Phantom study of augmented reality framework to assist epicardial punctures.

Purpose: The objective of this study is to evaluate the accuracy of an augmented reality (AR) system in improving guidance, accuracy, and visualization during the subxiphoidal approach for epicardial ablation. Approach: An AR application was developed to project real-time needle trajectories and patient-specific 3D organs using the Hololens 2. Additionally, needle tracking was implemented to offer real-time feedback to the operator, facilitating needle navigation. The AR application was evaluated through three different experiments: examining overlay accuracy, assessing puncture accuracy, and performing pre-clinical evaluations on a phantom. Results: The results of the overlay accuracy assessment for the AR system yielded 2.36±2.04 mm. Additionally, the puncture accuracy utilizing the AR system yielded 1.02±2.41 mm. During the pre-clinical evaluation on the phantom, needle puncture with AR guidance showed 7.43±2.73 mm, whereas needle puncture without AR guidance showed 22.62±9.37 mm. Conclusions: Overall, the AR platform has the potential to enhance the accuracy of percutaneous epicardial access for mapping and ablation of cardiac arrhythmias, thereby reducing complications and improving patient outcomes. The significance of this study lies in the potential of AR guidance to enhance the accuracy and safety of percutaneous epicardial access.

OPG-based dental age estimation using a data-technical exploration of deep learning techniques.

Dental age estimation, a cornerstone in forensic age assessment, has been extensively tried and tested, yet manual methods are impeded by tedium and interobserver variability. Automated approaches using deep transfer learning encounter challenges like data scarcity, suboptimal training, and fine-tuning complexities, necessitating robust training methods. This study explores the impact of convolutional neural network hyperparameters, model complexity, training batch size, and sample quantity on age estimation. EfficientNet-B4, DenseNet-201, and MobileNet V3 models underwent cross-validation on a dataset of 3896 orthopantomograms (OPGs) with batch sizes escalating from 10 to 160 in a doubling progression, as well as random subsets of this training dataset. Results demonstrate the EfficientNet-B4 model, trained on the complete dataset with a batch size of 160, as the top performer with a mean absolute error of 0.562 years on the test set, notably surpassing the MAE of 1.01 at a batch size of 10. Increasing batch size consistently improved performance for EfficientNet-B4 and DenseNet-201, whereas MobileNet V3 performance peaked at batch size 40. Similar trends emerged in training with reduced sample sizes, though they were outperformed by the complete models. This underscores the critical role of hyperparameter optimization in adopting deep learning for age estimation from complete OPGs. The findings not only highlight the nuanced interplay of hyperparameters and performance but also underscore the potential for accurate age estimation models through optimization. This study contributes to advancing the application of deep learning in forensic age estimation, emphasizing the significance of tailored training methodologies for optimal outcomes.

USE-Evaluator: Performance metrics for medical image segmentation models supervised by uncertain, small or empty reference annotations in neuroimaging.

Performance metrics for medical image segmentation models are used to measure the agreement between the reference annotation and the predicted segmentation. Usually, overlap metrics, such as the Dice, are used as a metric to evaluate the performance of these models in order for results to be comparable. However, there is a mismatch between the distributions of cases and the difficulty level of segmentation tasks in public data sets compared to clinical practice. Common metrics used to assess performance fail to capture the impact of this mismatch, particularly when dealing with datasets in clinical settings that involve challenging segmentation tasks, pathologies with low signal, and reference annotations that are uncertain, small, or empty. Limitations of common metrics may result in ineffective machine learning research in designing and optimizing models. To effectively evaluate the clinical value of such models, it is essential to consider factors such as the uncertainty associated with reference annotations, the ability to accurately measure performance regardless of the size of the reference annotation volume, and the classification of cases where reference annotations are empty. We study how uncertain, small, and empty reference annotations influence the value of metrics on a stroke in-house data set regardless of the model. We examine metrics behavior on the predictions of a standard deep learning framework in order to identify suitable metrics in such a setting. We compare our results to the BRATS 2019 and Spinal Cord public data sets. We show how uncertain, small, or empty reference annotations require a rethinking of the evaluation. The evaluation code was released to encourage further analysis of this topic https://github.com/SophieOstmeier/UncertainSmallEmpty.git.

Reversibility of Diffusion-Weighted Imaging Lesions in Patients With Ischemic Stroke in the WAKE-UP Trial.

BACKGROUND: Reversibility of the diffusion-weighted imaging (DWI) lesion means that not all of the DWI lesion represents permanently injured tissue. We investigated DWI reversibility and the association with thrombolysis, reperfusion and functional outcome in patients from the WAKE-UP trial (Efficacy and Safety of Magnetic Resonance Imaging-Based Thrombolysis in Wake-Up Stroke). METHODS: In this retrospective analysis of WAKE-UP, a randomized controlled trial (RCT) between September 2012 and June 2017 in Belgium, Denmark, France, Germany, Spain and United Kingdom, a convolutional neural network segmented the DWI lesions (b=1000 s/mm2) at baseline and follow-up (24 hours). We calculated absolute and relative DWI reversibility in 2 ways: first, a volumetric (baseline volume-24-hour volume >0) and second, a voxel-based (part of baseline lesion not overlapping with 24-hour lesion) approach. We additionally defined relative voxel-based DWI-reversibility >50% to account for coregistration inaccuracies. We calculated the odds ratio for reversibility according to treatment arm. We analyzed the association of reversibility with excellent functional outcome (modified Rankin Scale score of 0-1), in a multivariable model. RESULTS: In 363 patients, the median DWI volume was 3 (1-10) mL at baseline and 6 (2-20) mL at follow-up. Volumetric DWI reversibility was present in 19% (69/363) with a median absolute reversible volume of 1 mL (0-2) or 28% (14-50) relatively. Voxel-based DWI reversibility was present in 358/363 (99%) with a median absolute volume of 1 mL (0-2), or 22% (9-38) relatively. In 18% of the patients (67/363), relative voxel-based DWI reversibility >50% was present. Volumetric DWI reversibility and relative voxel-based DWI reversibility >50% was more frequent in patients treated with alteplase versus placebo (OR, 1.86 [95% CI, 1.09-3.17] and OR, 2.03 [95% CI, 1.18-3.50], respectively). Relative voxel-based DWI reversibility >50% was associated with excellent functional outcome (OR, 2.30 [95% CI, 1.17-4.51]). CONCLUSIONS: Small absolute volumes of DWI reversibility were present in a large proportion of randomized patients in the WAKE-UP trial. Reversibility was more often present after thrombolysis.

Deep learning-based dental implant recognition using synthetic X-ray images.

A novel algorithm for generating artificial training samples from triangulated three-dimensional (3D) surface models within the context of dental implant recognition is proposed. The proposed algorithm is based on the calculation of two-dimensional (2D) projections (from a number of different angles) of 3D volumetric representations of computer-aided design (CAD) surface models. A fully convolutional network (FCN) is subsequently trained on the artificially generated X-ray images for the purpose of automatically identifying the connection type associated with a specific dental implant in an actual X-ray image. Semi-automated and fully automated systems are proposed for segmenting questioned dental implants from the background in actual X-ray images. Within the context of the semi-automated system, suitable regions of interest (ROIs), which contain the dental implants, are manually specified. However, as part of the fully automated system, suitable ROIs are automatically detected. It is demonstrated that a segmentation/detection accuracy of 94.0% and a classification/recognition accuracy of 71.7% are attainable within the context of the proposed fully automated system. Since the proposed systems utilise an ensemble of techniques that has not been employed for the purpose of dental implant classification/recognition on any previous occasion, the above-mentioned results are very encouraging.

Cross-Modal Distillation to Improve MRI-Based Brain Tumor Segmentation With Missing MRI Sequences.

Convolutional neural networks (CNNs) for brain tumor segmentation are generally developed using complete sets of magnetic resonance imaging (MRI) sequences for both training and inference. As such, these algorithms are not trained for realistic, clinical scenarios where parts of the MRI sequences which were used for training, are missing during inference. To increase clinical applicability, we proposed a cross-modal distillation approach to leverage the availability of multi-sequence MRI data for training and generate an enriched CNN model which uses only single-sequence MRI data for inference but outperforms a single-sequence CNN model. We assessed the performance of the proposed method for whole tumor and tumor core segmentation with multi-sequence MRI data available for training but only T1-weighted ([Formula: see text]) sequence data available for inference, using BraTS 2018, and in-house datasets. Results showed that cross-modal distillation significantly improved the Dice score for both whole tumor and tumor core segmentation when only [Formula: see text] sequence data were available for inference. For the evaluation using the in-house dataset, cross-modal distillation achieved an average Dice score of 79.04% and 69.39% for whole tumor and tumor core segmentation, respectively, while a single-sequence U-Net model using [Formula: see text] sequence data for both training and inference achieved an average Dice score of 73.60% and 62.62%, respectively. These findings confirmed cross-modal distillation as an effective method to increase the potential of single-sequence CNN models such that segmentation performance is less compromised by missing MRI sequences or having only one MRI sequence available for segmentation.

Theoretical analysis and experimental validation of volume bias of soft Dice optimized segmentation maps in the context of inherent uncertainty.

The clinical interest is often to measure the volume of a structure, which is typically derived from a segmentation. In order to evaluate and compare segmentation methods, the similarity between a segmentation and a predefined ground truth is measured using popular discrete metrics, such as the Dice score. Recent segmentation methods use a differentiable surrogate metric, such as soft Dice, as part of the loss function during the learning phase. In this work, we first briefly describe how to derive volume estimates from a segmentation that is, potentially, inherently uncertain or ambiguous. This is followed by a theoretical analysis and an experimental validation linking the inherent uncertainty to common loss functions for training CNNs, namely cross-entropy and soft Dice. We find that, even though soft Dice optimization leads to an improved performance with respect to the Dice score and other measures, it may introduce a volume bias for tasks with high inherent uncertainty. These findings indicate some of the method's clinical limitations and suggest doing a closer ad-hoc volume analysis with an optional re-calibration step.

Optimization for Medical Image Segmentation: Theory and Practice When Evaluating With Dice Score or Jaccard Index.

In many medical imaging and classical computer vision tasks, the Dice score and Jaccard index are used to evaluate the segmentation performance. Despite the existence and great empirical success of metric-sensitive losses, i.e. relaxations of these metrics such as soft Dice, soft Jaccard and Lovász-Softmax, many researchers still use per-pixel losses, such as (weighted) cross-entropy to train CNNs for segmentation. Therefore, the target metric is in many cases not directly optimized. We investigate from a theoretical perspective, the relation within the group of metric-sensitive loss functions and question the existence of an optimal weighting scheme for weighted cross-entropy to optimize the Dice score and Jaccard index at test time. We find that the Dice score and Jaccard index approximate each other relatively and absolutely, but we find no such approximation for a weighted Hamming similarity. For the Tversky loss, the approximation gets monotonically worse when deviating from the trivial weight setting where soft Tversky equals soft Dice. We verify these results empirically in an extensive validation on six medical segmentation tasks and can confirm that metric-sensitive losses are superior to cross-entropy based loss functions in case of evaluation with Dice Score or Jaccard Index. This further holds in a multi-class setting, and across different object sizes and foreground/background ratios. These results encourage a wider adoption of metric-sensitive loss functions for medical segmentation tasks where the performance measure of interest is the Dice score or Jaccard index.

Whole liver segmentation based on deep learning and manual adjustment for clinical use in SIRT.

PURPOSE: In selective internal radiation therapy (SIRT), an accurate total liver segmentation is required for activity prescription and absorbed dose calculation. Our goal was to investigate the feasibility of using automatic liver segmentation based on a convolutional neural network (CNN) for CT imaging in SIRT, and the ability of CNN to reduce inter-observer variability of the segmentation. METHODS: A multi-scale CNN was modified for liver segmentation for SIRT patients. The CNN model was trained with 139 datasets from three liver segmentation challenges and 12 SIRT patient datasets from our hospital. Validation was performed on 13 SIRT datasets and 12 challenge datasets. The model was tested on 40 SIRT datasets. One expert manually delineated the livers and adjusted the liver segmentations from CNN for 40 test SIRT datasets. Another expert performed the same tasks for 20 datasets randomly selected from the 40 SIRT datasets. The CNN segmentations were compared with the manual and adjusted segmentations from the experts. The difference between the manual segmentations was compared with the difference between the adjusted segmentations to investigate the inter-observer variability. Segmentation difference was evaluated through dice similarity coefficient (DSC), volume ratio (RV), mean surface distance (MSD), and Hausdorff distance (HD). RESULTS: The CNN segmentation achieved a median DSC of 0.94 with the manual segmentation and of 0.98 with the manually corrected CNN segmentation, respectively. The DSC between the adjusted segmentations is 0.98, which is 0.04 higher than the DSC between the manual segmentations. CONCLUSION: The CNN model achieved good liver segmentations on CT images of good image quality, with relatively normal liver shapes and low tumor burden. 87.5% of the 40 CNN segmentations only needed slight adjustments for clinical use. However, the trained model failed on SIRT data with low dose or contrast, lesions with large density difference from their surroundings, and abnormal liver position and shape. The abovementioned scenarios were not adequately represented in the training data. Despite this limitation, the current CNN is already a useful clinical tool which improves inter-observer agreement and therefore contributes to the standardization of the dosimetry. A further improvement is expected when the CNN will be trained with more data from SIRT patients.

Towards fully automated third molar development staging in panoramic radiographs.

Staging third molar development is commonly used for age assessment in sub-adults. Current staging techniques are, at most, semi-automated and rely on manual interactions prone to operator variability. The aim of this study was to fully automate the staging process by employing the full potential of deep learning, using convolutional neural networks (CNNs) in every step of the procedure. The dataset used to train the CNNs consisted of 400 panoramic radiographs (OPGs), with 20 OPGs per developmental stage per sex, staged in consensus between three observers. The concepts of transfer learning, using pre-trained CNNs, and data augmentation were used to mitigate the issues when dealing with a limited dataset. In this work, a three-step procedure was proposed and the results were validated using fivefold cross-validation. First, a CNN localized the geometrical center of the lower left third molar, around which a square region of interest (ROI) was extracted. Second, another CNN segmented the third molar within the ROI. Third, a final CNN used both the ROI and the segmentation to classify the third molar into its developmental stage. The geometrical center of the third molar was found with an average Euclidean distance of 63 pixels. Third molars were segmented with an average Dice score of 93%. Finally, the developmental stages were classified with an accuracy of 54%, a mean absolute error of 0.69 stages, and a linear weighted Cohen's kappa coefficient of 0.79. The entire automated workflow on average took 2.72 s to compute, which is substantially faster than manual staging starting from the OPG. Taking into account the limited dataset size, this pilot study shows that the proposed fully automated approach shows promising results compared with manual staging.

Effect of Lower Third Molar Segmentations on Automated Tooth Development Staging using a Convolutional Neural Network.

Staging third molar development is commonly used for age estimation in subadults. Automated developmental stage allocation to the mandibular left third molar in panoramic radiographs has been examined in a pilot study. This method used an AlexNet Deep Convolutional Neural Network (CNN) approach to stage lower left third molars, which had been selected by manually drawn bounding boxes around them. This method (bounding box AlexNet = BA) still contained parts of surrounding structures which may have affected the automated stage allocation performance. We hypothesize that segmenting only the third molar could further improve the automated stage allocation performance. Therefore, the current study aimed to determine and validate the effect of lower third molar segmentations on automated tooth development staging. Retrospectively, 400 panoramic radiographs were collected, processed and segmented in three ways: bounding box (BB), rough (RS), and full (FS) tooth segmentation. A DenseNet201 CNN was used for automated stage allocation. Automated staging results were compared with reference stages - allocated by human observers - overall and per stage. FS rendered the best results with a stage allocation accuracy of 0.61, a mean absolute difference of 0.53 stages and a Cohen's linear κ of 0.84. Misallocated stages were mostly neighboring stages, and DenseNet201 rendered better results than AlexNet by increasing the percentage of correctly allocated stages by 3% (BA compared to BB). FS increased the percentage of correctly allocated stages by 7% compared to BB. In conclusion, full tooth segmentation and a DenseNet CNN optimize automated dental stage allocation for age estimation.

Feedback on prescribing errors to junior doctors: exploring views, problems and preferred methods.

BACKGROUND: Prescribing errors are common in hospital inpatients. However, the literature suggests that doctors are often unaware of their errors as they are not always informed of them. It has been suggested that providing more feedback to prescribers may reduce subsequent error rates. Only few studies have investigated the views of prescribers towards receiving such feedback, or the views of hospital pharmacists as potential feedback providers. OBJECTIVES: Our aim was to explore the views of junior doctors and hospital pharmacists regarding feedback on individual doctors' prescribing errors. Objectives were to determine how feedback was currently provided and any associated problems, to explore views on other approaches to feedback, and to make recommendations for designing suitable feedback systems. SETTING: A large London NHS hospital trust. METHODS: To explore views on current and possible feedback mechanisms, self-administered questionnaires were given to all junior doctors and pharmacists, combining both 5-point Likert scale statements and open-ended questions. MAIN OUTCOME MEASURES: Agreement scores for statements regarding perceived prescribing error rates, opinions on feedback, barriers to feedback, and preferences for future practice. RESULTS: Response rates were 49% (37/75) for junior doctors and 57% (57/100) for pharmacists. In general, doctors did not feel threatened by feedback on their prescribing errors. They felt that feedback currently provided was constructive but often irregular and insufficient. Most pharmacists provided feedback in various ways; however some did not or were inconsistent. They were willing to provide more feedback, but did not feel it was always effective or feasible due to barriers such as communication problems and time constraints. Both professional groups preferred individual feedback with additional regular generic feedback on common or serious errors. CONCLUSION: Feedback on prescribing errors was valued and acceptable to both professional groups. From the results, several suggested methods of providing feedback on prescribing errors emerged. Addressing barriers such as the identification of individual prescribers would facilitate feedback in practice. Research investigating whether or not feedback reduces the subsequent error rate is now needed.