Journal of the National Cancer Institute Advance Access originally published online on September 11, 2007
JNCI Journal of the National Cancer Institute 2007 99(18):1395-1400; doi:10.1093/jnci/djm119
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Published by Oxford University Press 2007.
ARTICLES |
Detection of Prostate Cancer via Biopsy in the Medicare–SEER Population During the PSA Era
Affiliations of authors: Department of Veterans Affairs Medical Center, White River Junction, VT (HGW, ESF); Institute for Health Policy and Clinical Practice at Dartmouth, Hanover, NH (HGW, ESF, DJG); Massachusetts General Hospital, Harvard Medical School, Boston, MA (MJB)
Correspondence to: H. Gilbert Welch, MD, MPH, VA Outcomes Group (111B), Department of Veterans Affairs Medical Center, White River Junction, VT 05009 (e-mail: h.gilbert.welch{at}dartmouth.edu).
 |
ABSTRACT |
|---|
 Top Abstract Context and CaveatsSubjects and MethodsResultsDiscussionFundingReferencesNotes |
|---|
Background: Despite the considerable attention given to the prostate-specific antigen (PSA) as a screening test for prostate cancer, it is needle biopsy—and not the PSA test result—that actually establishes the diagnosis of prostate cancer. We sought national estimates on the proportion of men found to have prostate cancer after a needle biopsy of the prostate and the risk of subsequent biopsies among those not found to have prostate cancer.
Methods: We linked Medicare claims data to Surveillance, Epidemiology, and End Results (SEER) data to analyze outcomes after 10429 needle biopsies performed in 1993 through 2001 in 8273 men aged 65 years and older enrolled in Medicare Part B who resided in a SEER area. We determined the proportion of needle biopsies that were followed by a diagnosis of prostate cancer, the cumulative risk of prostate cancer following multiple biopsies, and the risk of subsequent biopsy among men not found to have prostate cancer in the previous biopsy. All statistical tests were two-sided.
Results: The overall proportion of needle biopsies found to contain prostate cancer was 32% (95% confidence interval [CI] = 31% to 33%). The yield increased with age (26% for men aged 65–69 years, 31% for men aged 70–74 years, 35% for men aged 75–79 years, and 41% for men aged 80 years and older; Ptrend<.001). The cumulative risk of prostate cancer diagnosis increased with repeated biopsy, with 50% of men receiving a prostate cancer diagnosis after two biopsies, 62% after three biopsies, and 68% after four biopsies. Among men whose first recorded biopsy did not detect prostate cancer, the risk of having a subsequent biopsy was 11.6% (95% CI = 11% to 12%) at 1 year and 38% (95% CI = 36% to 40%) at 5 years.
Conclusions: About one-third of prostate biopsies identified prostate cancer in this population. Men not found to have prostate cancer on a first biopsy frequently undergo repeat biopsies, which raise the cumulative risk of prostate cancer diagnosis.
Prior knowledge To fully characterize the effects of prostate-specific antigen screening for prostate cancer, it is important to develop estimates of biopsy yield—i.e., the proportion of men whose prostate biopsies are found to contain prostate cancer. Study design Medicare claims data were linked to the National Cancer Institute's Surveillance, Epidemiology, and End Results to analyze outcomes after needle biopsies of the prostate in men aged 65 years and older. Contribution One-third of needle biopsies overall were found to contain prostate cancer, and the yield increased with age. The yield of second and subsequent biopsies was less than that of the first recorded biopsy, but the cumulative risk of prostate cancer increased with repeat biopsies. More than one-third of men whose first recorded biopsy did not detect prostate cancer had a subsequent biopsy within 5 years. Implications Multiple biopsies increase the likelihood that a man will be diagnosed with prostate cancer. Limitations The study was restricted to men aged 65 years and older, and its relevance to younger men is not known. Information on why biopsies were performed was not available. The number of biopsy cores sampled was not known.
|
The incidence of prostate cancer was rising relatively slowly between the early 1970s to the late 1980s. However, after prostate-specific antigen (PSA) screening became available in the United States in 1987, the incidence of prostate cancer rose sharply to a peak in 1992 and then fell to a new, higher baseline (1). As a result, the estimated lifetime probability of being diagnosed with prostate cancer for American men increased from about 7% in 1975 (2) to about 18% by 2002 (1). Whether the additional detection during the PSA era has resulted in lower morbidity and mortality from prostate cancer or simply more overdiagnosis is unknown.
However, it is a prostate biopsy, and not the PSA test, that actually establishes the diagnosis of prostate cancer. And several questions about the biopsy process have surfaced during the PSA era: What is the right PSA threshold for performing a biopsy? How many biopsy cores should be taken? Should a biopsy be repeated if the original set is negative? In general, growing appreciation of the limited sensitivity of both the PSA test and prostate biopsy has led advocates of increased detection to recommend that the PSA threshold for biopsy be lowered below the current threshold of 4 ng/mL (3,4), that more cores be taken at each biopsy (5–7), and that biopsies be repeated for selected patients whose original biopsies were negative (8–10). To the extent that these recommendations are followed, they would be expected to substantially increase the cumulative risk of prostate cancer diagnosis.
How these recommendations have affected the performance and yield of prostate biopsies in the broad spectrum of older men being evaluated for prostate cancer in community settings in the United States is unclear. To address this issue, we examined the yield of prostate biopsies (i.e., the percentage of biopsies that were found to contain cancer) and the likelihood and yield of subsequent biopsies in a large population-based cohort of men aged 65 years and older who underwent a prostate biopsy in the PSA era following the incidence peak associated with the introduction of the test.
|
Subjects and Methods |
|---|
|
TopAbstractContext and CaveatsSubjects and MethodsResultsDiscussionFundingReferencesNotes |
|---|
Overview
We used the Medicare National Claims History System (source: Health Care Financing Agency, Baltimore, MD) and the National Cancer Institute's Surveillance, Epidemiology, and End Results (SEER) data to identify a cohort of men who both resided in a SEER area and underwent a needle biopsy of the prostate from January 1, 1993, through December 30, 2001. We determined both the biopsy yield (i.e., the proportion of biopsies that resulted in a cancer diagnosis) and the risk of subsequent biopsy among men in whom cancer was not found.
Study Population
From the 5% Medicare Part B sample of claims for physician services, we identified all men with a claim for a needle biopsy of the prostate (identified using Current Procedure Terminology code 55700) during the period 1993–2001. To ensure that we could capture the diagnosis of prostate cancer, we needed to link the Medicare beneficiary identification number to the SEER data. Although a linked Medicare–SEER dataset already exists, it includes data on only beneficiaries with cancer and a sample of beneficiaries without cancer. Because we wanted to determine yield, we needed data on an entire population of men undergoing prostate biopsy (i.e., both those who did and those who did not receive a prostate cancer diagnosis). Thus, we needed to create a new Medicare–SEER linkage. We used the Medicare denominator file to restrict the analysis to men who resided in one of the SEER-11 areas (which include five states [Connecticut, Hawaii, Iowa, New Mexico, and Utah] and six metropolitan areas [Atlanta, Detroit, Los Angeles, San Francisco, San Jose, and Seattle]) both in the year of the biopsy and the subsequent year. In other words, we created a dataset of prostate biopsies that were performed in men who we could ensure resided in a SEER-11 area at the time of the biopsy—a restriction that was necessary to capture the diagnosis of prostate cancer in SEER. This study was approved by the Dartmouth Medical School Institutional Review Board.
We then excluded three groups of biopsies using demographic exclusion criteria. First, we excluded biopsies in men who were not aged 65 years and older at the start of each calendar year (both because individuals who become eligible for Medicare for reasons other than age represent a unique group and because those who become eligible midyear will not have complete data). Second, we excluded biopsies in men who were not entitled to Part B for the entire year (because diagnostic services provided to them are not reported in the claims). Third, we excluded biopsies in men enrolled in risk-contract managed care plans (for the same reason).
We then excluded two more groups of biopsies using clinical exclusion criteria. First, we excluded biopsies in men with a concomitant transurethral resection of the prostate, defined as one within a 5-month window (1 month before to 3 months after) surrounding the prostate biopsy, because it would not be clear whether to attribute a prostate cancer diagnosis to the surgery or to the biopsy. Second, we excluded biopsies in men with a prior prostate cancer diagnosis. These exclusions and their net effect on the number of patients in the cohort are illustrated in Fig. 1.
|
Measures
Patient Attributes and Area Biopsy Rate. Patient age and race were determined from the Medicare denominator file. The average biopsy rate in each SEER area over the period was determined as the total number of biopsies in 1993–2001 divided by the total person-years of observation among eligible men in the 5% Medicare sample who resided in that area. We then categorized the SEER areas into approximate terciles based on their biopsy rate: low (<20 biopsies per 1000 man-years), medium (20–25 biopsies per 1000 man-years), and high (>25 biopsies per 1000 man-years). To examine trends across time, we also calculated age- and race-adjusted annual rates in each area and in all areas combined.
Diagnosis of Prostate Cancer. A member of the cohort was determined to have developed prostate cancer if he appeared in the SEER data as having been diagnosed with incident cancer with a primary site International Classification of Disease for Oncology code of C619. We also used SEER data to obtain the date of diagnosis. Because SEER reports only the month of diagnosis, we considered a prostate cancer diagnosis with a date that was within 1 month before and 3 months after the biopsy date as being attributable to the biopsy (e.g., given a biopsy date of December 4, 1993, a prostate cancer diagnosis entered as occurring in November or December 1993 or January, February, or March 1994 would be attributed to the biopsy). We included the month before the biopsy because we often found that more men than expected (3%–6% of the total) were diagnosed with prostate cancer during this period (by contrast, <1% were diagnosed 2, 3, or 4 months before the biopsy). (It should be noted that the apparent diagnosis a month before biopsy may represent SEER linking the date of diagnosis to the PSA test that prompted the biopsy or a data entry error.) If two or more biopsies occurred during the 5-month window, the cancer diagnosis was attributed to the last biopsy performed. The preceding biopsies were thus considered not to have detected prostate cancer.
Statistical Analysis
Yield Analysis.
Overall yield was determined as the proportion of biopsies that resulted in a prostate cancer diagnosis. We also performed yield analyses stratified by age (65–69, 70–74, 75–79, 
80 years), race (black, other), and SEER area and examined the yield of repeat biopsies and calculated the cumulative risk of prostate cancer diagnosis following multiple biopsies. Each of the stratified analyses was adjusted for the other covariates (e.g., the yield of repeat biopsy was adjusted for age, race, and SEER area). Univariate and bivariate analyses (e.g., two-sample comparisons, tests of trend, simple correlation) were performed with STATA 9.0 (StataCorp LP, College Station, TX); multivariable analyses were performed with SAS 9.1 (SAS Institute Inc, Cary, NC). All statistical tests were two-sided.
Risk of Subsequent Biopsy. We also performed a survival analysis on the risk of a subsequent biopsy among men whose first recorded biopsy was negative for prostate cancer. In this analysis, the event of interest was another needle biopsy of the prostate. Two censoring events reflected a patient no longer being at risk for a subsequent biopsy: death or a diagnosis of prostate cancer. In addition, three other censoring events reflected being lost to follow-up: moving out of a SEER area (making us unable to identify a prostate cancer diagnosis, a censoring event), dropping Medicare Part B, or moving into a risk-contract managed care plan (both of which make us unable to detect needle biopsy, the event of interest).
We analyzed the risk of subsequent biopsy by using two time horizons: 1 year and 5 years. We also stratified the analysis for various subgroups: three age groups (65–69, 70–74, 75 years), two race groups (blacks, all others), three biopsy rate groups (low, medium, high), and biopsy number (first, second, third recorded biopsy). For each subgroup, we used a multivariable Cox model to predict the risk of subsequent biopsy while adjusting simultaneously for year of diagnosis (three categories: 1993–1995, 1996–1998, 1999–2001), age group, race, area biopsy rate, and biopsy number.
|
Results |
|---|
|
TopAbstractContext and CaveatsSubjects and MethodsResultsDiscussionFundingReferencesNotes |
|---|
Characteristics of the 10429 needle biopsies in our analysis are shown in Table 1. Overall, prostate cancer was detected in 32% of the biopsies (95% CI = 31% to 33%). The yield increased statistically significantly with patient age (26% for men aged 65–69 years, 31% for men aged 70–74 years, 35% for men aged 75–79 years, and 41% for men aged 80 years and older; Ptrend<.001) and was higher in blacks than in others (37% versus 31%, P<.001). This finding might be expected given that prostate cancer incidence in blacks is higher than average (1).
|
There was no trend of either increasing or decreasing yield across the 9-year analysis period (Fig. 2) (Ptrend>.2), yet the yield did vary considerably across the SEER areas (data not shown). In two areas, yields were lower than the overall average (Atlanta [25%] and Los Angeles [26%], both P<.001), and in two areas yields were higher than the average (Iowa [37%, P = .006] and Seattle [38%, P = .001]). Although the biopsy rate declined over the first 3 years (Fig. 2), there was not a statistically significant trend over the 9-year analysis period (Ptrend>.2). There was little evidence of a relationship between biopsy rate and yield, either across years (Fig. 2) or across area–years (Fig. 3).
|
|
The proportion of biopsies that resulted in a prostate cancer diagnosis fell with repeat biopsy (34% [95% CI = 33% to 35%] for the first recorded biopsy, 25% [95% CI = 23% to 27%] for the second, 24% [95% CI = 20% to 28%] for the third, 21% [95% CI = 15% to 28%] for the fourth or greater; Fig. 4). If all men with a negative first recorded biopsy are equally likely to undergo a repeat biopsy, one can calculate the effect of a strategy of repeat biopsy on cumulative yield. The cumulative risk of a prostate cancer diagnosis increased if a strategy of repeat biopsy was employed (Fig. 5): the risk was 50% following two biopsies, 62% following three biopsies, and 68% following four biopsies.
|
|
Risk of Subsequent Biopsy
Among men in whom prostate cancer was not found on their first recorded biopsy, the risk of a subsequent biopsy after 1 year was 11.6% (95% CI = 11% to 12%). This risk decreased with age, was slightly higher for blacks than for others, and increased with the biopsy rate in the SEER area (Fig. 6). The 1-year risk also increased with the number of biopsies performed: the risk of a subsequent biopsy after the first recorded biopsy was 10% (95% CI = 9% to 11%), after the second biopsy was 14% (95% CI = 13% to 16%), and after the third biopsy was 18% (95% CI = 15% to 21%) (Ptrend<.001).
|
After 5 years, among all biopsies that did not detect prostate cancer, the risk of a subsequent biopsy was 38% (95% CI = 36% to 40%). The general patterns of the 5-year risk data for various subgroups were similar to the 1-year data (Fig. 7). Like the 1-year risk, the 5-year risk also increased with the number of biopsies performed: the risk of a subsequent biopsy after the first recorded biopsy was 35% (95% CI = 33% to 36%), after the second biopsy was 45% (95% CI = 41% to 48%), and after the third biopsy was 54% (95% CI = 48% to 60%) (Ptrend<.001).
|
|
Discussion |
|---|
|
TopAbstractContext and CaveatsSubjects and MethodsResultsDiscussionFundingReferencesNotes |
|---|
In this analysis of data for our linked Medicare–SEER population, we found that approximately one-third of the needle biopsies performed during 1993–2001 detected prostate cancer. We also found that yield increased with age, as has previously been reported (11). In addition, the yield was slightly higher in blacks than in all other racial groups combined (about a 6% absolute difference), although the difference was greater in previous reports (10%–14%) (12,13). Finally, the yield in a SEER area appeared to have little relationship to the area biopsy rate.
We also found that men with a negative index biopsy frequently underwent a subsequent biopsy. More than one-third of men whose first recorded biopsies were negative were ultimately rebiopsied in the next 5 years. For men younger than 70 years, the risk of a repeat biopsy was 44%. Moreover, men who received a second negative biopsy were even more likely to undergo a third biopsy. Although it is possible that men who undergo repeat biopsy examinations represent a group that is at unusually high risk of prostate cancer (e.g., they may have a persistent and substantial PSA elevation after a negative biopsy), our finding that yield did not increase with repeat biopsy suggests that this is not the case. Indeed, the finding that yield declined with repeat biopsies suggests that the more obvious cancers (i.e., those involving a large portion of the gland) are identified on the initial biopsy and are not available to be detected on subsequent biopsies. If we assume that the decision to repeat the biopsy is largely a random event, our data provide some insight on the cumulative risk of prostate cancer diagnosis when a strategy of repeat biopsy is employed: the more biopsies, the more likely it is that cancer will be found.
Our analysis has a number of limitations. First, our use of Medicare data means that our conclusions must be limited to those men aged 65 years and older. The relationship that we observed between age and subsequent biopsy, however, raises the possibility that the risk of repeat biopsy may well be higher in younger populations.
Second, we could not determine why biopsies were performed. Some biopsies may have been initiated by symptoms (e.g., a change in urinary function), some by signs (e.g., an abnormal digital rectal exam), and others by screening (e.g., an elevated PSA level). We suspect, however, that the vast majority were prompted by an elevated PSA level, given both the time period of the study and the fact that our yield almost exactly matches the 30% yield reported in a study of almost 30000 men with a PSA level greater than 4 ng/mL (14).
Third, we do not know the "intensity" of the biopsy process experienced by the men in our study. The proportion of men found to have prostate cancer is directly related to how aggressively the prostate is biopsied. Historically, prostate biopsies have consisted of six needle cores, but now many urologists advocate sampling 12 or more needle cores, noting that the more biopsies taken, the more cancer is found (15–18). Some even advocate "saturation biopsy" (i.e., 32–38 needle cores), pointing out that microscopic cancers can still be found in men who had been cancer-free on three or more previous biopsies (19). Given these changing views of how aggressively cancer in the prostate should be sought, it is likely that the average number of cores taken per biopsy began to increase over our study period, although we do not have data on this. The finding of a stable yield over the period, however, suggests either that there was little change in intensity or that counterbalancing trends were also present (e.g., the PSA level that triggered a biopsy was lowered).
Finally, we do not know whether any men in our cohort had biopsies before 1993. In other words, the biopsy that we label as being "first recorded" represents the first biopsy in our data, not necessarily the first biopsy that the patient received. Consequently, the estimated risks for subsequent biopsy should be seen as being applicable to the population of men undergoing biopsy in general and not simply to men undergoing their first biopsy. To the extent that our analysis labels repeat biopsies as first recorded biopsies, our yield estimates would be biased downward (because yields fall with repeat biopsies).
Our analysis highlights the close linkage between the decision to biopsy and the diagnosis of prostate cancer. Roughly one-third of the biopsies in our study detected prostate cancer. And although the yield of prostate cancer declined slightly with subsequent biopsy, the cumulative probability of prostate cancer diagnosis rose sharply. Because the goal of screening is to reduce deaths from the disease and not simply to find more cancer, the utility of repeat biopsy warrants careful reassessment.
Our prior work (20) indicates that one tactic to reduce overdiagnosis might be to raise the PSA threshold at which a repeat biopsy is performed following a negative biopsy, perhaps to 10 ng/mL. The findings of Carter et al. (21) would suggest an alternative tactic, namely, restricting repeat biopsy to those men with rapidly rising PSAs (in whom, because PSA velocity increases directly with the absolute PSA level, the precise threshold would have to be adjusted for the patient's baseline PSA level). Either approach would identify a number of men that more closely approximates the number at risk for prostate cancer death and would limit the number of men treated needlessly. Although some physicians might worry that a more conservative approach would miss lethal disease that could be cured by early therapy, it is incumbent on those making such an argument to demonstrate that the benefits for the few exceed the harms of exposing so many to biopsy and treatment.
|
Funding |
|---|
|
TopAbstractContext and CaveatsSubjects and MethodsResultsDiscussionFundingReferencesNotes |
|---|
National Cancer Institute (CA107124 [GenBank] ).
|
NOTES |
|---|
|
TopAbstractContext and CaveatsSubjects and MethodsResultsDiscussionFundingReferencesNotes |
|---|
The views expressed herein do not necessarily represent the views of the Department of Veterans Affairs or the US Government. The study sponsor had no role in the design, analysis, conduct, or writing of the study.
|
REFERENCES |
|---|
|
TopAbstractContext and CaveatsSubjects and MethodsResultsDiscussionFundingReferencesNotes |
|---|
(1) Ries LAG, Eisner MP, Kosary CL, Hankey BF, Miller BA, Clegg L, et al, eds. SEER Cancer Statistics Review, 1975–2002, based on November 2004 SEER data submission, posted to the SEER web site 2005. Bethesda (MD): National Cancer Institute. Available at: http://seer.cancer.gov/csr/1975_2002/. [Last accessed: April 15, 2006.].
(2) Merrill RM, Weed DL, Feuer EJ. The lifetime risk of developing prostate cancer in white and black men. Cancer Epidemiol Biomarkers Prev (1997) 6:763–8.
(3) Catalona WJ, Ramos CG, Carvalhal GF, Yan Y. Lowering PSA cutoffs to enhance detection of curable prostate cancer. Urology (2000) 55:791–5.[CrossRef][Web of Science][Medline]
(4) Gilbert SM, Cavallo CB, Kahane H, Lowe FC. Evidence suggesting PSA cutpoint of 2.5 ng/ml for prompting prostate biopsy: review of 36,316 biopsies. Urology (2005) 65:549–53.[CrossRef][Web of Science][Medline]
(5) Babaian RJ. Extended field prostate biopsy enhances cancer detection. Urology (2000) 55:453–6.[CrossRef][Web of Science][Medline]
(6) Presti JC, O'Dowd GJ, Miller MC, Mattu R, Veltri RW. Extended peripheral zone biopsy schemes increase cancer detection rates and minimize variance in prostate specific antigen and age related cancer rates: results of a community multi-practice study. J Urol (2003) 169:125–9.[CrossRef][Web of Science][Medline]
(7) Singh H, Canto EI, Shariat SF, Kadmon D, Miles BJ, Wheeler TM, et al. Improved detection of clinically significant, curable prostate cancer with systematic 12-core biopsy. J Urol (2004) 171:1089–92.[CrossRef][Web of Science][Medline]
(8) Roehl KA, Antenor JV, Catalona WJ. Serial biopsy results in prostate cancer screening study. J Urol (2002) 167:2435–9.[CrossRef][Web of Science][Medline]
(9) Singh H, Canto EI, Shariat SF, Kadmon D, Miles BJ, Wheeler TM, et al. Predictors of prostate cancer after initial negative systematic 12 core biopsy. J Urol (2004) 171:1850–4.[CrossRef][Web of Science][Medline]
(10) Stewart CS, Leibovich BC, Weaver AL, Lieber MM. Prostate cancer diagnosis using a saturation needle biopsy technique after previous negative sextant biopsies. J Urol (2001) 166:86–92.[CrossRef][Web of Science][Medline]
(11) Richie JP, Catalona WJ, Ahmann FR, Hudson MA, Scardino PT, Flanigan RC, et al. Effect of patient age on early detection of prostate cancer with serum prostate-specific antigen and digital rectal examination. Urology (1993) 42:365–74.[CrossRef][Web of Science][Medline]
(12) Latchamsetty KC, Kim J, Porten CR. Prostate specific antigen remains an independent predictor of cancer at prostate biopsy in Black American men but not white men: results from a consecutive series of 914 men. J Urol (2006) 175:913–7.[CrossRef][Web of Science][Medline]
(13) Smith DS, Bullock AD, Watalona WJ. Racial differences in operating characteristics of prostate cancer screening tests. J Urol (1997) 158:1861–6.[CrossRef][Web of Science][Medline]
(14) Gilbert S, Cavallo C, Kahane H, Lowe F. Evidence suggesting PSA cutpoint of 2.5 ng/mL for prompting prostate biopsy: review of 36,316 biopsies. Urology (2005) 65:549–53.[CrossRef][Web of Science][Medline]
(15) Eskew LA, Bare RL, McCullough DL. Systematic 5 region prostate biopsy is superior to sextant method for diagnosing carcinoma of the prostate. J Urol (1997) 157:199–202.[CrossRef][Web of Science][Medline]
(16) Babaian RJ, Toi A, Kamoi K, Troncoso P, Sweet J, Evans R, Johnston D, Chen M. A comparative analysis of sextant and an extended 11-core multisite directed biopsy strategy. J Urol (2000) 163:152–7.[CrossRef][Web of Science][Medline]
(17) Arnold PM, Niemann TH, Bahnson RR. Extended sector biopsy for detection of carcinoma of the prostate. Urol Oncol (2001) 6:91–3.[CrossRef][Web of Science][Medline]
(18) Durkan GC, Sheikh N, Johnson P, Hildreth AJ, Greene DR. Improving prostate cancer detection with an extended-core transrectal ultrasonography-guided prostate biopsy protocol. BJU Int (2002) 89:33–9.[Web of Science][Medline]
(19) Fleshner N, Klotz L. Role of "saturation biopsy" in the detection of prostate cancer among difficult diagnostic cases. Urology (2002) 60:93–7.[Web of Science][Medline]
(20) Welch HG, Schwartz LM, Woloshin S. Prostate-specific antigen levels in the United States: implications of various definitions for abnormal. J Natl Cancer Inst (2005) 97:1132–7.
(21) Carter HB, Ferrucci L, Kettermann A, Landis P, Wright EJ, Epstein JI, et al. Detection of life-threatening prostate cancer with prostate-specific antigen velocity during a window of curability. J Natl Cancer Inst (2006) 98:1521–7.
Manuscript received June 7, 2006; revised June 28, 2007; accepted July 20, 2007.
Related Article in JNCI
CiteULike 
Connotea 
Del.icio.us What's this?
J Natl Cancer Inst 2007 99: 1353.
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||