Stochastic search for semiparametric linear regression models

This paper introduces and analyzes a stochastic search method for parameter estimation in linear regression models in the spirit of Beran and Millar [Ann. Statist. 15(3) (1987) 1131–1154]. The idea is to generate a random finite subset of a parameter space which will automatically contain points which are very close to an unknown true parameter. The motivation for this procedure comes from recent work of Dümbgen et al. [Ann. Statist. 39(2) (2011) 702–730] on regression models with log-concave error distributions.

Quantile-Based Optimization of Noisy Computer Experiments with Tunable Precision

This article addresses the issue of kriging-based optimization of stochastic simulators. Many of these simulators depend on factors that tune the level of precision of the response, the gain in accuracy being at a price of computational time. The contribution of this work is two-fold: first, we propose a quantile-based criterion for the sequential design of experiments, in the fashion of the classical expected improvement criterion, which allows an elegant treatment of heterogeneous response precisions. Second, we present a procedure for the allocation of the computational time given to each measurement, allowing a better distribution of the computational effort and increased efficiency. Finally, the optimization method is applied to an original application in nuclear criticality safety. This article has supplementary material available online. The proposed criterion is available in the R package DiceOptim.

Tumor Regression Grading of Gastrointestinal Carcinomas after Neoadjuvant Treatment

Multimodal therapy concepts have been successfully implemented in the treatment of locally advanced gastrointestinal malignancies. The effects of neoadjuvant chemo- or radiochemotherapy such as scarry fibrosis or resorptive changes and inflammation can be determined by histopathological investigation of the subsequent resection specimen. Tumor regression grading (TRG) systems which aim to categorize the amount of regressive changes after cytotoxic treatment mostly refer onto the amount of therapy induced fibrosis in relation to residual tumor or the estimated percentage of residual tumor in relation to the previous tumor site. Commonly used TRGs for upper gastrointestinal carcinomas are the Mandard grading and the Becker grading system, e.g., and for rectal cancer the Dworak or the Rödel grading system, or other systems which follow similar definitions. Namely for gastro-esophageal carcinomas these TRGs provide important prognostic information since complete or subtotal tumor regression has shown to be associated with better patient's outcome. The prognostic value of TRG may even exceed those of currently used staging systems (e.g., TNM staging) for tumors treated by neoadjuvant therapy. There have been some limitations described regarding interobserver variability especially in borderline cases, which may be improved by standardization of work up of resection specimen and better training of histopathologic determination of regressive changes. It is highly recommended that TRG should be implemented in every histopathological report of neoadjuvant treated gastrointestinal carcinomas. The aim of this review is to disclose the relevance of histomorphological TRG to accomplish an optimal therapy for patients with gastrointestinal carcinomas.

Tumor regression grade of urothelial bladder cancer after neoadjuvant chemotherapy: a novel and successful strategy to predict survival.

Histopathologic tumor regression grades (TRGs) after neoadjuvant chemotherapy predict survival in different cancers. In bladder cancer, corresponding studies have not been conducted. Fifty-six patients with advanced invasive urothelial bladder cancer received neoadjuvant chemotherapy before cystectomy and lymphadenectomy. TRGs were defined as follows: TRG1: complete tumor regression; TRG2: >50% tumor regression; TRG3: 50% or less tumor regression. Separate TRGs were assigned for primary tumors and corresponding lymph nodes. The prognostic impact of these 2 TRGs, the highest (dominant) TRG per patient, and competing tumor features reflecting tumor regression (ypT/ypN stage, maximum diameter of the residual tumor) were determined. Tumor characteristics in initial transurethral resection of the bladder specimens were tested for response prediction. The frequency of TRGs 1, 2, and 3 in the primary tumors were n=16, n=19, and n=21; corresponding data from the lymph nodes were n=31, n=9, and n=16. Interobserver agreement in determination of the TRG was strong (κ=0.8). Univariately, all evaluated parameters were significantly (P≤0.001) related to overall survival; however, the segregation of the Kaplan-Meier curves was best for the dominant TRG. In multivariate analysis, only dominant TRG predicted overall survival independently (P=0.035). In transurethral resection specimens of the chemotherapy-naive bladder cancer, the only tumor feature with significant (P<0.03) predictive value for therapy response was a high proliferation rate. In conclusion, among all parameters reflecting tumor regression, the dominant TRG was the only independent risk factor. A favorable chemotherapy response is associated with a high proliferation rate in the initial chemotherapy-naive bladder cancer. This feature might help personalize neoadjuvant chemotherapy.

Comparison of partial least squares regression and principal component regression for pelvic shape prediction.

This paper studied two different regression techniques for pelvic shape prediction, i.e., the partial least square regression (PLSR) and the principal component regression (PCR). Three different predictors such as surface landmarks, morphological parameters, or surface models of neighboring structures were used in a cross-validation study to predict the pelvic shape. Results obtained from applying these two different regression techniques were compared to the population mean model. In almost all the prediction experiments, both regression techniques unanimously generated better results than the population mean model, while the difference on prediction accuracy between these two regression methods is not statistically significant (α=0.01).

3D volumetric intensity reconsturction from 2D x-ray images using partial least squares regression

Reconstruction of shape and intensity from 2D x-ray images has drawn more and more attentions. Previously introduced work suffers from the long computing time due to its iterative optimization characteristics and the requirement of generating digitally reconstructed radiographs within each iteration. In this paper, we propose a novel method which uses a patient-specific 3D surface model reconstructed from 2D x-ray images as a surrogate to get a patient-specific volumetric intensity reconstruction via partial least squares regression. No DRR generation is needed. The method was validated on 20 cadaveric proximal femurs by performing a leave-one-out study. Qualitative and quantitative results demonstrated the efficacy of the present method. Compared to the existing work, the present method has the advantage of much shorter computing time and can be applied to both DXA images as well as conventional x-ray images, which may hold the potentials to be applied to clinical routine task such as total hip arthroplasty (THA).

Robust Proximal Femur Segmentation in Conventional X-Ray Images via Random Forest Regression on Multi-resolution Gradient Features

In this paper, we propose a fully automatic, robust approach for segmenting proximal femur in conventional X-ray images. Our method is based on hierarchical landmark detection by random forest regression, where the detection results of 22 global landmarks are used to do the spatial normalization, and the detection results of the 59 local landmarks serve as the image cue for instantiation of a statistical shape model of the proximal femur. To detect landmarks in both levels, we use multi-resolution HoG (Histogram of Oriented Gradients) as features which can achieve better accuracy and robustness. The efficacy of the present method is demonstrated by experiments conducted on 150 clinical x-ray images. It was found that the present method could achieve an average point-to-curve error of 2.0 mm and that the present method was robust to low image contrast, noise and occlusions caused by implants.

Fully Automatic Segmentation of AP Pelvis X-rays via Random Forest Regression and Hierarchical Sparse Shape Composition

Knowledge of landmarks and contours in anteroposterior (AP) pelvis X-rays is invaluable for computer aided diagnosis, hip surgery planning and image-guided interventions. This paper presents a fully automatic and robust approach for landmarking and segmentation of both pelvis and femur in a conventional AP X-ray. Our approach is based on random forest regression and hierarchical sparse shape composition. Experiments conducted on 436 clinical AP pelvis x-rays show that our approach achieves an average point-to-curve error around 1.3 mm for femur and 2.2 mm for pelvis, both with success rates around 98%. Compared to existing methods, our approach exhibits better performance in both the robustness and the accuracy.

Plotting regression coefficients and other estimates in Stata

Graphical presentation of regression results has become increasingly popular in the scientific literature, as graphs are much easier to read than tables in many cases. In Stata such plots can be produced by the -marginsplot- command. However, while -marginsplot- is very versatile and flexible, it has two major limitations: it can only process results left behind by -margins- and it can only handle one set of results at the time. In this article I introduce a new command called -coefplot- that overcomes these limitations. It plots results from any estimation command and combines results from several models into a single graph. The default behavior of -coefplot- is to plot markers for coefficients and horizontal spikes for confidence intervals. However, -coefplot- can also produce various other types of graphs. The capabilities of -coefplot- are illustrated in this article using a series of examples.

COEFPLOT: Stata module to plot regression coefficients and other results

coefplot plots results from estimation commands or Stata matrices. Results from multiple models or matrices can be combined in a single graph. The default behavior of coefplot is to draw markers for coefficients and horizontal spikes for confidence intervals. However, coefplot can also produce various other types of graphs.

Regression of the inflammatory microenvironment of the peritoneal cavity in women with endometriosis by GnRHa treatment.

OBJECTIVE To investigate the effect of gonadotropin-releasing hormone analogues (GnRHa) on the peritoneal fluid microenvironment in women with endometriosis. STUDY DESIGN Peritoneal fluid was collected from 85 women with severe endometriosis (rAFS stage III and IV) during laparoscopic surgery during the proliferative phase. Prior to surgery clinical data were collected. The concentrations of specific markers for endometriosis in the peritoneal fluid were determined using an ELISA and a comparison between peritoneal fluid markers in women using GnRHa and no hormonal treatment was performed using a non-parametric Mann-Whitney U test. RESULTS The study included peritoneal fluid from 39 patients who had been administered GnRHa (Zoladex(®)) in the three months prior to surgery and 46 from women with no hormonal treatment in this period. Concentrations of IL-8, PAPP-A, glycodelin-A and midkine were significantly reduced in the GnRHa treatment group compared to women receiving no hormonal treatment. RANTES, MCP-1, ENA-78, TNF-α, OPG, IP-10 and defensin showed no significant change between the two groups. CONCLUSIONS GnRHa mediate a significant regression in the inflammatory nature of the peritoneal microenvironment in women with endometriosis.