Vision for the 12 LABOURS Digital Twin Platform.

Clinical translation of personalised computational physiology workflows and digital twins can revolutionise healthcare by providing a better understanding of an individual's physiological processes and any changes that could lead to serious health consequences. However, the lack of common infrastructure for developing these workflows and digital twins has hampered the realisation of this vision. The Auckland Bioengineering Institute's 12 LABOURS project aims to address these challenges by developing a Digital Twin Platform to enable researchers to develop and personalise computational physiology models to an individual's health data in clinical workflows. This will allow clinical trials to be more efficiently conducted to demonstrate the efficacy of these personalised clinical workflows. We present a demonstration of the platform's capabilities using publicly available data and an existing automated computational physiology workflow developed to assist clinicians with diagnosing and treating breast cancer. We also demonstrate how the platform facilitates the discovery and exploration of data and the presentation of workflow results as part of clinical reports through a web portal. Future developments will involve integrating the platform with health systems and remote-monitoring devices such as wearables and implantables to support home-based healthcare. Integrating outputs from multiple workflows that are applied to the same individual's health data will also enable the generation of their personalised digital twin.Clinical Relevance- The proposed 12 LABOURS Digital Twin Platform will enable researchers to 1) more efficiently conduct clinical trials to assess the efficacy of their computational physiology workflows and support the clinical translation of their research; 2) reuse primary and derived data from these workflows to generate novel workflows; and 3) generate personalised digital twins by integrating the outputs of different computational physiology workflows.

Atrous spatial pyramid pooling and multi-image data fusion for smooth muscle segmentation in upper gastrointestinal sphincters.

Gastrointestinal (GI) sphincters provide critical roles in regulating the transport of contents along the GI tract. Dysfunctions of GI sphincters are associated with a range of major digestive disorders. Despite their importance, the microstructures of GI sphincters are not well investigated. While micro-computed tomography (µ-CT) provides detailed 3D images, conventional segmentation methods rely on manual correction, which is both time-consuming and prone to human error. This study proposes a segmentation method using atrous spatial pyramid pooling (ASPP), which helps in capturing different effective fields of view from a given input feature map, thereof providing finer local and global information for a given pixel. Additionally, we explored the use of multi-species data fusion to make the model more generalized. The proposed segmentation network incorporating ASPP and multi-species data fusion improved the segmentation of sphincter muscle images. Specifically, it increased the dice score and Jaccard index by 3.7% and 5.8%, respectively, while reducing the variance compared to conventional methods.Clinical relevance- Techniques developed in this study will inform µ-CT segmentation of human upper GI sphincters for detailed structural analysis of muscular dysfunction.

Reconstruction of the human lower esophageal sphincter based on ultra-mill imaging for biomechanical analysis.

Introduction: The lower esophageal sphincter (LES) controls the passage into the stomach and prevents reflex of contents into the esophagus. Dysfunctions of this region typically involves impairment of muscular function, leading to diseases including gastro-esophageal reflux disease and achalasia. The main objective of this study was to develop a finite element model from a unique human LES dataset reconstructed from an ultra-mill imaging setup, and then to investigate the effect of anatomical characteristics on intraluminal pressures. Methods: A pipeline was developed to generate a mesh from a set of input images, which were extracted from a unique ultra-mill sectioned human LES. A total of 216 nodal points with cubic Hermite basis function was allocated to reconstruct the LES, including the longitudinal and circumferential muscles. The resultant LES mesh was used in biomechanical simulations, utilizing a previously developed LES mathematical model based on the Visible Human data to calculate intraluminal pressures. Anatomical and functional comparisons were made between the Ultra-mill and Visible human models. Results: Overall, the Ultra-mill model contained lower cavity (1,796 vs. 5,400 mm3) and muscle (1,548 vs. 15,700 mm3) volumes than the Visible Human model. The Ultra-mill model also developed a higher basal pressure (13.8 vs. 14.7 mmHg) and magnitude of pressure (19.8 vs. 18.9 mmHg) during contraction. Out of all the geometric transformations (i.e., uniform enlargement of volume, lengthening along the center-axis, dilation of the diameter, and increasing muscle thickness), the muscle volume was found to be the main contributor of basal and magnitude of pressures. Increases in length also caused proportional increases to pressures, while dilation of diameter had a less influential reverse effect. Discussion: The findings provide information on interindividual variability in LES pressure and demonstrates that anatomy has a large influence on pressures. This model forms the basis of more complex simulations involving food bolus transport and predicting LES dysfunctions.