContext and CaveatsSubjects and MethodsResultsDiscussionFundingReferences |
|---|
Dutch Cancer Society (KWF 94-869, 98-1657, 2002-277, 2006-3518 to F. H. S.); The Netherlands Organization for Health Research and Development (002822820, 22000106, 50-50110-98-311 to F. H. S.); 6th Framework Program of the EU: P-Mark (LSHC-CT-2004-503011 to F. H. S.); The Swedish Cancer Society (3792-B96-01AXB to J. H.); Wallac Oy; Hybritech Inc; Schering Plough, Sweden; Abbott pharmaceuticals, Sweden.
|
REFERENCES |
|---|
|
TopAbstractContext and CaveatsSubjects and MethodsResultsDiscussionFundingReferences |
|---|
(1) Jemal A, Murray T, Ward E, Samuels A, Tiwari RC, Ghafoor A. Cancer statistics, 2005. CA Cancer J Clin (2005) 55:10–30.
(2) Jones GW, Mettlin C, Murphy GP, Guinan P, Herr HW, Hussey DH, et al. Patterns of care for carcinoma of the prostate gland: results of a national survey of 1984 and 1990. J Am Coll Surg (1995) 180:545–54.[Web of Science][Medline]
(3) Draisma G, Boer R, Otto SJ, van der Cruijsen-Koeter IW, Damhuis RA, Schröder FH, et al. Lead times and overdetection due to prostate-specific antigen screening: estimates from the European Randomized Study of Screening for Prostate Cancer. J Natl Cancer Inst. 95:868–78.
(4) Sennfalt K, Sandblom G, Carlsson P, Varenhorst E. Costs and effects of prostate cancer screening in Sweden—a 15-year follow-up of a randomized trial. Scand J Urol Nephrol (2004) 38:291–8.[CrossRef][Web of Science][Medline]
(5) Roobol MJ, Schröder FH. European Randomized Study of Screening for Prostate Cancer: achievements and presentation. BJU Int (2003) 92(Suppl 2):117–22.[CrossRef][Web of Science][Medline]
(6) Schröder FH, Damhuis RA, Kirkels WJ, De Koning HJ, Kranse R, Nijs HG, et al. European Randomized Study of Screening for Prostate Cancer—the Rotterdam pilot studies. Int J Cancer (1996) 65:145–51.[CrossRef][Web of Science][Medline]
(7) Schröder FH, Roobol MJ, Damhuis RA, de Koning HJ, Blijenberg BG, van der Kwast TH, et al. Rotterdam randomized pilot studies of screening for prostate cancer—an overview after 10 years. J Natl Cancer Inst (2005) 97:696.
(8) Roobol MJ, Kirkels WJ, Schröder FH. Features and preliminary results of the Dutch centre of the ERSPC (Rotterdam, The Netherlands). BJU Int (2003) 92(Suppl 2):48–54.
(9) Hugosson J, Aus G, Lilja H, Lodding P, Pihl CG. Results of a randomized, population-based study of biennial screening using serum prostate-specific antigen measurement to detect prostate carcinoma. Cancer (2004) 100:1397–405.[CrossRef][Web of Science][Medline]
(10) Eskew LA, Bare RL, McCullough DL. Systematic 5 region prostate biopsy is superior to sextant method for diagnosing carcinoma of the prostate. J Urol (1997) 157:199–202.[CrossRef][Web of Science][Medline]
(11) Schröder FH, Hermanek P, Denis L, Fair WR, Gospodarowicz MK, Pavone-Macaluso M. The TNM classification of prostate cancer. Prostate Suppl (1992) 4:129–38.[Medline]
(12) Gleason DF, Mellinger GT. Prediction of prognosis for prostatic adenocarcinoma by combined histological grading and clinical staging. J Urol (1974) 111:58–64.[Web of Science][Medline]
(13) Ciatto S. The ERSPC Quality Control Committee. BJU Int (2003) 92(Suppl 2):57–61.[CrossRef][Web of Science][Medline]
(14) Aus G, Bergdahl S, Lodding P, Lilja H, Hugosson J. Prostate cancer screening decreases the absolute risk of being diagnosed with advanced prostate cancer-results from a prospective, population-based randomized controlled trial. Eur Urol (2007) 51:659–64.[CrossRef][Web of Science][Medline]
(15) Otto SJ, Schröder FH, de Koning HJ. Low all-cause mortality in the volunteer-based Rotterdam section of the European Randomised Study of Screening for Prostate Cancer: self-selection bias? J Med Screen (2004) 11:89–92.[CrossRef][Web of Science][Medline]
(16) Wang H, Bjurstam N, Bjorndal H, Braaten A, Eriksen L, Skaane P, et al. Interval cancers in the Norwegian breast cancer screening program: frequency, characteristics and use of HRT. Int J Cancer (2001) 94:594–8.[CrossRef][Web of Science][Medline]
(17) Fraucheboud J, de Koning HJ, Beemsterboer PM, Boer R, Verbeek AL, Hendriks JH, et al. Interval cancers in the Dutch breast cancer screening programme. Br J Cancer (1999) 81:912–7.[CrossRef][Web of Science][Medline]
(18) Beemsterboer PM, de Koning HJ, Kranse R, Trienekens PH, van der Maas PJ, Schröder FH. Prostate specific antigen testing and digital rectal examination before and during a randomized trial of screening for prostate cancer: European Randomized Study of Screening for Prostate Cancer, Rotterdam. J Urol (2000) 164:1216–20.[CrossRef][Web of Science][Medline]
(19) Otto SJ, van der Cruijsen-Koeter IW, Liem MK, Korfage IJ, Lous JJ, Schröder FH, et al. Effective PSA contamination in the Rotterdam section of the European Randomized Study of Screening for Prostate Cancer. Int J Cancer (2003) 105:394–9.[CrossRef][Web of Science][Medline]
(20) Hugosson J, Aus G, Lilja H, Lodding P, Pihl CG, Pileblad E. Prostate specific antigen based biennial screening is sufficient to detect almost all prostate cancers while still curable. J Urol (2003) 169:1720–3.[CrossRef][Web of Science][Medline]
(21) van der Cruijsen-Koeter IW, Vis AN, Roobol MJ, Wildhagen MF, de Koning HJ, van der Kwast TH, et al. Comparison of screen detected and clinically diagnosed prostate cancer in the European randomized study of screening for prostate cancer, section Rotterdam. J Urol (2005) 174:121–5.[CrossRef][Web of Science][Medline]
(22) Regionalt Onkologiskt Centrum i Uppsala/Örebroregionen. Available at: www.roc.se. [Last accessed: May 30, 2007.].
(23) Centraal Bureau voor de Statistiek. Available at: www.cbs.nl. [Last accessed: May 30, 2007.].
(24) Vis AN, Hoedemaeker RF, Roobol M, van der Kwast TH, Schröder FH. Tumor characteristics in screening for prostate cancer with and without rectal examination as an initial screening test at low PSA (0.0–3.9 ng/ml). Prostate (2001) 47:252–61.[CrossRef][Web of Science][Medline]
(25) Gosselaar C, Roobol MJ, Roemeling S, de Vries SH, van der Cruijsen-Koeter IW, van der Kwast TH, et al. Screening for prostate cancer without digital rectal examination and transrectal ultrasound: results after four years in the European Randomized Study of Screening for Prostate Cancer (ERSPC), Rotterdam. Prostate (2006) 66:625–31.[CrossRef][Web of Science][Medline]
(26) van der Cruijsen-Koeter IW, Roobol MJ, Wildhagen MF, van der Kwast TH, Kirkels WJ, Schröder FH. Tumor characteristics and prognostic factors in two subsequent screening rounds with 4-year interval within prostate cancer screening trial, ERSPC, Rotterdam. Urology (2006) 68:615–20.[CrossRef][Web of Science][Medline]
(27) Carter HB, Epstein JI, Chan DW, Fozard JL, Pearson JD. Recommended prostate-specific antigen testing intervals for the detection of curable prostate cancer. JAMA (1997) 277:1456–60.
(28) Ito K, Yamamoto T, Ohi M, Takechi H, Kurokawa K, Suzuki K, et al. Natural history of PSA increase with and without prostate cancer. Urology (2003) 62:64–9.[CrossRef][Web of Science][Medline]
(29) Ito K, Yamamoto T, Ohi M, Takechi H, Kurokawa K, Suzuki K, et al. Possibility of rescreening intervals of more than one year in men with PSA levels of 4.0 ng/mL or less. Prostate (2003) 57:8–13.[CrossRef][Web of Science][Medline]
(30) Aus G, Damber J-E, Khatami A, Lilja H, Stranne J, Hugosson J. Individualized screening interval for prostate cancer based on prostate-specific antigen level. Arch Intern Med (2005) 165:1857–61.
(31) Schröder FH, Raaijmakers R, Postma R, van der Kwast TH, Roobol MJ. 4-year prostate specific antigen progression and diagnosis of prostate cancer in the European Randomized Study of Screening for Prostate Cancer, Rotterdam. J Urol (2005) 174:489–94.[CrossRef][Web of Science][Medline]
(32) Roobol MJ, Roobol DW, Schröder FH. Is additional testing necessary in men with prostate-specific antigen levels of 1.0 ng/ml or less in a population based screening setting? (ERSPC, section Rotterdam). Urology (2005) 65:343–6.[CrossRef][Web of Science][Medline]
(33) Draisma G, Boer R, Otto SJ, van der Cruijsen-Koeter IW, Damhuis RA, Schröder FH, et al. Lead times and over detection due to prostate-specific antigen screening: estimates from the European Randomized Study of Screening for Prostate Cancer. J Natl Cancer Inst (2003) 95:868–78.
(34) Steyerberg EW, Roobol MJ, Kattan MW, van der Kwast TH, de Koning HJ, Schröder FH. Prediction of indolent prostate cancer: validation and updating of a prognostic nomogram. J Urol (2007) 177:107–12. discussion 112.[CrossRef][Web of Science][Medline]
(35) Roobol MJ, Schröder FH, Kranse R, ERSPC Rotterdam. A comparison of first and repeat (four years later) prostate cancer screening in a randomized cohort of a symptomatic men aged 55–75 years using a biopsy indication of 3.0 ng/ml (results of ERSPC, Rotterdam). Prostate (2006) 66:604–12.[CrossRef][Web of Science][Medline]
Manuscript received March 20, 2007; revised June 21, 2007; accepted July 6, 2007.
Editorial about this Article
Related Articles in JNCI
CiteULike 
Connotea 
Del.icio.us What's this?
J Natl Cancer Inst 2007 99: 1279-1280.
J Natl Cancer Inst 2007 99: 1277.
J Natl Cancer Inst 2007 99: 1277.
J Natl Cancer Inst 2007 99: 1277.
J Natl Cancer Inst 2007 99: 1277.
This article has been cited by other articles:
|
|
 |
|
  F. H. Schroder, J. Hugosson, M. J. Roobol, T. L.J. Tammela, S. Ciatto, V. Nelen, M. Kwiatkowski, M. Lujan, H. Lilja, M. Zappa, et al. Screening and Prostate-Cancer Mortality in a Randomized European Study N. Engl. J. Med., March 26, 2009; 360(13): 1320 - 1328. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 U.S. Preventive Services Task Force Screening for Prostate Cancer: U.S. Preventive Services Task Force Recommendation Statement Ann Intern Med, August 5, 2008; 149(3): 185 - 191. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. J. Roobol and F. H. Schroder Re: Is a Screening Interval of Every 4 Years for Prostate Cancer Acceptable? J Natl Cancer Inst, February 6, 2008; 100(3): 222 - 222. [Full Text] [PDF]
|
|
|
|
 |
|
 Is Less-Intensive Prostate Cancer Screening Appropriate? Journal Watch Oncology and Hematology, October 2, 2007; 2007(1002): 2 - 2. [Full Text]
|
|
|
|
 |
|
 What's an Appropriate Interval for PSA Screening? Journal Watch (General), September 27, 2007; 2007(927): 4 - 4. [Full Text]
|
|
|
|
|
|
E. D. Crawford Is a Screening Interval of Every 4 Years for Prostate Cancer Acceptable? J Natl Cancer Inst, September 5, 2007; 99(17): 1279 - 1280. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||