SEARCH

SEARCH BY CITATION

Keywords:

  • prostate cancer;
  • screening;
  • PSA;
  • DRE

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. SUBJECTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGEMENTS
  8. CONFLICT OF INTEREST
  9. REFERENCES

OBJECTIVE

To describe the results of the first four rounds (T0-T3) of prostate cancer screening in the Prostate, Lung, Colorectal and Ovarian (PLCO) cancer screening trial (designed to determine the value of screening in the four cancers), that for prostate cancer is evaluating whether annual screening with prostate-specific antigen (PSA) and a digital rectal examination (DRE) reduces prostate cancer-specific mortality.

SUBJECTS AND METHODS

In all, 38 349 men aged 55–74 years were randomized to undergo annual screening with PSA (abnormal >4.0 ng/mL) and a DRE. The follow-up of abnormal screening results was at the discretion of subjects’ physicians. PLCO staff obtained records related to diagnostic follow-up of positive screen results.

RESULTS

Compliance with screening decreased slightly from 89% at baseline to 85% at T3. Both PSA positivity rates (range 7.7–8.8% at T0-T3) and DRE positivity rates (range 6.8–7.6% at T0-T3) were relatively constant over time. The positive predictive value (PPV) of a PSA level of >4.0 ng/mL decreased from 17.9% at T0 to 10.4–12.3% at T1-T3; the PPV for DRE (in the absence of a positive PSA test) was constant over time (2.9–3.6%). Cancer was diagnosed in 1902 men (4.9%). Screen-detected cancers at T0 (549) were more likely to be clinical stage III/IV (5.8%) and to have a Gleason score of 7–10 (34%) than screen-detected cancers at T1-T3 (1.5–4.2% stage III/IV and 24–27% Gleason score 7–10 among 1054 cases).

CONCLUSION

The present findings on serial prostate screening are similar to those reported from other multi-round screening studies. Determining the effect of PSA screening on prostate cancer mortality awaits further follow-up.


Abbreviations
PLCO

Prostate, Lung, Colorectal and Ovarian

ERSPC

European Randomized Study of Screening for Prostate Cancer

PPV

positive predictive value

INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. SUBJECTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGEMENTS
  8. CONFLICT OF INTEREST
  9. REFERENCES

The benefit of screening for prostate cancer with serum PSA is unknown. Despite concern about its efficacy, PSA screening has been widely accepted by both patients and physicians in the USA. Despite many reports on PSA screening, no studies have shown a clear reduction of prostate cancer-specific mortality from screening for prostate cancer [1–4]. Nonetheless, the American Cancer Society, the AUA and the National Comprehensive Cancer Network recommend annual PSA and DRE screening to begin at age 50 years in normal-risk men [5–7]. By contrast, the USA Preventive Services Task Force and the American College of Physicians-American Society of Internal Medicine do not recommend screening, based on the lack of a clear benefit [8,9].

To date, mortality data from only one randomized trial of PSA-based screening has been reported, but the results have been widely discounted due to methodological concerns [10]. Currently, there are two ongoing randomized, controlled trials of PSA-based prostate cancer screening designed to determine the effect of screening on prostate cancer mortality: the Prostate, Lung, Colorectal and Ovarian (PLCO) Cancer Screening Trial in the USA and the European Randomized Study of Screening for Prostate Cancer (ERSPC) in Europe [11–14].

The prostate component of the PLCO trial is designed to determine the effect of annual PSA and DRE screening on prostate cancer-specific mortality by comparing a screened arm with a control arm of men undergoing ‘usual’ medical care [12]. The results of the baseline round of screening were reported previously [15]. In this report we update these findings with results from the first three (of five) follow-up screening rounds for men randomized to the screening arm of the PLCO trial. These data on serial screening in a prostate cancer screening trial allow an examination of long-term compliance in a screening study, the performance of PSA and DRE as screening tests over time, and the characteristics of cancers discovered during four annual rounds of screening.

SUBJECTS AND METHODS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. SUBJECTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGEMENTS
  8. CONFLICT OF INTEREST
  9. REFERENCES

The PLCO cancer screening trial is a multicentre, randomized, two-arm trial designed to evaluate the effect of screening for prostate, lung, colorectal and ovarian cancer on disease-specific mortality. Randomization began in November 1993 and concluded in June 2001, with 154 934 men and women aged 55–74 years enrolled. The study design, methods and exclusion criteria were described previously [12,15]. All participants signed informed consent documents approved by both the National Cancer Institute and local institutional review boards.

Briefly, men in the intervention arm had a DRE and serum PSA determination. All PSA tests were assessed in a centralized laboratory [15]. Until 1 January 2004 PSA was assayed using the Tandem-R PSA assay (Beckman-Coulter, Fullerton, CA, USA); after that date the Access Hybritech PSA (Beckman-Coulter, Fullerton, CA, USA) assay was used (>99.5% of all PSA tests reported here were assayed by the Tandem-R). PSA was assessed at baseline and then annually for 5 years, while a DRE was done at baseline and then annually for 3 years. Exclusion criteria specifically related to prostate cancer screening included: previous prostate cancer or surgical removal of the prostate, and use of finasteride during the previous 6 months. Beginning in April 1995, the PLCO trial also excluded men reporting more than one PSA blood test in the previous 3 years.

A serum PSA level of >4.0 ng/mL was considered a positive test. DREs were performed by physicians, qualified nurses, or physician assistants. DREs were characterized as suspicious for cancer if there was nodularity or induration of the prostate, or if the examiner judged the prostate to be suspicious for cancer on the basis of other criteria, including asymmetry. Men with positive PSA or DRE results were notified and advised to seek further diagnostic evaluation through their primary-care provider, who was also notified of the test results. The follow-up diagnostic evaluation was left to the patients and their primary physicians; the PLCO trial did not specify a diagnostic algorithm, nor oversee the diagnostic process. PLCO trial staff obtained medical records related to diagnostic follow-up of positive screens, and medical record abstractors recorded information on relevant diagnostic procedures. Information on cancers diagnosed in the absence of a positive screen was obtained from annual study update forms sent to each participant. Certified tumour registrars ascertained the stage, Gleason score, and histology of all diagnosed prostate cancer cases. Reported Gleason scores are from biopsy specimens. Clinical stage was determined using the TNM staging system. The tumour stage was categorized according to the fifth edition of the American Joint Committee on Cancer’s Cancer Staging Manual [16]. Clinical information for nodal and metastatic staging was recorded when available.

For the purposes of this report, screen-detected cancers were defined as those diagnosed as a result of investigations initiated after a positive screening test, with no lapse in the diagnostic evaluation of >9 months. Non-screen-detected cancers in men who received previous PLCO screens were denoted as interval cancers. ‘Never screened’ cancers were defined as cancers diagnosed in men who did not receive any PLCO prostate cancer screening tests.

Compliance with screening was defined as the number of participants screened divided by the number expected for the test. The positive predictive value (PPV) of a test (or combination of tests) was defined as the proportion of those men positive on the test who had a screen-detected cancer diagnosed in the same screening round. Groups were compared using chi-squared tests with P < 0.05 considered to indicate statistical significance. Cumulative incidence curves were derived using the Kaplan-Meier method.

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. SUBJECTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGEMENTS
  8. CONFLICT OF INTEREST
  9. REFERENCES

In all, 38 349 men were randomized to the intervention arm of the PLCO trial; the demographics of this group are shown in Table 1. Briefly, the study population was predominantly non-Hispanic white (86%), and just over 60% were age <65 years at enrolment. Nearly 25% had a history of an enlarged or inflamed prostate or problems with the prostate, 7.1% had a family history of prostate cancer, 34.6% had one PSA test in the 3 years before study entry and 9.4% had more than one.

Table 1.  The baseline demographic characteristics of 38 349 men randomized to the intervention arm of the PLCO cancer screening trial
FactorFactor N (% of category)
Race/ethnicity 
 White, non-Hispanic33 044 (86.2)
 Black, non-Hispanic  1 713 (4.5)
 Hispanic   816 (2.1)
 Asian 1 534 (4.0)
 Pacific Islander or American Indian   322 (0.8)
 Missing   920 (2.4)
Age, years 
 55–5912 389 (32.3)
 60–6412 015 (31.3)
 65–69 8 879 (23.2)
 70–74 5 066 (13.2)
Education 
 Less than high school  3 101 (8.1)
 High school 11 347 (29.6)
 Some college 7 643 (19.9)
 College graduate 7 038 (18.4)
 Post-graduate 8 258 (21.5)
 Unknown   962 (2.5)
Previous prostate biopsy 
 No34 632 (90.3)
 Yes 1 650 (4.3)
 Unknown 2 067 (5.4)
History of benign prostate problems 
 No27 833 (72.7)
 Yes 9 525 (24.8)
 Unknown   941 (2.5)
Family history of prostate cancer 
 No33 780 (88.1)
 Yes 2 735 (7.1)
 Unknown 1 834 (4.8)
Had PSA test within 3 years of study entry 
 No17 272 (45.0)
 Once13 252 (34.6)
 More than once 3 588 (9.4)
 Unknown 4 237 (11.0)
Had DRE within 3 years of study entry 
 No15 362 (40.1)
 Once12 568 (32.8)
 More than once 8 528 (22.2)
 Unknown 1 891 (4.9)

Compliance was similar with the PSA test and DRE; the overwhelming majority of men undergoing screening with either test received both tests (Table 2). Compliance (with either test) decreased slightly over time, from 89.4% at baseline to 85.1% at T3.

Table 2.  Compliance with screening, screen results, biopsies, and prostate cancers detected by testing, for each study year
VariableStudy year
T0 (baseline)T1T2T3
Number eligible38 32037 49836 70335 898
% compliant with:    
 PSA    89.4    87.2    86.3    85.1
 DRE    89.1    86.5    85.7    84.3
 both    89.0    86.5    85.7    84.2
 either    89.4    87.2    86.4    85.1
Screen results    
Number tested (PSA or DRE)34 26232 69631 69730 544
Screen result, %    
PSA +ve (>4.0 ng/mL)     7.9     7.7     8.2     8.8
First +ve PSA    –     3.0     2.7     2.7
DRE positive     7.2     6.8     7.3     7.6
First DRE positive    –     4.5     4.0     3.7
Either test positive    14.0    13.5    14.4    15.1
Both tests positive     1.2     1.0    1.1     1.2
Biopsies and prostate cancers detected    
PSA > 4 ng/mL    
No. 2 718 2 502 2 593 2 676
% biopsied    40.2    32.8    31.2    30.1
No. of cancers   487   308   270   281
% cancers/biopsied    44.5    37.6    33.4    34.9
% cancers/positives, PPV (95% CI)    17.9 (16.4–19.4)    12.3 (11.0–13.6)    10.4 (9.2–11.6)    10.5 (9.3–11.7)
Cancers/1000 screened (yield)    14.2     9.4     8.5     9.3
DRE abnormal and PSA ≤ 4 ng/mL    
No. 2 083 1 923 1 973 1 943
% biopsied    18.9    18.6    15.8    14.5
No. of cancers    62    67    71    57
% cancers/biopsied    15.7    18.8    22.8    20.2
% cancers/positives, PPV (95% CI)     3.0 (2.3–3.7)     3.5 (2.7–4.3)     3.6 (2.8–4.4)     2.9 (2.1–3.7)
Cancers/1000 screened (yield)     1.8     2.1     2.3     1.9
PSA > 4 ng/mL or DRE abnormal (either test positive)    
No. 4 801 4 425 4 566 4 619
% biopsied    31.0    26.6    24.5    23.5
No. of cancers   549   375   341   338
% cancers/biopsied    36.9       31.9    30.5    31.1
% cancers/positives, PPV (95% CI)    11.4 (10.5–12.3)     8.5 (7.7–9.3)     7.5 (6.7–8.3)     7.3 (6.5–8.1)
Cancers/1000 screened (yield)    16.0    11.5    10.8    11.1

Table 2 shows the results of screening tests by study year. At baseline (T0), 7.9% of participants had a PSA level of >4.0 ng/mL, and this remained relatively stable through the first 3 years of follow-up. The percentage of men who had their first positive screening test (PSA or DRE) during each follow-up year is also shown; for PSA this was ≈3% each year. The rate of positive DREs was also stable over time, with rates over T0-T3 of 6.8–7.6%. Of the men, 14–15% had either a PSA or DRE positive at each screening round.

For men with a PSA level of >4 ng/mL, both the proportion receiving biopsy and the proportion of those biopsies that were positive were significantly lower during the subsequent rounds than at the baseline round (Table 2). Specifically, the proportion having a biopsy decreased from 40.2% at T0 to 30.1–32.8% at T1-T3, while the proportion of biopsies that were positive decreased from 44.5% at T0 to 33.4–37.6% at T1-T3. As a result, the PPV of a PSA level of >4 ng/mL was significantly lower at T1-T3 (10.4–12.3%) than at baseline (17.9%). The yield of the PSA test (cancers per 1000 men screened) was also significantly lower at T1-T3 (8.5–9.4%) than at T0 (14.2%).

Compared with men with a positive PSA test, men with only a positive DRE (and PSA level ≤4.0 ng/mL) had lower biopsy rates (14.5–18.9%), lower rates of biopsies being positive (15.7–22.8%) and accordingly, substantially lower PPVs (2.9–3.6%) and cancer yields (1.8–2.3 per 1000). Among the DRE (only)-positive men, there were generally smaller differences in the outcomes between the baseline and subsequent rounds than among the PSA-positive men, e.g. the PPV was 3.0%, 3.5%, 3.6% and 2.9% at T0-T3, respectively.

Among men with a positive test on either PSA or DRE, the PPV decreased from 11.4% at baseline to 7.3–8.5% at T1-T3. Similarly, the cancer yield decreased from 16.0 per 1000 men screened at T0 to 10.8–11.5 per 1000 at T1-T3. For men with a positive PSA and DRE screen, the PPVs were higher than for either test alone, i.e. 37.7% at T0 and 18–23% at T1-T3 (data not shown in Table 2).

The above results pertain to findings at each of the individual screening rounds. Figure 1 and Table 2 show longitudinal results over all rounds of screening. As seen in Fig. 1, the result of the baseline screen determined four distinct trajectories of cumulative prostate cancer incidence curves; after 4 years the cumulative incidences were 52%, 32%, 7.5% and 2.3% for men whose baseline screens were positive on both PSA and DRE, positive on PSA only, positive on DRE only, and positive on neither test, respectively. Of interest, among men with positive baseline screens, the cumulative incidence continued to increase steadily beyond the end of the baseline screening round, roughly doubling from year 1 to year 4 in the PSA and DRE (only)-positive groups and increasing by 50% in the group positive on both screens.

image

Figure 1. The cumulative percentage of men with confirmed prostate cancer, according to baseline PSA/DRE status, diagnosed within 4 years of the baseline screen. Numbers at risk at baseline are 398 (PSA and DRE positive), 2320 (PSA only positive), 2083 (DRE only positive) and 29 461 (neither positive). All pair-wise differences among curves were statistically significant (< 0.001).

Download figure to PowerPoint

Table 3 shows the outcome at the end of the 4-year study period for men with a positive PSA test at baseline and for men converting to a positive PSA test at either the T1 or T2 round. For the 2718 men with a positive PSA test (irrespective of DRE findings) at baseline, 34.7% had a cancer diagnosis through the 4 years of follow-up and 34.3% had a negative biopsy (and no subsequent cancer diagnosis). An additional 12.5% had PSA levels consistently lower (through the T3 screening round) than their initial PSA; thus in all, 82% of all men had some clinical resolution to their initial positive screen through the 4 years. For men first positive at the T1 and T2 screens, only 3 and 2 years of follow-up were available, respectively. Although the proportions with a cancer diagnosis decreased, to 24.0% (T1) and 18.0% (T2), the proportion with a resolution of their positive screen nonetheless remained at nearly 80%, due to an increasing proportion with lower subsequent PSA values.

Table 3.  The follow-up and clinical outcome of men with first positive PSA screens
VariableStudy year
T0 (baseline)T1T2
  • *

    On all subsequent PSA determinations through the T3 round;

  • †On any subsequent PSA determination through the T3 round.

Final status2718 (100)922 (100)796 (100)
Resolved   
 Cancer 944 (34.7)221 (24.0)143 (18.0)
 Negative biopsy 933 (34.3)251 (27.2)217 (27.3)
 Lower PSA* 340 (12.5)262 (28.4) 311 (39.1)
 Subtotal2217 (81.6)734 (79.6) 671 (84.3)
Unresolved   
 Higher PSA 433 (15.9)147 (15.9) 95 (11.9)
 No subsequent PSA  68 (2.5) 41 (4.4) 30 (3.8)
 Subtotal 501 (18.4)188 (20.4)125 (15.7)

Among the 38 349 men in the intervention arm, 1902 (4.9%) were diagnosed with prostate cancer through 4 years of follow-up. Among these cancers, 1603 (84.2%) were classified as screen-detected, 204 (10.7%) as interval, and 95 (5.0%) occurred in men who never had a PLCO screen.

Table 4 shows the distribution of Gleason score and clinical stage. Overall, 69% of all cancers had a Gleason score of 2–6, 21% had a score of 7 and 7% had a score of 8–10. The vast majority of cancers (96%) were clinically localized (stage I/II) tumours; <2% were locally advanced with no distant metastases (stage III) and 2.1% were locally invasive or metastatic (stage IV). Men with screen-detected cancer at baseline (T0) were significantly more likely to have Gleason 7–10 tumours and clinically advanced tumours than men with screen-detected cancers at T1-T3; among T0 screen-detected cancers, 34% were Gleason ≥7 and 5.8% were clinical stage III or IV, compared with 24.6–27.2% Gleason ≥7 and 1.5–4.2% clinical stage III or IV among T1-T3 screen detected cancers. The clinical stage and Gleason score distribution for the interval cancers generally resembled those of the T1-T3 screen-detected cancers.

Table 4.  Gleason score and clinical stage of prostate cancers by method of detection
CancersNGleason score, %Clinical stage, %
2–678–10UnknownI–IIIIIIV
  • *

    Clinical stage not available for 3 (1.5%) interval cancers.

Screen-detected (year)        
T0 54962.524.010.03.594.22.23.6
T1 37572.819.55.12.797.91.60.5
T2 34173.619.45.61.595.92.12.1
T3 33870.120.46.82.798.50.90.6
Interval* 20473.019.14.43.496.10.52.0
Never screened  9562.124.28.45.393.72.14.2
All190269.021.17.02.996.21.62.1

Among screen-detected cancers, the method of detection, whether through PSA, DRE, or both tests, was strongly associated with Gleason score and stage (data not shown). At baseline, 48.3% of cancers with both an abnormal PSA and DRE had a Gleason score of 7–10, whereas only 30.1% and 20.0% of cancers with an abnormal PSA only and abnormal DRE only, respectively, had a Gleason score of 7–10. During the follow-up (T1-T3) the pattern was similar, as 36.6% of men with both tests abnormal, 24.0% of men with only PSA abnormal and 18.5% of men with only DRE abnormal had Gleason scores of 7–10. Similarly, for clinical stage among T0 screen-detected cancers, 11.9% of men with an abnormal PSA and DRE were clinical stage III or IV, compared to 3.6% of men with only an abnormal PSA and 3.2% of men with only an abnormal DRE. During the follow-up rounds the trend was less clear, as 2.3%, 3.0% and 1.5% of cancers among men with both tests abnormal, PSA abnormal only, and DRE abnormal only, respectively, had clinical stage III or IV.

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. SUBJECTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGEMENTS
  8. CONFLICT OF INTEREST
  9. REFERENCES

This report on the first four rounds of screening in the PLCO trial follows the report published in 2005 on the results of the baseline screening round [15]. The additional data described here on three subsequent rounds of screening after baseline allow an examination of changes in compliance, screening results, diagnostic follow-up, PPV, cancer yield, and cancer characteristics from the baseline round to later screening rounds.

Compliance with screening, both for PSA and for DRE, remained high over the 4 years of screening, decreasing only minimally from the baseline to later rounds. Similarly, screen positivity rates on both PSA and DRE also showed little change over the four screening rounds. By contrast, the PPV of a positive screen, specifically of a positive PSA test, decreased substantially from the baseline round (17.9%) to the later rounds (10.5–12.4%). This decrease is probably a ‘saturation’ effect of PLCO screening on the population, whereby the initial screen identified many of the cancers, leaving relatively more men with benign PSA increases in subsequent screening rounds and thus leading to a lower PPV for PSA. Interestingly, the PPV for a positive DRE (only) was relatively stable over the screening rounds.

The PPVs of the PSA test and DRE reported here were influenced by the study protocol, which specified that decisions on the diagnostic follow-up of positive tests were to be made by the subject in consultation with his private physician, not by the PLCO trial. As such, diagnostic evaluation patterns observed in the PLCO might reflect community standards for each of the PLCO centres. Because of the known variability of PSA tests, the relative subjectivity of DRE and other factors, most men with abnormal screens either were not recommended for, or chose not to undergo, immediate biopsy. However, the large majority of men with an abnormal PSA, as shown in Table 3, had achieved a clinical resolution of their status at the end of the 4-year screening period, by virtue of undergoing a biopsy or showing a declining PSA level. However, the frequency of delayed biopsies resulted in many cancers being diagnosed more than a year after an initial positive screening test. As shown in Fig. 1, of all cancers diagnosed throughout the 4-year period after a positive baseline screen, only about half were diagnosed in the first year. At the baseline screening round, the 549 screen-diagnosed cancers gave a prevalence rate in this cohort of 1.4%. Now, after three additional screening rounds and 4 years of overall follow-up, the cumulative prevalence of all cancers, including those not detected by screening, was 4.9%.

In addition to the PPV, the other important change from baseline to subsequent screening rounds was in the characteristics of the detected cancers. Specifically, cancers detected at baseline were more likely to have Gleason scores of 7–10 (34%) and to be of advanced clinical stage (5.8%) than those detected at subsequent rounds, where overall 25.5% were Gleason 7–10 and 2.9% were of advanced clinical stage. Interestingly, the Gleason score and stage distribution for the cancers defined as ‘interval’ were similar to those of the screen-detected cancers diagnosed at later rounds, and actually more favourable than those of the baseline screen-detected cases; this is contrary to the general pattern in screening studies, whereby interval cases have a worse clinical profile than screen-detected cases. However, data collected by PLCO on the reasons for seeking the diagnostic evaluation that eventually led to a cancer diagnosis suggest that many of these interval cases were in fact not symptomatic but were evaluated due to an increasing PSA level or a PSA level considered by a clinician to be ‘elevated’ but still <4.0 ng/mL. This might help to explain the present findings in relation to the Gleason score and stage distribution of the interval cases. Note that the baseline results reported here differ slightly from PLCO data published previously for Gleason score and clinical stage [15]. In the previous report the pathological Gleason scores and staging were reported when available from radical prostatectomy specimens. This resulted in higher reported rates of Gleason 7–10 tumours and a higher rate of stage III tumours.

Several of the key findings presented here with serial screening, specifically the decrease in the PPV of PSA level from the baseline to subsequent screening rounds, the less favourable distribution of Gleason score and clinical stage at baseline than in subsequent rounds, and the cumulative diagnosis rate over 4 years, are generally consistent with findings reported at several sites of the ERSPC trial in Europe. Three ERSPC sites, Rotterdam, Sweden, and Finland, have now reported second or later screening round results, which allows for comparisons with the PLCO trial [17–19]. However, the ERSPC trial differed from the PLCO in that the second screening round was at either a 4-year (Rotterdam and Finland) or a 2-year interval (Sweden), instead of annually, as in the PLCO.

As with the PLCO trial, the PPV of a PSA level of >4.0 ng/mL decreased moderately and statistically significantly from the baseline round to the second or later rounds at all three of these ERSPC sites. In Rotterdam, the PPV (of a PSA level of >4 ng/mL) decreased from 28% at baseline to 17% at round 2, while in Finland it decreased from 26% at baseline to 18% at round 2 [17,18]. In Sweden the PPV decreased from 27% at baseline to 20% at round 2, 16% at round 3 and 12% at round 4 [19]. Although the trends are the same, the absolute PPV reported from these ERSPC sites was about 50% higher than in the PLCO trial. This reflects in part the higher biopsy rates after positive PSA tests at these sites (≈90%) than in the PLCO trial (≈40%). This might also be due to a much higher prevalence of PSA testing before trial entry in the PLCO trial. Nearly 45% of men in the PLCO trial had at least one PSA test before trial enrolment, while previous PSA testing in the ERSPC was rare. The PPV (at T0) for PLCO subjects with no previous PSA testing, at 24.4%, was much closer to the PPV for first-round screening in the ERSPC than was the overall T0 PPV.

The ERSPC sites also reported similar findings to the PLCO in terms of stage and Gleason score findings over time, with all showing a pattern of more favourable Gleason score and/or clinical stage at subsequent screening rounds than at baseline. In Rotterdam, 18.7% of 1014 baseline cases had advanced clinical stage (T3-4) and 35.5% had a Gleason score of 7–10, compared with rates of 3.7% for advanced clinical stage and 20.7% for Gleason 7–10 disease among 550 cases diagnosed at round 2 [17]. In Sweden, among 202 baseline cases, 8% had advanced clinical stage and 22% had Gleason 7–10 disease; the corresponding rates among the 348 cases diagnosed at rounds 2–4 (at 2, 4 and 6 years after baseline) were 1.7% advanced stage and 10.4% Gleason 7–10 [19]. Finally, at the Finnish site, the rate of advanced clinical stage was 13% among 105 baseline cases but only 6% among 79 cases detected at re-screening during round 2 [18]. However, the Gleason score distribution of these cases were similar, with 17% Gleason 7–10 at baseline and 21% at round 2.

In Sweden and Rotterdam, the shift toward more favourable Gleason score and clinical stage distributions in later rounds was accompanied by a decrease in the PSA levels of diagnosed cancers. For example, in Rotterdam, three-quarters of cases detected at baseline but only about half of cases detected at round 2 had a PSA level of >4.0 ng/mL [17]. Note that this PSA shift at these two sites was in part due to lowering of the threshold levels from baseline to later rounds; Rotterdam used a threshold of 4.0 ng/mL for half of the baseline screens and of 3.0 ng/mL for the other half of the baseline screens and all of the round 2 screens, while Sweden changed their threshold from 3.0 to 2.5 ng/mL for the third round of screening [17,19]. In the PLCO trial, while the PSA threshold remained fixed, the decrease in Gleason score and clinical stage was also accompanied by a decrease in PSA level, with the median PSA level of screen-diagnosed cases decreasing from 7.0 ng/mL at baseline to 5.3–5.8 ng/mL at T1-T3.

Finally, the cumulative diagnosis rate of 4.9% seen over 4 years and four rounds of screening in the PLCO was generally similar to the cumulative rates seen over a comparable period, but with fewer screening rounds, at the ERSPC sites. After two rounds of screening 4 years apart, Finland reported a cumulative rate of 4.6%, while in Rotterdam the cumulative rate was 7.9%[17,18]. In Sweden, over three rounds of screening 2 years apart, the cumulative rate was ≈5%[19].

In conclusion, the PLCO trial is the largest cancer screening trial ever conducted in the USA. Through four rounds of screening with PSA and DRE, the PLCO trial has continued to maintain excellent compliance with prostate screening. Compared with the baseline screening round, the PPV of an abnormal PSA level (but not of an abnormal DRE) decreased at subsequent screening rounds. In addition, the clinical stage and Gleason score were significantly more favourable during subsequent screening rounds than at baseline. The findings here over several screening rounds were similar to those observed with serial screening in the ERSPC trial. Subsequent comparison of prostate cancer mortality between the intervention and control arms will determine if there is a benefit to PSA and DRE prostate cancer screening.

ACKNOWLEDGEMENTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. SUBJECTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGEMENTS
  8. CONFLICT OF INTEREST
  9. REFERENCES

This work was supported by individual contracts from the National Cancer Institute to each of the 10 screening centres, the coordinating centre, and the clinical laboratory at UCLA.

CONFLICT OF INTEREST

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. SUBJECTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGEMENTS
  8. CONFLICT OF INTEREST
  9. REFERENCES

Gerald L. Andriole is a consultant/advisor to Aeterna Zentaris, Nema Steba and Onconome, an inverstor in Envisioneering and Viking Medical, and an investigator for Ferring Pharmaceutical, Glaxo Smith Kline, Veridex LLC and Zeneca. Source of funding: NCI.

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. SUBJECTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. ACKNOWLEDGEMENTS
  8. CONFLICT OF INTEREST
  9. REFERENCES
  • 1
    Gann PH, Hennekens CH, Stampfer MJ. A prospective evaluation of plasma prostate-specific antigen for detection of prostatic cancer. J Am Med Assoc 1995; 273: 28994
  • 2
    Smith DS, Catalona WJ, Herschman JD. Longitudinal screening for prostate cancer with prostate-specific antigen. J Am Med Assoc 1996; 276: 130915
  • 3
    Bartsch G, Horninger W, Klocker H et al. Prostate cancer mortality after introduction of prostate-specific antigen mass screening in the Federal State of Tyrol. Austria Urol 2001; 58: 41724
  • 4
    Efstathiou JA, Chen MH, Catalona WJ et al. Prostate-specific antigen-based serial screening may decrease prostate cancer-specific mortality. Urology 2006; 68: 3427
  • 5
    Smith RA, Cokkinides V, Eyre HJ. American Cancer Society guidelines for the early detection of cancer, 2004. CA Cancer J Clin 2004; 54: 4152
  • 6
    American Urological Association. Prostate-specific antigen (PSA) best practice policy. Oncology (Williston Park) 2000; 14: 778
  • 7
    Mohler J, Bahnson RR, Boston B et al. National Comprehensive Cancer Network. Practice Guidelines in Oncology, Prostate Cancer. 2007. Available at http://www.nccn.org/professionals/physician_gls/PDF/prostate.pdf
  • 8
    US Preventive Services Task Force. Screening for prostate cancer: U.S. Preventive Services Task Force recommendation statement. Ann Intern Med 2008; 143: 18591
  • 9
    American College of Physicians. Screening for prostate cancer. Ann Intern Med 1997; 126: 4804
  • 10
    Labrie F, Candas B, Dupont A et al. Screening decreases prostate cancer death. first analysis of the 1988 Quebec prospective randomized controlled trial. Prostate 1999; 38: 8391
  • 11
    Gohagan JK, Prorok PC, Kramer BS, Cornett JE. Prostate cancer screening in the prostate, lung, colorectal and ovarian cancer screening trial of the National Cancer Institute. J Urol 1994; 152: 19059
  • 12
    Prorok PC, Andriole GL, Bresalier RS et al. Design of the Prostate, Lung, Colorectal and Ovarian (PLCO) Cancer Screening Trial. Control Clin Trials 2000; 21: 273S309S
  • 13
    Schroder FH, Kranse R, Rietbergen J, Hoedemaeke R, Kirkels W. The European Randomized Study of Screening for Prostate Cancer (ERSPC): an update. Members of the ERSPC, Section Rotterdam. Eur Urol 1999; 35: 53943
  • 14
    Auvinen A, Rietbergen JB, Denis LJ, Schroder FH, Prorok PC. Prospective evaluation plan for randomised trials of prostate cancer screening. The International Prostate Cancer Screening Trial Evaluation Group. J Med Screen 1996; 3: 97104
  • 15
    Andriole GL, Levin DL, Crawford ED et al. Prostate Cancer Screening in the Prostate, Lung, Colorectal and Ovarian (PLCO) Cancer Screening Trial: findings from the initial screening round of a randomized trial. J Natl Cancer Inst 2005; 97: 4338
  • 16
    American Joint Committee on Cancer, American Cancer Society and American College of Surgeons. AJCC Cancer Staging Manual, 5th edn. Philadelphia: Lippincott-Raven, 1997
  • 17
    Postma R, Schroder FH, Van Leenders G et al. Cancer detection and cancer characteristics in the European Randomized Study of Screening for Prostate Cancer (ERSPC) – Section Rotterdam a comparison of two rounds of screening. Eur Urol 2007; 52: 8997
  • 18
    Makinen T, Tammela T, Stenman U et al. Second round results of the Finnish population-based prostate cancer screening trial. Clin Can Res 2004; 10: 22316
  • 19
    Hugosson J, Aus G, Lilja H, Lodding P, Pihl C. Results of a randomized, population –based study of biennial screening using serum prostate-specific antigen measurement to detect prostate carcinoma. Cancer 2004; 100: 1397405