Objective

No abstract

The goal of this study was to evaluate the maturity of current Deep Learning classification techniques for their application in a real maternal-fetal clinical environment. A large dataset of routinely acquired maternal-fetal screening ultrasound images (which will be made publicly available) was collected from two different hospitals by several operators and ultrasound machines. All images were manually labeled by an expert maternal fetal clinician. Images were divided into 6 classes: four of the most widely used fetal anatomical planes (Abdomen, Brain, Femur and Thorax), the mother's cervix (widely used for prematurity screening) and a general category to include any other less common image plane. fetal brain images were further categorized into the 3 most common fetal brain planes (Trans-thalamic, Trans-cerebellum, Trans-ventricular) to judge fine grain categorization performance. The final dataset is comprised of over 12,400 images from 1,792 patients, making it the largest ultrasound dataset to date. We then evaluated a wide variety of state-of-the-art deep convolutional neural networks on this dataset and analyzed results in depth, comparing the computational models to research technicians, which are the ones currently performing the task daily. Results indicate for the first time that computational models have similar performance compared to humans when classifying common planes in human fetal examination. However, the dataset leaves the door open on future research to further improve results, especially on fine-grained plane categorization.

Background Prediction of neonatal respiratory morbidity may be useful to plan delivery in complicated pregnancies. The limited predictive performance of the current diagnostic tests together with the risks of an invasive procedure restricts the use of fetal lung maturity assessment. Objective The objective of this study was to evaluate the performance of quantitative ultrasound texture analysis of the fetal lung (quantusFLM) to predict neonatal respiratory morbidity in preterm and early-term (<39.0 weeks) deliveries. Study Design This was a prospective multicenter study conducted in 20 centers worldwide. Fetal lung ultrasound images were obtained at 25.0-38.6 weeks of gestation within 48 hours of delivery, stored in Digital Imaging and Communication in Medicine format, and analyzed with quantusFLM. Physicians were blinded to the analysis. At delivery, perinatal outcomes and the occurrence of neonatal respiratory morbidity, defined as either respiratory distress syndrome or transient tachypnea of the newborn, were registered. The performance of the ultrasound texture analysis test to predict neonatal respiratory morbidity was evaluated. Results A total of 883 images were collected, but 17.3% were discarded because of poor image quality or exclusion criteria, leaving 730 observations for the final analysis. The prevalence of neonatal respiratory morbidity was 13.8% (101/730). The quantusFLM predicted neonatal respiratory morbidity with a sensitivity, specificity, and positive and negative predictive values of 74.3% (75/101), 88.6% (557/629), 51.0% (75/147), and 95.5% (557/583), respectively. Accuracy was 86.5% (632/730), and the positive and negative likelihood ratios were 6.5 and 0.3, respectively. Conclusion The quantusFLM predicted neonatal respiratory morbidity with an accuracy similar to that previously reported for other tests with the advantage of being a non-invasive technique.

BackgroundWhilst multi-parametric magnetic resonance imaging (mp-MRI) has been a significant advance in the diagnosis of prostate cancer, scanning all patients with elevated prostate specific antigen (PSA) levels is considered too costly for widespread National Health Service (NHS) use, as the predictive value of PSA levels for significant disease is poor. Despite the fact that novel blood and urine tests are available which may predict aggressive disease better than PSA, they are not routinely employed due to a lack of clinical validity studies.Furthermore approximately 40 % of mp-MRI studies are reported as indeterminate, which can lead to repeat examinations or unnecessary biopsy with associated patient anxiety, discomfort, risk and additional costs.Methods/DesignWe aim to clinically validate a panel of minimally invasive promising blood and urine biomarkers, to better select patients that will benefit from a multiparametric prostate MRI. We will then test whether the performance of the mp-MRI can be improved by the addition of an advanced diffusion-weighted MRI technique, which uses a biophysical model to characterise tissue microstructure called VERDICT; Vascular and Extracellular Restricted Diffusion for Cytometry in Tumours.INNOVATE is a prospective single centre cohort study in 365 patients. mp-MRI will act as the reference standard for the biomarker panel. A clinical outcome based reference standard based on biopsy, mp-MRI and follow-up will be used for VERDICT MRI.DiscussionWe expect the combined effect of biomarkers and VERDICT MRI will improve care by better detecting aggressive prostate cancer early and make mp-MRI before biopsy economically viable for universal NHS adoption.Trial registrationINNOVATE is registered on ClinicalTrials.gov, with reference NCT02689271.Electronic supplementary materialThe online version of this article (doi:10.1186/s12885-016-2856-2) contains supplementary material, which is available to authorized users.

espite the merits of the apparent diffusion coefficient (ADC), reporting quantitative ADC values is not a routine part of clinical practice. This is partially due to lack of biologic specificity (1). Recently, our group presented the feasibility of Vascular, Extracellular, and Restricted Diffusion for Cytometry in Tumors (VERDICT) MRI as a quantitative microstructural imaging tool for prostate cancer (2). VERDICT combines a diffusion-weighted MRI acquisition with a mathematical model and assigns the diffusion-weighted MRI signal to three principal components: (a) intracellular water, (b) water in the extracellular extravascular space, and (c) water in the microvasculature. Because the fraction of each of these compartments differs between each Gleason grade (3), we hypothesized that

VERDICT (vascular, extracellular and restricted diffusion for cytometry in tumours) estimates and maps microstructural features of cancerous tissue non‐invasively using diffusion MRI. The main purpose of this study is to address the high computational time of microstructural model fitting for prostate diagnosis, while retaining utility in terms of tumour conspicuity and repeatability. In this work, we adapt the accelerated microstructure imaging via convex optimization (AMICO) framework to linearize the estimation of VERDICT parameters for the prostate gland. We compare the original non‐linear fitting of VERDICT with the linear fitting, quantifying accuracy with synthetic data, and computational time and reliability (performance and precision) in eight patients. We also assess the repeatability (scan‐rescan) of the parameters. Comparison of the original VERDICT fitting versus VERDICT‐AMICO showed that the linearized fitting (1) is more accurate in simulation for a signal‐to‐noise ratio of 20 dB; (2) reduces the processing time by three orders of magnitude, from 6.55 seconds/voxel to 1.78 milliseconds/voxel; (3) estimates parameters more precisely; (4) produces similar parametric maps and (5) produces similar estimated parameters with a high Pearson correlation between implementations, r 2 > 0.7. The VERDICT‐AMICO estimates also show high levels of repeatability. Finally, we demonstrate that VERDICT‐AMICO can estimate an extra diffusivity parameter without losing tumour conspicuity and retains the fitting advantages. VERDICT‐AMICO provides microstructural maps for prostate cancer characterization in seconds.

Objectives: We tested the hypothesis whether a texture analysis (TA) algorithm applied to MRI brain images identified different patterns in small for gestational age (SGA) fetuses as compared with adequate for gestational age (AGA). Study Design: MRI was performed on 83 SGA and 70 AGA at 37 weeks' GA. Texture features were quantified in the frontal lobe, basal ganglia, mesencephalon, cerebellum and cingulum. A classification algorithm based on discriminative models was used to correlate texture features with clinical diagnosis. Results: Region of interest delineation in all areas was achieved in 61 SGA (12 vasodilated) and 52 AGA; this was the sample for TA feature extraction which allowed classifying SGA from AGA withaccuracies ranging from 90.9 to 98.9% in SGA versus AGA comparison and from 93.6 to 100% in vasodilated SGA versus AGA comparison. Conclusions: This study demonstrates that TA can detect brain differences in SGA fetuses. This supports the existence of brain microstructural changes in SGA fetuses.