White matter plasticity following cataract surgery in congenitally blind patients.

The visual system develops abnormally when visual input is absent or degraded during a critical period early in life. Restoration of the visual input later in life is generally thought to have limited benefit because the visual system will lack sufficient plasticity to adapt to and utilize the information from the eyes. Recent evidence, however, shows that congenitally blind adolescents can recover both low-level and higher-level visual function following surgery. In this study, we assessed behavioral performance in both a visual acuity and a face perception task alongside longitudinal structural white matter changes in terms of fractional anisotropy (FA) and mean diffusivity (MD). We studied congenitally blind patients with dense bilateral cataracts, who received cataract surgery at different stages of adolescence. Our goal was to differentiate between age- and surgery-related changes in both behavioral performance and structural measures to identify neural correlates which might contribute to recovery of visual function. We observed surgery-related long-term increases of structural integrity of late-visual pathways connecting the occipital regions with ipsilateral fronto-parieto-temporal regions or homotopic contralateral areas. Comparison to a group of age-matched healthy participants indicated that these improvements went beyond the expected changes in FA and MD based on maturation alone. Finally, we found that the extent of behavioral improvement in face perception was mediated by changes in structural integrity in late visual pathways. Our results suggest that sufficient plasticity remains in adolescence to partially overcome abnormal visual development and help localize the sites of neural change underlying sight recovery.

Automated prostate multi-regional segmentation in magnetic resonance using fully convolutional neural networks.

OBJECTIVE: Automatic MR imaging segmentation of the prostate provides relevant clinical benefits for prostate cancer evaluation such as calculation of automated PSA density and other critical imaging biomarkers. Further, automated T2-weighted image segmentation of central-transition zone (CZ-TZ), peripheral zone (PZ), and seminal vesicle (SV) can help to evaluate clinically significant cancer following the PI-RADS v2.1 guidelines. Therefore, the main objective of this work was to develop a robust and reproducible CNN-based automatic prostate multi-regional segmentation model using an intercontinental cohort of prostate MRI. METHODS: A heterogeneous database of 243 T2-weighted prostate studies from 7 countries and 10 machines of 3 different vendors, with the CZ-TZ, PZ, and SV regions manually delineated by two experienced radiologists (ground truth), was used to train (n = 123) and test (n = 120) a U-Net-based model with deep supervision using a cyclical learning rate. The performance of the model was evaluated by means of dice similarity coefficient (DSC), among others. Segmentation results with a DSC above 0.7 were considered accurate. RESULTS: The proposed method obtained a DSC of 0.88 ± 0.01, 0.85 ± 0.02, 0.72 ± 0.02, and 0.72 ± 0.02 for the prostate gland, CZ-TZ, PZ, and SV respectively in the 120 studies of the test set when comparing the predicted segmentations with the ground truth. No statistically significant differences were found in the results obtained between manufacturers or continents. CONCLUSION: Prostate multi-regional T2-weighted MR images automatic segmentation can be accurately achieved by U-Net like CNN, generalizable in a highly variable clinical environment with different equipment, acquisition configurations, and population. KEY POINTS: • Deep learning techniques allows the accurate segmentation of the prostate in three different regions on MR T2w images. • Multi-centric database proved the generalization of the CNN model on different institutions across different continents. • CNN models can be used to aid on the diagnosis and follow-up of patients with prostate cancer.

Modified radial-search algorithm for segmentation of tibiofemoral cartilage in MR images of patients with subchondral lesion.

PURPOSE: The quantitative analysis of weight-bearing articular cartilage superficial to subchondral abnormality is important in osteoarthritis (OA) progression studies. The current study aimed to address the challenges of a semi-automatic segmentation of tibiofemoral cartilage in MR images of OA patient with and without subchondral bone abnormalities (SBA). METHODS: In this study, knee MRI data [fat-suppressed proton density-weighted, multi-echo T2-weighted (CartiGram) images] of 29 OA patients, acquired at 3.0T MR scanner, were retrospectively collected. Out of 29 data, 9 had SBA in femur bone. Initially, a semi-automatic femur cartilage segmentation based on radial intensity search approach by Akhtar et al. was implemented in-house. This algorithm was considered as the radial-search method for further comparison. In this current study, the reported radial-search (RS)-based semi-automatic cartilage segmentation method was modified using thresholding, connected component labelling, convex-hull operation and spline-based curve fitting for the improved segmentation of tibiofemoral cartilage. Cartilage was manually segmented by two experienced radiologists, and inter-reader variability was estimated using coefficient of variation (CV). The segmentation results were validated using dice coefficient (DC), Jaccard coefficient (JC) and sensitivity index measurements. RESULTS: DC values for segmented femur cartilage in patients with SBA were 64.6 ± 7.8% and 81.4 ± 2.8% using reported RS method and modified radial-search method, respectively. DC values for segmented femur cartilage in patients without SBA were 82.5 ± 4.5% and 84.8 ± 2.0% using RS method and modified radial method, respectively. Similarly, DC values for tibial cartilage in all OA patients were 80.4 ± 1.6% and 81.9 ± 2.4% using RS method and modified radial method, respectively. Similar segmentation results were also obtained from the T2-weighted images. Inter-reader variability result based on CV in femur cartilage was 3.40 ± 2.12% (without SBA) and 4.18 ± 3.18% (with SBA). CONCLUSION: In the current study, a semi-automated segmentation of tibiofemoral cartilage was presented. Modified radial-search approach can successfully segment tibiofemoral cartilage, and the results were tested and validated on knee MRI data of OA patients with and without SBA.

Unboxing AI - Radiological Insights Into a Deep Neural Network for Lung Nodule Characterization.

RATIONALE AND OBJECTIVES: To explain predictions of a deep residual convolutional network for characterization of lung nodule by analyzing heat maps. MATERIALS AND METHODS: A 20-layer deep residual CNN was trained on 1245 Chest CTs from National Lung Screening Trial (NLST) trial to predict the malignancy risk of a nodule. We used occlusion to systematically block regions of a nodule and map drops in malignancy risk score to generate clinical attribution heatmaps on 103 nodules from Lung Image Database Consortium image collection and Image Database Resource Initiative (LIDC-IDRI) dataset, which were analyzed by a thoracic radiologist. The features were described as heat inside nodule -bright areas inside nodule, peripheral heat continuous/interrupted bright areas along nodule contours, heat in adjacent plane -brightness in scan planes juxtaposed with the nodule, satellite heat - a smaller bright spot in proximity to nodule in the same scan plane, heat map larger than nodule bright areas corresponding to the shape of the nodule seen outside the nodule margins and heat in calcification. RESULTS: These six features were assigned binary values. This feature vector was fedinto a standard J48 decision tree with 10-fold cross-validation, which gave an 85 % weighted classification accuracy with a 77.8% True Positive (TP) rate, 8% False Positive (FP) rate for benign cases and 91.8% TP and 22.2% FP rates for malignant cases. Heat Inside nodule was more frequently observed in nodules classified as malignant whereas peripheral heat, heat in adjacent plane, and satellite heat were more commonly seen in nodules classified as benign. CONCLUSION: We discuss the potential ability of a radiologist to visually parse the deep learning algorithm generated "heat map" to identify features aiding classification.