Advertisement

Changes in mortality and incidence of prostate cancer by risk class in different periods in Italy: the possible effects of PSA spread

Abstract

Purpose

In Italy, the spread of prostate-specific antigen (PSA) testing varies in different areas. A peak of incidence was reached in 2003-2004 in some areas, while in others the incidence is still increasing. Mortality has declined since 1999 in some areas, while it remains stable in others. We compared mortality and the risk of advanced cancer over 2 periods (1996-1998; 2005-2007) and by geographic area characterized by a different spread of PSA, to understand the possible impact of PSA on the epidemiology of prostate cancer.

Methods

In 8 Italian Cancer Registries (CRs), 4,632 cases diagnosed over 2 periods, 1996-1998 and 2005-2007, were sampled to assess risk class. The CRs were classified into late and early phase of PSA testing depending on whether an incidence peak had been reached by 2008. Incidence by risk class was estimated based on overall incidence in each CR and on risk class distribution in the sample. We calculated standardized mortality (MRR) and risk class-specific incidence rate ratios (IRR) to compare the 2 periods.

Results

Incidence increased from 1996-1998 to 2005-2007 (IRR 1.5; 95% CI 1.4, 1.6). High-risk and metastatic cancer incidence decreased only in late-phase areas (IRR 0.78; 95% CI 0.69, 0.88; and 0.40; 95% CI 0.30, 0.54, respectively), while in early-phase areas, incidence remained virtually stable (IRR 1.2; 95% CI 1.0, 1.4; and 0.77; 95% CI 0.59, 1.0, respectively). Mortality decreased only in late-phase areas (MRR 0.81; 95% CI 0.85, 0.97; vs 1.1; 95% CI 0.92, 1.2) in early-phase areas.

Conclusions

Mortality reduction and a decrease in high-risk and metastatic cases occurred simultaneously only in areas in late phase of PSA spread.

Tumori 2017; 103(3): 292 - 298

Article Type: ORIGINAL RESEARCH ARTICLE

DOI:10.5301/tj.5000613

Authors

Massimo Vicentini, Claudio Sacchettini, Annalisa Trama, Nicola Nicolai, Gemma Gatta, Laura Botta, Riccardo Valdagni, Paolo Giorgi Rossi, Prostate Cancer High-Resolution Study Working Group

Article History

Disclosures

Financial support: Supported by the Associazione Italiana per la Ricerca sul Cancro (grant 9293, IG 2009) and by AMGEN Dompé. The sponsors had no role in the study design; in the collection, analysis, and interpretation of data; in writing the report; or in the decision to submit the article for publication.
Conflict of interest: None of the authors has conflict of interest with this submission.

This article is available as full text PDF.

Download any of the following attachments:

Introduction

An estimated 1.1 million men worldwide were diagnosed with prostate cancer in 2012 and 307,000 died, accounting for 15% of the cancers diagnosed in men and 6.6% of cancer deaths. Prostate cancer incidence varies more than 25-fold worldwide; the rates are highest in Australia/New Zealand and North America (age-standardized rate 111.6 and 97.2 per 100,000, respectively) and in Western and Northern Europe (1). Less variation is observed in mortality rates worldwide (10-fold from approximately 3 to 30 per 100,000) than in incidence.

In Italy, in 2014, approximately 36,000 new cases and 7,400 prostate cancer deaths were estimated. While mortality has decreased steadily (about -1.8% per year) over the last 20 years, incidence increased rapidly from the 1990s to 2004 and then decreased slightly. Despite large differences in incidence, there are no substantial differences in mortality between the different areas of the country, with levels of about 17-20 deaths per 100,000 inhabitants/year (2).

More than other cancers, prostate cancer incidence must be interpreted within the context of diagnostic intensity and screening behavior. In fact, the prevalence of latent prostate cancer is high: 15% to 30% of men over 50 will have a prostatic adenocarcinoma that is thought to have low potential for growth and metastasis (3).

Screening for prostate-specific antigen (PSA) has dramatically affected incidence. Prostate-specific antigen testing was introduced in 1986 to monitor disease progression. Widespread opportunistic screening for PSA has since been introduced in several Western countries, thus enabling the detection of a high proportion of latent lesions. Consequently, incidence rates in some countries reflect the sum of clinical disease and latent disease, but in other countries reflect only clinical disease (4). The lower variability in mortality compared to incidence is consistent with the fact that PSA testing has a much greater effect on incidence than on mortality (5).

Opportunistic screening with the PSA test was introduced in Italy (albeit with geographic differences) around 1990, and is now widespread in Italy with a coverage of more than 40% of men aged between 50 and 70; coverage is also high in older age groups, where the well-known effect on overdiagnosis is not counterbalanced by any effect on mortality (5-6-7). From 1990 to 2005, prostate cancer incidence rates more than doubled (43,000 new cases in 2005) (8, 9).

This rapid increase, followed by a slow decrease to levels that are still higher than in the pre-PSA era, have been interpreted as the consequences of the different phases of the spread of PSA screening: initially a so-called prevalence round, when the test is mainly given to people who have never been tested before, with incidence continuously increasing due to detection because of the finding of prevalent asymptomatic cancers; and a second phase in which the PSA test is performed on men who have been tested already and in whom mostly incident cases are detected. This shape of curve has been observed in several industrialized countries (10) and is consistent with data from experimental studies, even if in practice it occurs over a longer timeframe.

We compared estimated advanced prostate cancer incidence in the general population between the 2 periods 1996-1998 and 2005-2007 in Cancer Registries (CRs) during their early phase and in those in their late phase of PSA testing spread, in order to understand how changes in risk classes at diagnosis influenced mortality rates.

Methods

Eight Italian CRs that participated in the Prostate Cancer High-Resolution Study are included in the analyses: Alto Adige, Naples, Latina, Genoa, Ragusa, Reggio Emilia, Veneto, and Varese.

Classification of the areas according to prostate cancer incidence and mortality

Data from the Italian CR database were retrieved using the ITACAN queryable application to calculate standardized rates of incidence and mortality. Supplementary Figure 1 (Available online at www.tumorijournal.com) shows incidence and mortality trends (calculated with 2 join point and standardized rates) by year and CRs using the period from 1996 to the most recent available year for each registry (see Supplementary Fig. 1). We used these data only to classify CRs in the early and late phase of PSA testing. Specifically, CRs were grouped into CRs in their early phase of PSA testing, if the incidence constantly increased up to 2008, and CRs in their late phase of PSA testing, if a peak and a consequent decrease in incidence occurred before 2008.

The Prostate Cancer High-Resolution Study Protocol

The methods of the Prostate Cancer High-Resolution Study are described in detail elsewhere (11). Briefly, each participating CR was required to provide 600 primary prostate cancer cases, age 18-99 years, diagnosed in 2 periods: 1996-1998 and 2005-2007. Malignant, both histologically and not histologically verified, primary and second primary prostate cancer tumors were included (defined by the ICD-O3 topography code C61.9); sarcomas (ICD-O3 morphology codes 8800, 9120, 9220) were excluded. The second primary cancers were included only if the first tumor was not in the prostate. Prostate cancer cases diagnosed at autopsy or identified only by death certificate were excluded.

The cases were randomly sampled, 300 for each period in each CR, at central level from the EUROCARE (www.eurocare.it) database. Thus, 4,632 cases were analyzed. Most CRs collected data for 600 cases, except Naples CR, which reviewed 414 because incident cases in the period 1996-1998 were too few to reach the required number. Contrariwise, Ragusa CR decided to participate with 660 cases.

Stage at diagnosis was recorded according to the 6th edition of the TNM manual.

Data relating to the Gleason score/pattern were also collected. Risk groups were defined as published by Trama et al (11), as follows:

Low risk: Gleason score ≤6 and PSA <10 ng/mL and T1-T2

Intermediate risk: Gleason score 7, T1-T2, and/or PSA ≤20 ng/mL or Gleason score ≤6, T1-T2, and PSA 10-20 ng/mL

High risk: PSA >20 ng/mL and/or Gleason ≥8 and/or T3-T4 or N1

Systemic disease: Cases with metastasis at diagnosis (any T, any N, M+) were kept separate (referred to here as metastatic [M1])

Statistical analysis

Standardized incidence and mortality rates were calculated by period (1996-1998, 2005-2007) and compared by the incidence rate ratio (IRR) and mortality rate ratio (MRR), with their relative 95% confidence interval (95% CI). Comparisons were stratified by the phase of PSA testing, late or early.

We estimated the number of high-risk cases in the population applying the proportion obtained by the Prostate Cancer High-Resolution Study sample (11) by age class, period, and CR. The standardized incidence rates were calculated for 2 different thresholds of risk: 1) high-risk or more severe (i.e., high-risk class and metastatic tumors), thereafter called high-risk + M1; 2) metastatic only, thereafter called M1. These 2 thresholds were chosen because the first includes the vast majority of fatal cancers (about 90% of deaths occurring in the first 5 years after diagnosis) and the second is composed of mostly fatal cancer (more than 80% dead 5 years after diagnosis) (11). Furthermore, we looked at unknown risk classes in order to exclude bias in the comparison between the 2 periods due to an improvement in cancer registration accuracy. We compared high risk + M1 and M1 cancer incidence rates between the 2 periods computing IRR, overall and stratified by age class and spread of PSA testing phase. To estimate IRR 95% CI for high risk + M1 and M1 cancers, we took into account the uncertainty due to the sample size of the Prostate Cancer High-Resolution Study (11); we used Poisson distribution. Analyses were repeated by the following age subgroups: 0-69, 70-79, and 80+. Results by age group are reported in Supplementary Table S1, a and b (Available online at www.tumorijournal.com).

Results

In order to obtain proportion of risk class by period and CR, we employed 2,174 cases and 2,458 cases in 1996-1998 and 2005-2007, respectively, from the 8 CRs participating in the Prostate Cancer High-Resolution Study (11).

The proportion of high-risk cancers ranged from 20% to 42% in the first period and from 21% to 36% in the second period; metastatic cancers ranged from 10% to 24% and from 3% to 11% in the 2 periods, respectively (Tab. I).

Sampled case by registry, risk group, presence of metastases, and period

Period Cancer registry Low risk Intermediate risk High risk Metastatic Unknown Overall no.
No. % No. % No. % No. % No. %
↑ = early phase of prostate-specific antigen testing; ↓ = late phase of prostate-specific antigen testing.
1996-1998 ↓Alto Adige 60 21 57 20 84 29 37 13 54 18 292
↓Genoa 48 16 67 22 107 36 36 12 42 14 300
↓Latina 20 7 40 14 73 26 48 17 105 37 286
↑Naples 10 9 13 11 23 20 27 24 41 36 114
↑Ragusa 38 13 39 14 118 42 35 12 54 19 284
↓Reggio Emilia 34 11 69 23 69 23 57 19 69 23 298
↑Trento 51 17 39 13 117 39 72 24 21 7 300
↓Varese 86 29 70 23 94 31 30 10 20 7 300
Total 347 16 394 18 685 32 342 16 406 19 2,174
2005-2007 ↓Alto Adige 127 43 84 28 61 20 16 5 10 3 298
↓Genoa 123 41 96 32 64 21 10 3 7 2 300
↓Latina 68 24 77 27 83 29 27 9 30 11 285
↑Naples 71 24 67 22 107 36 33 11 22 7 300
↑Ragusa 109 29 86 23 98 26 29 8 54 14 376
↓Reggio Emilia 114 38 78 26 63 21 10 3 34 11 299
↑Trento 94 31 76 25 99 33 30 10 1 0 300
↓Varese 127 42 81 27 62 21 10 3 20 7 300
Total 833 34 645 26 637 26 165 7 178 7 2,458

From 1996 to 2008, prostate cancer incidence increased in all registries. In the last period, a plateau or decreasing phase can be identified in some CRs (Alto Adige, Genoa, Reggio Emilia, Varese, Latina). These registries were classified as in the late phase of PSA testing. All these registries have a very high peak incidence, i.e., about 100/100,000 or more, with the exception of Latina, which has an incidence peak of 70/100,000 (Supplementary Fig. 1). Mortality shows trends with greater stability over the period, with a slight decrease in all 5 registries in the late phase of PSA testing. On the contrary, the other registries (Naples, Trento, Ragusa), which are classified as early phase, showed a lower incidence, but still steeply rising. Mortality in these CRs was almost stable, except for a slight increase in Naples (Tab. II).

Overall incidence and mortality (per 100,000 person-years) in the 2 periods by registry

Cancer registry 1996-1998 2005-2007
Incidence 95% CI Mortality 95% CI Incidence 95% CI Mortality 95% CI
Data from Cancer Registries, total population.
↑ = early phase of PSA testing; ↓ = late phase of PSA testing; CI = confidence interval; PSA = prostate-specific antigen.
↓Alto Adige 93.3 23 118.4 19.2
↓Genoa 76.3 21 97.8 17.2
↓Latina 41 19.6 68.4 19.7
↑Naples 21.7 14.5 58.9 18.7
↑Ragusa 40.8 16 62.7 18.8
↓Reggio Emilia 57.4 21 101.6 16.2
↑Trento 56 23.2 86.7 21.2
↓Varese 84.9 26.8 124.7 22.3
Total 65.2 63.4, 67.0 21.9 20.8, 23 95.5 93.4, 97.6 19.7 18.8, 20.6
Late phase of PSA testing 72.8 75.1, 79.9 103.3 106.6, 112.1
Early phase of PSA testing 41.8 39.1, 44.0 69.22 65.9, 72.0

We grouped the late phase of PSA testing registries, i.e., Alto Adige, Genoa, Varese, Reggio Emilia, and Latina, and the early phase of PSA testing registries, Naples, Ragusa, and Trento, in the following analyses according to the shape of the incidence curve (Tab. III).

Comparing the 1996-1998 period with the 2005-2007 period, the IRR was 1.46 (95% CI 1.38, 1.55) for all CRs and higher for the early-phase CRs, where no decrease in recent years was observed (IRR 1.66; 95% CI 1.51, 1.83). The MRR of the 2005-2007 vs 1996-1998 period in the pool of registries was 0.90 (95% CI 0.86, 0.94) and 0.85 (95% CI 0.84, 0.96) in late-phase CRs and 1.09 (95% CI 0.94, 1.25) in early-phase CRs (Tab. IV).

High-risk and metastatic cancer incidence (per 100,000 person-years) in the 2 periods by registry

Cancer registry 1996-1998 95% CI 2005-2007 95% CI
Incidence of the specific group has been estimated applying the age-specific proportion of high-risk and metastatic cancer observed in the study sample to the overall incidence by age class (see Methods for a detailed description of the computing).
↑ = early phase of PSA testing; ↓ = late phase of PSA testing; CI = confidence interval; PSA = prostate-specific antigen.
a) Standardized incidence rate (×100,000) for high risk class + M1 (synthetic indicator)
 ↓Alto Adige 38.2 33.7, 42.8 30.5 26.8, 34.2
 ↓Genoa 36.5 33.9, 39.0 23.9 21.8, 26.0
 ↓Latina 17.3 14.3, 20.2 25.5 22.4, 28.7
 ↑Naples 9.7 7.1, 12.3 27.4 23.6, 31.2
 ↑Ragusa 21.6 17.7, 25.5 20 16.5, 23.5
 ↓Reggio Emilia 24.3 21.1, 27.4 22.6 19.7, 25.4
 ↑Trento 35.2 31.1, 39.3 36.2 31.5, 40.9
 ↓Varese 35.2 32.1, 38.4 31 28.4, 33.6
 Total 30 28.8, 31.2 26.8 25.7, 27.8
 Late phase of PSA testing 32 32.5, 35.7 26.4 25.3, 27.9
 Early phase of PSA testing 24 20.2, 23.8 28.4 25.4, 29.2
b) Standardized incidence rate (×100,000) for M1 (metastatic tumors)
 ↓Alto Adige 11.7 9.2, 14.2 6.2 4.5, 7.8
 ↓Genoa 8.8 7.6, 10.1 3.1 2.4, 3.9
 ↓Latina 6.8 5.0, 8.7 5.9 4.4, 7.3
 ↑Naples 5.4 3.5, 7.4 6.6 4.7, 8.4
 ↑Ragusa 4.8 3.0, 6.6 4.5 2.8, 6.1
 ↓Reggio Emilia 10.5 8.4, 12.5 2.7 1.8, 3.7
 ↑Trento 13.4 10.9, 15.9 8.2 6.0, 10.4
 ↓Varese 8.6 7.0, 10.1 4.3 3.3, 5.3
 Total 9 8.4, 9.7 4.7 4.2, 5.1
 Late phase of PSA testing 9.2 8.7, 10.3 4.1 3.3, 4.3
 Early phase of PSA testing 8.6 7.0, 9.2 6.5 5.4, 7.1

Mortality rate ratios and incidence rate ratios for the 2005-2007 period vs 1996-1998

Incidence IRR 95% CI
CI = confidence interval; IRR = incidence rate ratio MRR = mortality rate ratio; PSA = prostate-specific antigen.
Overall cases
 Late phase of PSA testing 1.41 1.31, 1.52
 Early phase of PSA testing 1.66 1.51, 1.83
 Total 1.46 1.38, 1.55
High-risk + M1 cancers (including metastatic)
 Late phase of PSA testing 0.83 0.73, 0.94
 Early phase of PSA testing 1.18 1.02, 1.35
 Total 0.89 0.81, 0.98
Metastatic cancers
 Late phase of PSA testing 0.45 0.34, 0.59
 Early phase of PSA testing 0.76 0.58, 1.00
 Total 0.52 0.43, 0.63
Mortality MRR 95% CI
Overall cases
 Late phase of PSA testing 0.85 0.84, 0.96
 Early phase of PSA testing 1.09 0.94, 1.25
 Total 0.90 0.86, 0.94

We estimated the incidence of high-risk plus metastatic (M1) cancers and only metastatic cancer (Tab. III, a and b) for CR and period. The incidence of high-risk + M1 cancers decreased in Alto Adige, Genoa, and Varese, and remained almost stable in Reggio Emilia and Ragusa, while it increased in Latina and Naples. Overall, the incidence of high-risk + M1 cancers decreased slightly (Tab. IV), but the opposite trend was noted in the late and early phase of PSA testing CRs: a decrease in CRs in their late phase (IRR 0.83; 95% CI 0.73, 0.94) and a slight increase in those in the early phase (IRR 1.18; 95% CI 1.02, 1.35). With these trends, the CR in the early phase had a higher incidence for high-risk + M1 cancer than the late phase of PSA testing CRs (28.4 vs 26.4/100,000) in 2005-2007, even though they had a lower incidence of high-risk + M1 cancers in 1996-1998 (24.0 vs 32.0/100,000). The decrease in incidence observed in the late phase of PSA testing was almost completely due to a decrease in men over age 80 (Supplementary Tables, high risk + M1 S1a and metastatic S1b cancer incidence in the 2 periods by registry and age class).

The standardized incidence rate of M1 decreased in the late phase of PSA testing CRs and for Trento. For other registries, the incidence of M1 is almost stable. The IRR of M1 cancers in CR in the late phase of PSA testing is 0.45 (95% CI 0.34, 0.59); for those in the early-phase PSA testing, the IRR is 0.76 (95% CI 0.58, 1.00) (Tab. IV). In this case as well, the largest changes were observed in men over the age of 80, both for late- and early-phase PSA testing registries (Supplementary Tables, high risk + M1 S1a and metastatic S1b cancer incidence in the 2 periods by registry and age class).

In the same period, unclassifiable cancers decreased overall both as a proportion of diagnosed cancers and in incidence, even though differences among the CRs exist (Tabs. I and V).

Incidence of cancer with unknown risk characteristics in the 2 periods by registry (standardized incidence rate × 100,000 for unknown risk)

Cancer Registry 1996-1998 95% CI 2005-2007 95% CI
Incidence (per 100,000 person-years) of the specific group has been estimated applying the age-specific proportion of cancer with unknown risk observed in the study sample to the overall incidence (see Methods for a detailed description of the computing).
↑ = early phase of PSA testing; ↓ = late phase of PSA testing; CI = confidence interval; PSA = prostate-specific antigen.
↓Alto Adige 16.7 13.7, 19.7 3.9 2.6, 5.2
↓Genoa 10.1 8.8, 11.4 2.1 1.5, 2.6
↓Latina 14.7 12.0, 17.5 7.3 5.6, 9.0
↑Naples 7.6 5.3, 9.9 4.1 2.7, 5.6
↑Ragusa 8 5.6, 10.4 8.9 6.5, 11.4
↓Reggio Emilia 8.9 6.5, 11.4 13.5 11.1, 15.9
↑Trento 3.9 2.5, 5.2 0.3 0.0, 0.7
↓Varese 5.7 4.5, 7.0 8 6.6, 9.3
Total 9.8 9.2, 10.5 5.7 5.2, 6.2
Late phase of PSA testing 6.1 5.9, 6.4 4.5 3.9, 5.1
Early phase of PSA testing 11.1 8.4, 13.8 6.1 3.0, 9.3

Discussion

Incidence increased in all areas, but started at different levels and showed different slopes. On the other hand, mortality decreased only in areas with the highest incidence, which reached a peak of incidence before 2008, i.e., those that probably concluded a prevalence round of PSA testing, which we called areas in the late phase of PSA testing. In all areas, the increase in incidence was mostly due to low-risk cancers, but in areas in the late phase of PSA testing, we observed a net decrease in incidence of advanced cancers.

Our results, with the exception of the Trento area, concur with the scanty data on the spread of PSA testing in that period that are available, in which a higher and earlier spread was shown in the north as compared with the south (6, 7, 12).

These observations are consistent with the results from the European Randomized Study of Screening for Prostate Cancer (ERSPC) trial. In that trial, a reduction in mortality subsequent to a decrease in the incidence of advanced cancers and a strong increase in overall incidence were noted (13, 14).

The reduction in advanced cancers and in mortality occurred only when the increase in overall incidence reached a peak or decreased. This probably means that most of the people tested in this late phase of PSA testing had been already tested and thus were not being tested for the first time; i.e., test coverage reached a plateau. Thus the proportion of prevalent cancer detected started to decrease.

The results are also consistent with PLCO data, where in the intervention arm only a modest increase in incidence did not correspond to any decrease in mortality (5, 15).

Cancer registries that have a late phase of PSA testing also showed a reduction in mortality in the second period. Interestingly, the CRs with highest incidence level, i.e., Alto Adige, Genoa, Varese, and Reggio Emilia, showed a decrease in mortality of 4%-5% from the first to the second period; on the other hand, the 2 CRs in the late phase but with a medium level incidence rate (70-100/100,000), Latina and Varese, showed almost stable mortality in the 2 periods. The same trend can be observed in the early-phase registries, where Naples and Ragusa show the lowest incidence rate and even an increase in mortality over time (Naples and Ragusa).

Overall, there was a reduction in metastatic cases on the order of 4/100,000, and a corresponding reduction in mortality of 4/100,000. The mortality for prostate cancer is not incidence-based, so its reduction is due to a decrease in metastatic cancer diagnosed in previous years, rather than those incident in 2005-2007. Nevertheless, this correspondence between the magnitude of the reduction in metastatic cancers and mortality strongly suggests that the improvements in mortality are due to early diagnosis and are consistent with the results observed in randomized trials (16, 17).

The CRs in the late phase of PSA testing showed a marked reduction in metastatic cancers (M1) in the second period (9.2/100,000 in 1996-1998 compared to 4.1/100,000 in 2005-2007), but those in the early phase showed a slight reduction in metastatic cancer over time (8.6/100,000 vs 6.5/100,000).

High-risk cancer incidence did not show an overall decrease; such a decrease could be observed only for those areas in the late phase of PSA testing.

This picture is consistent with areas where a prevalent round of screening ended during the study period. Therefore, in the second period, what we observed is a lower incidence of advanced cancers that have been previously diagnosed. In the early phase, the prevalence round had not concluded and consequently we observed an overall stronger increase in incidence, since such increase was not mitigated by a decrease after the end of the prevalence peak. We also observed no reduction in advanced cancer, because prevalent advanced cancers are still the vast majority of incident cancers.

Looking separately at the different age groups, in areas in the late phase of PSA testing, we observe a reduction in metastatic tumors in the second period in all age groups, although the reduction is greater for the 85+ age group. Otherwise, in early-phase areas there was a reduction in M1 cases in the second period in the 85+ age group only.

A similar effect can be observed in areas in the late phase of PSA testing on the impact on high risk + M1 cancers: the most marked incidence reduction is shown in the most advanced age group.

Given that the average lead time of PSA, i.e., the length of anticipation of diagnosis due to PSA testing compared to symptomatic onset of the cancer, is 6.9 years (18), and that PSA testing in Italy is prolonged over the age of 80 (6, 7), it is not surprising that the effects of reductions in metastatic cancer incidence are first observed in men aged 85 and over. Furthermore, for a cancer with long survival and lead time, overdiagnosis due to competitive mortality is significant, even without taking indolent cancers into account.

Therefore, the question is whether and to what extent the lead time offered by PSA-based diagnosis could improve quality of life.

Strengths and limitations

The main limit of this study is that we do not have a direct measurement of actual PSA testing in the population during the 2 study periods in the 8 CRs included. Therefore we cannot validate our classification of the CRs as early and late PSA spread. We found only 3 studies measuring PSA uptake in the population in areas that are not included in our study (6, 7, 12). These data confirm that PSA uptake in the late 1990s and early 2000s was higher in northern Italy than in the south. Furthermore, the information collected for this high-resolution study did not enable us to distinguish between cancers detected through PSA screening and cancers diagnosed because of symptoms.

The overall trend of diagnostic accuracy and the test used for assessment/staging produced a stage migration that increased the probability of identifying metastases and lymph node involvement. This phenomenon introduces a bias toward the null hypothesis, i.e., not seeing a decrease in advanced and high-risk cancer when it actually occurred.

Furthermore, with many missing values for PSA and Gleason score in the first period, the trend in accuracy of registration can also mask a decrease in the incidence of advanced cancers. Cases with missing values tend to be very advanced cancers, in which there is no clinical advantage of further staging or diagnostic assessment. The use of multivariate imputation of missing values can only partially reduce this bias, because often many variables are missing at the same time. Thus, in these cases, imputation only reproduces the characteristics of the average population.

In conclusion, overall prostate cancer incidence increased from 1997 to 2005. In those areas that reached the highest incidence rates in the second period, we observed a decrease in high risk + M1 cancer incidence. A simultaneous decrease in mortality was only observed in areas in the late phase of PSA testing. Therefore, our data suggest that the incidence and mortality of prostate cancer in Italy changed in a way that can be interpreted as an effect of PSA spread, and that the magnitude of mortality reduction and incidence increase in those areas where the spread of PSA occurred sufficiently early between the 1990s and early 2000s is similar to the magnitude observed in the ERSPC trial (13, 14, 16).

Furthermore, we observed an almost identical absolute reduction in mortality and incidence of metastatic cancers, which suggests that the improvement in mortality is mostly due to earlier diagnosis leading to better prognosis.

The above observations are consistent with the hypothesis of an effect of a different PSA test spread during the study period, thus suggesting that a significant portion of the reduction in mortality is due to earlier diagnosis and better prognosis, but also suggesting that the benefits in terms of lower mortality can only be achieved with an enormous increase in low-risk cancer incidence.

Prostate Cancer High-Resolution Study Working Group

Fabio Pannozzo (Latina Cancer Registry, Italy), Paolo Contiero (Varese Cancer Registry, Fondazione IRCCS Istituto Nazionale dei Tumori, Milan, Italy), Mario Fusco (Naples Cancer Registry, Italy), Michele Lodde (Urology Unit, Ospedale Centrale di Bolzano, Italy), Guido Mazzoleni (Alto Adige Cancer Registry, Provincia di Bolzano, Italy), Silvano Piffer (Evaluative Epidemiology Unit, Trento Cancer Registry, Italy), Rosario Tumino (Ragusa Cancer Registry, Italy), Antonella Puppo (Liguria Cancer Registry, IRCCS AOU San Martino IST Genoa, Italy), Andreas Seeber (Department for Haematology and Oncology, Tyrolean Cancer Research Institut, Innsbruck Medical University, Austria), Lucia Mangone (Reggio Emilia Cancer Registry, AUSL Reggio Emilia, and Arcispedale Santa Maria Nuova, IRCCS, Reggio Emilia, Italy).

Disclosures

Financial support: Supported by the Associazione Italiana per la Ricerca sul Cancro (grant 9293, IG 2009) and by AMGEN Dompé. The sponsors had no role in the study design; in the collection, analysis, and interpretation of data; in writing the report; or in the decision to submit the article for publication.
Conflict of interest: None of the authors has conflict of interest with this submission.
References
  • 1. Matsuda T Saika K Comparison of time trends in prostate cancer incidence (1973 1997) in East Asia, Europe and USA, from Cancer Incidence in Five Continents Vols IV VIII. Jpn J Clin Oncol 2007 37 7 556 557 Google Scholar
  • 2. AIRTUM Working Group. Italian cancer figures, report 2009: Cancer trend (19982005). Epidemiol Prev 2009 33 4-5 Suppl 1 1 168 Google Scholar
  • 3. Chan JM Stampfer MJ Giovannucci EL What causes prostate cancer? A brief summary of the epidemiology. Semin Cancer Biol 1998 8 4 263 273 Google Scholar
  • 4. Adami H-O Hunter D Trichopoulos D eds. Textbook of Cancer Epidemiology. Oxford University Press, 2008. Google Scholar
  • 5. Andriole GL Crawford ED Grubb RL III et al; PLCO Project Team. Mortality results from a randomized prostate-cancer screening trial. N Engl J Med 2009 360 13 1310 1319 Google Scholar
  • 6. Leoni M Falcini F Ravaioli A et al. [Estimating standard performance measures of opportunistic screening for prostate cancer]. Epidemiol Prev 2008 32 6 285 293 Google Scholar
  • 7. Ciatto S Houssami N Martinelli F Giusti F Zappa M PSA use and incidence of prostate biopsy in the Tuscany region: is opportunistic screening discounting biopsy in subjects with PSA elevation? Tumori 2008 94 4 518 522 Google Scholar
  • 8. Quaglia A Parodi S Grosclaude P Martinez-Garcia C Coebergh JW Vercelli M Differences in the epidemic rise and decrease of prostate cancer among geographical areas in Southern Europe. an analysis of differential trends in incidence and mortality in France, Italy and Spain. Eur J Cancer 2003 39 5 654 665 Google Scholar
  • 9. Quaglia A Vercelli M Puppo A et al; Prostate Working Group. Prostate cancer in Italy before and during the PSA era: survival trend and prognostic determinants. Eur J Cancer Prev 2003 12 2 145 152 Google Scholar
  • 10. Crocetti E Ciatto S Buzzoni C Zappa M Prostate cancer incidence rates have started to decrease in central Italy. J Med Screen 2010 17 1 50 51 Google Scholar
  • 11. Trama A Botta L Nicolai N et al; Prostate Cancer High Resolution Study Working Group. Prostate cancer changes in clinical presentation and treatments in two decades: an Italian population-based study. Eur J Cancer 2016 67 91 98 Google Scholar
  • 12. D’Ambrosio G Samani F Cancian M De Mola C Practice of opportunistic prostate-specific antigen screening in Italy: data from the Health Search database. Eur J Cancer Prev 2004 13 5 383 386 Google Scholar
  • 13. Schröder FH Hugosson J Roobol MJ et al; ERSPC Investigators. Prostate-cancer mortality at 11 years of follow-up. N Engl J Med 2012 366 11 981 990 Google Scholar
  • 14. Schröder FH Hugosson J Roobol MJ et al; ERSPC Investigators. Screening and prostate cancer mortality: results of the European Randomised Study of Screening for Prostate Cancer (ERSPC) at 13 years of follow-up. Lancet 2014 384 9959 2027 2035 Google Scholar
  • 15. Andriole GL Crawford ED Grubb RL III et al; PLCO Project Team. Prostate cancer screening in the randomized Prostate, Lung, Colorectal, and Ovarian Cancer Screening Trial: mortality results after 13 years of follow-up. J Natl Cancer Inst 2012 104 2 125 132 Google Scholar
  • 16. Buzzoni C Auvinen A Roobol MJ et al. Metastatic Prostate Cancer Incidence and Prostate-specific Antigen Testing: New Insights from the European Randomized Study of Screening for Prostate Cancer. Eur Urol 2015 68 5 885 890 Google Scholar
  • 17. Bokhorst LP Zappa M Carlsson SV et al. Correlation between stage shift and differences in mortality in the European Randomised study of Screening for Prostate Cancer (ERSPC). BJU Int 2016 118 5 677 680 Google Scholar
  • 18. Finne P Fallah M Hakama M et al. Lead-time in the European Randomised Study of Screening for Prostate Cancer. Eur J Cancer 2010 46 17 3102 3108 Google Scholar

Authors

Affiliations

  • Interinstitutional Epidemiology Unit, AUSL Reggio Emilia, Reggio Emilia - Italy
  • Arcispedale Santa Maria Nuova, IRCCS, Reggio Emilia - Italy
  • Evaluative Epidemiology Unit, Fondazione IRCCS Istituto Nazionale Tumori, Milan - Italy
  • Urology Unit, Fondazione IRCCS Istituto Nazionale dei Tumori, Milan - Italy
  • Diagnostic Imaging and Radiotherapy, Università degli Studi di Milano, Milan - Italy
  • Prostate Cancer Program and Radiation Oncology 1, Fondazione IRCCS Istituto Nazionale dei Tumori, Milan - Italy

Article usage statistics

The blue line displays unique views in the time frame indicated.
The yellow line displays unique downloads.
Views and downloads are counted only once per session.

This article has supplementary materials available to download.