Dataset Information

Towards Intraoperative Quantification of Atrial Fibrosis Using Light-Scattering Spectroscopy and Convolutional Neural Networks.

ABSTRACT: Light-scattering spectroscopy (LSS) is an established optical approach for characterization of biological tissues. Here, we investigated the capabilities of LSS and convolutional neural networks (CNNs) to quantitatively characterize the composition and arrangement of cardiac tissues. We assembled tissue constructs from fixed myocardium and the aortic wall with a thickness similar to that of the atrial free wall. The aortic sections represented fibrotic tissue. Depth, volume fraction, and arrangement of these fibrotic insets were varied. We gathered spectra with wavelengths from 500-1100 nm from the constructs at multiple locations relative to a light source. We used single and combinations of two spectra for training of CNNs. With independently measured spectra, we assessed the accuracy of the CNNs for the classification of tissue constructs from single spectra and combined spectra. Combined spectra, including the spectra from fibers distal from the illumination fiber, typically yielded the highest accuracy. The maximal classification accuracy of the depth detection, volume fraction, and permutated arrangements was (mean ± standard deviation (stddev)) 88.97 ± 2.49%, 76.33 ± 1.51%, and 84.25 ± 1.88%, respectively. Our studies demonstrate the reliability of quantitative characterization of tissue composition and arrangements using a combination of LSS and CNNs. The potential clinical applications of the developed approach include intraoperative quantification and mapping of atrial fibrosis, as well as the assessment of ablation lesions.

SUBMITTER: Knighton NJ

PROVIDER: S-EPMC8471003 | biostudies-literature |

REPOSITORIES: biostudies-literature

ACCESS DATA

Similar Datasets

Project description:RATIONALE AND OBJECTIVES:We propose an automated segmentation pipeline based on deep learning for proton lung MRI segmentation and ventilation-based quantification which improves on our previously reported methodologies in terms of computational efficiency while demonstrating accuracy and robustness. The large data requirement for the proposed framework is made possible by a novel template-based data augmentation strategy. Supporting this work is the open-source ANTsRNet-a growing repository of well-known deep learning architectures first introduced here. MATERIALS AND METHODS:Deep convolutional neural network (CNN) models were constructed and trained using a custom multilabel Dice metric loss function and a novel template-based data augmentation strategy. Training (including template generation and data augmentation) employed 205 proton MR images and 73 functional lung MRI. Evaluation was performed using data sets of size 63 and 40 images, respectively. RESULTS:Accuracy for CNN-based proton lung MRI segmentation (in terms of Dice overlap) was left lung: 0.93 ± 0.03, right lung: 0.94 ± 0.02, and whole lung: 0.94 ± 0.02. Although slightly less accurate than our previously reported joint label fusion approach (left lung: 0.95 ± 0.02, right lung: 0.96 ± 0.01, and whole lung: 0.96 ± 0.01), processing time is <1 second per subject for the proposed approach versus ?30 minutes per subject using joint label fusion. Accuracy for quantifying ventilation defects was determined based on a consensus labeling where average accuracy (Dice multilabel overlap of ventilation defect regions plus normal region) was 0.94 for the CNN method; 0.92 for our previously reported method; and 0.90, 0.92, and 0.94 for expert readers. CONCLUSION:The proposed framework yields accurate automated quantification in near real time. CNNs drastically reduce processing time after offline model construction and demonstrate significant future potential for facilitating quantitative analysis of functional lung MRI.

Dataset Information

Towards Intraoperative Quantification of Atrial Fibrosis Using Light-Scattering Spectroscopy and Convolutional Neural Networks.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets