SEARCH

SEARCH BY CITATION

Keywords:

  • screening;
  • PSA;
  • prostate;
  • cancer

Abstract

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

OBJECTIVE

To evaluate the clinical and pathological characteristics of screen vs non-screen-detected prostate cancers, to determine if there is a difference in the same prostate-specific antigen (PSA) range.

PATIENTS AND METHODS

In all, 997 patients who had had a radical prostatectomy were evaluated; 806 were Tyrolean screening volunteers, and 191 were from outside Tyrol, representing the ‘referred prostate cancer’ group. PSA level, age, prostate volume and pathological characteristics were assessed, as was the amount of over- and under-diagnosis.

RESULTS

There were no statistically significant differences in patient age or PSA levels in the two groups. Even in the same PSA range there were statistically significantly more extraprostatic cancers in the referral group, at 31.7% and 17.4%, respectively. In the referred and screening groups there was over-diagnosis in 7.9% and 16.8%, and under-diagnosis in 40.8% and 27.8%, respectively.

CONCLUSION

This study suggests that screening volunteers have a statistically significantly higher rate of organ-confined prostate cancers, and a statistically significantly lower rate of extracapsular extension and positive surgical margins than their counterparts in the referral group even in the same PSA range. As the pathological stage and surgical margin status are significant predictors of recurrence, these findings support the concept of PSA screening.


Abbreviations
RRP

radical retropubic prostatectomy

GS

Gleason score.

INTRODUCTION

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

In 2006, prostate cancer will affect an estimated 232 090 men in both the USA and Europe, and 30 350 are expected to die as a result of the disease [1]. This incidence is the highest among all non-cutaneous malignancies, and the mortality among men is now second only to lung cancer.

The widespread implementation of prostate cancer screening in the USA led to stage migration, with more cancers being detected at a lower stage. There has been a 75% reduction reported in the proportion of cases presenting with metastatic disease at diagnosis and a corresponding 32.5% decrease in the age-adjusted mortality rate from 1993 to 2003 [2,3]. Opponents of screening suggest that the observed decrease in prostate cancer mortality might be due to better treatments or to misclassification of the cause of death, and stress that some men will be treated unnecessarily, because they would not have been diagnosed with prostate cancer without screening.

The principal goal of an early detection programme for prostate cancer is to detect clinically relevant tumours while they are still curable. To examine whether current screening protocols might accomplish this objective, we compared patients whose cancer was diagnosed through a PSA-based screening study with those whose cancer was detected during routine clinical practice. Pending the results of randomized clinical trials, there is increasing evidence suggesting that screening leads to improved outcomes [4–6].

The purpose of the present study was to evaluate the clinical and pathological characteristics of screen vs non-screen-detected prostate cancers, to determine if there is a difference even in the same total PSA range.

PATIENTS AND METHODS

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

From February 1999 to March 2006, after obtaining informed consent, 997 patients with prostate cancer who had a radical retropubic prostatectomy (RRP) at our institution and were included in the study; 806 were Tyrolean screening volunteers, and 191 were from outside Tyrol, mainly from the rest of Austria and southern Germany, representing the referral group. Patients who had had neoadjuvant hormonal or other systemic therapy or radiotherapy were excluded.

Since 1993, the Tyrol screening/early detection programme has been underway in the Federal State of Tyrol, Austria. Tyrol is an alpine region in Western Austria with, at the 1991 census, 631 410 inhabitants (307 249 of them men) in an area of 12 647 km2. Tyrol is dominated by the mountains of the Central Alps, and the distances to Innsbruck, the capital, where the central healthcare unit is located, are not great (infrequently >100 km). Of the 307 249 male inhabitants in 1991, 86 067 were aged 45–75 years. PSA testing has been made freely available at any time; in 2005, 86% of the men were tested once during the study period. The screening demonstration project was conducted in collaboration with GPs, medical examiners, urologists, medical laboratories, and the Tyrol Blood Bank of the Red Cross. Informed consent was obtained from all volunteers participating in the programme. If the PSA level was elevated, the volunteers were invited to undergo additional urological evaluations at no cost, and the men with normal PSA levels were invited to have a repeat PSA test 6 or 12 months later. A DRE was not a part of the screening protocol, but men had a DRE at prostate biopsy. PSA testing was provided free of charge to men aged 45–75 years, and to younger men (40 years) with a family history of prostate cancer. In all laboratories, the PSA concentration was assessed using the IMX assay (Abbott Laboratories, Abbott Park, IL, USA).

For diagnosis, from 1996 to 2000, 10 systematic TRUS-guided prostate biopsy cores were taken in a standard spatial distribution; from 2000 to 2006, an additional five Doppler-enhanced targeted biopsies cores were taken on the basis of age-specific PSA reference ranges [7]. These reference ranges were defined as one half the age-referenced PSA level reported by Oesterling [8], combined with a percentage free PSA level of <18%. TRUS was done using a Combison 530MT unit (Kretztechnik, Zipf, Austria) fitted with a biplanar probe operating at a grey-scale frequency of 10 MHz and at the locations previously described [7]. Each biopsy core was reviewed by one pathologist and assessed as cancerous with an assigned Gleason score (GS). After a diagnosis of prostate cancer, RRP was done ≥2 months later using an open retropubic approach.

The pathological stage and GS were assessed; possible over-diagnosis was defined as a GS of <7, and a pathological stage of pT2a and negative surgical margins; under-diagnosis was defined as ≥pT3 or positive surgical margins [9]. The results were assessed using standard methods, with P < 0.05 considered to indicate statistical significance.

RESULTS

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

The characteristics of the two groups are shown in Table 1; there were no statistically significant differences between the two groups in patient age or PSA levels. The two groups differed in other clinical variables; although the percentage free PSA values were similar (13.3% and 12.5%), there were significant differences in prostate volume, with higher volumes in the referred group (33.9 vs 43.9 mL).

Table 1.  The characteristics of the two groups
VariableOverallScreeningReferral
  • *

    P ≤ 0.05.

No. of patients 806 191
Mean (range)
 PSA, ng/mL  6.7 (0.7–30)  6.2 (0.7–30)  7.1 (1.7–24)
 Prostate volume, mL38.9 (10–160) 33.9 (10–140) 43.9 (20–160)
 % free PSA12.9 (6–39) 13.3 (9–34) 12.5 (6–39)
 Age, years60 (45–75) 60.0 (45–75) 59.9 (45–75)
Pathological stage, %
 pT2a17.8 22.6 12.9*
 pT2b29.4 35.2 23.6*
 pT2c28.3 24.8 31.7*
 ≥pT324.6 17.4 31.7*
GS, %
 <612.9 14.4  11.5
 ≥632.5 39.2 25.7*
 ≥754.6 46.4 62.8*
Positive surgical margins, %18.1  11.7 24.4*
Mean (range) GS 6.6 (4–10)  6.4 (4–10)  6.8 (4–9)
Over- and under-diagnosis
Over-diagnosis, % (pT2a + GS <7 + −ve margins)  16.8  7.9*
Under-diagnosis, % (≥pT3a or +ve margins)  27.8 40.8*

Data on the distribution of pathological stage in the two groups are also shown in Table 1; the distribution was significantly different between the groups (P < 0.05). Even in the same PSA range there were statistically significant more extraprostatic cancers in the referral group. Overall, there were positive surgical margins in 18.1%, with a statistically significantly higher proportion in the referral group (11.7% vs 24.4%; P < 0.05, Table 1).

There were statistical significant differences in the distribution of GS (Table 1), with similar proportions of patients with GS < 6 (14.4% vs 11.5%) and 6 (39.2% and 25.7%) but statistically significantly proportions in the referred group with GS ≥ 7 (46.4% vs 62.8%; P < 0.05). The proportion of men in the referral group was significantly higher for under-diagnosis and lower for over-diagnosis (Table 1) than in the screened group (P < 0.05).

DISCUSSION

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

There is general agreement among clinicians that PSA testing has a higher predictive value for prostate cancer than a DRE and therefore PSA screening can detect early-stage cancers. However, screening for prostate cancer remains controversial because of the lack of randomized controlled trials showing a reduction in mortality in screened populations, although the observed trends in PSA-induced clinical and pathological stage migration and declining mortality where screening is common provide inferential evidence that screening is beneficial [6]. Since the implementation of widespread PSA-based screening, the risk of over-diagnosis has become a concern. The rate of over-diagnosis can be expressed in several ways. Etzioni et al.[10] calculated the amount of over-diagnosis as a percentage of the total detection by considering the probability of over-detection to be equal to the probability of dying from other causes during the lead time. McGregor et al.[11] defined over-diagnosis as the detection of non-lethal cancer.

The estimates of over-diagnosis of prostate cancer (30–50%) have been exaggerated previously [10,12], possibly partly because much of the data have been derived from older men, in whom over-diagnosis is a greater concern because of their limited life-expectancy. In younger men, who are most likely to benefit from early diagnosis and treatment, the criteria for judging over-diagnosis are much less frequently met [9]. In the present study population there was over-diagnosis in the screening and referral groups in 16.8% and 7.9%, respectively, with very young patients (mean ages 60 and 59 years).

Here, over-diagnosis was based on the pathological features GS and pT stage after RRP. Both the pT stage and the Gleason grade of prostate cancer are established important predictors of biochemical failure after RRP [13,14]. There is a statistically significant correlation between disease-free survival after RRP and the preoperative PSA level and final pathological GS in most series. Patients with a pathological GS of 6 have an excellent progression-free survival rate of up to 90%. However, men with a GS of 7 have a 29–43% risk of death from prostate cancer within 20 years [15].

To our knowledge, the perception of under-diagnosis has not been carefully evaluated in patients with prostate cancer. In a recent study Graif et al.[16] quantified the rates of over- and under-diagnosis of prostate cancer after RRP in two large groups of patients. Under-diagnosis was defined by extraprostatic tumour with of ≥pT3, or positive surgical margins. They showed that the proportion with cancer that was under-diagnosed was 25–30% and concluded that under-diagnosis of prostate cancer continues to be more frequent than over-diagnosis. In the present study with the same definitions, under-diagnosis was 18.6% and 42.2% in the screening and referral groups, respectively.

In a recent study in our department, we evaluated patients in the low-PSA group (2.0–3.9 ng/mL) who would not have been detected if the threshold for a prostate biopsy was a PSA level of >4.0 ng/mL [9]. In that study population, there was over-diagnosis in 19.7% and 16.5% in the low-PSA and at higher PSA values, respectively. However, in the higher PSA range, there were 2.5 times more patients with extracapsular disease (i.e. extracapsular disease or positive margins) with nearly the same proportion of over-diagnosis. In that study we also showed a correlation of so-called biologically insignificant tumours according to the criteria of Epstein et al.[17] and our definition of over-diagnosis. The real definition of over-diagnosis can only evaluated with randomized, controlled trials evaluating the effectiveness of PSA level and a DRE in reducing prostate cancer mortality. These studies are underway [18–20], but the results will not be available for several years. The present referral group had 31.7% of extraprostatic cancers after RRP, vs 17.4% in the screened group, with a significantly higher proportion of GS ≥ 7 tumours than in the screened group (62.8% vs 46.4%, P = 0.05). Although on balance the present results and previous reports support a role for early cancer diagnosis and treatment, we think that patients should be carefully selected for aggressive therapy.

There are several possible limitations of the present study. As it is known that the estimates of under-diagnosis and over-diagnosis are highly dependent on the definition used, and subject to debate, the present results are valid in a screening population with the present definitions and with no measurement of the tumour volume, that was not available in all patients. In particular, for the definition of over- and under-diagnosis we included margin status. Although it cannot be stated that a positive margin is with certainty a biological factor (or a possible sign of surgical failure and surgeon variables) it is important in particular in patients with pT3 disease [20], i.e. under-diagnosis in the present study population. Also, it is only possible to truly address the effect of lead-time bias in a randomized controlled trial; lead-time bias refers to the possibility that indolent cancers remain in the asymptomatic phase for longer, thus making it temporally more likely that they will be detected through screening than a rapidly progressive aggressive tumour. Although it is certainly true that the application of widespread screening has led to a stage migration, there is substantial evidence that the great majority of cancers thus detected have the features commonly associated with clinically significant tumours [21,22]. Nevertheless, evidence from randomized controlled trials is needed to confirm that our results of under-diagnosis, over-diagnosis and pathological characteristics are the result of screening. Another limitation of our study is the potential for selection bias. Because the groups were not randomized it is possible that the referred patients were sent by their physicians for reasons that differentiate them from the general group of community men who participated in the screening study. Furthermore, misclassification bias is also a possibility, because the rate of screening among the referred patients is unknown. Because of the wide availability of prostate-cancer screening in Austria, it is likely that some of the referred patients were screened with periodical PSA measurements and/or a DRE by their referring physicians. However, this limitation would underestimate any under- or over-diagnosis advantage, making our estimates of the difference between groups more conservative.

There are several implications of the present study; it is impossible to say with certainty in a young man that his tumour has been over-diagnosed. At present we cannot estimate tumour volumes or locate the tumour accurately before treatment with RRP, even with the help of CT, MRI or ultrasonography. However, very little is reported about under-diagnosis, and (at least using our criteria) under-diagnosis is still more common than over-diagnosis. To avoid under-diagnosis, early screening with a ‘low PSA level’ approach to identify aggressive cancer by rapidly increasing PSA levels (PSA velocity), and volume of cancer in biopsies or PSA density, should be used and new biomarkers should be found. Possible consequences of over-diagnosis can be minimized by high-quality treatment, e.g. preservation of continence and potency by nerve-sparing in locally confined tumours. Tumours with high mortality that occur in the low PSA range must be identified sufficiently early to allow the opportunity to provide individually curative treatment. Our data also show that existing clinical markers (pT stage, PSA level and GS) are useful but not sufficiently predictive to allow the discrimination of a clinically significant tumour from an indolent tumour. Prospective analyses of markers will require assessment not only of their performance characteristics, but also of their interaction with constitutional variables.

The present data highlight concerns about the prevalence of over-diagnosis of prostate cancer. Estimates of rates of over-diagnosis vary considerably (15–84%) in recent studies [10–12], depending on the definition of over-diagnosis used, and on factors like the pattern and method of screening, the average lead-time between detection and expected clinical presentation, and secular trends in cancer incidence.

However, in the present patients, 27.8% and 40.8% were under-diagnosed in the screening and referral groups, respectively. Our data show that the reported estimates of over-diagnosis in the low PSA group are exaggerated in the screening group, and that there is still a high proportion of under-diagnosis in the referral group. Furthermore, the proportion of under-diagnosis in the referral group was much higher than that of over-diagnosis in the screening group. As the body of evidence from studies of all types continues to grow, we think that the widespread use of routine screening will result in lower rates of cancer progression in men with clinically significant prostate cancer.

In conclusion, the present results suggest that screening volunteers have a statistically significantly higher rate of organ-confined prostate cancers, and a statistically significantly lower rate of extracapsular extension and positive surgical margins than their counterparts in a referral population, even in the same PSA range. As the pathological stage and surgical margin status are significant predictors of recurrence, these findings support the concept of PSA-based prostate cancer screening.

REFERENCES

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