Search

Automated segmentation tools often encounter accuracy and adaptability issues when applied to images of different pathology. The purpose of this study is to explore the feasibility of building a workflow to efficiently route images to specifically trained segmentation models. By implementing a deep learning classifier to automatically classify the images and route them to appropriate segmentation models, we hope that our workflow can segment the images with different pathology accurately. The data we used in this study are 350 CT images from patients affected by polycystic liver disease and 350 CT images from patients presenting with liver metastases from colorectal cancer. All images had the liver manually segmented by trained imaging analysts. Our proposed adaptive segmentation workflow achieved a statistically significant improvement for the task of total liver segmentation compared to the generic single segmentation model (non-parametric Wilcoxon signed rank test, n=100, p-value << 0.001). This approach is applicable in a wide range of scenarios and should prove useful in clinical implementations of segmentation pipelines., Comment: J Digit Imaging. Inform. med. (2024)

Purpose: This study evaluated the out-of-domain performance and generalization capabilities of automated medical image segmentation models, with a particular focus on adaptation to new image acquisitions and disease type. Materials: Datasets from both non-contrast and contrast-enhanced abdominal CT scans of healthy patients and those with polycystic kidney disease (PKD) were used. A total of 400 images (100 non-contrast controls, 100 contrast controls, 100 non-contrast PKD, 100 contrast PKD) were utilized for training/validation of models to segment kidneys, livers, and spleens, and the final models were then tested on 100 non-contrast CT images of patients affected by PKD. Performance was evaluated using Dice, Jaccard, TPR, and Precision. Results: Models trained on a diverse range of data showed no worse performance than models trained exclusively on in-domain data when tested on in-domain data. For instance, the Dice similarity of the model trained on 25% from each dataset was found to be non-inferior to the model trained purely on in-domain data. Conclusions: The results indicate that broader training examples significantly enhances model generalization and out-of-domain performance, thereby improving automated segmentation tools' applicability in clinical settings. The study's findings provide a roadmap for future research to adopt a data-centric approach in medical image AI model development.

Methods for automatically flag poor performing-predictions are essential for safely implementing machine learning workflows into clinical practice and for identifying difficult cases during model training. We present a readily adoptable method using sub-models trained on different dataset folds, where their disagreement serves as a surrogate for model confidence. Thresholds informed by human interobserver values were used to determine whether a final ensemble model prediction would require manual review. In two different datasets (abdominal CT and MR predicting kidney tumors), our framework effectively identified low performing automated segmentations. Flagging images with a minimum Interfold test Dice score below human interobserver variability maximized the number of flagged images while ensuring maximum ensemble test Dice. When our internally trained model was applied to an external publicly available dataset (KiTS21), flagged images included smaller tumors than those observed in our internally trained dataset, demonstrating the methods robustness to flagging poor performing out-of-distribution input data. Comparing interfold sub-model disagreement against human interobserver values is an efficient way to approximate a model's epistemic uncertainty - its lack of knowledge due to insufficient relevant training data - a key functionality for adopting these applications in clinical practice., Comment: 16 pages, 6 figures

Evaluating automatically generated text is generally hard due to the inherently subjective nature of many aspects of the output quality. This difficulty is compounded in automatic consultation note generation by differing opinions between medical experts both about which patient statements should be included in generated notes and about their respective importance in arriving at a diagnosis. Previous real-world evaluations of note-generation systems saw substantial disagreement between expert evaluators. In this paper we propose a protocol that aims to increase objectivity by grounding evaluations in Consultation Checklists, which are created in a preliminary step and then used as a common point of reference during quality assessment. We observed good levels of inter-annotator agreement in a first evaluation study using the protocol; further, using Consultation Checklists produced in the study as reference for automatic metrics such as ROUGE or BERTScore improves their correlation with human judgements compared to using the original human note., Comment: Accepted for publication at EMNLP 2022

The de facto standard of dynamic histogram binning for radiomic feature extraction leads to an elevated sensitivity to fluctuations in annotated regions. This may impact the majority of radiomic studies published recently and contribute to issues regarding poor reproducibility of radiomic-based machine learning that has led to significant efforts for data harmonization; however, we believe the issues highlighted here are comparatively neglected, but often remedied by choosing static binning. The field of radiomics has improved through the development of community standards and open-source libraries such as PyRadiomics. But differences in image acquisition, systematic differences between observers' annotations, and preprocessing steps still pose challenges. These can change the distribution of voxels altering extracted features and can be exacerbated with dynamic binning., Comment: 5 pages, 1 figure, 2 tables

A growing body of work uses Natural Language Processing (NLP) methods to automatically generate medical notes from audio recordings of doctor-patient consultations. However, there are very few studies on how such systems could be used in clinical practice, how clinicians would adjust to using them, or how system design should be influenced by such considerations. In this paper, we present three rounds of user studies, carried out in the context of developing a medical note generation system. We present, analyse and discuss the participating clinicians' impressions and views of how the system ought to be adapted to be of value to them. Next, we describe a three-week test run of the system in a live telehealth clinical practice. Major findings include (i) the emergence of five different note-taking behaviours; (ii) the importance of the system generating notes in real time during the consultation; and (iii) the identification of a number of clinical use cases that could prove challenging for automatic note generation systems., Comment: Accepted for publication at NAACL 2022

In recent years, machine learning models have rapidly become better at generating clinical consultation notes; yet, there is little work on how to properly evaluate the generated consultation notes to understand the impact they may have on both the clinician using them and the patient's clinical safety. To address this we present an extensive human evaluation study of consultation notes where 5 clinicians (i) listen to 57 mock consultations, (ii) write their own notes, (iii) post-edit a number of automatically generated notes, and (iv) extract all the errors, both quantitative and qualitative. We then carry out a correlation study with 18 automatic quality metrics and the human judgements. We find that a simple, character-based Levenshtein distance metric performs on par if not better than common model-based metrics like BertScore. All our findings and annotations are open-sourced., Comment: To be published in proceedings of ACL 2022

Recent advances in Automatic Speech Recognition (ASR) have made it possible to reliably produce automatic transcripts of clinician-patient conversations. However, access to clinical datasets is heavily restricted due to patient privacy, thus slowing down normal research practices. We detail the development of a public access, high quality dataset comprising of57 mocked primary care consultations, including audio recordings, their manual utterance-level transcriptions, and the associated consultation notes. Our work illustrates how the dataset can be used as a benchmark for conversational medical ASR as well as consultation note generation from transcripts.

Clinical notes are an efficient way to record patient information but are notoriously hard to decipher for non-experts. Automatically simplifying medical text can empower patients with valuable information about their health, while saving clinicians time. We present a novel approach to automated simplification of medical text based on word frequencies and language modelling, grounded on medical ontologies enriched with layman terms. We release a new dataset of pairs of publicly available medical sentences and a version of them simplified by clinicians. Also, we define a novel text simplification metric and evaluation framework, which we use to conduct a large-scale human evaluation of our method against the state of the art. Our method based on a language model trained on medical forum data generates simpler sentences while preserving both grammar and the original meaning, surpassing the current state of the art.

Automatic summarisation has the potential to aid physicians in streamlining clerical tasks such as note taking. But it is notoriously difficult to evaluate these systems and demonstrate that they are safe to be used in a clinical setting. To circumvent this issue, we propose a semi-automatic approach whereby physicians post-edit generated notes before submitting them. We conduct a preliminary study on the time saving of automatically generated consultation notes with post-editing. Our evaluators are asked to listen to mock consultations and to post-edit three generated notes. We time this and find that it is faster than writing the note from scratch. We present insights and lessons learnt from this experiment.

Medical images such as 3D computerized tomography (CT) scans and pathology images, have hundreds of millions or billions of voxels/pixels. It is infeasible to train CNN models directly on such high resolution images, because neural activations of a single image do not fit in the memory of a single GPU/TPU, and naive data and model parallelism approaches do not work. Existing image analysis approaches alleviate this problem by cropping or down-sampling input images, which leads to complicated implementation and sub-optimal performance due to information loss. In this paper, we implement spatial partitioning, which internally distributes the input and output of convolutional layers across GPUs/TPUs. Our implementation is based on the Mesh-TensorFlow framework and the computation distribution is transparent to end users. With this technique, we train a 3D Unet on up to 512 by 512 by 512 resolution data. To the best of our knowledge, this is the first work for handling such high resolution images end-to-end.

In this work we approach the task of learning multilingual word representations in an offline manner by fitting a generative latent variable model to a multilingual dictionary. We model equivalent words in different languages as different views of the same word generated by a common latent variable representing their latent lexical meaning. We explore the task of alignment by querying the fitted model for multilingual embeddings achieving competitive results across a variety of tasks. The proposed model is robust to noise in the embedding space making it a suitable method for distributed representations learned from noisy corpora., Comment: Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics

Image segmentation plays an essential role in medicine for both diagnostic and interventional tasks. Segmentation approaches are either manual, semi-automated or fully-automated. Manual segmentation offers full control over the quality of the results, but is tedious, time consuming and prone to operator bias. Fully automated methods require no human effort, but often deliver sub-optimal results without providing users with the means to make corrections. Semi-automated approaches keep users in control of the results by providing means for interaction, but the main challenge is to offer a good trade-off between precision and required interaction. In this paper we present a deep learning (DL) based semi-automated segmentation approach that aims to be a "smart" interactive tool for region of interest delineation in medical images. We demonstrate its use for segmenting multiple organs on computed tomography (CT) of the abdomen. Our approach solves some of the most pressing clinical challenges: (i) it requires only one to a few user clicks to deliver excellent 2D segmentations in a fast and reliable fashion; (ii) it can generalize to previously unseen structures and "corner cases"; (iii) it delivers results that can be corrected quickly in a smart and intuitive way up to an arbitrary degree of precision chosen by the user and (iv) ensures high accuracy. We present our approach and compare it to other techniques and previous work to show the advantages brought by our method.

The well-known 3V architectural paradigm for Big Data introduced by Laney (2011), provides a simplified framework for defining the architecture of a big data platform to be deployed in various scenarios tackling processing of massive datasets. While additional components such as Variability and Veracity have been discussed as an extension to the 3V model, the basic components (volume, variety, velocity) provide a quantitative framework while variability and veracity target a more qualitative approach. In this paper we argue why the basic 3V's are not equal due to the different requirements that need to be covered in case higher demands for a particular "V". Similar to other conjectures such as the CAP theorem 3V based architectures differ on their implementation. We call this paradigm heterogeneity and we provide a taxonomy of the existing tools (as of 2013) covering the Hadoop ecosystem from the perspective of heterogeneity. This paper contributes on the understanding of the Hadoop ecosystem from the perspective of different workloads and aims to help researchers and practitioners on the design of scalable platforms targeting different operational needs.

With the continuous increase of online services as well as energy costs, energy consumption becomes a significant cost factor for the evaluation of data center operations. A significant contributor to that is the performance of database servers which are found to constitute the backbone of online services. From a software approach, while a set of novel data management technologies appear in the market e.g. key-value based or in-memory databases, classic relational database management systems (RDBMS) are still widely used. In addition from a hardware perspective, the majority of database servers is still using standard magnetic hard drives (HDDs) instead of solid state drives (SSDs) due to lower cost of storage per gigabyte, disregarding the performance boost that might be given due to high cost. In this study we focus on a software based assessment of the energy consumption of a database server by running three different and complete database workloads namely TCP-H, Star Schema Benchmark -SSB as well a modified benchmark we have derived for this study called W22. We profile the energy distribution among the ost important server components and by using different resource allocation we assess the energy consumption of a typical open source RDBMS (PostgreSQL) on a standard server in relation with its performance (measured by query time). Results confirm the well-known fact that even for complete workloads, optimization of the RDBMS results to lower energy consumption., Comment: 10 pages, 12 figures

An invariant description of Bianchi Homogeneous (B.H.) 3-spaces is presented, by considering the action of the Automorphism Group on the configuration space of the real, symmetric, positive definite, $3\times 3$ matrices. Thus, the gauge degrees of freedom are removed and the remaining (gauge invariant) degrees, are the --up to 3-- curvature invariants. An apparent discrepancy between this Kinematics and the Quantum Hamiltonian Dynamics of the lower Class A Bianchi Types, occurs due to the existence of the Outer Automorphism Subgroup. This discrepancy is satisfactorily removed by exploiting the quantum version of some classical integrals of motion (conditional symmetries) which are recognized as corresponding to the Outer Automorphisms., Comment: 18 pages, LaTeX2e, no figures, one table, to appear in Communications in Mathematical Physics

We investigate the set of spacetime general coordinate transformations (G.C.T.) which leave the line element of a generic Bianchi Type Geometry, quasi-form invariant; i.e. preserve manifest spatial Homogeneity. We find that these G.C.T.'s, induce special time-dependent automorphic changes, on the spatial scale factor matrix $\gamma_{\alpha\beta}(t)$ -along with corresponding changes on the lapse function $N(t)$ and the shift vector $N^{\alpha}(t)$. These changes, which are Bianchi Type dependent, form a group and are, in general, different from those induced by the group SAut(G) -advocated in earlier investigations as the relevant symmetry group-, they are used to simplify the form of the line element -and thus simplify Einstein's equations as well-, without losing generality. As far as this simplification procedure is concerned, the transformations found, are proved to be essentialy unique. For the case of Bianchi Types II and V, where the most general solutions are known -Taub's and Joseph's, respectively-, it is explicitly verified that our transformations and only those, suffice to reduce the generic line element, to the previously known forms. It becomes thus possible, -for these Types- to give in closed form, the most general solution, containing all the necessary ``gauge'' freedom., Comment: 60 pages, Typeset using Latex2e, no figures, 1 table

We consider the quantization of a general spatially homogeneous space-time belonging to an arbitrary but fixed Class A Bianchi type. Exploiting the information furnished by the quantum version of the momentum constraints, we use as variables the two simplest contractions of $C^{\alpha}_{\beta \gamma}$ and $\gamma_{\alpha \beta}$ as well as the determinant of $\gamma_{\alpha \beta}$ ; We thus arrive at an equation for the wave function in terms of these quantities. This fact enables us to treat in a uniform manner all Class A cosmologies. For these spacetimes the Langrangian used correctly reproduces Einstein's equations. We also discuss the imposition of simplifying ans\"{a}tzen at the quantum level in a way that respects the scalings of the quadratic Hamiltonian constraint., Comment: Revised in 2002, LaTeX, 12 pages, G&C style, published version

This study presents the results of a pilot learning intervention for improving children's ideas about plant protection. The research was executed in two phases. The first phase aimed at exploring children's ideas about plant protection. These ideas were taken into account for the design and development of a digital learning environment. The second phase concerned the development, implementation and evaluation of the learning environment. In the first phase, 262 participants (10-12 years old) were asked to write a short text about "Environmental protection". During the second phase, twenty-six participants (10-12 years old) were asked to write a text on "How we can protect the endemic plants", before and after the implementation of the learning activities. Responses were content analysed and their meaning was classified according to the coding scheme of Blanchet-Cohen, Ragan, and Amsden ("Children, Youth and Environments" 2003; 13: 1-7). The results of the first phase showed that participants mentioned individual lifestyle actions of environmental protection and, therefore, the learning environment should emphasise the importance of other types of environmental protection ("eg" political actions, legislation, social actions). The proposed learning activities enriched participants' awareness about the importance of social and legal actions for plant protection.