Search

To achieve a fully automatic artificial pancreas (AP), i.e., an AP without the need for meal announcements, the intraperitoneal (IP) route is explored. This route has faster dynamics than the typical subcutaneous (SC) route. Model predictive control (MPC) is the most promising control algorithm, but it requires a predictive and identifiable model. This paper presents the design of such a model for MPC-based dual hormone IP APs. This model is trained and tested on recorded data from anesthetized pigs. Animal experiments show that the saturation of the hepatic first-pass effect is essential in how IP insulin and IP glucagon affect glucose levels. These physiological phenomena must be modeled to estimate the system behavior for various conditions. This, in turn, increases the number of parameters and complicates system identification. The availability of rich experimental data from 26 animal trials motivated the design of a technique to exploit this prior information to ensure the identifiability of our model. Through this technique, most parameters were either modeled as body weight functions or common among animals. The correlation between parameter values and body weight is discovered utilizing prior data from various animal experiments, such as blood glucose, plasma insulin, and glucagon levels, in which hormones were administered intraperitoneally or intravenously. This method simplifies the system identification for every new subject while keeping the model’s essential details that improve the prediction capability relative to comparable models. The model can be exploited in MPC or any other model-based controller of a bi-hormonal IP AP. It can also be used as a simulator to develop control approaches for single and bi-hormonal IP APs.

Aims: To conduct a pooled analysis to assess the performance of intermittently scanned continuous glucose monitoring (isCGM) in association with the rate of change in sensor glucose in a cohort of children, adolescents, and adults with type 1 diabetes. Material and methods: In this pooled analysis, isCGM system accuracy was assessed depending on the rate of change in sensor glucose. Clinical studies that have been investigating isCGM accuracy against blood glucose, accompanied with collection time points were included in this analysis. isCGM performance was assessed by means of median absolute relative difference (MedARD), Parkes error grid (PEG) and Bland-Altman plot analyses. Results: Twelve studies comprising 311 participants were included, with a total of 15 837 paired measurements. The overall MedARD (interquartile range) was 12.7% (5.9-23.5) and MedARD differed significantly based on the rate of change in glucose (P < 0.001). An absolute difference of -22 mg/dL (-1.2 mmol/L) (95% limits of agreement [LoA] 60 mg/dL (3.3 mmol/L), -103 mg/dL (-5.7 mmol/L)) was found when glucose was rapidly increasing (isCGM glucose minus reference blood glucose), while a -32 mg/dL (1.8 mmol/L) (95% LoA 116 mg/dL (6.4 mmol/L), -51 mg/dL (-2.8 mmol/L)) absolute difference was observed in periods of rapidly decreasing glucose. Conclusions: The performance of isCGM was good when compared to reference blood glucose measurements. The rate of change in glucose for both increasing and decreasing glucose levels diminished isCGM performance, showing lower accuracy during high rates of glucose change. Keywords: continuous glucose monitoring (CGM); type 1 diabetes.

Sex differences, tumor vascularity, and blood flow seen on multiecho dynamic contrast–based MRI scans enabled prediction of treatment response and overall survival in patients with rectal cancer. Background: MRI is the standard tool for rectal cancer staging. However, more precise diagnostic tests that can assess biologic tumor features decisive for treatment outcome are necessary. Tumor perfusion and hypoxia are two important features; however, no reference methods that measure these exist in clinical use. Purpose: To assess the potential predictive and prognostic value of MRI-assessed rectal cancer perfusion, as a surrogate measure of hypoxia, for local treatment response and survival. Materials and Methods: In this prospective observational cohort study, 94 study participants were enrolled from October 2013 to December 2017 (ClinicalTrials.gov: NCT01816607). Participants had histologically confirmed rectal cancer and underwent routine diagnostic MRI, an extended diffusion-weighted sequence, and a multiecho dynamic contrast agent–based sequence. Predictive and prognostic values of dynamic contrast-enhanced, dynamic susceptibility contrast (DSC), and intravoxel incoherent motion MRI were investigated with response to neoadjuvant treatment, progression-free survival, and overall survival as end points. Secondary objectives investigated potential sex differences in MRI parameters and relationship with lymph node stage. Statistical methods used were Cox regression, Student t test, and Mann-Whitney U test.Results: A total of 94 study participants (mean age, 64 years ± 11 [standard deviation]; 61 men) were evaluated. Baseline tumor blood flow from DSC MRI was lower in patients who had poor local tumor response to neoadjuvant treatment (96 mL/min/100 g ± 33 for ypT2–4, 120 mL/min/100 g ± 21 for ypT0–1; P = .01), shorter progression-free survival (hazard ratio = 0.97; 95% confidence interval: 0.96, 0.98; P < .001), and shorter overall survival (hazard ratio = 0.98; 95% confidence interval: 0.98, 0.99; P < .001). Women had higher blood flow (125 mL/min/100 g ± 27) than men (74 mL/min/100 g ± 26, P < .001) at stage 4. Volume transfer constant and plasma volume from dynamic contrast-enhanced MRI as well as ΔR2* peak and area under the curve for 30 and 60 seconds from DSC MRI were associated with local malignant lymph nodes (pN status). Median area under the curve for 30 seconds was 0.09 arbitrary units (au) ± 0.03 for pN1–2 and 0.19 au ± 0.12 for pN0 (P = .001). Conclusion:Low tumor blood flow from dynamic susceptibility contrast MRI was associated with poor treatment response in study participants with rectal cancer.

The design of a model-based control method for an Artificial Pancreas requires a relatively simple and identifiable mathematical model to control glucose levels through hormone delivery. In this work we introduce new, simple nonlinear models to simulate data from experiments where insulin boluses are administrated in the peritoneal cavity. The models account for the delay between insulin administration and its nonlinear transport to other compartments. They were calibrated using experimental data from pigs. The results show that the suggested models are able to describe the data well, with average BIC value of 145. Moreover, the new models were compared with a common linear model which was not able to describe the data well, with BIC value of 920. They were also compared with a common nonlinear model which failed to represent insulin increases in the data and had BIC value of 637. Finally, profile likelihoods were applied for assessing the identifiability of one of the new models. © 2019, IFAC (International Federation of Automatic Control) Hosting by Elsevier Ltd.

Background The analysis of abdominal sounds can help to diagnose gastro-intestinal diseases. Sounds originating from the stomach and the intestine, the so-called bowel sounds, occur in various forms. They are described as loose successions or clusters of rather sudden bursts. Realistic recordings of abdominal sounds are contaminated with noise and artifacts from which the bowel sounds must be differentiated. Methods The proposed intrinsic mode function-fractal dimension (IMF-FD) filtering utilizes the property of the multivariate empirical mode decomposition (MEMD) to behave as a series of band pass filters. The MEMD decomposes the abdominal signal into its different frequency components. The resulting intrinsic mode functions (IMFs) are modulated in amplitude and frequency where transient sonic events occur. Based on the complexity of the IMFs, measured by their fractal dimension (FD) in sliding windows, the information-carrying IMFs are selected. The filtered signal is formed as the superposition of all selected IMFs. The IMF-FD filter not only enhances the non-linear components of the original signal but also segments them from the rest. Another important aspect of this work is that typical artifacts that occur in the same frequency range as bowel sounds can be subsequently eliminated by heuristic rules. Conclusions The method is tested on a realistic, contaminated data set with promising performance: close to 100% of the manually labeled bowel sounds are identified. © The Author(s) 2019. This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/)

To decrease colon polyp miss-rate during colonoscopy, a real-time detection system with high accuracy is needed. Recently, there have been many efforts to develop models for real-time polyp detection, but work is still required to develop real-time detection algorithms with reliable results. We use single-shot feed-forward fully convolutional neural networks (F-CNN) to develop an accurate real-time polyp detection system. F-CNNs are usually trained on binary masks for object segmentation. We propose the use of 2D Gaussian masks instead of binary masks to enable these models to detect different types of polyps more effectively and efficiently and reduce the number of false positives. The experimental results showed that the proposed 2D Gaussian masks are efficient for detection of flat and small polyps with unclear boundaries between background and polyp parts. The masks make a better training effect to discriminate polyps from the polyp-like false positives. The proposed method achieved state-of-the-art results on two polyp datasets. On the ETIS-LARIB dataset we achieved 86.54% recall, 86.12% precision, and 86.33% F1-score, and on the CVC-ColonDB we achieved 91% recall, 88.35% precision, and F1-score 89.65%. © 2020 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/)