Abstract
Backgrount and Objective
Studies evaluating the role of baseline midlife PSA as a predictor of development and progression of prostate cancer relied predominately on cohorts from the pre-PSA screening introduction era. The aim of our study was to examine the role of baseline PSA prior the age of 60 as a predictor of developing lethal prostate cancer using a contemporary North American cohort.
Methods
Our cohort included all men aged 40–59 years, who received their first PSA through our health system between the years 1995 and 2019. Patients were divided into 4 categories based on age: 40–44, 45–49, 50–54, and 55–59 years. Baseline PSA was the predictor of interest. Lethal disease was defined as death from prostate cancer or development of metastatic disease either at diagnosis or during follow-up. Cancer-specific mortality and overall mortality were obtained by linking our database to the Michigan Vital Records registry. Competing-risk regression was used to evaluate the association between PSA and lethal prostate cancer.
Key findings and limitations
A total of 129,067 men met inclusion criteria during the study period. Median follow-up for patients free from cancer was 7.4 years. For men aged 40 to 44, 45 to 49, 50 to 54, and 55 to 59 years, the estimated rate of lethal prostate cancer at 20 years was 0.02%, 0.14%, 0.33%, 0.51% in men with PSA < median, and 0.79%, 0.16%, 2.5%, 5.4% in men with PSA ≥90th percentile. For the same age category, the estimated rate of any prostate cancer at 20 years was 1.6%, 2.9%, 3.9%, 5.8% in men with PSA < median, and 25%, 28%, 38% and 39% in men with PSA ≥90th percentile. On multivariable analysis, men with a PSA≥90th percentile had a hazard ratio of 7.48 (95% CI: 6.20 – 9.03) for lethal disease, when compared to those with PSA < median. On multivariable analysis, men with a PSA ≥90th percentile had an HR of 20.47-fold (95% CI: 18.58 – 22.55) for prostate cancer incidence, when compared to those with PSA < median at first. Limitations included shorter median follow-up than prior literature.
Conclusions and clinical implications
Baseline PSA is a very strong predictor of the subsequent risk of developing lethal prostate cancer in a large contemporary diverse North American cohort, which was exposed to opportunistic PSA screening. The association was far larger than that found for polygenic risk scores, confirming that baseline PSA prior the age of 60 is the most effective tool for adjusting subsequent screening. Compared to studies on unscreened cohorts, there was a smaller difference in discrimination between incident and lethal disease, reflecting the influence of screening.
Keywords: Prostate-Specific Antigen, Prostatic Neoplasms, Middle Aged, Mortality, Healthcare Disparities
Patient summary:
In this study, we found that a single baseline PSA is strongly predictive of the subsequent risk of developing metastatic prostate cancer, as well as the risk of dying from prostate cancer. The initial PSA level can therefore be used to adjust the frequency of subsequent PSA testing.
Introduction
Prostate cancer screening using prostate-specific antigen (PSA) has been shown to reduce prostate cancer mortality1 but only at the cost of a large number of men being tested, diagnosed and treated to save one life2,3. A baseline midlife PSA baseline value has been proposed as a method to risk stratify PSA screening, with the potential of decreasing overdiagnosis and overtreatment4,5.
Previous reports addressing the role of baseline midlife PSA focused mainly on populations providing blood samples before the PSA screening era. Moreover, the landmark Malmö study focused on a Scandinavian cohort that was far more homogenous in terms of genetic ancestry and access to health care than US patients5. Likewise, studies that originated from US cohorts focused exclusively on distinct subgroup of the population, such as physicians or Black patients6,7, which might not be necessarily applicable to the general US population. Finally, PSA screening affects the natural history of disease: a man with an elevated PSA is more likely to be diagnosed with prostate cancer, but accordingly less likely to die from disease, if he is screened compared to being unscreened. Recent studies have examined the influence of screening on the prognostic significance of initial PSA levels, although they suffered of limited sample sizes and concentrated exclusively on narrow age groups, such as men aged 50–54 or 55–608,9.
Our aim was to examine the value of time PSA values prior the age of 60 in predicting lethal prostate cancer among a large contemporary North American cohort, which was exposed to PSA screening. We also hypothesized that the difference in discrimination for prostate cancer specific mortality (PCSM) versus prostate cancer incidence will be smaller in comparison to historical studies on unscreened cohorts.
Materials & Methods
Data Sources
We utilized our institutional database which was built by interrogating our electronic medical records for all men receiving care in Henry Ford Health (HFH) between 1995 and 2019. This is a system of over 50 medical centers and hospitals that serves the Detroit metropolitan population, which is highly diverse in terms of race and socioeconomic status. Patients who did not undergo any PSA test at HFHS were excluded (484,198). We included men aged between 40 years and 59 years (92,590 excluded). Only the first PSA test result for men within this age range was included in the study. Our selection criteria yielded a total of 129,067 assessable patients (see Consort diagram, supplementary figure 1).
Variables
We extracted the following variables for each patient: PSA, age, race (white, Black, other, unknown), Charlson Comorbidity Index (CCI) [0, 1, 2, ≥3], education level by census tract, and median household income by census tract.
Within each age group, the first PSA in the system for each patient was considered as the baseline value and then, for illustrative purposes, categorized by quartile10,11 with the fourth quartile subdivided into 75th to 90th percentile and 90th to 100th percentile. PSA- and age categorization was based on previously published methodology5–7,12. Patient data regarding prostate cancer diagnosis and date of diagnosis were recorded. For patients with prostate cancer, the clinical stage and grade (ISUP score) of the disease were also extracted.
Endpoints
Our primary endpoint was the likelihood of developing lethal disease, defined as death from prostate cancer or development of metastatic disease either at diagnosis or during follow-up, as previously described7. Time to lethal disease was defined as the date at which distant metastasis were diagnosed or death form prostate cancer (if no previous metastasis were detected). Our secondary endpoints were the likelihood of developing any prostate cancer within the study period, and PSCM. The date and cause of death were obtained by linking our database with the Michigan Vital Records registry. To address the limitation regarding the lack of information on patients moving or dying outside Michigan, follow-up was defined as their last encounter at HFH and patients were censored at that point. Conversely, death due to a cause other than prostate cancer was defined as a competing event. Patients who experienced a death due to a cause other than prostate cancer were considered to be no longer at risk, though they were not censored. Of note, 53,251 men in the final cohort had follow-up greater than or equal to 10 years. Numbers of patients lost to follow up at 5 years, 10 years, 15 years and 20 years for each hemi-decade are displayed in supplementary table 1.
Statistical Analysis
We estimated the risk of developing lethal disease using competing-risk cumulative incidence. Multivariable, competing-risk regression was used to calculate hazard ratios for developing lethal disease based on baseline PSA. We included the following predictors measured at time of baseline PSA: age at first PSA test, race (Black, Other, white as reference), comorbidity, level of education (Bachelor’s degree) and median family income (above vs below median). Death from causes other than prostate cancer was considered as a competing risk. For illustrative purposes, the relationship between baseline PSA and the risk of developing lethal disease was graphically depicted using restricted cubic splines. Harrell’s concordance index was used to assess the discrimination of baseline PSA in predicting mortality and incidence, and the results were compared with historical data to examine whether the predictive value of the first PSA has changed compared to reports from unscreened cohorts. Similar steps were repeated to calculate the risk of developing our secondary endpoints, namely developing any prostate cancer, and PCSM. Alpha was set at two-sided 0.05. All analyses were performed using SAS 9.4 (SAS Institute, Cary, North Carolina). Patients were retrospectively enrolled in HFH-IRB#12641–29. An institutional review board waiver for informed consent was obtained prior to conducting this study, in accordance with institutional regulations when dealing with de-identified previously collected data.
Results
Cohort characteristics are shown in table 1. We included 24,051 men who received their first PSA between 40 and 44 years old, 32,320 aged 45–49, 40,761 aged 50–54, and 31,935 aged 55–59. Our cohort had a good representation of non-white patients with 24% identifying as Black, and 16,100 (12%) as “other” race. Median (IQR) follow-up of patients who did not develop prostate cancer was 7.4 (3 – 15) years. Median PSA values were consistent with previous reports, ranging between 0.7 and 0.9 ng/ml across all age categories. The 90th percentile value was almost double the median across all age categories. Other descriptive statistics are reported in Table 1. Overall, we recorded 6136 (4.8%) prostate cancer cases, of whom 930 (0.72%) were lethal disease and 360 (0,28%) died from prostate cancer. Clinically significant prostate cancer was found in 3165 (2.5%) men.
Table 1.
Comparative statistics of 129,067 men aged 40 to 59 years, who received at least one PSA testing within the Henry Ford Health system, within the study period from 1995 to 2018, by age group.
| Age Group | |||||
|---|---|---|---|---|---|
| 40 – 44 (N=24051) |
45 – 49 (N=32320) |
50 – 54 (N=40761) |
55 – 59 (N=31935) |
Total | |
| Baseline first-time PSA (ng/mL): | |||||
| Median (IQR) | 0.6 (0.4, 0.9) | 0.7 (0.4, 1.0) | 0.7 (0.5, 1.2) | 0.9 (0.5, 1.6) | 0.7 (0.5, 1.2) |
| Charlson Comorbidity Index, n(%): | |||||
| 0 | 17403 (76%) | 22995 (75%) | 27558 (73%) | 19559 (66%) | 87515 (72%) |
| 1 | 3902 (17%) | 5590 (18%) | 7344 (19%) | 6583 (22%) | 23419 (19%) |
| 2 | 791 (4.0%) | 1167 (3.8%) | 1650 (4.4%) | 1736 (5.9%) | 5344 (4.4%) |
| 3+ | 653 (3.0%) | 1060 (3.2%) | 1367 (3.6%) | 1563 (5.1%) | 4643 (3.8%) |
| Missing | 1302 | 1508 | 2842 | 2494 | 8146 |
| Race, n (%) | |||||
| White | 12305 (51%) | 19694 (61%) | 27867 (68%) | 22218 (70%) | 82084 (64%) |
| Black | 8522 (35%) | 8346 (26%) | 7898 (20%) | 6117 (19%) | 30883 (24%) |
| Other | 3224 (14%) | 4280 (13%) | 4996 (12%) | 3600 (11%) | 16100 (12%) |
| Median Family Income by Census Tract (1000$) | |||||
| Median (IQR) | 75 (52–103) | 75 (53, 102) | 75 (55, 101) | 73 (53, 96) | 74 (54, 101) |
| Prostate Cancer, n | |||||
| No Prostate Cancer | 23433 (97.4%) | 31115 (96.3%) | 38705 (94.9%) | 29678 (92.9%) | 122931 (95.2%) |
| Prostate Cancer | 618 (2.6%) | 1205 (3.7%) | 2056 (5.1%) | 2257 (7.1%) | 6136 (4.8%) |
| ISUP Grade Group, n | |||||
| 1 | 78 (20.2%) | 185 (23.6%) | 277 (20.4%) | 333 (22.0%) | 873 (21.6%) |
| 2 | 220 (57.0%) | 393 (50.1%) | 691 (51.0%) | 759 (50.2%) | 2063 (51.1%) |
| 3 | 60 (15.5%) | 138 (17.6%) | 259 (19.1%) | 294 (19.4%) | 751 (18.6%) |
| 4 | 15 (3.9%) | 28 (3.6%) | 67 (45.0%) | 52 (3.4%) | 162 (4.0%) |
| 5 | 13 (3.4%) | 40 (5.1%) | 61 (4.5%) | 75 (5.0%) | 189 (4.7%) |
| Missing | 232 | 421 | 701 | 744 | 2098 |
| No cancer | 23433 | 31115 | 38705 | 29678 | 122931 |
| Age at PCa diagnosis (years): | |||||
| Median (IQR) | 54 (49–58) | 58 (51–62) | 59 (53–65) | 60 (57–67) | 59 (54–65) |
| Baseline PSA at diagnosis (ng/mL): | |||||
| Median (IQR) | 5.8 (4.5–10.0) | 6.0 (4.6–11.0) | 6.0 (4.5–9.4) | 5.9 (4.6–10.2) | 5.9 (4.5–10.1) |
Table 2 showed the cumulative incidence estimate for lethal disease according to baseline PSA value. The risk of lethal disease at 20 years of follow-up for PSA value below the median was consistently low across all age categories (<1.3%). The increased risk for lethal disease was particularly evident for the group with a baseline value within top 10th percentile, group who exhibited a risk at 20 years ranging between 3.1%−7.5% across all age categories (Supplementary figures 2–3). Concentrations of PSA were also significantly associated with risk of lethal prostate cancer, with C index of 0.703 for men aged 40–44, 0.652 for age 45–49, 0.679 for age 50–54 and 0.729 for age 55–59, respectively. On multivariable analysis, men with a PSA ≥90th percentile had a hazard ratio (HR) of 7.48 (95% CI: 6.20 – 9.03) for lethal prostate cancer, when compared to those with PSA below the median at first testing (Table 3). This hazard ratio decreased with increasing age (Supplementary tables 2a–d). Being Non-Hispanic Black and having a Charlson Comorbidity Index > 3 were other predictor of lethality. The 10-,15- and 20-years predicted risk of developing lethal prostate cancer based on baseline PSA are graphically displayed in Figures 1–2.
Table 2.
Cumulative incidence of lethal prostate of 129,067 men aged 40 to 59 years, who received at least one PSA testing within the Henry Ford Health system, within the study period from 1995 to 2018, by Baseline PSA percentile group.
| Covariate | PSA value (ng/ml) |
% Risk (95% CI) | ||
|---|---|---|---|---|
| 10 years | 15 years | 20 years | ||
| All ages | ||||
| Below Median | < 0.7 | 0.07 (0.052–0.11) | 0.26 (0.20–0.33) | 0.94 (0.77–1.1) |
| Quartile 3 | 0.7–1.2 | 0.15 (0.10–0.23) | 0.45 (0.33–0.59) | 1.6 (1.3–1.9) |
| 75th to 90th percentile | 1.2–2.1 | 0.47 (0.36–0.62) | 1.2 (0.99–1.48) | 3.1 (2.7–3.7) |
| Top 10th percentile | >2.1 | 2.3 (1.9–2.6) | 3.3 (2.9–3.8) | 4.98 (4.3–5.7) |
| 40–44: | ||||
| Below Median | <0.6 | 0 (NA) | 0.11 (0.049–0.24) | 0.50 (0.26–0.89) |
| Quartile 3 | 0.6–0.9 | 0.05 (0.01–0.17) | 0.17 (0.063–0.39) | 1.6 (0.96–2.5) |
| 75th to 90th percentile | 0.9–1.4 | 0.15 (0.052–0.38) | 0.65 (0.35–1.1) | 1.8 (1.1–2.7) |
| Top 10th percentile | >1.4 | 0.63 (0.32–1.1) | 1.1 (0.65–2.0) | 4.2 (2.5–6.6) |
| 45–49: | ||||
| Below Median | <0.7 | 0.04 (0.015–0.89) | 0.19 (0.10–0.32) | 0.92 (0.62–1.3) |
| Quartile 3 | 0.7–1.2 | 0.03 (0.023–0.18) | 0.19 (0.06–0.47) | 1.6 (0.99–2.4) |
| 75th to 90th percentile | 1.2–1.6 | 0.17 (0.064–0.40) | 0.70 (0.41–1.2) | 2.5 (1.7–3.5) |
| Top 10th percentile | >1.6 | 0.91 (0.56–1.4) | 2.2 (1.5–3.1) | 3.1(2.2–4.3) |
| 50–54: | ||||
| Below Median | <0.7 | 0.02 (0.0050–0.08) | 0.24 (0.14–0.40) | 1.1 (0.75–1.5) |
| Quartile 3 | 0.7–1.6 | 0.19 (0.10–0.36) | 0.60 (0.38–0.90) | 1.3 (0.91–1.8) |
| 75th to 90th percentile | 1.6–2.1 | 0.45 (0.26–0.73) | 1.5 (1.1–2.1) | 4.2 (3.2–5.4) |
| Top 10th percentile | >2.1 | 2.7 (2.1–3.4) | 3.8 (3.0–4.8) | 5.4 (4.2–6.8) |
| 55–59: | ||||
| Below Median | <0.9 | 0.30 (0.19–0.45) | 0.59 (0.40–0.83) | 1.3 (0.90–1.8) |
| Quartile 3 | 0.9–1.2 | 0.37 (0.20–0.65) | 0.87 (0.54–1.4) | 2.1 (1.4–3.0) |
| 75th to 90th percentile | 1.2–3.0 | 1.4 (0.94–2.0) | 2.3 (1.6–3.1) | 4.0 (2.9–5.2) |
| Top 10th percentile | >3.0 | 4.6 (3.7–5.6) | 5.6 (4.5–6.8) | 7.5 (6.0–9.3) |
Table 3.
Fine-Gray Cumulative Incidence Regression for lethal prostate cancer of 129,067 men aged 40 to 59 years, who received at least one PSA testing within the Henry Ford Health system, within the study period from 1995 to 2018, by PSA cutoff
| Follow-up Time in Years (Lethal PCa) | ||
|---|---|---|
| ---------------------------------------- | ||
| Covariate | Hazard Ratio | HR P-value |
| Baseline PSA percentile Group: | ||
| 75th to 90th percentile | 3.82 (3.13–4.65) | <.001 |
| Quartile 3 | 1.90 (1.54–2.35) | |
| Top 10th percentile | 7.48 (6.20–9.03) | |
| Below Median | Ref. | |
| Age Group: | ||
| 40–44 | 0.49 (0.39–0.61) | <.001 |
| 45–49 | 0.57 (0.47–0.69) | |
| 50–54 | 0.78 (0.66–0.92) | |
| 55–59 | Ref. | |
| Race: | ||
| Black | 1.77 (1.52–2.07) | <.001 |
| Other | 1.05 (0.83–1.34) | |
| White | Ref. | |
| CCI: | ||
| 1 | 1.09 (0.92–1.29) | <.001 |
| 2 | 0.84 (0.60–1.20) | |
| 3+ | 1.96 (1.51–2.54) | |
| 0 | Ref. | |
| Proportion of People with A Bachelor’s Degree by Census Tract: | ||
| Bottom 50% | 1.07 (0.90–1.28) | 0.44 |
| Top 50% | Ref. | |
| Median Family Income by Census Tract: | ||
| Bottom 50% | 0.85 (0.71–1.02) | 0.078 |
| Top 50% | Ref. | |
Number of observations in the original data set = 129067. Number of observations used = 114046.
Figure 1:
Predictive risk of lethal prostate cancer at 10-, 15- and 20 years by baseline PSA value in 129,067 men aged 40 to 59 years, who received at least one PSA testing within the Henry Ford Health system, within the study period from 1995 to 2018.
Figures 2a-d:
Predictive risk of lethal prostate cancer at 10-, 15- and 20 years by baseline PSA value in 129,067 men aged 40 to 59 years, who received at least one PSA testing within the Henry Ford Health system, within the study period from 1995 to 2018, by age category.
The association of baseline PSA concentration and prostate cancer incidence is displayed in Supplementary table 3, with a C-Index of 0.707, 0.641, 0.656 and 0.658 for the four age cohorts. On multivariable analysis, men with a PSA ≥90th percentile had an HR of 20.47-fold (95% CI: 18.58 – 22.55) for prostate cancer incidence, when compared to those with PSA < median at first testing (Supplementary table 4). The HR for prostate incidence in men aged 40–44 was the higher observed (25.88; 95% CI: 19.21 – 34.87), but differently from lethal prostate cancer, a decreasing trend with the increase of age was not observed for this outcome (Supplementary tables 5a–d).
The HR of baseline PSA for PCSM is given in Supplementary table 6. Men with a baseline baseline PSA ≥90th percentile had an HR of 12.75-fold (95% CI: 9.55 – 17.02) for PCSM, when compared to those with value below the median (Supplementary table 7). Concentrations of PSA were associated with incidence with a C-Index of 0.937, 0.790, 0.764 and 0.770 for the four age cohorts.
Discussion
The role of baseline midlife PSA as a long-term predictor for lethal prostate cancer has been evaluated in several US6,7,10,11,13,14 and Swedish5,12,15,16 cohorts, in which midlife baseline PSA was assessed by retrospectively testing blood samples collected more than thirty years previously. For example, both the Physicians’ Health Study7 and the Malmo Preventive Project5,15,16 cohorts showed how midlife PSA value was able to identify a small percentage of men at higher risk for lethal disease to be intensively screened, while avoiding further PSA testing in more than a half of the population. Our aim was to retest the role of baseline PSA as a predictor of lethal prostate cancer, as well as any prostate cancer and PCSM, using a large contemporary North American cohort which was exposed to PSA screening.
Our findings are severalfold. First, our results corroborated previous literature from pre-PSA screening introduction era, showing that concentration of baseline PSA subsequently predicted lethal cancer, PCSM or any prostate cancer diagnosis. Second, race, level of education and income were associated with PCSM. While Non-Hispanic Black men were persistently at higher risk of developing all the outcomes evaluated (regardless of age group), the impact of education and income seems highly dependent on age. Specifically, men aged 40–49 who were less educated and had lower income showed higher hazard for lethal prostate cancer, while these effects seemed to disappear with age. Third, the risk of developing lethal prostate cancer at 20-years for PSA value below 2 ng/ml ranged between 1.5% and 2% for men across all ages. Conversely, an increased risk for lethal disease was particular evident for PSA value above the same cut-off, even for men aged 55–59. This suggested that continuing screening at age 60 might be beneficial for these men, as previously showed4.
Our study markedly differs from the seminal Malmo study primarily because the former one was conducted on an unscreened cohort originating from the era before the introduction of PSA screening. When stratified for hemi-decades, median PSA values found in the Malmo cohort were comparable to our findings. On the other hand, PSA screening significantly influenced the rates of PCSM, which were significantly higher in Malmo than what is reported here. Moreover, the discrimination of baseline PSA is lower in our cohort, while the association with prostate cancer incidence increased. This observation was highlighted by the fact that only 35% of lethal cases were found for values above the 10th percentile, versus more than 50% in the previous Malmo study. With the exception of men in the very youngest age group, the difference between discrimination for incident vs. lethal cancer was generally lower in our study than in the Malmo cohorts, reflecting the impact of screening: men undergoing PSA testing are more likely to have a PSA-detected cancer but less likely to die of disease if cancer is found early. Indeed, our assessment of the screening rate within our cohort showed that only 3% of them met the stringent PLCO criteria for screening17 (annual PSA testing for 6 consecutive years). Moreover, a substantial proportion (~30%) underwent just a single PSA test in 5 years. That said, after accounting for race, our white patient subset exhibited comparable PCSM rate with the Goteborg population, suggesting that screening reduced the mortality regardless of PSA testing rate11. Previous authors have already explored the role of baseline midlife PSA according to racial diversity, showing that, although Non-Hispanic Black patients exhibited both higher incidence and lethality due to prostate cancer, PSA levels were similar to those among control6. Although we believe this topic deserves a separate analysis outside this manuscript, our preliminary findings have confirmed previous findings from unscreened era.
In the last few years, new tools have been developed to improve prostate cancer risk stratification, with the aim of reducing overdiagnosis and consequent overtreatment. For instance, the polygenic hazard score (PHS290) has been demonstrated to identify men at the highest risk of developing fatal prostate cancer18,19. However, the associations are far weaker than demonstrated here. For example, in the Million Veteran Program, men with a PHS290 score above the 95th percentile had a HR of 3.0 for lethal prostate cancer and 3.39 for any prostate cancer, compared to those below the median percentile20. We report hazard ratios more than 4-fold larger, even while using a lower percentile cut-off (i.e. 90th vs. 95th).
Although our study expands upon the previous literature, it is not without limitations. First, The median follow-up is slightly shorter than those in prior literature on midlife PSA, meaning that we captured fewer prostate cancer cases and deaths. For these reasons, we truncated our risk analyses to 20 years, as the tail of the curves would be unreliable and overestimate the PCSM risk. Second, we do not have any information on how many men moved or died outside Michigan, thus follow-up was defined as their last encounter at HFH and patients were censored at that point. Similarly, the lack of complete data on the distribution of Gleason Grade disease detected during follow-up limited the conclusion on whether PSA screening helped in detecting clinically significant disease or low-grade prostate cancer. Third, we acknowledge that other factors, which we could not adjust for in the multivariable analysis, likely modify the PSA value, such as obesity or 5α-reductase inhibitors.
While current international guidelines recommend the use of baseline PSA to refine subsequent screening and diminish overdiagnosis of indolent cancers21,22, its clinical application remains scarce. Indeed, a strategy aimed at optimizing PSA screening based on the baseline midlife PSA value should intensify screening for individuals with a baseline PSA above the 90th percentile. On the other hand, focusing exclusively on these patients would miss a significant amount of lethal cases. However, because our cohort was exposed to opportunistic screening, the prognostic characteristics of the PSA test are influenced by prior screenings. This fundamentally alters its properties, making not possible to draw definitive conclusions regarding the optimal baseline PSA cut-off.
Despite the altered association between PSA values and prostate cancer incidence and mortality due to screening practices, baseline PSA stillis a strong predictor of prostate cancer lethality. These findings confirm previous reports from screened cohorts8,9,23, although we are the first to utilize a contemporary diverse North American cohort, which was larger and comprehensively included all men aged 40–59 years old.
Conclusion
Our report examined the role of first-time PSA in a contemporary and large North American cohort of men aged 40–59. Our findings corroborate prior studies by suggesting that midlife first-time PSA is a strong predictor of the subsequent risk of developing lethal prostate cancer, in a cohort which was exposed to PSA screening. The association was far larger than that found for polygenic risk scores, confirming that baseline PSA is the most effective tool to adjust the intensity of prostate cancer screening.
Supplementary Material
Funding:
The VUI Center for Outcomes Research, Analysis, and Evaluation is supported by a fund, which was started by a contribution from Menon Foundation and VUI Foundation. A.J. Vickers was supported in part by the National Institutes of Health/National Cancer Institute (NIH/NCI) with a Cancer Center Support Grant to Memorial Sloan Kettering Cancer Center [P30-CA008748], a SPORE grant in Prostate Cancer to Dr. H. Scher [P50-CA92629).
Footnotes
Conflicts of interest: A.J. Vickers is named on a patent for a statistical method to detect prostate cancer that has been licensed to and commercialized as the 4Kscore test by OPKO Health and receives royalties from sales of this test. A.J. Vickers has stock options in OPKO Health.
References
- 1.Hugosson J, Roobol MJ, Månsson M, et al. A 16-yr Follow-up of the European Randomized study of Screening for Prostate Cancer. Eur Urol. 2019;76(1):43–51. doi: 10.1016/j.eururo.2019.02.009 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Ilic D, Djulbegovic M, Jung JH, et al. Prostate cancer screening with prostate-specific antigen (PSA) test: a systematic review and meta-analysis. BMJ. 2018;362:k3519. doi: 10.1136/bmj.k3519 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Hayes JH, Barry MJ. Screening for prostate cancer with the prostate-specific antigen test: a review of current evidence. JAMA. 2014;311(11):1143–1149. doi: 10.1001/jama.2014.2085 [DOI] [PubMed] [Google Scholar]
- 4.Carlsson S, Assel M, Sjoberg D, et al. Influence of blood prostate specific antigen levels at age 60 on benefits and harms of prostate cancer screening: population based cohort study. BMJ. 2014;348:g2296. doi: 10.1136/bmj.g2296 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Vickers AJ, Ulmert D, Sjoberg DD, et al. Strategy for detection of prostate cancer based on relation between prostate specific antigen at age 40–55 and long term risk of metastasis: case-control study. BMJ. 2013;346:f2023. doi: 10.1136/bmj.f2023 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Preston MA, Gerke T, Carlsson SV, et al. Baseline Prostate-specific Antigen Level in Midlife and Aggressive Prostate Cancer in Black Men. Eur Urol. 2019;75(3):399–407. doi: 10.1016/j.eururo.2018.08.032 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Preston MA, Batista JL, Wilson KM, et al. Baseline Prostate-Specific Antigen Levels in Midlife Predict Lethal Prostate Cancer. J Clin Oncol Off J Am Soc Clin Oncol. 2016;34(23):2705–2711. doi: 10.1200/JCO.2016.66.7527 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Kovac E, Carlsson SV, Lilja H, et al. Association of Baseline Prostate-Specific Antigen Level With Long-term Diagnosis of Clinically Significant Prostate Cancer Among Patients Aged 55 to 60 Years: A Secondary Analysis of a Cohort in the Prostate, Lung, Colorectal, and Ovarian (PLCO) Cancer Screening Trial. JAMA Netw Open. 2020;3(1):e1919284. doi: 10.1001/jamanetworkopen.2019.19284 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Carlsson S, Assel M, Ulmert D, et al. Screening for Prostate Cancer Starting at Age 50–54. A Population-based Cohort Study. Eur Urol. 2017;71(1):46–52. doi: 10.1016/j.eururo.2016.03.026 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Kuller LH, Thomas A, Grandits G, Neaton JD, Multiple Risk Factor Intervention Trial Research Group. Elevated prostate-specific antigen levels up to 25 years prior to death from prostate cancer. Cancer Epidemiol Biomark Prev Publ Am Assoc Cancer Res Cosponsored Am Soc Prev Oncol. 2004;13(3):373–377. [PubMed] [Google Scholar]
- 11.Whittemore AS, Cirillo PM, Feldman D, Cohn BA. Prostate specific antigen levels in young adulthood predict prostate cancer risk: results from a cohort of Black and White Americans. J Urol. 2005;174(3):872–876; discussion 876. doi: 10.1097/01.ju.0000169262.18000.8a [DOI] [PubMed] [Google Scholar]
- 12.Stattin P, Vickers AJ, Sjoberg DD, et al. Improving the Specificity of Screening for Lethal Prostate Cancer Using Prostate-specific Antigen and a Panel of Kallikrein Markers: A Nested Case-Control Study. Eur Urol. 2015;68(2):207–213. doi: 10.1016/j.eururo.2015.01.009 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Loeb S, Roehl KA, Antenor JAV, Catalona WJ, Suarez BK, Nadler RB. Baseline prostate-specific antigen compared with median prostate-specific antigen for age group as predictor of prostate cancer risk in men younger than 60 years old. Urology. 2006;67(2):316–320. doi: 10.1016/j.urology.2005.08.040 [DOI] [PubMed] [Google Scholar]
- 14.Whittemore AS, Lele C, Friedman GD, Stamey T, Vogelman JH, Orentreich N. Prostate-specific antigen as predictor of prostate cancer in black men and white men. J Natl Cancer Inst. 1995;87(5):354–360. doi: 10.1093/jnci/87.5.354 [DOI] [PubMed] [Google Scholar]
- 15.Lilja H, Cronin AM, Dahlin A, et al. Prediction of significant prostate cancer diagnosed 20 to 30 years later with a single measure of prostate-specific antigen at or before age 50. Cancer. 2011;117(6):1210–1219. doi: 10.1002/cncr.25568 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Vickers AJ, Cronin AM, Björk T, et al. Prostate specific antigen concentration at age 60 and death or metastasis from prostate cancer: case-control study. BMJ. 2010;341:c4521. doi: 10.1136/bmj.c4521 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Pinsky PF, Miller E, Prorok P, Grubb R, Crawford ED, Andriole G. Extended follow-up for prostate cancer incidence and mortality among participants in the Prostate, Lung, Colorectal and Ovarian randomized cancer screening trial. BJU Int. 2019;123(5):854–860. doi: 10.1111/bju.14580 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Huynh-Le MP, Karunamuni R, Fan CC, et al. Prostate cancer risk stratification improvement across multiple ancestries with new polygenic hazard score. Prostate Cancer Prostatic Dis. 2022;25(4):755–761. doi: 10.1038/s41391-022-00497-7 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Seibert TM, Fan CC, Wang Y, et al. Polygenic hazard score to guide screening for aggressive prostate cancer: development and validation in large scale cohorts. BMJ. 2018;360:j5757. doi: 10.1136/bmj.j5757 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Pagadala MS, Lynch J, Karunamuni R, et al. Polygenic risk of any, metastatic, and fatal prostate cancer in the Million Veteran Program. J Natl Cancer Inst. 2023;115(2):190–199. doi: 10.1093/jnci/djac199 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.EAU Guidelines on Prostate Cancer - Uroweb. Uroweb - European Association of Urology. Accessed June 7, 2023. https://uroweb.org/guidelines/prostate-cancer
- 22.Wei JT, Barocas D, Carlsson S, et al. Early Detection of Prostate Cancer: AUA/SUO Guideline Part I: Prostate Cancer Screening. J Urol. 2023;210(1):46–53. doi: 10.1097/JU.0000000000003491 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Remmers S, Bangma CH, Godtman RA, et al. Relationship Between Baseline Prostate-specific Antigen on Cancer Detection and Prostate Cancer Death: Long-term Follow-up from the European Randomized Study of Screening for Prostate Cancer. Eur Urol. 2023;84(5):503–509. doi: 10.1016/j.eururo.2023.03.031 [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.