Elena B. Elkin

Decision Curve Analysis: A Novel Method for Evaluating Prediction Models

Background. Diagnostic and prognostic models are typically evaluated with measures of accuracy that do not address clinical consequences. Decision-analytic techniques allow assessment of clinical outcomes but often require collection of additional information and may be cumbersome to apply to models that yield a continuous result. The authors sought a method for evaluating and comparing prediction models that incorporates clinical consequences, requires only the data set on which the models are tested, and can be applied to models that have either continuous or dichotomous results. Method. The authors describe decision curve analysis, a simple, novel method of evaluating predictive models. They start by assuming that the threshold probability of a disease or event at which a patient would opt for treatment is informative of how the patient weighs the relative harms of a false-positive and a false-negative prediction. This theoretical relationship is then used to derive the net benefit of the model across different threshold probabilities. Plotting net benefit against threshold probability yields the “decision curve.” The authors apply the method to models for the prediction of seminal vesicle invasion in prostate cancer patients. Decision curve analysis identified the range of threshold probabilities in which a model was of value, the magnitude of benefit, and which of several models was optimal. Conclusion. Decision curve analysis is a suitable method for evaluating alternative diagnostic and prognostic strategies that has advantages over other commonly used measures and techniques.

Partial nephrectomy versus radical nephrectomy in patients with small renal tumors--is there a difference in mortality and cardiovascular outcomes?

PURPOSE Compared with partial nephrectomy, radical nephrectomy increases the risk of chronic kidney disease, which is a significant risk factor for cardiovascular events and death. Given equivalent oncological efficacy in patients with small renal tumors, radical nephrectomy may result in overtreatment. We analyzed a population based cohort of patients to determine whether radical nephrectomy is associated with an increase in cardiovascular events and mortality compared with partial nephrectomy. MATERIALS AND METHODS Using Surveillance, Epidemiology and End Results cancer registry data linked with Medicare claims we identified 2,991 patients older than 66 years who were treated with radical or partial nephrectomy for renal tumors 4 cm or less between 1995 and 2002. The primary end points of cardiovascular events and overall survival were assessed using Kaplan-Meier survival estimation, Cox proportional hazards regression and negative binomial regression. RESULTS A total of 2,547 patients (81%) underwent radical nephrectomy and 556 (19%) underwent partial nephrectomy. During a median followup of 4 years 609 patients experienced a cardiovascular event and 892 died. When adjusting for preoperative demographic and comorbid variables, radical nephrectomy was associated with an increased risk of overall mortality (HR 1.38, p <0.01) and a 1.4 times greater number of cardiovascular events after surgery (p <0.05). However, radical nephrectomy was not significantly associated with time to first cardiovascular event (HR 1.21, p = 0.10) or with cardiovascular death (HR 0.95, p = 0.84). CONCLUSIONS Radical nephrectomy, which is currently the most common treatment for small renal tumors, may be associated with significant, adverse treatment effects compared with partial nephrectomy. Partial nephrectomy should be considered in most patients with small renal tumors.

Strategies to Identify the Lynch Syndrome Among Patients With Colorectal Cancer: A Cost-Effectiveness Analysis

BACKGROUND Testing has been advocated for all persons with newly diagnosed colorectal cancer to identify families with the Lynch syndrome, an autosomal dominant cancer-predisposition syndrome that is a paradigm for personalized medicine. OBJECTIVE To estimate the effectiveness and cost-effectiveness of strategies to identify the Lynch syndrome, with attention to sex, age at screening, and differential effects for probands and relatives. DESIGN Markov model that incorporated risk for colorectal, endometrial, and ovarian cancers. DATA SOURCES Published literature. TARGET POPULATION All persons with newly diagnosed colorectal cancer and their relatives. TIME HORIZON Lifetime. PERSPECTIVE Third-party payer. INTERVENTION Strategies based on clinical criteria, prediction algorithms, tumor testing, or up-front germline mutation testing, followed by tailored screening and risk-reducing surgery. OUTCOME MEASURES Life-years, cancer cases and deaths, costs, and incremental cost-effectiveness ratios. RESULTS OF BASE-CASE ANALYSIS The benefit of all strategies accrued primarily to relatives with a mutation associated with the Lynch syndrome, particularly women, whose life expectancy could increase by approximately 4 years with hysterectomy and salpingo-oophorectomy and adherence to colorectal cancer screening recommendations. At current rates of germline testing, screening, and prophylactic surgery, the strategies reduced deaths from colorectal cancer by 7% to 42% and deaths from endometrial and ovarian cancer by 1% to 6%. Among tumor-testing strategies, immunohistochemistry followed by BRAF mutation testing was preferred, with an incremental cost-effectiveness ratio of

HER-2 Testing and Trastuzumab Therapy for Metastatic Breast Cancer: A Cost-Effectiveness Analysis

36,200 per life-year gained. RESULTS OF SENSITIVITY ANALYSIS The number of relatives tested per proband was a critical determinant of both effectiveness and cost-effectiveness, with testing of 3 to 4 relatives required for most strategies to meet a threshold of

Extensions to decision curve analysis, a novel method for evaluating diagnostic tests, prediction models and molecular markers

50,000 per life-year gained. Immunohistochemistry followed by BRAF mutation testing was preferred in 59% of iterations in probabilistic sensitivity analysis at a threshold of

Adjuvant Chemotherapy and Survival in Older Women With Hormone Receptor–Negative Breast Cancer: Assessing Outcome in a Population-Based, Observational Cohort

100,000 per life-year gained. Screening for the Lynch syndrome with immunohistochemistry followed by BRAF mutation testing only up to age 70 years cost

Patterns of care in elderly glioblastoma patients.

44,000 per incremental life-year gained compared with screening only up to age 60 years, and screening without an upper age limit cost

Physician visits prior to treatment for clinically localized prostate cancer

88,700 per incremental life-year gained compared with screening only up to age 70 years. LIMITATION Other types of cancer, uncertain family pedigrees, and genetic variants of unknown significance were not considered. CONCLUSION Widespread colorectal tumor testing to identify families with the Lynch syndrome could yield substantial benefits at acceptable costs, particularly for women with a mutation associated with the Lynch syndrome who begin regular screening and have risk-reducing surgery. The cost-effectiveness of such testing depends on the participation rate among relatives at risk for the Lynch syndrome. PRIMARY FUNDING SOURCE National Institutes of Health.

Achieving Optimal Outcomes After Radical Prostatectomy

PURPOSE Trastuzumab therapy has been shown to benefit metastatic breast cancer patients whose tumors exhibit HER-2 protein overexpression or gene amplification. Several tests of varying accuracy and cost are available to identify candidates for trastuzumab. We estimated the cost-effectiveness of alternative HER-2 testing and trastuzumab treatment strategies. PATIENTS AND METHODS We performed a decision analysis using a state-transition model to simulate clinical practice in a hypothetical cohort of 65-year-old metastatic breast cancer patients. Outcomes were quality-adjusted life-years (QALYs), lifetime cost, and incremental cost-effectiveness ratio (ICER). Interventions included testing with the HercepTest (DAKO, Carpinteria, CA) immunohistochemical assay alone, fluorescence in situ hybridization (FISH) alone, and both tests, followed by trastuzumab and chemotherapy for patients with positive test results and chemotherapy alone for patients with negative test results. RESULTS In the base case, initial HercepTest with FISH confirmation of all positive results had an ICER of

Cancer's Next Frontier: Addressing High and Increasing Costs

125,000 per QALY gained. The incremental cost-effectiveness of initial FISH was