Matthew D. DiFranco

Ensemble based system for whole-slide prostate cancer probability mapping using color texture features

We present a tile-based approach for producing clinically relevant probability maps of prostatic carcinoma in histological sections from radical prostatectomy. Our methodology incorporates ensemble learning for feature selection and classification on expert-annotated images. Random forest feature selection performed over varying training sets provides a subset of generalized CIEL*a*b* co-occurrence texture features, while sample selection strategies with minimal constraints reduce training data requirements to achieve reliable results. Ensembles of classifiers are built using expert-annotated tiles from training images, and scores for the probability of cancer presence are calculated from the responses of each classifier in the ensemble. Spatial filtering of tile-based texture features prior to classification results in increased heat-map coherence as well as AUC values of 95% using ensembles of either random forests or support vector machines. Our approach is designed for adaptation to different imaging modalities, image features, and histological decision domains.

An Intra-Individual Comparison of MRI, [18F]-FET and [18F]-FLT PET in Patients with High-Grade Gliomas

Objectives Intra-individual spatial overlap analysis of tumor volumes assessed by MRI, the amino acid PET tracer [18F]-FET and the nucleoside PET tracer [18F]-FLT in high-grade gliomas (HGG). Methods MRI, [18F]-FET and [18F]-FLT PET data sets were retrospectively analyzed in 23 HGG patients. Morphologic tumor volumes on MRI (post-contrast T1 (cT1) and T2 images) were calculated using a semi-automatic image segmentation method. Metabolic tumor volumes for [18F]-FET and [18F]-FLT PETs were determined by image segmentation using a threshold-based volume of interest analysis. After co-registration with MRI the morphologic and metabolic tumor volumes were compared on an intra-individual basis in order to estimate spatial overlaps using the Spearmans rank correlation coefficient and the Mann-Whitney U test. Results [18F]-FLT uptake was negative in tumors with no or only moderate contrast enhancement on MRI, detecting only 21 of 23 (91%) HGG. In addition, [18F]-FLT uptake was mainly restricted to cT1 tumor areas on MRI and [18F]-FLT volumes strongly correlated with cT1 volumes (r = 0.841, p<0.001). In contrast, [18F]-FET PET detected 22 of 23 (96%) HGG. [18F]-FET uptake beyond areas of cT1 was found in 61% of cases and [18F]-FET volumes showed only a moderate correlation with cT1 volumes (r = 0.573, p<0.001). Metabolic tumor volumes beyond cT1 tumor areas were significantly larger for [18F]-FET compared to [18F]-FLT tracer uptake (8.3 vs. 2.7 cm3, p<0.001). Conclusion In HGG [18F]-FET but not [18F]-FLT PET was able to detect metabolic active tumor tissue beyond contrast enhancing tumor on MRI. In contrast to [18F]-FET, blood-brain barrier breakdown seems to be a prerequisite for [18F]-FLT tracer uptake.

A Study on the Basic Criteria for Selecting Heterogeneity Parameters of F18-FDG PET Images

Textural analysis might give new insights into the quantitative characterization of metabolically active tumors. More than thirty textural parameters have been investigated in former F18-FDG studies already. The purpose of the paper is to declare basic requirements as a selection strategy to identify the most appropriate heterogeneity parameters to measure textural features. Our predefined requirements were: a reliable heterogeneity parameter has to be volume independent, reproducible, and suitable for expressing quantitatively the degree of heterogeneity. Based on this criteria, we compared various suggested measures of homogeneity. A homogeneous cylindrical phantom was measured on three different PET/CT scanners using the commonly used protocol. In addition, a custom-made inhomogeneous tumor insert placed into the NEMA image quality phantom was imaged with a set of acquisition times and several different reconstruction protocols. PET data of 65 patients with proven lung lesions were retrospectively analyzed as well. Four heterogeneity parameters out of 27 were found as the most attractive ones to characterize the textural properties of metabolically active tumors in FDG PET images. These four parameters included Entropy, Contrast, Correlation, and Coefficient of Variation. These parameters were independent of delineated tumor volume (bigger than 25–30 ml), provided reproducible values (relative standard deviation< 10%), and showed high sensitivity to changes in heterogeneity. Phantom measurements are a viable way to test the reliability of heterogeneity parameters that would be of interest to nuclear imaging clinicians.

High-Resolution Peripheral Quantitative CT Imaging: Cortical Porosity, Poor Trabecular Bone Microarchitecture, and Low Bone Strength in Lung Transplant Recipients

PURPOSE To characterize bone microarchitecture and quantify bone strength in lung transplant (LT) recipients by using high-resolution (HR) peripheral quantitative computed tomographic (CT) imaging of the ultradistal radius. MATERIALS AND METHODS After study approval by the local ethics committee, all participants provided written informed consent. Included were 118 participants (58 LT recipients [mean age, 46.8 years ± 1.9; 30 women, 28 men] and 60 control participants [mean age, 39.9 years ± 1.9; 41 women, 19 men]) between April 2010 and May 2012. HR peripheral quantitative CT of the ultradistal radius was performed and evaluated for bone mineral density and trabecular and cortical bone microarchitecture. Mechanical competence was quantified by microfinite element analysis. Differences between LT recipients and control participants were determined by using two-way factorial analysis of covariance with age adjustment. RESULTS Total and trabecular bone mineral density were significantly lower (-13.4% and -16.4%, respectively; P = .001) in LT recipients than in healthy control participants. LT recipients had lower trabecular number (-9.7%; P = .004) and lower trabecular thickness (-8.1%; P = .025). Trabecular separation and trabecular network heterogeneity were higher (+24.3% and +63.9%, respectively; P = .007 and P = .012, respectively) in LT recipients. Moreover, there was pronounced cortical porosity (+31.3%; P = .035) and lower cortical thickness (-10.2%, P = .005) after LT. In addition, mechanical competence was impaired, which was reflected by low stiffness (-15.0%; P < .001), low failure force (-14.8%; P < .001), and low bone strength (-14.6%; P < .001). CONCLUSION Men and women with recent LT showed severe deficits in cortical and trabecular bone microarchitecture. Poor bone microarchitecture and low bone strength are likely to contribute to high fracture susceptibility observed in LT recipients.

Reproducibility of MRI Dixon-based attenuation correction in combined PET/MR with applications for lean body mass estimation

The aim of this study was to assess the reproducibility of standard, Dixon-based attenuation correction (MR-AC) in PET/MR imaging. A further aim was to estimate a patient-specific lean body mass (LBM) from these MR-AC data. Methods: Ten subjects were positioned in a fully integrated PET/MR system, and 3 consecutive multibed acquisitions of the standard MR-AC image data were acquired. For each subject and MR-AC map, the following compartmental volumes were calculated: total body, soft tissue (ST), fat, lung, and intermediate tissue (IT). Intrasubject differences in the total body and subcompartmental volumes (ST, fat, lung, and IT) were assessed by means of coefficients of variation (CVs) calculated across the 3 consecutive measurements and, again, across these measurements but excluding those affected by major artifacts. All subjects underwent a body composition measurement using air displacement plethysmography (ADP) that was used to calculate a reference LBMADP. A second LBM estimate was derived from available MR-AC data using a formula incorporating the respective tissue volumes and densities as well as the subject-specific body weights. A third LBM estimate was obtained from a sex-specific formula (LBMFormula). Pearson correlation was calculated for LBMADP, LBMMR-AC, and LBMFormula. Further, linear regression analysis was performed on LBMMR-AC and LBMADP. Results: The mean CV for all 30 scans was 2.1 ± 1.9% (TB). When missing tissue artifacts were excluded, the CV was reduced to 0.3 ± 0.2%. The mean CVs for the subcompartments before and after exclusion of artifacts were 0.9 ± 1.1% and 0.7 ± 0.7% for the ST, 2.9 ± 4.1% and 1.3 ± 1.0% for fat, and 3.6 ± 3.9% and 1.3 ± 0.7% for the IT, respectively. Correlation was highest for LBMMR-AC and LBMADP (r = 0.99). Linear regression of data excluding artifacts resulted in a scaling factor of 1.06 for LBMMR-AC. Conclusion: LBMMR-AC is shown to correlate well with standard LBM measurements and thus offers routine LBM-based SUV quantification in PET/MR. However, MR-AC images must be controlled for systematic artifacts, including missing tissue and tissue swaps. Efforts to minimize these artifacts could help improve the reproducibility of MR-AC.

Ensemble Prostate Tumor Classification in H&E Whole Slide Imaging via Stain Normalization and Cell Density Estimation

Classification of prostate tumor regions in digital histology images requires comparable features across datasets. Here we introduce adaptive cell density estimation and apply H&E stain normalization into a supervised classification framework to improve inter-cohort classifier robustness. The framework uses Random Forest feature selection, class-balanced training example subsampling and support vector machine SVM classification to predict the presence of high- and low-grade prostate cancer HG-PCa and LG-PCa on image tiles. Using annotated whole-slide prostate digital pathology images to train and test on two separate patient cohorts, classification performance, as measured with area under the ROC curve AUC, was 0.703 for HG-PCa and 0.705 for LG-PCa. These results improve upon previous work and demonstrate the effectiveness of cell-density and stain normalization on classification of prostate digital slides across cohorts.

Performance assessment of automated tissue characterization for prostate H and E stained histopathology

Reliable automated prostate tumor detection and characterization in whole-mount histology images is sought in many applications, including post-resection tumor staging and as ground-truth data for multi-parametric MRI interpretation. In this study, an ensemble-based supervised classification algorithm for high-resolution histology images was trained on tile-based image features including histogram and gray-level co-occurrence statistics. The algorithm was assessed using different combinations of H and E prostate slides from two separate medical centers and at two different magnifications (400x and 200x), with the aim of applying tumor classification models to new data. Slides from both datasets were annotated by expert pathologists in order to identify homogeneous cancerous and non-cancerous tissue regions of interest, which were then categorized as (1) low-grade tumor (LG-PCa), including Gleason 3 and high-grade prostatic intraepithelial neoplasia (HG-PIN), (2) high-grade tumor (HG-PCa), including various Gleason 4 and 5 patterns, or (3) non-cancerous, including benign stroma and benign prostatic hyperplasia (BPH). Classification models for both LG-PCa and HG-PCa were separately trained using a support vector machine (SVM) approach, and per-tile tumor prediction maps were generated from the resulting ensembles. Results showed high sensitivity for predicting HG-PCa with an AUC up to 0.822 using training data from both medical centres, while LG-PCa showed a lower sensitivity of 0.763 with the same training data. Visual inspection of cancer probability heatmaps from 9 patients showed that 17/19 tumors were detected, and HG-PCa generally reported less false positives than LG-PCa.

A New Nasopharyngeal Dynamic Reference Frame Improves Accuracy in Navigated Skull Base Targets

Objective. We questioned whether the position of the dynamic reference frame (DRF) influences the application accuracy in electromagnetically navigated cranial procedures. A carrier for an electromagnetic DRF was developed, which could be fixed at the posterior edge of the vomer near the center of the head. This nasopharyngeal DRF was compared with a standard DRF fixed to the surface of the forehead. Methods. Image coordinates and real-world coordinates were co-registered and the total target error (TTE) was measured in the frontal and the lateral skull base of formalin fixed human head. At each anatomical site, 10 targets served for TTE determinations and 5 different fiducial combinations were used for registration. Results. With the nasopharyngeal DRF, lower TTE values (2.8 ± 1.4 mm; mean ± SD) were observed when compared with the forehead DRF (3.7 ± 2.8 mm; P = .004). TTEs of both anatomical sites investigated were significantly lower when using the nasopharyngeal DRF (frontal skull base 3.4 vs 2.1 mm, P = .005 and lateral skull base 3.9 vs 3.5 mm, P = .013) than with the standard forehead mounted one. Conclusion. Positioning the DRF in the center of the head significantly improved the application accuracy of targets in the skull base with electromagnetic navigation by 25%.

Trabecular bone class mapping across resolutions: translating methods from HR-pQCT to clinical CT

Quantitative assessment of 3D bone microarchitecture in high-resolution peripheral quantitative computed tomography (HR-pQCT) has shown promise in fracture risk assessment and biomechanics, but is limited to the distal radius and tibia. Trabecular microarchitecture classes (TMACs), based on voxel-wise clustering texture and structure tensor features in HRpQCT, is extended in this paper to quantify trabecular bone classes in clinical multi-detector CT (MDCT) images. Our comparison of TMACs in 12 cadaver radii imaged using both HRpQCT and MDCT yields a mean Dice score of up to 0.717±0.40 and visually concordant bone quality maps. Further work to develop clinically viable bone quantitative imaging using HR-pQCT validation could have a significant impact on overall bone health assessment.