Clinical practice requires the production of a time-and resource-consuming great amount of notes. They contain relevant information, but their secondary use is almost impossible, due to their unstructured nature. Researchers are trying to address this problems, with traditional and promising novel techniques. Application in real hospital settings seems not to be possible yet, though, both because of relatively small and dirty dataset, and for the lack of language-specific pre-trained models. Aim: Our aim is to demonstrate the potential of the above techniques, but also raise awareness of the still open challenges that the scientific communities of IT and medical practitioners must jointly address to realize the full potential of unstructured content that is daily produced and digitized in hospital settings, both to improve its data quality and leverage the insights from data-driven predictive models. Methods: To this extent, we present a narrative literature review of the most recent and relevant contributions to leverage the application of Natural Language Processing techniques to the free-text content electronic patient records. In particular, we focused on four selected application domains, namely: data quality, information extraction, sentiment analysis and predictive models, and automated patient cohort selection. Then, we will present a few empirical studies that we undertook at a major teaching hospital specializing in musculoskeletal diseases. Results: We provide the reader with some simple and affordable pipelines, which demonstrate the feasibility of reaching literature performance levels with a single institution non-English dataset. In such a way, we bridged literature and real world needs, performing a step further toward the revival of notes fields.

Finite element modeling is a precious tool for the investigation of the biomechanics of the musculoskeletal system. A key element for the development of anatomically accurate, state-of-the art finite element models is medical imaging. Indeed, the workflow for the generation of a finite element model includes steps which require the availability of medical images of the subject of interest: segmentation, which is the assignment of each voxel of the images to a specific material such as bone and cartilage, allowing for a three-dimensional reconstruction of the anatomy; meshing, which is the creation of the computational mesh necessary for the approximation of the equations describing the physics of the problem; assignment of the material properties to the various parts of the model, which can be estimated for example from quantitative computed tomography for the bone tissue and with other techniques (elastography, T1rho, and T2 mapping from magnetic resonance imaging) for soft tissues. This paper presents a brief overview of the techniques used for image segmentation, meshing, and assessing the mechanical properties of biological tissues, with focus on finite element models of the musculoskeletal system. Both consolidated methods and recent advances such as those based on artificial intelligence are described.

In this work we propose to use Deep Learning to automatically calculate the coordinates of the vertebral corners in sagittal x-rays images of the thoracolumbar spine and, from those landmarks, to calculate relevant radiological parameters such as L1–L5 and L1–S1 lordosis and sacral slope. For this purpose, we used 10,193 images annotated with the landmarks coordinates as the ground truth. We realized a model that consists of 2 steps. In step 1, we trained 2 Convolutional Neural Networks to identify each vertebra in the image and calculate the landmarks coordinates respectively. In step 2, we refined the localization using cropped images of a single vertebra as input to another convolutional neural network and we used geometrical transformations to map the corners to the original image. For the localization tasks, we used a differentiable spatial to numerical transform (DSNT) as the top layer. We evaluated the model both qualitatively and quantitatively on a set of 195 test images. The median localization errors relative to the vertebrae dimensions were 1.98% and 1.68% for x and y coordinates respectively. All the predicted angles were highly correlated with the ground truth, despite non-negligible absolute median errors of 1.84°, 2.43° and 1.98° for L1–L5, L1–S1 and SS respectively. Our model is able to calculate with good accuracy the coordinates of the vertebral corners and has a large potential for improving the reliability and repeatability of measurements in clinical tasks.

Background: Aortic dilatation is common in hypertensive patients and is associated with higher risk of cardiovascular events. Parameters predicting further dilatation during lifetime are poorly understood. Aim: To predict the midterm aortic diameter evolution in a cohort of hypertensive patients with known aortic dilatation at Sinus of Valsalva (SOV) level. Methods: We prospectively analyzed a cohort of essential hypertensive outpatients without any other known risk factor for aortic dilatation. They underwent serial echocardiographic evaluations from 2003 to 2016. Results: Two hundred and forty-two hypertensive outpatients with a mild-to-moderate (37–53 mm) aortic dilatation were followed up for at least 5 years. Mean growth rate was 0.08 ± 0.35 mm/year. No clinical or anthropometric parameters were significantly different in patients with and without aortic diameter increase. Aortic z score (number of standard deviations from the average value observed in the general population) at baseline was inversely associated with growth rate (R 2 0.04, P < 0.05). Aortic diameter at first visit, demographic and echocardiographic variables were major determinants of aortic diameter at second visit, accounting for about 90% of its total variability. Conclusion: Mean growth rate of proximal aorta in hypertensive patients with known aortic dilatation was of about 0.1 mm/year. Dilatation over time is slower in patients with increased rather than normal aortic z score. Eventually, it could be possible to reliably predict aortic diameter at few months from first visit.

Study Design: Retrospective study. Objectives: Huge amounts of images and medical reports are being generated in radiology departments. While these datasets can potentially be employed to train artificial intelligence tools to detect findings on radiological images, the unstructured nature of the reports limits the accessibility of information. In this study, we tested if natural language processing (NLP) can be useful to generate training data for deep learning models analyzing planar radiographs of the lumbar spine. Methods: NLP classifiers based on the Bidirectional Encoder Representations from Transformers (BERT) model able to extract structured information from radiological reports were developed and used to generate annotations for a large set of radiographic images of the lumbar spine (N = 10 287). Deep learning (ResNet-18) models aimed at detecting radiological findings directly from the images were then trained and tested on a set of 204 human-annotated images. Results: The NLP models had accuracies between 0.88 and 0.98 and specificities between 0.84 and 0.99; 7 out of 12 radiological findings had sensitivity >0.90. The ResNet-18 models showed performances dependent on the specific radiological findings with sensitivities and specificities between 0.53 and 0.93. Conclusions: NLP generates valuable data to train deep learning models able to detect radiological findings in spine images. Despite the noisy nature of reports and NLP predictions, this approach effectively mitigates the difficulties associated with the manual annotation of large quantities of data and opens the way to the era of big data for artificial intelligence in musculoskeletal radiology.