A nonparametric method of comparing the partial areas under two correlated receiver operating characteristic curves

A non-parametric method was developed and tested to compare the partial areas under two correlated Receiver Operating Characteristic curves. Based on the theory of generalized U-statistics the mathematical formulas have been derived for computing ROC area, and the variance and covariance between the portions of two ROC curves. A practical SAS application also has been developed to facilitate the calculations. The accuracy of the non-parametric method was evaluated by comparing it to other methods. By applying our method to the data from a published ROC analysis of CT image, our results are very close to theirs. A hypothetical example was used to demonstrate the effects of two crossed ROC curves. The two ROC areas are the same. However each portion of the area between two ROC curves were found to be significantly different by the partial ROC curve analysis. For computation of ROC curves with large scales, such as a logistic regression model, we applied our method to the breast cancer study with Medicare claims data. It yielded the same ROC area computation as the SAS Logistic procedure. Our method also provides an alternative to the global summary of ROC area comparison by directly comparing the true-positive rates for two regression models and by determining the range of false-positive values where the models differ. ^

Tumor heterogeneity dictates sensitivity to gefitinib (Iressa(TM)/ZD1839) in solid tumor models

The epidermal growth factor receptor (EGFR) and its ligands are overexpressed in many human tumors, including bladder and pancreas, correlating with a more aggressive tumor phenotype and poor patient prognosis. We initiated the present study to characterize the heterogeneity of gefitinib responsiveness in a panel of human bladder and pancreatic cancer cell lines in order to identify the biological characteristics of EGFR-dependent proliferation that could be used to prospectively identify drug-sensitive tumors. A second objective was to elucidate how to best exploit these results by utilizing gefitinib in combination therapy. To these ends, we examined the effects of the EGFR antagonist gefitinib on proliferation and apoptosis in a panel of 18 human bladder cancer cell lines and 9 human pancreatic cancer cell lines. Our data confirmed the existence of marked heterogeneity in Iressa responsiveness with less than half of the cell lines displaying significant growth inhibition by clinically relevant concentrations of the drug. Gefitinib responsiveness was found to be p27 kip1 dependent as DNA synthesis was restored following exposure to p27siRNA. Unfortunately, Iressa responsiveness was not closely linked to surface EGFR or TGF-α expression in the bladder cancer cells, however, cellular TGF-α expression correlated directly with Iressa sensitivity in the pancreatic cancer cell lines. These findings provide the potential for prospectively identifying patients with drug-sensitive tumors. ^ Further studies aimed at exploiting gefitinib-mediated cell cycle effects led us to investigate if gefitinib-mediated TRAIL sensitization correlated with increased p27kip1 accumulation. We observed that increased TRAIL sensitivity following gefitinib exposure was not dependent on p27 kip1 expression. Additional studies initiated to examine the role(s) of Akt and Erk signaling demonstrated that exposure to PI3K or MEK inhibitors significantly enhanced TRAIL-induced apoptosis at concentrations that block target phosphorylation. Furthermore, combinations of TRAIL and the PI3K or MEK inhibitors increased procaspase-8 processing above levels observed with TRAIL alone, indicating that the effects were exerted at the level of caspase-8 activation, considered the earliest step in the TRAIL pathway. ^

Analysis of life expectancy using linear and nonlinear models

Life expectancy has consistently increased over the last 150 years due to improvements in nutrition, medicine, and public health. Several studies found that in many developed countries, life expectancy continued to rise following a nearly linear trend, which was contrary to a common belief that the rate of improvement in life expectancy would decelerate and was fit with an S-shaped curve. Using samples of countries that exhibited a wide range of economic development levels, we explored the change in life expectancy over time by employing both nonlinear and linear models. We then observed if there were any significant differences in estimates between linear models, assuming an auto-correlated error structure. When data did not have a sigmoidal shape, nonlinear growth models sometimes failed to provide meaningful parameter estimates. The existence of an inflection point and asymptotes in the growth models made them inflexible with life expectancy data. In linear models, there was no significant difference in the life expectancy growth rate and future estimates between ordinary least squares (OLS) and generalized least squares (GLS). However, the generalized least squares model was more robust because the data involved time-series variables and residuals were positively correlated. ^