The diagnostic likelihood ratio function and modified test for trend: Identifying, evaluating, and validating nontraditional biomarkers in case-control studies.

The ROC curve and its associated summary statistic, the AUC, are used to identify informative diagnostic biomarkers under the assumption that risk of disease is a monotone function of the biomarker. We refer to biomarkers that meet this assumption as traditional, and those that do not as nontraditional. Nontraditional biomarkers most often arise when both low and high biomarker values are associated with an outcome of interest, such as blood pressure with medical complications or leukocyte count with ICU prognosis. Since nontraditional biomarkers do not meet the assumptions for ROC-based analyses, we propose using the discrete diagnostic likelihood ratio (DLR) function to evaluate a wider class of informative biomarkers. We obtain the DLR function using the multinomial logistic regression (MLR) model to improve upon existing estimation techniques, and implement a likelihood ratio test to identify candidate informative traditional and nontraditional biomarkers. We propose a modification of the Cochran-Armitage test for trend that separates biomarkers deemed informative into traditional and nontraditional categories. The statistical properties of the likelihood ratio test and modified test for trend are explored under simulation. Together, these methods achieve the identification, evaluation, and validation of biomarkers from early discovery research. Finally, we show that incorporating covariates into the MLR model results in a covariate-adjusted DLR function that is useful for integrating multiple sources of information in clinical decision making. The methods are applied to gene expression data from subjects with high grade serous ovarian cancer, where stage, early stage vs late stage, is the outcome of interest.

Dialysis Nonadherence and Kidney Transplant Outcomes: A Retrospective Cohort Study.

RATIONALE & OBJECTIVE: Concerns about nonadherent behaviors often prevent dialysis patients from entering waitlists for transplant even though there is an inconsistent association of these behaviors with posttransplant outcomes. We examined the association between plausible metrics of nonadherence related to dialysis treatment and posttransplant outcomes. STUDY DESIGN: Retrospective cohort. We linked national dialysis treatment data with transplant registry data. SETTING AND PARTICIPANTS: Adult patients receiving maintenance hemodialysis from January 1, 2004, through December 31, 2014, who received a kidney transplant at a US center. EXPOSURES: We examined 5 nonadherence metrics: serum potassium level (≥5.2 mEq/L), serum phosphorus level (>5.5 mg/dL), interdialytic weight gain (IDWG; ≥5 L), shortened treatments (≥30 min), and missed treatments (≥1); missed treatment data were available only for 2004-2009. These metrics were characterized per proportion of time under observation. Dialysis observation time was divided into 3-month intervals (quarters), and the number of nonadherent measurements in each domain was calculated for each quarter. OUTCOMES: Allograft loss, mortality, and acute rejection in the first posttransplant year. ANALYTICAL APPROACH: Using Cox proportional hazards and logistic regression, we estimated the hazard ratios for graft loss and mortality and odds ratios for rejection. RESULTS: 9,543 patients met inclusion criteria. In our primary model, hyperphosphatemia (adjusted hazard ratio [aHR], 1.27 [95% CI, 1.08-1.49]), large IDWG (aHR, 1.39 [95% CI, 1.23-1.59]), and shortened treatments (aHR, 1.54 [95% CI, 1.12-2.13]) were associated with greater rates of allograft loss, but hyperkalemia was not. Large IDWG (aHR, 1.49 [95% CI, 1.29-1.73]) and shortened treatments (aHR, 1.34 [95% CI, 1.13-1.58]) were associated with mortality, whereas hyperkalemia and hyperphosphatemia were not. Only shortened treatments were associated with an increased risk of acute rejection (adjusted odds ratio, 3.88 [95% CI, 1.98-7.58]). In models limited to the years 2004-2009 that included missed treatments, missed treatments were associated only with mortality. LIMITATIONS: Unmeasured confounding (eg, dietary data); adherence metrics used may have multiple, complex causes. CONCLUSIONS: Plausible measures of dialysis nonadherence have long-term associations with allograft and patient survival. Behavioral metrics were more closely associated with outcomes than laboratory markers were. The implications of nonadherent behaviors for dialysis patients must be carefully considered before patients are excluded from transplantation.

Urethral involvement is associated with higher mortality and local recurrence in vulvar melanoma: a single institutional experience.

Vulvar malignant melanoma (VMM), although uncommon, comprises 5-10% of all vulvar malignancies. Local control is notoriously poor in VMM with recurrence rates of 30-50% compared with approximately 3% in cutaneous melanomas. We studied clinicopathologic features of 37 women with VMM, after reviewing three decades of clinical follow-up data in our institutional databases. Most patients were Caucasian (n = 35) with an average age at diagnosis of 60.6 years (range 23-83). The most common subtype was mucosal lentiginous melanoma (n = 25). We compared Kaplan-Meier survival curves of 31 patients defined by clinical and microscopic attributes using exact log-rank tests. Younger patients at diagnosis (23-64 years), those with thin melanomas (≤1 mm), and those with Clark's level II or III tumors had better 5-year survival rates than older patients (65-83 years) and those with thick melanomas (>1 mm) and those with Clark's level IV or V (P ≤ 0.05), respectively, by exact log-rank test. Local recurrence of melanoma occurred in 15 patients. Nine patients (24%) had eventual urethral involvement by malignant melanoma, and this feature was associated with significantly shorter survival (P = 0.036). Patients with urethral involvement had shorter median time to death and worse 5-year survival rates. Given that spread to the urethra is common in VMM and urethral recurrence is also associated with mortality, pathology excision specimens should be carefully reviewed with attention to urethral involvement as a potentially important prognostic factor.

Increasing Urologic Care Ratios: Implications of Male Patient Care in Florida.

Ongoing trends have revealed an inverse relationship between population growth and the number of practicing urologists in the U.S. per capita, which threatens urologic care accessibility. Furthermore, different regions in the United States may be more negatively impacted due to higher population growth rates. The state of Florida witnessed over a 10% higher growth rate compared with national figures between 2000 and 2015. Coupled with data suggesting that since the 1980s, the number of U.S. urologists per capita has been decreasing, the foreseeable future presents many challenges regarding health equity and accessibility. This secondary analysis aimed to investigate the implications of forecasted urologic care decline within a growing population and how it can contribute to adverse male health outcomes. National- and state-level data were collected to calculate a series of urologic care ratios as defined by the number of urologists compared with population sizes. Analyses revealed that national-level urologic care ratios and prostate cancer incidence rates have a significant positive relationship, lending to the conclusion that with fewer urologists, the number of cases identified will decrease. State-level forecasted models indicated that the urologic care ratio will decrease approximately 30% in Florida from 6.23 per 100,000 in 2010 to 4.39 per 100,000 by the year 2030. As growth in demand for urologic care will increase in the next decade, a dire public health scenario is potentially unfolding. Future implications of undiagnosed prostate cancer due to the lack of access will drive an increase in mortality rates as well as health equity concerns for men.