Receive a weekly summary and discussion of the top papers of the week by leading researchers in the field.

General General

Stain color translation of multi-domain OSCC histopathology images using attention gated cGAN.

In Computerized medical imaging and graphics : the official journal of the Computerized Medical Imaging Society

Oral Squamous Cell Carcinoma (OSCC) is the most prevalent type of oral cancer across the globe. Histopathology examination is the gold standard for OSCC examination, where stained histopathology slides help in studying and analyzing the cell structures under a microscope to determine the stages and grading of OSCC. One of the staining methods popularly known as H&E staining is used to produce differential coloration, highlight key tissue features, and improve contrast, which makes cell analysis easier. However, the stained H&E histopathology images exhibit inter and intra-variation due to staining techniques, incubation times, and staining reagents. These variations negatively impact computer-aided diagnosis (CAD) and Machine learning algorithm's accuracy and development. A pre-processing procedure called stain normalization must be employed to reduce stain variance's negative impacts. Numerous state-of-the-art stain normalization methods are introduced. However, a robust multi-domain stain normalization approach is still required because, in a real-world situation, the OSCC histopathology images will include more than two color variations involving several domains. In this paper, a multi-domain stain translation method is proposed. The proposed method is an attention gated generator based on a Conditional Generative Adversarial Network (cGAN) with a novel objective function to enforce color distribution and the perpetual resemblance between the source and target domains. Instead of using WSI scanner images like previous techniques, the proposed method is experimented on OSCC histopathology images obtained by several conventional microscopes coupled with cameras. The proposed method receives the L* channel from the L*a*b* color space in inference mode and generates the G(a*b*) channel, which are color-adapted. The proposed technique uses mappings learned during training phases to translate the source domain to the target domain; mapping are learned using the whole color distribution of the target domain instead of one reference image. The suggested technique outperforms the four state-of-the-art methods in multi-domain OSCC histopathological translation, the claim is supported by results obtained after assessment in both quantitative and qualitative ways.

Barua Barun, Bora Kangkana, Kr Das Anup, Ahmed Gazi N, Rahman Tashnin

2023-Feb-24

Attention gated generator, Conditional generative adversarial network (cGAN), H&E stain normalization, Histopathology, OSCC, Stain translation

General General

Dual-domain accelerated MRI reconstruction using transformers with learning-based undersampling.

In Computerized medical imaging and graphics : the official journal of the Computerized Medical Imaging Society

Acceleration in MRI has garnered much attention from the deep-learning community in recent years, particularly for imaging large anatomical volumes such as the abdomen or moving targets such as the heart. A variety of deep learning approaches have been investigated, with most existing works using convolutional neural network (CNN)-based architectures as the reconstruction backbone, paired with fixed, rather than learned, k-space undersampling patterns. In both image domain and k-space, CNN-based architectures may not be optimal for reconstruction due to its limited ability to capture long-range dependencies. Furthermore, fixed undersampling patterns, despite ease of implementation, may not lead to optimal reconstruction. Lastly, few deep learning models to date have leveraged temporal correlation across dynamic MRI data to improve reconstruction. To address these gaps, we present a dual-domain (image and k-space), transformer-based reconstruction network, paired with learning-based undersampling that accepts temporally correlated sequences of MRI images for dynamic reconstruction. We call our model DuDReTLU-net. We train the network end-to-end against fully sampled ground truth dataset. Human cardiac CINE images undersampled at different factors (5-100) were tested. Reconstructed images were assessed both visually and quantitatively via the structural similarity index, mean squared error, and peak signal-to-noise. Experimental results show superior performance of DuDReTLU-net over state-of-the-art methods (LOUPE, k-t SLR, BM3D-MRI) in accelerated MRI reconstruction; ablation studies show that transformer-based reconstruction outperformed CNN-based reconstruction in both image domain and k-space; dual-domain reconstruction architectures outperformed single-domain reconstruction architectures regardless of reconstruction backbone (CNN or transformer); and dynamic sequence input leads to more accurate reconstructions than single frame input. We expect our results to encourage further research in the use of dual-domain architectures, transformer-based architectures, and learning-based undersampling, in the setting of accelerated MRI reconstruction. The code for this project is made freely available at https://github.com/william2343/dual-domain-mri-recon-nets (Hong et al., 2022).

Hong Guan Qiu, Wei Yuan Tao, Morley William A W, Wan Matthew, Mertens Alexander J, Su Yang, Cheng Hai-Ling Margaret

2023-Feb-23

Acceleration, Cardiac, Magnetic resonance imaging (MRI), Neural network, Transformer, Undersampling

Radiology Radiology

SynthSeg: Segmentation of brain MRI scans of any contrast and resolution without retraining.

In Medical image analysis

Despite advances in data augmentation and transfer learning, convolutional neural networks (CNNs) difficultly generalise to unseen domains. When segmenting brain scans, CNNs are highly sensitive to changes in resolution and contrast: even within the same MRI modality, performance can decrease across datasets. Here we introduce SynthSeg, the first segmentation CNN robust against changes in contrast and resolution. SynthSeg is trained with synthetic data sampled from a generative model conditioned on segmentations. Crucially, we adopt a domain randomisation strategy where we fully randomise the contrast and resolution of the synthetic training data. Consequently, SynthSeg can segment real scans from a wide range of target domains without retraining or fine-tuning, which enables straightforward analysis of huge amounts of heterogeneous clinical data. Because SynthSeg only requires segmentations to be trained (no images), it can learn from labels obtained by automated methods on diverse populations (e.g., ageing and diseased), thus achieving robustness to a wide range of morphological variability. We demonstrate SynthSeg on 5,000 scans of six modalities (including CT) and ten resolutions, where it exhibits unparallelled generalisation compared with supervised CNNs, state-of-the-art domain adaptation, and Bayesian segmentation. Finally, we demonstrate the generalisability of SynthSeg by applying it to cardiac MRI and CT scans.

Billot Benjamin, Greve Douglas N, Puonti Oula, Thielscher Axel, Van Leemput Koen, Fischl Bruce, Dalca Adrian V, Iglesias Juan Eugenio

2023-Feb-25

CNN, Contrast and resolution invariance, Domain randomisation, Segmentation

Surgery Surgery

Prediction of epilepsy surgery outcome using foramen ovale EEG - A machine learning approach.

In Epilepsy research

INTRODUCTION : Patients with drug-resistant focal epilepsy may benefit from ablative or resective surgery. In presurgical work-up, intracranial EEG markers have been shown to be useful in identification of the seizure onset zone and prediction of post-surgical seizure freedom. However, in most cases, implantation of depth or subdural electrodes is performed, exposing patients to increased risks of complications.

METHODS : We analysed EEG data recorded from a minimally invasive approach utilizing foramen ovale (FO) and epidural peg electrodes using a supervised machine learning approach to predict post-surgical seizure freedom. Power-spectral EEG features were incorporated in a logistic regression model predicting one-year post-surgical seizure freedom. The prediction model was validated using repeated 5-fold cross-validation and compared to outcome prediction based on clinical and scalp EEG variables.

RESULTS : Forty-seven patients (26 patients with post-surgical 1-year seizure freedom) were included in the study, with 31 having FO and 27 patients having peg onset seizures. The area under the receiver-operating curve for post-surgical seizure freedom (Engel 1A) prediction in patients with FO onset seizures was 0.74 ± 0.23 using electrophysiology features, compared to 0.66 ± 0.22 for predictions based on clinical and scalp EEG variables (p < 0.001). The most important features for prediction were spectral power in the gamma and high gamma ranges. EEG data from peg electrodes was not informative in predicting post-surgical outcomes.

CONCLUSION : In this hypothesis-generating study, a data-driven approach based on EEG features derived from FO electrodes recordings outperformed the predictive ability based solely on clinical and scalp EEG variables. Pending validation in future studies, this method may provide valuable post-surgical prognostic information while minimizing risks of more invasive diagnostic approaches.

Miron Gadi, Müller Paul Manuel, Holtkamp Martin, Meisel Christian

2023-Feb-17

Epilepsy surgery, Foramen ovale, Presurgical evaluation, Surgical outcome prediction

General General

Liquid-solid ratio during hydrothermal carbonization affects hydrochar application potential in soil: Based on characteristics comparison and economic benefit analysis.

In Journal of environmental management

Returning straw-like agricultural waste to the field by converting it into hydrochar through hydrothermal carbonization (HTC) is an important way to realize resource utilization of waste, soil improvement, and carbon sequestration. However, the large-scale HTC is highly limited by the large water consumption and waste liquid pollution. Here, we propose strategies to optimize the liquid-solid ratio (LSR) of HTC, and comprehensively evaluate the stability, soil application potential, and economic benefits of corn stover-based hydrochar under different LSRs. The results showed that the total amount of dissolved organic carbon of hydrochars increased by 55.0% as LSR reducing from 10:1 to 2:1, while the element content, thermal stability, carbon fixation potential, specific surface area, pore volume, and functional group type were not obviously affected. The specific surface area and pore volume of hydrochar decreased by 61.8% and 70.9% as LSR reduced to 1:1, due to incomplete carbonization. According to the gray relation, hydrochar derived at LSR of 10:1 and followed by 2:1 showed greatest relation degree of 0.80 and 0.70, respectively, indicating better soil application potential. However, reducing LSR from 10:1 to 2:1 made the income of single process production increased from -388 to 968 ¥, and the wastewater generation decreased by 80%. Considering the large-scale application of HTC in fields for farmland improvement and environmental remediation, the comprehensive advantages of optimized LSR will be further highlighted.

Si Hongyu, Zhao Changkai, Wang Bing, Liang Xiaohui, Gao Mingjie, Jiang Zhaoxia, Yu Hewei, Yang Yuanyuan, Gu Zhijie, Ogino Kenji, Chen Xiuxiu

2023-Feb-27

Corn straw hydrochar, Cost-benefit analysis, Hydrothermal carbonization, Liquid-solid ratio, Soil application

General General

Identifying subtypes of chronic kidney disease with machine learning: development, internal validation and prognostic validation using linked electronic health records in 350,067 individuals.

In EBioMedicine

BACKGROUND : Although chronic kidney disease (CKD) is associated with high multimorbidity, polypharmacy, morbidity and mortality, existing classification systems (mild to severe, usually based on estimated glomerular filtration rate, proteinuria or urine albumin-creatinine ratio) and risk prediction models largely ignore the complexity of CKD, its risk factors and its outcomes. Improved subtype definition could improve prediction of outcomes and inform effective interventions.

METHODS : We analysed individuals ≥18 years with incident and prevalent CKD (n = 350,067 and 195,422 respectively) from a population-based electronic health record resource (2006-2020; Clinical Practice Research Datalink, CPRD). We included factors (n = 264 with 2670 derived variables), e.g. demography, history, examination, blood laboratory values and medications. Using a published framework, we identified subtypes through seven unsupervised machine learning (ML) methods (K-means, Diana, HC, Fanny, PAM, Clara, Model-based) with 66 (of 2670) variables in each dataset. We evaluated subtypes for: (i) internal validity (within dataset, across methods); (ii) prognostic validity (predictive accuracy for 5-year all-cause mortality and admissions); and (iii) medications (new and existing by British National Formulary chapter).

FINDINGS : After identifying five clusters across seven approaches, we labelled CKD subtypes: 1. Early-onset, 2. Late-onset, 3. Cancer, 4. Metabolic, and 5. Cardiometabolic. Internal validity: We trained a high performing model (using XGBoost) that could predict disease subtypes with 95% accuracy for incident and prevalent CKD (Sensitivity: 0.81-0.98, F1 score:0.84-0.97). Prognostic validity: 5-year all-cause mortality, hospital admissions, and incidence of new chronic diseases differed across CKD subtypes. The 5-year risk of mortality and admissions in the overall incident CKD population were highest in cardiometabolic subtype: 43.3% (42.3-42.8%) and 29.5% (29.1-30.0%), respectively, and lowest in the early-onset subtype: 5.7% (5.5-5.9%) and 18.7% (18.4-19.1%).

MEDICATIONS : Across CKD subtypes, the distribution of prescription medication classes at baseline varied, with highest medication burden in cardiometabolic and metabolic subtypes, and higher burden in prevalent than incident CKD.

INTERPRETATION : In the largest CKD study using ML, to-date, we identified five distinct subtypes in individuals with incident and prevalent CKD. These subtypes have relevance to study of aetiology, therapeutics and risk prediction.

FUNDING : AstraZeneca UK Ltd, Health Data Research UK.

Dashtban Ashkan, Mizani Mehrdad A, Pasea Laura, Denaxas Spiros, Corbett Richard, Mamza Jil B, Gao He, Morris Tamsin, Hemingway Harry, Banerjee Amitava

2023-Feb-27

CKD subtype, Cluster analysis, Machine learning, Survival analysis, Unsupervised clustering