Statistical methods for handling missing data to align with treatment policy strategy.

The International Council for Harmonization (ICH) E9(R1) addendum recommends choosing an appropriate estimand based on the study objectives in advance of trial design. One defining attribute of an estimand is the intercurrent event, specifically what is considered an intercurrent event and how it should be handled. The primary objective of a clinical study is usually to assess a product's effectiveness and safety based on the planned treatment regimen instead of the actual treatment received. The estimand using the treatment policy strategy, which collects and analyzes data regardless of the occurrence of intercurrent events, is usually utilized. In this article, we explain how missing data can be handled using the treatment policy strategy from the authors' viewpoint in connection with antihyperglycemic product development programs. The article discusses five statistical methods to impute missing data occurring after intercurrent events. All five methods are applied within the framework of the treatment policy strategy. The article compares the five methods via Markov Chain Monte Carlo simulations and showcases how three of these five methods have been applied to estimate the treatment effects published in the labels for three antihyperglycemic agents currently on the market.

Informatics in Medical Product Regulation: The Right Drug at the Right Dose for the Right Patient.

This chapter describes the role of regulatory medical product review in furthering precision medicine. Efficient data processing and appropriate analyses are needed to synthesize information and provide directions for use in a medical product label. We describe opportunities and challenges in outcome assessment through informatics, as bioengineered therapeutics are increasingly developed for the unmet needs of molecularly defined diseases. Data submission requirements and analytic principles are outlined, and regulatory resources and foundational law and statute are cited for the reader.

Considerations to address missing data when deriving clinical trial endpoints from digital health technologies.

Digital health technologies (DHTs) enable us to measure human physiology and behavior remotely, objectively and continuously. With the accelerated adoption of DHTs in clinical trials, there is an unmet need to identify statistical approaches to address missing data to ensure that the derived endpoints are valid, accurate, and reliable. It is not obvious how commonly used statistical methods to handle missing data in clinical trials can be directly applied to the complex data collected by DHTs. Meanwhile, current approaches used to address missing data from DHTs are of limited sophistication and focus on the exclusion of data where the quantity of missing data exceeds a given threshold. High-frequency time series data collected by DHTs are often summarized to derive epoch-level data, which are then processed to compute daily summary measures. In this article, we discuss characteristics of missing data collected by DHT, review emerging statistical approaches for addressing missingness in epoch-level data including within-patient imputations across common time periods, functional data analysis, and deep learning methods, as well as imputation approaches and robust modeling appropriate for handling missing data in daily summary measures. We discuss strategies for minimizing missing data by optimizing DHT deployment and by including the patients' perspectives in the study design. We believe that these approaches provide more insight into preventing missing data when deriving digital endpoints. We hope this article can serve as a starting point for further discussion among clinical trial stakeholders.

Biomarkers of exposure among U.S. cigar smokers: an analysis of 1999-2012 National Health and Nutrition Examination Survey (NHANES) data.

BACKGROUND: Cigar consumption is increasing in the United States, but little information is available about exposure to toxic constituents from cigar smoking. METHODS: We conducted a cross-sectional analysis of biomarkers of tobacco exposure among 25,522 participants from the National Health and Nutrition Examination Survey (NHANES, 1999-2012). The biomarkers analyzed were serum cotinine, urinary 4-(methylnitrosamino)-1-(3-pyridyl)-1-butanol (NNAL), blood lead, blood cadmium, and urinary arsenic. We calculated geometric mean concentrations for each biomarker by tobacco use category and geometric mean ratios controlling for demographic factors. RESULTS: Cigar smokers had higher cotinine, NNAL, and lead concentrations than nontobacco users. The geometric mean concentration [95% confidence interval (CI)] of cotinine for primary cigar smokers (i.e., current cigar/never cigarette smokers) was 6.2 (4.2-9.2) ng/mL versus 0.045 (0.043-0.048) ng/mL for nontobacco users, and the NNAL concentration was 19.1 (10.6-34.3) pg/mg creatinine for primary cigar smokers versus 1.01 (0.95-1.07) pg/mg creatinine for nontobacco users. Secondary cigar smokers (i.e., current cigar/former cigarette smokers) and dual cigar/cigarette smokers had higher cadmium concentrations than nontobacco users. Cigar smoking was associated with significantly higher concentrations of cotinine, NNAL, cadmium, and lead, after adjusting for demographic factors. Secondary cigar smokers had significantly higher cotinine and NNAL concentrations than primary cigar smokers. The NNAL concentrations in daily cigar smokers were comparable with those in daily cigarette smokers. CONCLUSIONS: Cigar smokers have higher concentrations of several toxic and carcinogenic substances than nontobacco users. IMPACT: Our results are consistent with epidemiologic evidence demonstrating cigar smoking as a cause of disease and premature death.

Cigarette smokers' use of unconventional tobacco products and associations with quitting activity: findings from the ITC-4 U.S. cohort.

INTRODUCTION: The purpose of this study was to evaluate the prevalence and correlates of use of nicotine-containing tobacco products such as cigars, pipe tobacco, and cigarettes that promise less exposure to toxins; e-cigarettes; and smokeless tobacco products among a cohort of conventional cigarette smokers followed over the past decade. We also evaluated associations between use of such products and cigarette quitting. METHODS: Participants were 6,110 adult smokers in the United States, who were interviewed as part of the International Tobacco Control Four Country Survey between 2002 and 2011. Respondents reported their concurrent use of other smoked tobacco products (including cigars, pipe tobacco, and cigarillos), smokeless tobacco products (including chewing tobacco, snus, and snuff), unconventional cigarettes (including Omni, Accord, and Eclipse), and electronic cigarettes. Prevalence and correlates of use and associations between use and cigarette quitting were assessed using regression analyses via generalized estimating equations. RESULTS: Most cigarette smokers did not use unconventional tobacco products, although use of any of these products started to rise at the end of the study period (2011). For each type of tobacco product evaluated, use was most prevalent among those aged 18-24 years. Smokers who did use unconventional tobacco products did not experience a clear cessation advantage. CONCLUSIONS: During the past decade, relatively few cigarette smokers reported also using other tobacco products. Those that did use such products were no more likely to stop using conventional cigarettes compared with those who did not use such products.

Distribution of PSA velocity by total PSA levels: data from the Baltimore Longitudinal Study of Aging.

OBJECTIVES: To describe the distribution and implications of prostate-specific antigen velocity (PSAV) by prostate-specific antigen (PSA) in an unselected population. A PSAV >0.35 and >2.0 ng/mL/y have been associated with an increased risk of prostate cancer (CaP) death more than 10 years and 1 year before diagnosis, respectively. It is unknown how frequently PSAVs of this magnitude occur in community men. METHODS: From the Baltimore Longitudinal Study of Aging, we examined the PSAV distribution in 786 men with serial PSA measurements (3474 PSAV observations) at total PSA levels <10 ng/mL. We also determined whether PSAV altered the probability of overall and life-threatening CaP at PSA levels <3 and 3-10 ng/mL. RESULTS: Overall, the mean PSA and PSAV were 1.3 ng/mL and 0.05 ng/mL/y, respectively. PSAV rose continuously with increasing PSA (P <.0001), and was significantly higher in cancers than controls for observations at PSA levels <3 ng/mL (P = .02) and 3-10 ng/mL (P = .0008). The probability of life-threatening CaP was 3% at a PSA <3 ng/mL, but increased to 13.6% with PSAV >0.4 ng/mL/y. At PSA levels of 3-10 ng/mL, the probability of life-threatening CaP was 9.8% based on PSA alone vs 12% with PSAV >0.4 ng/mL/y. CONCLUSIONS: PSAV was significantly higher in CaP observations than controls in all PSA ranges studied and altered the risk of overall and life-threatening CaP at a given PSA level. Because the value of PSAV is PSA-dependent, the PSA level should be taken into account when interpreting PSAV.

Interpretation of the prostate-specific antigen history in assessing life-threatening prostate cancer.

OBJECTIVE: To present an effective approach to the early detection of lethal prostate cancer using longitudinal data on prostate-specific antigen (PSA) and its rate of change, i.e. PSA velocity (PSAV). This longitudinal approach might also be extendible to other biomarkers. SUBJECTS AND METHODS: PSAV was calculated using five techniques for 634 subjects with at least three PSA measurements in a longitudinal ageing study, censoring PSA levels of > 10 ng/mL. The efficacy for predicting death from prostate cancer was assessed with concordance indices and by using net reclassification improvement (NRI), which indicated the net increase in sensitivity and specificity when adding a biomarker to a base Cox proportional hazards model. The PSAV techniques were compared for the 5-10 years before the clinical diagnosis of prostate cancer. The most effective technique was then applied at the transition point when each man's PSA history curve transformed from linear to exponentially increasing, and its predictive value was compared to that of concurrent PSA level. RESULTS: A PSA transition point was found in 522 (82%) of the 634 men, including all 11 who died from prostate cancer. At the transition point, the mean PSA level was 1.4 ng/mL, and PSAV but not PSA level was significantly higher among men who died from prostate cancer than among men who did not (P = 0.021 vs P = 0.112; Wilcoxon two-sample test). At the transition point, adding PSAV to a base model consisting of age and date of diagnosis improved the concordance index by 0.05, and significantly improved the overall sensitivity and specificity (NRI, P = 0.028), while adding PSA level to the same base model resulted in little improvement (concordance index increase < 0.01 and NRI P = 0.275). CONCLUSION: When the shape of a man's PSA history curve changes from linear to exponential, PSAV might help in the early identification of life-threatening prostate cancer at a time when PSA values are still low in most men.

Prostate-specific antigen kinetics during follow-up are an unreliable trigger for intervention in a prostate cancer surveillance program.

PURPOSE: To assess the predictive ability of prostate-specific antigen (PSA) velocity (PSAV) and doubling time (PSADT) for biopsy progression and adverse pathology at prostatectomy among men with low-risk prostate cancer enrolled on an active-surveillance program. METHODS: We evaluated 290 men who met criteria for active surveillance (ie, PSA density < 0.15 ng/mL/cm(3) and Gleason score < or = 6 with no pattern > or = 4, involving < or = 2 cores with cancer, and < or = 50% involvement of any core by cancer) with two or more serial PSA measurements after diagnosis from 1994 to 2008. Follow-up included twice-yearly digital rectal exam and PSA measurements and yearly surveillance biopsy. Treatment was recommended for biopsy progression (ie, Gleason score > or = 7, or > 2 positive cores, or > 50% core involvement). Sensitivity and specificity of postdiagnostic PSAV and PSADT were explored by using receiver operating characteristic (ROC) analysis. RESULTS: Overall, 188 (65%) men remained on active surveillance, and 102 (35%) developed biopsy progression at a median follow-up of 2.9 years. PSADT was not significantly associated with subsequent adverse biopsy findings (P = .83), and PSAV was marginally significant (P = .06). No PSAV or PSADT cut point had both high sensitivity and specificity (area under the curve, 0.61 and 0.59, respectively) for biopsy progression. In those who eventually underwent radical prostatectomy, PSAV (P = .79) and PSADT (P = .87) were not associated with the presence of unfavorable surgical pathology. CONCLUSION: Postdiagnostic PSA kinetics do not reliably predict adverse pathology and should not be used to replace annual surveillance biopsy for monitoring men on active surveillance.

Bone mineral content and prostate cancer risk: data from the Baltimore Longitudinal Study of Aging.

STUDY TYPE: Aetiology (inception cohort) Level of Evidence 2b. OBJECTIVE: To determine whether there might be differences in bone mineral content (BMC) between men who develop life-threatening prostate cancer and those who do not, as bone is a common site of prostate cancer metastases. SUBJECTS AND METHODS: From 1973 to 1984, BMC was serially measured in 519 participants (778 observations) as part of a longitudinal study of ageing. We examined the association between serial BMC measurements with the development of overall and high-risk prostate cancer over the next one to three decades. For all prostate cancer cases, BMC was censored at the time of diagnosis. RESULTS: During a median (range) overall follow-up of 21.1 (0.2-35.0) years after the last BMC measurement, 76 (14.6%) men were later diagnosed with prostate cancer (18 high-risk and 58 not high-risk). BMC declined with age to a greater extent in healthy controls than among men diagnosed with prostate cancer (P = 0.018, likelihood ratio test), and tended to decline less in high-risk than non-high-risk cases. CONCLUSION: The distribution of BMC was significantly different between men who did and did not develop prostate cancer, over an extended follow-up. Specifically, BMC appeared to decline to a greater extent with age among healthy controls than in men with prostate cancer, especially high-risk disease. The biology underlying the lesser decline in BMC among men with prostate cancer remains unclear, but suggests that host factors in the bony milieu might be associated with prostate cancer development and progression.

Serum testosterone is associated with aggressive prostate cancer in older men: results from the Baltimore Longitudinal Study of Aging.

STUDY TYPE: Prognosis (inception cohort). LEVEL OF EVIDENCE: 1b. OBJECTIVE: To evaluate the relationship between testosterone levels and the development of high-risk prostate cancer, by prospectively examining serum androgen concentrations in a well-studied cohort, as the role of testosterone in prostate cancer progression is debated. PATIENTS AND METHODS: The study comprised 781 men in the Baltimore Longitudinal Study of Aging who had sex steroid measurements before a diagnosis of prostate cancer, or at their last visit for those without cancer (no cancer, 636; cancer, not high risk, 109; cancer, high risk, 36). High-risk cancer was defined as death from prostate cancer, a prostate specific antigen (PSA) level of > or =20 ng/mL at diagnosis, or a Gleason score of > or =8. The hazard ratio (HR) of high-risk disease was determined using a Cox proportional hazards regression model with simple updating, and risk rates were stratified by age and tercile for androgens of interest based on the proportional hazards analyses. RESULTS: The likelihood of high-risk prostate cancer doubled per unit (0.1) increase in the free testosterone index (FTI) for patients aged >65 years (HR 2.07, 95% confidence interval, CI, 1.01-4.23; P = 0.047); the likelihood for men aged < or =65 years was inversely related to the FTI (HR 0.96, 95% CI 0.35-2.6; P = 0.9). The risk rate per person-years increased from lowest to highest tercile of FTI for the oldest men (age >70 years) but this trend was not apparent among younger men. CONCLUSION: Higher levels of serum free testosterone are associated with an increased risk of aggressive prostate cancer among older men. These data highlight the importance of prospective trials to insure the safety of testosterone-replacement therapy.

Accuracy of PCA3 measurement in predicting short-term biopsy progression in an active surveillance program.

PURPOSE: PCA3 is a prostate specific noncoding mRNA that is significantly over expressed in prostate cancer tissue. Urinary PCA3 levels have been associated with prostate cancer grade and extent, suggesting a possible role in monitoring patients on active surveillance. We assessed the relationship between PCA3 and prostate biopsy results in men in a surveillance program. MATERIALS AND METHODS: Urine specimens were obtained from 294 men with prostate cancer enrolled in the Johns Hopkins surveillance program. The followup protocol included semiannual free and total prostate specific antigen measurements, digital rectal examination and annual surveillance prostate biopsy. Cox proportional hazards regression was used to evaluate the association between PCA3 results and progression on surveillance biopsy (defined as Gleason pattern 4 or 5, more than 2 positive biopsy cores or more than 50% involvement of any core with cancer). RESULTS: Patients with progression on biopsy (12.9%) had a mean PCA3 score similar to that of those without progression (60.0 vs 50.8, p = 0.131). ROC analysis suggested that PCA3 alone could not be used to identify men with progression on biopsy (AUC 0.589, 95% CI 0.496-0.683, p = 0.076). After adjustment for age and date of diagnosis PCA3 was not significantly associated with progression on biopsy (p = 0.15). CONCLUSIONS: In men with low risk prostate cancer who were carefully selected for surveillance the PCA3 score was not significantly associated with short-term biopsy progression. Further analysis is necessary to assess the usefulness of PCA3 in combination with other biomarkers or in selected subsets of patients undergoing surveillance.

Should prostate specific antigen be adjusted for body mass index? Data from the Baltimore Longitudinal Study of Aging.

PURPOSE: Obesity may be associated with lower prostate specific antigen through hemodilution. We examined the relationship between body mass index and prostate specific antigen by age in men without prostate cancer in a longitudinal aging study to determine whether prostate specific antigen must be adjusted for body mass index. MATERIALS AND METHODS: The study population included 994 men (4,937 observations) without prostate cancer in the Baltimore Longitudinal Study of Aging. Mixed effects models were used to examine the relationship between prostate specific antigen and body mass index in kg/m(2) by age. Separate models were explored in men with prostate cancer censored at diagnosis, for percent body fat measurements, for weight changes with time and adjusting for initial prostate size in 483 men (2,523 observations) with pelvic magnetic resonance imaging measurements. RESULTS: In men without prostate cancer body mass index was not significantly associated with prostate specific antigen after adjusting for age (p = 0.06). A 10-point body mass index increase was associated with a prostate specific antigen difference of -0.03 ng/ml (95% CI -0.40-0.49). Results were similar when men with prostate cancer were included, when percent body fat was substituted for body mass index, and after adjusting for prostate volume. Longitudinal weight changes also had no significant association with prostate specific antigen. CONCLUSIONS: Consistent with prior studies, we found an inverse relationship between obesity and serum prostate specific antigen. However, the magnitude of the difference was small. Thus, adjusting prostate specific antigen for body mass index does not appear warranted.

Prostate volume changes over time: results from the Baltimore Longitudinal Study of Aging.

PURPOSE: According to a 1944 publication by Swyer benign prostatic hyperplasia develops in some men after age 45 with further prostatic growth whereas in other men prostate size remains stable or decreases with advancing age. Although there is an abundance of literature describing prostatic enlargement in association with benign prostatic hyperplasia, less is known about the phenomenon of prostate atrophy. MATERIALS AND METHODS: In the Baltimore Longitudinal Study of Aging serial pelvic magnetic resonance imaging was performed in men without prostate cancer beginning in 1993. From this population we retrospectively identified 278 men with 2 or more magnetic resonance imaging determined prostate volume measurements to examine differential growth rates in a cohort of community men over time. RESULTS: Median age was 58 years and median prostate size was 28 cc at study entry. At a median followup of 4.3 years prostate size increased in 61.9% and remained stable or decreased in 38.1% of men. The median rate of volume change was 0.6 cc per year (range -9.9 to 62.1), corresponding to a median growth rate of 2.5% per year (range -29.2 to 176.4%). During followup 64.6% of men with an initial prostate size less than 40 cc had prostate growth compared to only 50.9% of men with an initial prostate size of 40 cc or greater. CONCLUSIONS: These results suggest that changes in prostate size are highly variable among aging men. Although benign prostatic hyperplasia is common, a considerable proportion of aging men have a stable or decreasing prostate size. Further research is needed to identify the underlying mechanism for such differences in prostate growth.

Single nucleotide polymorphisms and the likelihood of prostate cancer at a given prostate specific antigen level.

PURPOSE: Prostate specific antigen is used for prostate cancer screening but its specificity is limited. Specificity might be increased by considering genotype associated prostate specific antigen levels. MATERIALS AND METHODS: We examined associations between single nucleotide polymorphisms on chromosomes 10 and 19 (previously shown to be associated with prostate specific antigen) with prostate specific antigen and prostate cancer in 505 men from the Baltimore Longitudinal Study of Aging. RESULTS: In a model with age and date the risk ratio for prostate cancer was 1.18 (95% CI 1.13-1.23) per unit increase in prostate specific antigen. Including the interaction between alleles and prostate specific antigen significantly altered the risk ratio for prostate cancer (Cox proportional hazards p <0.001). Specifically prostate cancer risk per unit increase in prostate specific antigen was significantly different in carriers than in noncarriers of a minor allele (1.28 vs 1.10, respectively, Cox proportional hazards p <0.001), whereas men with a minor allele had a significantly higher risk of prostate cancer at prostate specific antigen levels greater than 6 ng/ml. CONCLUSIONS: Our data suggest that genotype influences the risk of prostate cancer per unit increase in prostate specific antigen. Prostate cancer risk stratification using prostate specific antigen and genotype could improve prostate specific antigen test performance.

Prostate specific antigen testing among the elderly--when to stop?

PURPOSE: Prostate specific antigen testing is common in the elderly despite evidence that older men without aggressive prostate cancer are unlikely to benefit from diagnosis and treatment. We evaluated the relationship between prostate specific antigen and the risk of aggressive prostate cancer developing in men of various ages. MATERIALS AND METHODS: This longitudinal cohort study consisted of 849 men (122 with and 727 without prostate cancer) with serial prostate specific antigen measurements participating in the Baltimore Longitudinal Study of Aging. The primary outcome measure was the proportion of men by prostate specific antigen and age who died of prostate cancer or in whom aggressive prostate cancer developed (death from prostate cancer, a prostate specific antigen 20 ng/ml or greater, or Gleason score 8 or greater). RESULTS: No participants between 75 and 80 years old with a prostate specific antigen less than 3.0 ng/ml died of prostate cancer. In contrast, men of all ages with a prostate specific antigen of 3.0 ng/ml or greater had a continually increasing probability of death from prostate cancer (Fisher's exact test p <0.001). The time to death or diagnosis of aggressive prostate cancer after age 75 years was not significantly different between the prostate specific antigen categories of 3 to 3.9 and 4 to 9.9 ng/ml (p = 0.634), whereas the time to death or diagnosis of high risk prostate cancer was significantly longer for the prostate specific antigen category of less than 3 vs 3 ng/ml or greater (p = 0.019). CONCLUSIONS: Men 75 to 80 years old with a prostate specific antigen less than 3 ng/ml are unlikely to die of or experience aggressive prostate cancer during their remaining life, suggesting that prostate specific antigen testing might be safely discontinued for these men.

Does prostate growth confound prostate specific antigen velocity? Data from the Baltimore longitudinal study of aging.

PURPOSE: Although prostate specific antigen velocity was proposed to increase the specificity of prostate specific antigen-based screening, there are little published data on the effect of differential prostate growth on prostate specific antigen velocity. If a patient presents with rising prostate specific antigen over a year or more, it would be useful to know whether such a change in prostate specific antigen could be explained by prostate growth. Thus, we investigated the relationship between changes in prostate size and prostate specific antigen changes in a large cohort of men without prostate cancer. MATERIALS AND METHODS: We identified 242 men without prostate cancer from the Baltimore Longitudinal Study of Aging who had 2 or greater serial pelvic magnetic resonance imaging studies and contemporaneous prostate specific antigen measurements. In this population we used the t test, correlation coefficients, and regression analysis to examine the relationship between prostate specific antigen changes and prostate volume changes, as assessed by magnetic resonance imaging. RESULTS: The mean age was 55 years. During 4.2 years of median followup, the median rate of volume change was 0.6 cc per year (range -9.9 to 11.8), and the median prostate specific antigen change was 0.03 ng/ml per year. There was no correlation between prostate specific antigen changes and prostate growth, as measured in cc per year (r = -0.01, p = 0.9) or the percent change per year (r = 0.07, p = 0.3). On multivariate analysis, there was no significant relationship between changes in prostate volume and prostate specific antigen changes. CONCLUSIONS: Our data suggest that volume increases alone do not cause a high prostate specific antigen velocity. Despite growth rates as high as 10 cc per year, prostate specific antigen velocity was less than 0.1 ng/ml per year in most men without prostate cancer. Thus, differential rates of prostatic growth should not confound the use of prostate specific antigen velocity for prostate cancer detection and prognostication.

PSA doubling time versus PSA velocity to predict high-risk prostate cancer: data from the Baltimore Longitudinal Study of Aging.

BACKGROUND: Our group has previously shown that prostate-specific antigen (PSA) velocity (PSAV) is associated with the presence of life-threatening prostate cancer. Less is known about the relative utility of pretreatment PSA doubling time (PSA DT) to predict tumor aggressiveness. OBJECTIVE: To compare the utility of PSAV and PSA DT for the prediction of life-threatening prostate cancer. DESIGN, SETTING, AND PARTICIPANTS: From the Baltimore Longitudinal Study of Aging, we identified 681 men with serial PSA measurements. MEASUREMENTS: Receiver operating characteristic analysis was used to evaluate the relationship between PSAV, PSA DT, and the presence of high-risk disease. RESULTS AND LIMITATIONS: Within the period of 5 yr prior to diagnosis, PSAV was significantly higher among men with high-risk or fatal prostate cancer than men without it. By contrast, PSA DT was not significantly associated with high-risk or fatal disease. On multivariate analysis, including age, date of diagnosis, and PSA, the addition of PSAV significantly improved the concordance index from 0.85 to 0.88 (p<0.001), whereas PSA DT did not. CONCLUSIONS: These data suggest that PSAV is more useful than PSA DT in the pretreatment setting to help identify those men with life-threatening disease.

Prostate-specific antigen velocity risk count assessment: a new concept for detection of life-threatening prostate cancer during window of curability.

OBJECTIVES: To determine whether the number of times the prostate-specific antigen (PSA) velocity (PSAV) exceeds a threshold (PSAV risk count) is predictive of high-risk prostate cancer. METHODS: The PSAV was determined in 717 men (606 without prostate cancer; 32 with high-risk prostate cancer defined as death from cancer, PSA level of 20 ng/mL or more, or a Gleason score of 8 or more; and 79 with prostate cancer who were alive or dead of another cause). Multiple PSAVs determined from three repeated measures from each subject during 10 to 20 years were used to determine the risk count by summing the number of times a subject exceeded a PSAV threshold. Cox proportional hazards regression analysis was used to evaluate the associations between the risk count and the probability of high-risk disease. The statistical tests were two-sided. RESULTS: The probability of high-risk disease increased directly with the risk count. After adjusting for age, PSA level, PSAV, and date of diagnosis, the PSAV risk count was significantly associated with the development of high-risk prostate cancer (relative risk 1.41, 95% confidence interval 1.25 to 1.59 for a PSAV cutpoint 0.2 ng/mL/yr; relative risk 1.49, 95% confidence interval 1.29 to 1.71 for a PSAV cutpoint of 0.4 ng/mL/yr; P <0.001). CONCLUSIONS: The PSAV risk count could be a useful method of interpreting a PSA history to help identify those men who will benefit from a diagnosis of prostate cancer at PSA levels associated with curable disease.