Although numerous risk-prediction models exist in patients presenting with acute coronary syndromes (ACS), they are subject to important short-comings, including lack of contemporary information. Short-term models are frequently biased by in-hospital events. Accordingly, we sought to create contemporary risk-prediction models for clinical outcomes following ACS up to 1 year following discharge. Models were constructed for death at 30 days and 1 year, death/myocardial infarction (MI)/revascularization at 30 days and death/MI at 1 year in consecutive patients presenting with ACS at our institution between 2006 and 2008, and discharged alive. Logistic regression was used to model the 30 day outcomes and Cox proportional hazards were used to model the 1 year outcomes. No linearity assumptions were made for continuous variables. The final model coefficients were used to create a prediction nomogram, which was incorporated into an online risk calculator. A total of 2,681 patients were included, of which about 9.5% presented with ST-elevation MI. All-cause mortality was 2.6% at 30 days and 13% at 1 year. Demographic, past medical history, laboratory, pharmacological and angiographic parameters were identified as being predictive of adverse ischemic outcomes at 30 days and 1 year. The c-indices for these models ranged from 0.73 to 0.82. Our study thus identified risk factors that are predictive of short- and long-term ischemic and revascularization outcomes in contemporary patients with ACS, and incorporated them into an easy-to-use online calculator, with equal or better discriminatory power than currently available models.

New markers may improve prediction of diagnostic and prognostic outcomes. We review various measures to quantify the incremental value of markers over standard, readily available characteristics. Widely used traditional measures include the improvement in model fit or in the area under the receiver operating characteristic (ROC) curve (AUC). New measures include the net reclassification index (NRI) and decision–analytic measures, such as the fraction of true positive classifications penalized for false positive classifications (‘net benefit’, NB).

For illustration we discuss a case study on the presence of residual tumor versus benign tissue in 544 patients with testicular cancer. We assessed 3 tumor markers (AFP, HCG, and LDH) for their incremental value over currently standard clinical predictors. AUC and R2 values suggested adding continuous LDH and AFP whereas NB only favored HCG as a potentially promising marker at a clinically defendable decision threshold of 20% risk. Results based on the NRI fell in the middle, suggesting reclassification potential of all three markers.

We conclude that improvement in standard discrimination measures, which focus on finding variables that might be promising across all decision thresholds, may not detect the most informative markers at a specific threshold of particular clinical relevance. When a marker is intended to support decision making, calculation of the improvement in a decision–analytic measure, such as NB, is preferable over an overall judgment as obtained from the AUC in ROC analysis.

The performance of prediction models can be assessed using a variety of different methods and metrics. Traditional measures for binary and survival outcomes include the Brier score to indicate overall model performance, the concordance (or c) statistic for discriminative ability (or area under the receiver operating characteristic (ROC) curve), and goodness-of-fit statistics for calibration.

Several new measures have recently been proposed that can be seen as refinements of discrimination measures, including variants of the c statistic for survival, reclassification tables, net reclassification improvement (NRI), and integrated discrimination improvement (IDI). Moreover, decision–analytic measures have been proposed, including decision curves to plot the net benefit achieved by making decisions based on model predictions.

We aimed to define the role of these relatively novel approaches in the evaluation of the performance of prediction models. For illustration we present a case study of predicting the presence of residual tumor versus benign tissue in patients with testicular cancer (n=544 for model development, n=273 for external validation).

We suggest that reporting discrimination and calibration will always be important for a prediction model. Decision-analytic measures should be reported if the predictive model is to be used for making clinical decisions. Other measures of performance may be warranted in specific applications, such as reclassification metrics to gain insight into the value of adding a novel predictor to an established model.

Background

We have previously demonstrated that there is a learning curve for open radical prostatectomy. In this study we sought to determine whether the effects of the learning curve are modified by patient risk as defined by preoperative tumor characteristics.

Methods

The study included 7,683 eligible prostate cancer patients treated with open radical prostatectomy by one of 72 surgeons. Surgeon experience was coded as the total prior number of radical prostatectomies conducted by the surgeon prior to a patient’s surgery. Multivariable survival-time regression models were used to evaluate the association between surgeon experience and biochemical recurrence, separately for each preoperative risk group.

Results

We saw no evidence that patient risk affects the learning curve: there was a statistically significant association between biochemical recurrence and surgeon experience in all analyses. The absolute risk difference for a patient receiving treatment from a surgeon with 10 compared to 250 prior radical prostatectomies was 6.6% (95% C.I. 3.4%, 10.3%), 12.0% (6.9%, 18.2%) and 9.7% (1.2%, 18.2%) for patients at low, medium and high preoperative risk patients. Recurrence-free probability for patients with low risk disease approached 100% for the most experienced surgeons

Conclusions

Cancer control after radical prostatectomy improves with increasing surgeon experience irrespective of patient risk. Excellent rates of cancer control for patients with low risk disease treated by the most experienced surgeons suggests that the primary reason such patients recur is inadequate surgical technique. The results have significant implications for clinical care.

Purpose: To identify and examine polymorphisms of genes associated with aggressive and clinical significant forms of prostate cancer among a screening cohort.

 

 

Experimental Design: We conducted a genome-wide association study among patients with aggressive forms of prostate cancer and biopsy-proven normal controls ascertained from a prostate cancer screening program. We then examined significant associations of specific polymorphisms among a prostate cancer screened cohort to examine their predictive ability in detecting prostate cancer.

Results: We found significant associations between aggressive prostate cancer and five single nucleotide polymorphisms (SNPs) in the 10q26 (rs10788165, rs10749408, and rs10788165, p value for association 1.3 × 10−10 to 3.2 × 10−11) and 15q21 (rs4775302 and rs1994198, p values for association 3.1 × 10−8 to 8.2 × 10−9) regions. Results of a replication study done in 3439 patients undergoing a prostate biopsy, revealed certain combinations of these SNPs to be significantly associated not only with prostate cancer but with aggressive forms of prostate cancer using an established classification criterion for prostate cancer progression (odds ratios for intermediate to high-risk disease 1.8–3.0, p value 0.003–0.001). These SNP combinations were also important clinical predictors for prostate cancer detection based on nomogram analysis that assesses prostate cancer risk.

Conclusions: Five SNPs were found to be associated with aggressive forms of prostate cancer. We demonstrated potential clinical applications of these associations.

BACKGROUND

Accurate preoperative and postoperative risk assessment has been critical for counseling patients regarding radical prostatectomy for clinically localized prostate cancer. In addition to other treatment modalities, neoadjuvant or adjuvant therapies have been considered. The growing literature suggested that the experience of the surgeon may affect the risk of prostate cancer recurrence. The purpose of this study was to develop and internally validate nomograms to predict the probability of recurrence, both preoperatively and postoperatively, with adjustment for standard parameters plus surgeon experience.

METHODS

The study cohort included 7724 eligible prostate cancer patients treated with radical prostatectomy by 1 of 72 surgeons. For each patient, surgeon experience was coded as the total number of cases conducted by the surgeon before the patient’s operation. Multivariable Cox proportional hazards regression models were developed to predict recurrence. Discrimination and calibration of the models was assessed following bootstrapping methods, and the models were presented as nomograms.

RESULTS

In this combined series, the 10-year probability of recurrence was 23.9%. The nomograms were quite discriminating (preoperative concordance index, 0.767; postoperative concordance index, 0.812). Calibration appeared to be very good for each. Surgeon experience seemed to have a quite modest effect, especially postoperatively.

CONCLUSIONS

Nomograms have been developed that consider the surgeon’s experience as a predictor. The tools appeared to predict reasonably well but were somewhat little improved with the addition of surgeon experience as a predictor variable.

Introduction

The prognosis of men with clinically localized prostate cancer is highly variable, and it is difficult to counsel a man who may be considering avoiding, or delaying, aggressive therapy. After collecting data on a large cohort of men who received no initial active prostate cancer therapy, we sought to develop, and to internally validate, a nomogram for prediction of disease-specific survival.

Methods

Working with 6 cancer registries within England and numerous hospitals in the region, we constructed a population-based cohort of men diagnosed with prostate cancer between 1990 and 1996. All men had baseline serum prostate specific antigen (PSA) measurements, centralized pathologic grading, and centralized review of clinical stage assignment. Based upon the clinical and pathological data from 1,911 men, we developed and validated a statistical model that served as the basis for the nomogram. The discrimination and calibration of the nomogram were assessed with use of one third of the men, who were omitted from modeling and used as a test sample.

Results

The median age of the included men was 70.4 years. The 25th and 75th percentiles of PSA were 7.3 and 32.6 ng/ml respectively, and the median was 15.4 ng/ml. Forty-two percent of the men had high grade disease. The nomogram predicted well with a concordance index of 0.73 and had good calibration.

Conclusions

We have developed an accurate tool for predicting the probability that a man with clinically localized prostate cancer will survive his disease for 120 months if the cancer is not treated with curative intent immediately. The tool should be helpful for patient counseling and clinical trial design.

Statistical models predicting cancer recurrence after surgery are based on biologic variables. We have previously shown that prostate cancer recurrence is related both to tumor biology and to surgical technique. Here we evaluate the association between several biological predictors and biochemical recurrence across varying surgical experience. The study included two separate cohorts: 6091 patients treated by open radical prostatectomy and an independent replication set of 2298 patients treated laparoscopically. We calculated the odds ratios for biological predictors of biochemical recurrence– stage, Gleason grade and prostate-specific antigen (PSA) – and also the predictive accuracy (AUC) of a multivariable model, for subgroups of patients defined by the experience of their surgeon. In the open cohort, the odds ratio for Gleason score 8+ and advanced pathologic stage, though not PSA or Gleason score 7, increased dramatically when patients treated by surgeons with lower levels of experience were excluded (Gleason 8+: odds ratios 5.6 overall vs. 13.0 for patients treated by surgeons with 1000+ prior cases; locally advanced disease: odds ratios of 6.6 vs. 12.2 respectively). The AUC of the multivariable model was 0.750 for patients treated by surgeons with 50 or fewer cases compared to 0.849 for patients treated by surgeons with 500 or more. Although predictiveness was overall lower for the independent replication set cohort, the main findings were replicated. Surgery confounds biology. Although our findings have no direct clinical implications, studies investigating biological variables as predictors of outcome after curative resection of cancer should consider the impact of surgeon specific factors.

OBJECTIVE—The objective of this study was to create a tool that predicts the risk of mortality in patients with type 2 diabetes.

RESEARCH DESIGN AND METHODS—This study was based on a cohort of 33,067 patients with type 2 diabetes identified in the Cleveland Clinic electronic health record (EHR) who were initially prescribed a single oral hypoglycemic agent between 1998 and 2006. Mortality was determined in the EHR and the Social Security Death Index. A Cox proportional hazards regression model was created using medication class and 20 other predictor variables chosen for their association with mortality. A prediction tool was created using the Cox model coefficients. The tool was internally validated using repeated, random subsets of the cohort, which were not used to create the prediction model.

RESULTS—Follow-up in the cohort ranged from 1 day to 8.2 years (median 28.6 months), and 3,661 deaths were observed. The prediction tool had a concordance index (i.e., c statistic) of 0.752.

CONCLUSIONS—We successfully created a tool that accurately predicts mortality risk in patients with type 2 diabetes. The incorporation of medications into mortality predictions in patients with type 2 diabetes should improve treatment decisions.

Background

Although the American Board of Internal Medicine (ABIM) certification is valued as a reflection of physicians’ experience, education, and expertise, limited methods exist to predict performance in the examination.

Purpose

The objective of this study was to develop and validate a predictive tool based on variables common to all residency programs, regarding the probability of an internal medicine graduate passing the ABIM certification examination.

Methods

The development cohort was obtained from the files of the Cleveland Clinic internal medicine residents who began training between 2004 and 2008. A multivariable logistic regression model was built to predict the ABIM passing rate. The model was represented as a nomogram, which was internally validated with bootstrap resamples. The external validation was done retrospectively on a cohort of residents who graduated from two other independent internal medicine residency programs between 2007 and 2011.

Results

Of the 194 Cleveland Clinic graduates used for the nomogram development, 175 (90.2%) successfully passed the ABIM certification examination. The final nomogram included four predictors: In-Training Examination (ITE) scores in postgraduate year (PGY) 1, 2, and 3, and the number of months of overnight calls in the last 6 months of residency. The nomogram achieved a concordance index (CI) of 0.98 after correcting for over-fitting bias and allowed for the determination of an estimated probability of passing the ABIM exam. Of the 126 graduates from two other residency programs used for external validation, 116 (92.1%) passed the ABIM examination. The nomogram CI in the external validation cohort was 0.94, suggesting outstanding discrimination.

Conclusions

A simple user-friendly predictive tool, based on readily available data, was developed to predict the probability of passing the ABIM exam for internal medicine residents. This may guide program directors’ decision-making related to program curriculum and advice given to individual residents regarding board preparation.

Introduction: 5-alpha reductase inhibitors can reduce the risk of prostate cancer (PCa) but can be associated with significant side effects. A library of nomograms which predict the risk of clinical endpoints relevant to dutasteride treatment may help determine if chemoprevention is suited to the individual patient. Methods: Data from the REDUCE trial was used to identify predictive factors for 9 endpoints relevant to dutasteride treatment. Using the treatment and placebo groups from the biopsy cohort, Cox proportional hazards (PH) and competing risks regression (CRR) models were used to build 18 nomograms, whose predictive ability was measured by concordance index (CI) and calibration plots. Results: A total of 18 nomograms assessing the risks of cancer, high grade cancer, high grade prostatic intraepithelial neoplasia (HGPIN), atypical small acinar proliferation (ASAP), erectile dysfunction (ED), acute urinary retention (AUR), gynecomastia, urinary tract infection (UTI) and BPH-related surgery either on or off dutasteride were created. The nomograms for cancer, high grade cancer, ED, AUR, and BPH-related surgery demonstrated good discrimination and calibration while those for gynecomastia, UTI, HGPIN, and ASAP predicted no better than random chance. Conclusions: To aid patients in determining whether the benefits of dutasteride use outweigh the risks, we have developed a comprehensive metagram that can generate individualized risks of 9 outcomes relevant to men considering chemoprevention. Better models based on more predictive markers are needed for some of the endpoints but the current metagram demonstrates potential as a tool for patient counseling and decision-making that is accessible, intuitive, and clinically relevant.

Objectives

To compare the predictive performance and potential clinical usefulness of risk calculators of the European Randomized Study of Screening for Prostate Cancer (ERSPC RC) with and without information on prostate volume.

Methods

We studied 6 cohorts (5 European and 1 US) with a total of 15,300 men, all biopsied and with pre-biopsy TRUS measurements of prostate volume. Volume was categorized into 3 categories (25, 40, and 60 cc), to reflect use of digital rectal examination (DRE) for volume assessment. Risks of prostate cancer were calculated according to a ERSPC DRE-based RC (including PSA, DRE, prior biopsy, and prostate volume) and a PSA + DRE model (including PSA, DRE, and prior biopsy). Missing data on prostate volume were completed by single imputation. Risk predictions were evaluated with respect to calibration (graphically), discrimination (AUC curve), and clinical usefulness (net benefit, graphically assessed in decision curves).

Results

The AUCs of the ERSPC DRE-based RC ranged from 0.61 to 0.77 and were substantially larger than the AUCs of a model based on only PSA + DRE (ranging from 0.56 to 0.72) in each of the 6 cohorts. The ERSPC DRE-based RC provided net benefit over performing a prostate biopsy on the basis of PSA and DRE outcome in five of the six cohorts.

Conclusions

Identifying men at increased risk for having a biopsy detectable prostate cancer should consider multiple factors, including an estimate of prostate volume.

An existing preoperative nomogram predicts the probability of prostate cancer recurrence, defined by prostate-specific antigen (PSA), at 5 years after radical prostatectomy based on clinical stage, serum PSA, and biopsy Gleason grade. In an updated and enhanced nomogram, we have extended the predictions to 10 years, added the prognostic information of systematic biopsy results, and enabled the predictions to be adjusted for the year of surgery. Cox regression analysis was used to model the clinical information for 1978 patients treated by two high-volume surgeons from our institution. The nomogram was externally validated on an independent cohort of 1545 patients with a concordance index of 0.79 and was well calibrated with respect to observed outcome. The inclusion of the number of positive and negative biopsy cores enhanced the predictive accuracy of the model. Thus, a new preoperative nomogram provides robust predictions of prostate cancer recurrence up to 10 years after radical prostatectomy.

Purpose

Physicians often order periodic bone scans (BS) to check for metastases in patients with an increasing prostate-specific antigen (PSA; biochemical recurrence [BCR]) after radical prostatectomy (RP), but most scans are negative. We studied patient characteristics to build a predictive model for a positive scan.

Patients and Methods

From our prostate cancer database we identified all patients with detectable PSA after RP. We analyzed the following features at the time of each bone scan for association with a positive BS: preoperative PSA, time to BCR, pathologic findings of the RP, PSA before the BS (trigger PSA), PSA kinetics (PSA doubling time, PSA slope, and PSA velocity), and time from BCR to BS. The results were incorporated into a predictive model.

Results

There were 414 BS performed in 239 patients with BCR and no history of androgen deprivation therapy. Only 60 (14.5%) were positive for metastases. In univariate analysis, preoperative PSA (P = .04), seminal vesicle invasion (P = .02), PSA velocity (P < .001), and trigger PSA (P < .001) predicted a positive BS. In multivariate analysis, only PSA slope (odds ratio [OR], 2.71; P = .03), PSA velocity (OR, 0.93; P = .003), and trigger PSA (OR, 1.022; P < .001) predicted a positive BS. A nomogram for predicting the bone scan result was constructed with an overfit-corrected concordance index of 0.93.

Conclusion

Trigger PSA, PSA velocity, and slope were associated with a positive BS. A highly discriminating nomogram can be used to select patients according to their risk for a positive scan. Omitting scans in low-risk patients could reduce substantially the number of scans ordered.

OBJECTIVE

Sulfonylureas have historically been analyzed as a medication class, which may be inappropriate given the differences in properties inherent to the individual sulfonylureas (hypoglycemic risk, sulfonylurea receptor selectivity, and effects on myocardial ischemic preconditioning). The purpose of this study was to assess the relationship of individual sulfonylureas and the risk of overall mortality in a large cohort of patients with type 2 diabetes.

RESEARCH DESIGN AND METHODS

A retrospective cohort study was conducted using an academic health center enterprise-wide electronic health record (EHR) system to identify 11,141 patients with type 2 diabetes (4,279 initiators of monotherapy with glyburide, 4,325 initiators of monotherapy with glipizide, and 2,537 initiators of monotherapy with glimepiride), ≥18 years of age with and without a history of coronary artery disease (CAD) and not on insulin or a noninsulin injectable at baseline. The patients were followed for mortality by documentation in the EHR and Social Security Death Index. Multivariable Cox models were used to compare cohorts.

RESULTS

No statistically significant difference in the risk of overall mortality was observed among these agents in the entire cohort, but we did find evidence of a trend toward an increased overall mortality risk with glyburide versus glimepiride (hazard ratio 1.36 [95% CI 0.96–1.91]) and glipizide versus glimepiride (1.39 [0.99–1.96]) in those with documented CAD.

CONCLUSIONS

Our results did not identify an increased mortality risk among the individual sulfonylureas but did suggest that glimepiride may be the preferred sulfonylurea in those with underlying CAD.

OBJECTIVE

To design a decision-support tool to facilitate evidence-based treatment decisions in clinically localized prostate cancer, as individualized risk assessment and shared decision-making can decrease distress and decisional regret in patients with prostate cancer, but current individual models vary or only predict one outcome of interest.

METHODS

We searched Medline for previous reports and identified peer-reviewed articles providing pretreatment predictive models that estimated pathological stage and treatment outcomes in men with biopsy-confirmed, clinical T1-3 prostate cancer. Each model was entered into a spreadsheet to provide calculated estimates of extracapsular extension (ECE), seminal vesicle invasion (SVI), and lymph node involvement (LNI). Estimates of the prostate-specific antigen (PSA) outcome after radical prostatectomy (RP) or radiotherapy (RT), and clinical outcomes after RT, were also entered. The data are available at http://www.capcalculator.org.

RESULTS

Entering a patient’s 2002 clinical T stage, Gleason score and pretreatment PSA level, and details from core biopsy findings, into the CaP Calculator provides estimates from predictive models of pathological extent of disease, four models for ECE, four for SVI and eight for LNI. The 5-year estimates of PSA relapse-free survival after RT and 10-year estimates after RP were available. A printout can be generated with individualized results for clinicians to review with each patient.

CONCLUSIONS

The CaP Calculator is a free, online ‘clearing house’ of several predictive models for prostate cancer, available in an accessible, user-friendly format. With further development and testing with patients, the CaP Calculator might be a useful decision-support tool to help doctors promote evidence-based shared decision-making in prostate cancer.

Background

Multivariable prediction models have been shown to predict cancer outcomes more accurately than cancer stage. The effects on clinical management are unclear. We aimed to determine whether a published multivariable prediction model for bladder cancer (“bladder nomogram”) improves medical decision making, using referral for adjuvant chemotherapy as a model.

Methods

We analyzed data from an international cohort study of 4462 patients undergoing cystectomy without chemotherapy 1969 – 2004. The number of patients eligible for chemotherapy was determined using pathologic stage criteria (lymph node positive or stage pT3 or pT4), and for three cut-offs on the bladder nomogram (10%, 25% and 70% risk of recurrence with surgery alone). The number of recurrences was calculated by applying a relative risk reduction to eligible patients' baseline risk. Clinical net benefit was then calculated by combining recurrences and treatments, weighting the latter by a factor related to drug tolerability.

Results

A nomogram cut-off outperformed pathologic stage for chemotherapy for every scenario of drug effectiveness and tolerability. For a drug with a relative risk of 0.80, where clinicians would treat no more than 20 patients to prevent one recurrence, use of the nomogram was equivalent to a strategy that resulted in 60 fewer chemotherapy treatments per 1000 patients without any increase in recurrence rates.

Conclusions

Referring cystectomy patients to adjuvant chemotherapy on the basis of a multivariable model is likely to lead to better patient outcomes than the use of pathological stage. Further research is warranted to evaluate the clinical effects of multivariable prediction models.

Summary

Prostate cancer is a very complex disease, and the decision-making process requires the clinician to balance clinical benefits, life expectancy, comorbidities, and potential treatment related side effects. Accurate prediction of clinical outcomes may help in the difficult process of making decisions related to prostate cancer. In this review, we discuss attributes of predictive tools and systematically review those available for prostate cancer. Types of tools include probability formulas, look-up and propensity scoring tables, risk-class stratification prediction tools, classification and regression tree analysis, nomograms, and artificial neural networks. Criteria to evaluate tools include discrimination, calibration, generalizability, level of complexity, decision analysis, and ability to account for competing risks and conditional probabilities. We describe the available predictive tools and their features, focusing on nomograms. While some tools are well-calibrated, few have been externally validated or directly compared to other tools. In addition, the clinical consequences of applying predictive tools need thorough assessment. Nevertheless, predictive tools can facilitate medical decision-making by showing patients tailored predictions of their outcomes with various alternatives. Additionally, accurate tools may improve clinical trial design.

Purpose

Prostate-specific antigen is a glycoprotein found almost exclusively in normal and neoplastic prostate cells. PSA doubling time, or the change in PSA level over time, has emerged as a useful predictive marker for assessing disease outcome in patients with prostate cancer. It is important to agree on definitions and values for the calculation of PSADT and to develop a common approach to outcome analysis and reporting.

Methods

In September 2006 a conference was held at the National Cancer Institute in Bethesda, Maryland to define these parameters and develop guidelines for their use.

Results

The PSA Working Group defined the following criteria regarding PSADT: (1) calculation of PSADT, (2) evidence to support PSADT as a predictive factor in the setting of biochemical recurrence, and (3) use of PSADT as a stratification factor in clinical trials.

Conclusions

We propose that investigators calculate PSADT prior to enrolling patients on clinical studies and calculate it as an additional measurement of therapeutic activity. We believe we have developed practical guidelines for the calculation of PSADT and its use as a measurement of prognosis and outcome. Furthermore, the use of common standards for PSADT in clinical trials is important as we determine which treatments should progress to randomized trials in which “hard” end points such as survival will be employed.

The act of diagnosis requires that patients be placed in a binary category of either having or not having a certain disease. Accordingly, the diseases of particular concern for industrialized countries—such as type 2 diabetes, obesity, or depression—require that a somewhat arbitrary cut-point be chosen on a continuous scale of measurement (for example, a fasting glucose level >6.9 mmol/L [>125 mg/dL] for type 2 diabetes). These cut-points do not adequately reflect disease biology, may inappropriately treat patients on either side of the cut-point as 2 homogenous risk groups, fail to incorporate other risk factors, and are invariable to patient preference. This article discusses risk prediction as an alternative to diagnosis: Patient risk factors (blood pressure, age) are combined into a single statistical model (risk for a cardiovascular event within 10 years) and the results are used in shared decision making about possible treatments. The authors compare and contrast the diagnostic and risk prediction approaches and attempt to identify the types of medical problem to which each is best suited.

BACKGROUND

There is wide interest in the use of molecular markers for the early detection of cancer, the prediction of disease outcome, and the selection of patients for chemotherapy. Despite significant and increasing research activity, to the authors’ knowledge only a small number of molecular markers have been successfully integrated into clinical practice. In the current study, the experimental designs and statistical methods used in contemporary molecular marker studies are reviewed, particularly with respect to whether these evaluated a marker’s clinical value.

METHODS

MEDLINE was searched for studies that analyzed an association between a cancer outcome and a marker involving chemical analysis of body fluid or tissue. For each article, data were extracted regarding patients, markers, type of statistical analysis, and principal results.

RESULTS

The 129 articles eligible for analysis included a very large variety of molecular markers; the total number of markers was larger than the number of articles. Only a minority of articles (47 articles; 36%) incorporated multivariate modeling in which the marker was added to standard clinical variables, and only a very small minority had any measure of predictive accuracy (14 articles; 11%). No article used decision analytic methods or experimentally evaluated the clinical value of a marker. Correction for overfit was also rare (3 articles).

CONCLUSIONS

Statistical methods in molecular marker research have not focused on the clinical value of a marker. Attention to sound statistical practice, in particular the use of statistical approaches that provide clinically relevant information, will help maximize the promise of molecular markers for care of the cancer patient.

Objectives

We previously demonstrated that there is a learning curve for open radical prostatectomy. We sought to determine whether the effects of the learning curve are modified by pathologic stage.

Methods

The study included 7765 eligible prostate cancer patients treated with open radical prostatectomy by one of 72 surgeons. Surgeon experience was coded as the total number of radical prostatectomies conducted by the surgeon prior to a patient’s surgery. Multivariable regression models of survival time were used to evaluate the association between surgeon experience and biochemical recurrence, with adjustment for PSA, stage, and grade. Analyses were conducted separately for patients with organ-confined and locally advanced disease.

Results

Five-year recurrence-free probability for patients with organ-confined disease approached 100% for the most experienced surgeons. Conversely, the learning curve for patients with locally advanced disease reached a plateau at approximately 70%, suggesting that about a third of these patients cannot be cured by surgery alone.

Conclusions

Excellent rates of cancer control for patients with organ-confined disease treated by the most experienced surgeons suggest that the primary reason such patients recur is inadequate surgical technique.

Purpose

An increasing serum prostate-specific antigen (PSA) level is the initial sign of recurrent prostate cancer among patients treated with radical prostatectomy. Salvage radiation therapy (SRT) may eradicate locally recurrent cancer, but studies to distinguish local from systemic recurrence lack adequate sensitivity and specificity. We developed a nomogram to predict the probability of cancer control at 6 years after SRT for PSA-defined recurrence.

Patients and Methods

Using multivariable Cox regression analysis, we constructed a model to predict the probability of disease progression after SRT in a multi-institutional cohort of 1,540 patients.

Results

The 6-year progression-free probability was 32% (95% CI, 28% to 35%) overall. Forty-eight percent (95% CI, 40% to 56%) of patients treated with SRT alone at PSA levels of 0.50 ng/mL or lower were disease free at 6 years, including 41% (95% CI, 31% to 51%) who also had a PSA doubling time of 10 months or less or poorly differentiated (Gleason grade 8 to 10) cancer. Significant variables in the model were PSA level before SRT (P < .001), prostatectomy Gleason grade (P < .001), PSA doubling time (P < .001), surgical margins (P < .001), androgen-deprivation therapy before or during SRT (P < .001), and lymph node metastasis (P = .019). The resultant nomogram was internally validated and had a concordance index of 0.69.

Conclusion

Nearly half of patients with recurrent prostate cancer after radical prostatectomy have a long-term PSA response to SRT when treatment is administered at the earliest sign of recurrence. The nomogram we developed predicts the outcome of SRT and should prove valuable for medical decision making for patients with a rising PSA level.

Objectives

Men with clinically detected localized prostate cancer treated without curative intent are at risk of complications from local tumor growth. We investigated rates of local progression and need for local therapy among such men.

Methods

Men diagnosed with prostate cancer during 1990–1996 were identified from cancer registries throughout the United Kingdom. Inclusion criteria were age ≤76 yr at diagnosis, PSA level ≤100 ng/ml, and, within 6 mo after diagnosis, no radiation therapy, radical prostatectomy, evidence of metastatic disease, or death. Local progression was defined as increase in clinical stage from T1/2 to T3/T4 disease, T3 to T4 disease, and/or need for transurethral resection of the prostate (TURP) to relieve symptoms >6 mo after cancer diagnosis.

Results

The study included 2333 men with median follow-up of 85 mo (range: 6–174). Diagnosis was by TURP in 1255 men (54%), needle biopsy in 1039 (45%), and unspecified in 39 (2%). Only 29% were treated with hormonal therapy within 6 mo of diagnosis. Local progression occurred in 335 men, including 212 undergoing TURP. Factors most predictive of local progression on multivariable analysis were PSA at diagnosis and Gleason score of the diagnostic tissue (detrimental), and early hormonal therapy (protective). We present a nomogram that predicts the likelihood of local progression within 120 mo after diagnosis.

Conclusions

Men with clinically detected localized prostate cancer managed without curative intent have an approximately 15% risk for local progression within 10 yr of diagnosis. Among those with progression, the need for treatment is common, even among men diagnosed by TURP. When counseling men who are candidates for management without curative intent, the likelihood of symptoms from local progression must be considered.