Machine learning for intraoperative prediction of viability in ischemic small intestine

Strand-Amundsen, Runar; Tronstad, Christian; Reims, Henrik M.; Reinholt, Finn P.; Høgetveit, Jan Olav; Tønnessen, Tor Inge

doi:10.1088/1361-6579/aae0ea

Cited by 9 publications

(9 citation statements)

References 46 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The classification performance using FNN and RNN-LSTM with bioimpedance data (Strand-Amundsen et al, 2018c), to predict intestinal tissue viability and histological grading based on the reference data (Strand-Amundsen et al, 2018a) indicates that good binary prediction of tissue viability is possible based on one bioimpedance measurement before reperfusion and an FNN model. When using the whole time-course of repeated bioimpedance measurement during the experiments a significantly higher accuracy was obtained by utilizing LSTM-RNN ( Figure 29).…”

Section: Machine Learning and Bioimpedance -Results (Paper V)mentioning

confidence: 99%

“…Prediction of tissue viability was assessed for different data points and segments, based on estimation of clinical application and relevance of the method. Classification using all or clinically relevant intervals of history of repeated measurements with different labels, was compared to classification using single measurement points at selected intervals, employing suitable machine learning approaches for both cases (Strand-Amundsen et al, 2018c).…”

Section: Machine Learning and Bioimpedance (Paper V)mentioning

confidence: 99%

“…Training a dataset using LSTM-RNN can be relatively time-and resource-consuming, especially because of the use of repeated-double cross validation, where models with all hyperparameter combinations are trained and validated in the inner loop, which is necessary to avoid overfitting. The time used for the classification training presented in (Strand-Amundsen et al, 2018c) was approximately 90 24-hour periods of calculation on a regular laptop. This can be done over a short time period when utilizing a series of cores on a supercomputer cluster.…”

Section: Bioimpedance Measurement and Neural Network Using Lstm Unitmentioning

confidence: 99%

See 2 more Smart Citations

In vivocharacterization of ischemic small intestine using bioimpedance measurements

et al. 2016

View full text Add to dashboard Cite

The standard clinical method for the assessment of viability in ischemic small intestine is still visual inspection and palpation. This method is non-specific and unreliable, and requires a high level of clinical experience. Consequently, viable tissue might be removed, or irreversibly damaged tissue might be left in the body, which may both slow down patient recovery. Impedance spectroscopy has been used to measure changes in electrical parameters during ischemia in various tissues. The physical changes in the tissue at the cellular and structural levels after the onset of ischemia lead to time-variant changes in the electrical properties. We aimed to investigate the use of bioimpedance measurement to assess if the tissue is ischemic, and to assess the ischemic time duration. Measurements were performed on pigs (n = 7) using a novel two-electrode setup, with a Solartron 1260/1294 impedance gain-phase analyser. After induction of anaesthesia, an ischemic model with warm, full mesenteric arterial and venous occlusion on 30 cm of the jejunum was implemented. Electrodes were placed on the serosal surface of the ischemic jejunum, applying a constant voltage, and measuring the resulting electrical admittance. As a control, measurements were done on a fully perfused part of the jejunum in the same porcine model. The changes in tan δ (dielectric parameter), measured within a 6 h period of warm, full mesenteric occlusion ischemia in seven pigs, correlates with the onset and duration of ischemia. Tan δ measured in the ischemic part of the jejunum differed significantly from the control tissue, allowing us to determine if the tissue was ischemic or not (P < 0.0001, F = (1,75.13) 188.19). We also found that we could use tan δ to predict ischemic duration. This opens up the possibility of real-time monitoring and assessment of the presence and duration of small intestinal ischemia.

show abstract

Section: Machine Learning and Bioimpedance -Results (Paper V)mentioning

confidence: 99%

Section: Machine Learning and Bioimpedance (Paper V)mentioning

confidence: 99%

Section: Bioimpedance Measurement and Neural Network Using Lstm Unitmentioning

confidence: 99%

See 1 more Smart Citation

In vivocharacterization of ischemic small intestine using bioimpedance measurements

et al. 2016

View full text Add to dashboard Cite

show abstract

“…While we do not aim to exhaustively assess the myriad of possibilities and methods within the realm of machine learning, we do intend to cast light on possibilities with neural network architectures for time-series data of bioimpedance. A recent study in our group compared the accuracy of using FNN, versus the accuracy when using LSTM-RNN with classification of intestinal viability following ischemia/reperfusion and found that accuracies in the range of what has been reported clinically can be achieved using FNN's on a single bioimpedance measurement, and higher accuracies can be achieved when employing LSTM-RNN on a sequence of data history (13).…”

Section: Discussionmentioning

confidence: 96%

Possibilities in the application of machine learning on bioimpedance time-series

Tronstad

Strand-Amundsen

2019

Journal of Electrical Bioimpedance

Self Cite

View full text Add to dashboard Cite

The relation between a biological process and the changes in passive electrical properties of the tissue is often non-linear, in which developing prediction models based on bioimpedance spectra is not trivial. Relevant information on tissue status may also lie in characteristic developments in the bioimpedance spectra over time, often neglected by conventional methods. The aim of this study was to explore possibilities in machine learning methods for time series of bioimpedance spectra, where we used organ ischemia as an example. Based on published data on the development of the bioimpedance spectrum during liver ischemia, a simulation model was made and used to generate sets of synthetic data with different levels of organ-to-organ variation, measurement noise and drift. Three types of artificial neural networks were employed in learning to predict the ischemic duration, based on the simulated datasets. The simulated prediction performance was very dependent on the amount of training examples, the organ-to-organ variation and the selection of input variables from the bioimpedance spectrum. The performance was also affected by noise and drift in the measurement, but a recurrent neural network with long short-term memory units could obtain good predictions even on noisy and drifting measurements. This approach may be relevant for further exploration on several applications of bioimpedance having the purpose of predicting a biological state based on spectra measured over time.

show abstract

“…Recent studies from our group demonstrated promising results using a recurrent neural network (RNN) with long short-term memory (LSTM). Through hours long measurements, they obtained high classification accuracies using repeated measurements in tissue related studies suggesting that this type of machine learning approach may be useful in a wider sense for impedance time series problems [16,17].…”

Section: Introductionmentioning

confidence: 99%

Machine learning for stem cell differentiation and proliferation classification on electrical impedance spectroscopy

Cunha

Hou

Schuelke

2019

Journal of Electrical Bioimpedance

View full text Add to dashboard Cite

Electrical impedance spectroscopy (EIS) measurements on cells is a proven method to assess stem cell proliferation and differentiation. Cell regenerative medicine (CRM) is an emerging field where the need to develop and deploy stem cell assessment techniques is paramount as experimental treatments reach pre-clinical and clinical stages. However, EIS measurements on cells is a method requiring extensive post-processing and analysis. As a contribution to address this concern, we developed three machine learning models for three different stem cell lines able to classify the measured data as proliferation or differentiation laying the stone for future studies on using machine learning to profile EIS measurements on stem cells spectra.

show abstract

Machine learning for intraoperative prediction of viability in ischemic small intestine

Cited by 9 publications

References 46 publications

In vivocharacterization of ischemic small intestine using bioimpedance measurements

In vivocharacterization of ischemic small intestine using bioimpedance measurements

Possibilities in the application of machine learning on bioimpedance time-series

Machine learning for stem cell differentiation and proliferation classification on electrical impedance spectroscopy

Contact Info

Product

Resources

About