Original Article: Clinical Investigation
Population-based screening for prostate cancer by measuring total serum prostate-specific antigen in Iran
Mohammah Reza Safarinejad md, Urology and Nephrology Research Center, Shaheed Beheshti University of Medical Sciences, PO Box 19395-1849, Tehran, Iran. Email: email@example.com
Objective: To report the results from an Iranian large population-based randomized study of screening using prostate-specific antigen (PSA) to detect prostate cancer.
Materials and Methods: A total of 3758 Iranian men older than 40 years were mass checked by PSA-based screening. Men with an abnormal digital rectal examination (DRE) and serum total PSA level of greater than 4 ng/mL, underwent transrectal ultrasonography (TRUS)-guided extended prostate biopsy.
Results: The PSA value (mean ± standard deviation, SD) in all men without prostate cancer was 1.6 ± 1.1 ng/mL and in those with cancer 18 ± 44.8 ng/mL (P = 0.001). PSA values increased with age. In those aged 40–49, 50–59, 60–69 and ≥70 years, the mean ± SD PSA values were 1.3 ± 0.7, 1.4 ± 0.8, 1.8 ± 1 and 2.2 ± 1.6 ng/mL, respectively. Among the screened men, 323 (8.6%) had a serum PSA concentration greater than 4 ng/mL. Of patients who underwent prostate biopsy (230, 71.2%), 129 (positive predictive value, 56.1%) had prostate cancer. Additionally, nine cancers were detected among 16 patients with PSA of less than 4 ng/mL who had a doubtful DRE finding. The overall cancer detection rate was 3.6%; 1.4% at 40–49, 1.6% at 50–59, 4.2% at 60–69 and 12.9% at ≥70 years. Conventional systematic sextant biopsies, which accounted for six of the 10 cores in our biopsy scheme, detected 98 (71%) of the cancers.
Conclusions: The Iranian male population develops prostate cancer quite commonly if their serum PSA levels are greater than 4.0 ng/mL. In this study, 65.9% of the detected cancers were clinically significant. The conventional systematic sextant technique may be inappropriate for detection of all prostate cancers. The results need to be confirmed in other randomized trials.
There were 679 000 new cases of prostate cancer worldwide in 2002, making this the fifth most common cancer in the world and the second most common in men (11.7% of new cancer cases overall, 19% in developed countries and 5.3% in developing countries).1 Several studies have shown different serum levels of prostrate-specific antigen (PSA) in different races, ethnic groups and populations.2–5 Prostate cancer incidence is characterized by a very large geographic variability. Variations in incidence between different ethnicities that have been observed in international statistics are due to ethnic disparity in risk within certain countries. The black population has the highest incidence (and mortality) rates in the US, some 70% higher than in the white population, who consecutively have rates significantly higher than populations of Asian ancestry (e.g. Chinese, Japanese and Korean males).1 The incidence of prostate cancer in African-American men has been reported to be 50% higher than that of age-matched US Caucasian men, and African-American men have the highest incidence of prostate cancer of any ethnic group in the world.6 Asian countries have much lower rates of occurrence of the disease than North American, and North and Western European countries, with Southern European and South American countries displaying an intermediate incidence rate.7 Very low rates are also observed in China (6.8 per 100 000) and Western Asia (5.8 per 100 000); the lowest recorded rate is 0.4 per 100 000 in Ardabil, north-west Iran.8 In Italy and Spain, prostate cancer incidence, according to estimates in 2000, was ranked third (approximately 10% of all new diagnosed cancers), while in France it was the most common male cancer (approximately 19%).9 In a recent study, the overall detection rate of prostate cancer in Iran was 3.5%.10 The reasons for these racial differences are poorly understood, although genetic, nutritional and socioeconomic factors have all been implicated.11 The average increase in the estimated age-adjusted incidence of prostate cancer worldwide between 1985 and 2002 was around 1.1% annually.1 Hsing et al.12 reviewed data on international trends in prostate cancer incidence and mortality. There were also large increases in low-risk countries between 1975 and 1990: 104% in Singapore; 84% in Miyagi, Japan; 55% in Hong Kong; and 44% in Shanghai, China.
Prostate cancer screening using PSA testing is rather peculiar among other screening interventions. PSA is a valid screening test for prostate cancer, which compares favorably with mammography for breast cancer.13 Several studies have confirmed that PSA-based screening is the most effective screening method; however, most of these studies were done in men referred to urological care settings because of signs and symptoms.14–16 Because Iranian men are ethnically and racially different from most of Asian countries' men (e.g. Japanese, Chinese and Arabic men) the biomedical parameters of prostate cancer should be different. To study this issue, we performed a mass screening for prostate cancer using serum PSA determination as a first-line screening method in Iranian men.
Materials and methods
Between March 2003 and October 2005, a total of 3758 volunteer Iranian men older than 40 years who agreed by informed consent were mass checked by PSA based screening and digital rectal examination (DRE) as the initial tests. The study population included residents of Metropolitan Tehran in Tehran County, Iran. Clinical evaluations comprised DRE, and serum levels of total PSA using a TSH-IRMA PSA assay (Immunotech, Marseille, France). Blood samples were collected at room temperature by venipuncture without anticoagulant. Serum was separated after complete clot retraction and stored at −70°C until further analysis. None of them had a history of prostate cancer and males with a history of prostatitis and of prostatectomy or other conditions that interfered with voiding, were excluded from the study. Use of any medication or food supplement that could influence the PSA value, such as finasteride, saw palmetto and any androgen or estrogen, dehydroepiandrosterone, herbs and vitamins, was recorded. If the DRE was clinically suspicious of malignancy, or the serum PSA was >4 ng/mL, the patient was referred for transrectal ultrasonography (TRUS)-guided 10-core prostate biopsy. In addition to the conventional sextant biopsies, four cores were addressed to far lateral and transition zones. Hypoechoic lesions were biopsied separately. Palpable abnormalities were characterized according to the International Union Against Cancer (UICC) 1992.17 For the prostate biopsy, with informed consent provided, TRUS was performed using General Electric Logic 500 and using 1.2 mm Trucut needles (Collibri Medical, Helsingborg Sweden). The histological criteria for prostate cancer used were those of the World Health Organization (WHO).18 The clinical stage was evaluated according to the tumor node metastasis (TNM) classification,19 UICC and the American Joint Committee on Cancer (AJCC). Cancers were graded according to the Gleason's score system.20 One urological pathologist did all pathological evaluations. Information on the patient's age, race, DRE result, serum PSA level, and biopsy result (cancer or no cancer) was entered into a computerized database for analysis. The initial and second screenings were performed at the same institution.
The distribution of PSA levels was highly skewed to the right, and therefore a logarithmic transformation was first applied to the data. Group comparisons were performed using two-sample t-tests (two groups) or anova (more than two groups). All values are reported as mean ± standard deviation (SD). Differences in incidence between groups were assessed using the χ2 and Fisher exact tests. Statistical analysis was performed using the computer statistical package SPSS version 10.0 (SPSS, Chicago, IL, USA) and SAS version 6.4 (SAS Institute Cary, NC, USA).
A total of 3758 subjects were recruited in this study. The mean age of these men was 63 years ranging 40–78 years old. Of these, 351 (9.3%) were between 40 and 49, 1810 (48.2%) were between 50 and 59, 1088 (29%) were between 60 and 69, and 509 (13.5%) were ≥70 years of age. The DRE was recorded as clinically suspicious of malignancy in 47 (1.3%) overall. Among the screened men, 323 (8.6%) had a serum PSA concentration >4 ng/mL. The number of males with PSA levels above cut-off value, were 12 (3.7%), 67 (20.7%), 99 (30.7%) and 145 (44.9%) in 40–49, 50–59, 60–69 and ≥70 age groups, respectively. The PSA value (mean ± SD) in all men without prostate cancer was 1.6 ± 1.1 ng/mL and in those with cancer 18 ± 44.8 ng/mL (P = 0.001). PSA values increased with age. In those aged 40–49, 50–59, 60–69 and ≥70 years, the mean ± SD PSA values were 1.3 ± 0.7, 1.4 ± 0.8, 1.8 ± 1 and 2.2 ± 1.6 ng/mL, respectively (Table 1). The resultant values for patients with prostate cancer were 14.4 ± 22.3, 22.6 ± 33.2, 36.5 ± 112 and 51.4 ± 124.2, respectively.
Table 1. Age distribution, mean value, number of biopsies and number of carcinomas stratified by prostrate-specific antigen (PSA) ranges in studied population
|40–49||356 (9.5)||1.3 ± 0.7||2 (0.5)||7 (2.0)||5 (1.4)||10 (71.4)||5 (50)||1.4|
|50–59||1856 (49.4)||1.4 ± 0.8||7 (0.4)||41 (2.2)||26 (1.4)||55 (74.3)||30 (54.5)||1.6|
|60–69||1113 (29.6)||1.8 ± 1.1||5 (0.5)||41 (3.7)||58 (5.2)||85 (81.7)||47 (55.3)||4.2|
|≥70||433 (11.5)||2.2 ± 1.6||2 (0.5)||47 (10.9)||98 (22.6)||96 (65.3)||56 (58.3)||12.9|
|Totals||3758 (100)||1.6 ± 1.0||16 (0.4)||136 (3.6)||187 (5)||246 (72.6)||138 (56.1)||3.6|
Of the 323 patients with a serum PSA level of >4 ng/mL, 230 (71.2%) underwent prostate biopsy. Biopsy was not done because of concomitant urinary tract infection in 12 (3.7%), risk of discontinuation of anticoagulants in 11 (3.4%), poor general condition in 16 (5%) and patient rejection in 54 (16.7%). There were no significant differences in age (61.5 vs 62 years, Mann–Whitney test P = 0.2), serum PSA (6.1 vs 6.2 ng/mL, P = 0.4), and prostate volume (48.3 vs 49.1 mL, P = 0.6) between patients who underwent biopsy and those who did not. Of patients who underwent prostate biopsy, 129 (positive predictive value [PPV] 56.1%) had prostate cancer. Additionally, nine cancers were detected among 16 patients with a PSA of <4 ng/mL who had a doubtful DRE finding. Overall prostate cancer was diagnosed in 138 men (PPV, 56.1%), whereas the remainder (108) had benign tissue alone in their biopsies. The overall cancer detection rate was 3.6%; 1.4% at 40–49 years, 1.6% at 50–59, 4.2% at 60–69 years and 12.9% at ≥70 years; the detailed characteristics are given in Table 1. Of 230 biopsied patients, 97 (42.2%) had a serum PSA level between 4 and 10 ng/mL. Thirty-four percent of these patients (33 from 97) had prostate cancer (PPV, 34%, detection rate 0.9%; Table 2). The positive predictive value and detection rate for serum PSA level >10 ng/mL were 72.2 and 2.5%, respectively. We found foci of high-grade prostatic intraepithelial neoplasia (PIN) concurrently with prostate cancer in 6 patients (4.3%). Among the patients biopsied for the first time 3 (1.2%) had high-grade PIN without having prostate cancer and 45 (18.3%) had microacinar proliferation suspicious for malignancy.
Table 2. Clinicopathological characteristics of patients diagnosed with prostate cancer stratified by PSA
|Mean age ± SD||61 ± 5.4||67 ± 5.7||71 ± 6.8|
|Mean percentage positive cores ± SD||8.7 ± 7.9||16.6 ± 15.8||28.8 ± 21.3|
|Mean max Ca length (mm) ± SD||2.0 ± 1.3||2.5 ± 1.4||5 ± 3.4|
|Mean PSA density (ng/mL/mL) ± SD||0.13 ± 0.03||0.22 ± 0.12||0.32 ± 0.16|
|No. Gleason's score (%):|
| <6||7 (78)||18 (55)||11 (12)|
| 6||1 (11)||12 (36)||58 (60)|
| 7||1 (11)||2 (6)||20 (21)|
| ≥8||0 (0)||1 (3)||7 (7)|
|No. clinical stage (%)†|
| T1c-2N0M0||9 (100)||27 (82)||6 (6)|
| T3N0M0||0 (0)||3 (9)||80 (83)|
| N1 or M1||0 (0)||3 (9)||10 (10)|
|No. clinically significant cancer (%)‡||1 (11)||17 (52)||73 (76)|
Conventional systematic sextant biopsies, which accounted for six of the 10 cores in our biopsy scheme, detected 98 (71%) of the cancers. Cancer was detected only in an alternate site in 29% of cases, including 25 in the anterior horn right and/or left and 15 in the transition zone. Only one core positive for cancer was seen in 58 (42%) men, 28 of whom had a positive alternate site. Gleason's scores were 6 in 45 patients, 7 (4 + 3) in nine, and 8 in four. The length of cancer in these biopsies was 3 mm or less in 36 patients with a Gleason's score of 6. Of 38 cases with two cores positive for cancer, the Gleason's score was 6 in 26, 7 (4 + 3) in eight and 8 in four. Of 42 cases with three or more cores positive for cancer, including one with 5 positive cores, Gleason's score was 7 in 14 (3 + 4 in four and 4 + 3 in one) and 8 in three (Table 2).
Of 129 males with prostate carcinoma, DRE or TRUS showed abnormal findings in 42 (32.6%) and 48 (37.2%), respectively. The proportion of males with PSA levels above the cut-off value increased with age. The numbers of cancer cases detected for age groups 40–49, 50–59, 60–69 and ≥70 years of age were five (3.6%), 30 (21.7%), 47 (34.1%) and 56 (40.6%), respectively (Table 1).
Detection rates with serum PSA levels of 4–10 ng/mL were 0.3, 0.5, 0.6 and 3.5% in 40–49, 50–59, 60–69 and ≥70 years age groups, respectively. For serum PSA levels of >10 ng/mL these were 0.6, 0.9, 3.4 and 9.0, respectively (Table 3). Of 138 detected cancers six were stage M1, seven were stage N+, 42 were stage T1c-2N0M0 and 83 were stage T3N0M0. At biopsy, 26.1% were Gleason's score <6, 51.4% were Gleason's score 6, 16.7% were Gleason's score 7 and 5.8% were Gleason's score ≥8 (Table 2). Clinically significant disease was defined based on the definition proposed by Terris et al. that is more than one positive core, Gleason's score ≥7 or a cancer focus in one core greater than 3 mm.20 Of 138 patients with prostate cancer, 91 (65.9%) had clinically significant disease. The mean Gleason's score was 7.5 (range, 7–9) for peripheral zone (PZ) cancers, 6.2 (range, 4–8) for transition zone (TZ) cancers, and 7.4 (range, 7–9) originating from the TZ as well as the PZ.
Table 3. Average value, number of biopsies and number of carcinomas in men stratified by age ranges
|40–49||7 (2.0)||4 (57.1)||1 (25)||5.6 ± 1.2||0.3||5 (1.4)||4 (80)||2 (50)||22.4 ± 46.2||0.6|
|50–59||41 (2.2)||24 (58.5)||10 (41.6)||6.4 ± 1.4||0.5||26 (1.4)||24 (92.3)||17 (70.8)||36.7 ± 56.8||0.9|
|60–69||41 (3.7)||30 (73.1)||7 (23.3)||7.4 ± 1.3||0.6||58 (5.2)||50 (86.2)||38 (76)||45.7 ± 124.6||3.4|
|≥70||47 (10.9)||39 (83)||15 (38.5)||8.4 ± 1.2||3.5||98 (22.6)||55 (56.1)||39 (70.9)||67.2 ± 128.2||9.0|
|Totals||136 (3.6)||97 (71.3)||33 (34)||7.1 ± 1.2||0.9||187 (5)||133 (71.1)||96 (72.2)||44.3 ± 86.7||2.5|
Comparative data from the world's published work on serum PSA reference values in healthy subjects of different populations have clearly demonstrated slight but remarkable discrepancies. The proportion of cancers among those with a positive screening test is different, at 56.1% (129/230) in the present study, to the 19.3% in Andriole and Catalona,22 21.9% in Labrie et al23 and 20.1% in Catalona et al.24 In Western countries, the proportion of PSA abnormalities has been reported to be approximately 10% of subjects when the cut-off value for PSA is 4.0 ng/mL.15–17 In a study from California, a total of 297 consecutive patients with an elevated serum PSA and/or abnormal DRE underwent transrectal ultrasound with lesion-directed and systematic biopsy. Ethnic analysis was performed for the entire group, patients with a normal DRE, and those with a normal DRE and PSA of 4–10 ng/mL. Of the 297 patients, 131 (44%) had cancer, including 48 of 97 black (50%) and 83 of 200 white (42%) patients.25 In our study, the overall cancer detection rate was 3.6%. The similar detection rates reported by Catalona et al.24 of 3.1% and Labrie et al.23 of 3.5%. Grubb et al. reviewed records of men 45–59 years old at study entry with a PSA less than 2.6 ng/mL and benign DRE who underwent annual DRE and PSA testing in a screening study between 1991 and 2001. Of 10 174 men with these characteristics, 232 (detection rate, 2.3%) were subsequently diagnosed with prostate cancer.26 From a prostate cancer screening study Nadler et al. studied 782 consecutive men who underwent prostate biopsy for PSA greater than 2.5 ng/mL or suspicious digital rectal examination. Biopsy results were evaluated as a function of patient age. Overall PSA between 2.6 and 4.0 ng/mL was associated with a cancer detection rate of 16.2% using a sextant biopsy technique.27 In a study from Japan, urological patients 70 years or younger tested for prostate cancer were studied. There were 550, 97, 112 and 52 patients with a PSA of less than 2.5, 2.5–4.0, 4.1–10.0 and more than 10.0 ng/mL, respectively. Transrectal 10-core prostate biopsy was performed in 80 (82%) of the 97 patients with a PSA of 2.5–4.0 ng/mL and 102 (91%) of the 112 patients with a PSA of 4.1–10.0 ng/mL. Cancer detection rates in patients who underwent biopsy were 26.3% and 34.3% at PSA levels 2.5–4.0 and 4.1–10.0 ng/mL, respectively.28 In a study from the University of Texas, men who had serum PSA levels between 2.5 and 4.0 ng/mL were asked to undergo prostate biopsy. Of 268 eligible men, 151 (56%) agreed to participate in this trial. All men underwent biopsy using an 11-core multisite-directed biopsy scheme. Cancer was identified in 24.5% (37 of 151) of the men biopsied.29 Our detection rate is lower than most of previous studies using the same or stricter contemporary biopsy scheme. A possible explanation is that the higher cancer detection rate in some studies merely reflects a higher biopsy rate compared with screening population; however, there is little evidence to explain it. In another study from Iran, 3670 Iranian men older than 40 years were mass checked by PSA-based screening. They were invited to have a DRE, serum PSA assay and TRUS-guided sextant prostate biopsy if the DRE was clinically suspicious of malignancy, the serum PSA was ≥2.1 ng/mL or free-to-total PSA (f/tPSA) ratio ≤15%. In 433 (11.8%) of screened males, tPSA levels exceeded the cut-off value of ≥2.1 ng/mL and 128 prostate cancers were diagnosed (PPV, 29.6%). Altogether, 138 cancers were detected (detection rate 3.8%).10 In a study from Sweden, 9973 men aged 50–66 years were randomized to screening and to future controls. Men with a total PSA of ≥3.0 ng/mL were investigated with a DRE, TRUS and sextant biopsies. Of those invited, 60% accepted PSA testing and 11.3% had a total PSA of ≥3.0 ng/mL. Altogether 145 cancers were detected (PPV, 24%). The overall cancer detection rate was 2.5%.30
In most screening studies, patient compliance with further evaluation including transrectal needle biopsy presented a problem, as can be seen from the biopsy rates reported in the published work, which range between 50 and 80%.14,31,32 Heyns et al. reported that there is a significant problem in getting men with an elevated serum PSA level to undergo a prostate biopsy in the primary healthcare setting in South Africa.33 In a study from the USA, where men with a serum PSA of 2.5–4.0 ng/mL were asked for a prostate biopsy, 56% of the eligible men agreed to participate.29 We were able to obtain biopsies 71.2% with elevated PSA. Catalona et al. reported that 83% of patients with PSA levels between 2.6 and 4.0 ng/mL who underwent prostatectomy had significant disease defined as greater than 0.5 cm3 tumor volume.34 In this screening study, 62.8% had clinically significantly disease. There are studies indicating that an extended core biopsy protocol gives significantly better detection rates for prostate cancer than the standard sextant biopsy alone.35–37 Comparison of the conventional systematic sextant biopsy approach and a 10-core multisite-directed method revealed cancer detection rates of 71% for systematic sextant biopsy.
The significance of our study is that approximately 71.2% of all eligible patients underwent biopsy, reducing the effect of bias on the result. There were no significant differences between men with negative and those with positive biopsy results in regard to abnormal findings on DRE. Furthermore, no patient in our study had ever taken finasteride.
Investigators at John Hopkins Hospital used a mean ± 3 standard deviations and reported an upper limit of normal as 2.0 ng/mL for men less than 40 years of age and 2.8 ng/mL for men older than 40 years without any prostatic disease.2,38 Whether lowering the PSA cut-off value to 2 ng/mL will raise the disease specific survival rate remains to be determined. Biopsy in men with PSA levels between 2.6 and 4.0 ng/mL may detect clinically significant prostate cancer more frequently at an organ-confined stage, with a lower PSA progression rate.39 An international multicenter study coordinated by Boerhinger Mannheim from Germany in 199640 covered more than 70 centers from 18 countries (Austria, Canada, Czechin, France, Germany, India, Israel, Italy, Japan, Saudi Arabia, Netherlands, New Zealand, Poland, Portugal, South Africa, Spain, Sweden, Turkey and the UK). The estimation of serum PSA in 1063 asymptomatic healthy men revealed a minimum serum PSA value as 0.001 and maximum as 14.8 ng/mL. The median total PSA concentration for all ages was 0.77 ng/mL. Mettlin et al.41 reported the mean serum PSA as 2.1 ng/mL in 2011 a healthy American white population with a mean average age of 64.5 ± 5.0 years. The average mean PSA concentration for all age groups in our study was 1.6 ng/mL for 3435 healthy men. Also, higher detection rates have been reported when other screening tools have been incorporated in the primary screening. Bangma et al.42 studied 1726 men aged 55–76 years. They combined total serum PSA with a DRE. Cancer detection rate was 4.0% from the biopsy of men with an abnormal DRE, TRUS or total PSA of ≥4.0 ng/mL. Gustafsson et al.43 used an abnormal DRE or TRUS, or total PSA of ≥10 ng/mL as indications for biopsy in 1782 men aged 55–70 years. The reported cancer detection rate was 3.6%.
In our study, the age limit was lowered to 40 years. Screening of younger men will produce a lower, but still significant yield of prostate cancers, but younger men have a greater potential to benefit because they have a longer life expectancy and are more likely to have organ-confined disease. The present study reveals that PSA alone can be used efficiently as a first-line and as a repeat screening test for prostate cancer. PSA levels >4 ng/mL deserve prostatic biopsy, even if DRE and TRUS findings are unremarkable.
The overall detection rates of an Iranian population with a PSA of >4.0 ng/mL are higher than those among community volunteers in most Western and Asian countries with the same PSA range. The majority of these cancers are clinically significant. Furthermore, if a decision is made to perform a biopsy when serum PSA is >4.0 ng/mL, the conventional systematic sextant technique may be inadequate for the detection of all clinically significant cancers. Given the results of this study, we hypothesize that racial differences in PSA production arise either prior to and/or independently of cancer. The possibility that the racial differences in cancer detection rate in our study may be due to socioeconomic factors, needs further study. Our results clearly need to be confirmed by other, larger studies.