A Novel and Efficient Digital Pathology Classifier for Predicting Cancer Biomarkers Using Sequencer Architecture.

In digital pathology tasks, transformers have achieved state-of-the-art results, surpassing convolutional neural networks (CNNs). However, transformers are usually complex and resource intensive. This study developed a novel and efficient digital pathology classifier called DPSeq to predict cancer biomarkers through fine-tuning a sequencer architecture integrating horizontal and vertical bidirectional long short-term memory networks. Using hematoxylin and eosin-stained histopathologic images of colorectal cancer from two international data sets (The Cancer Genome Atlas and Molecular and Cellular Oncology), the predictive performance of DPSeq was evaluated in a series of experiments. DPSeq demonstrated exceptional performance for predicting key biomarkers in colorectal cancer (microsatellite instability status, hypermutation, CpG island methylator phenotype status, BRAF mutation, TP53 mutation, and chromosomal instability), outperforming most published state-of-the-art classifiers in a within-cohort internal validation and a cross-cohort external validation. In addition, under the same experimental conditions using the same set of training and testing data sets, DPSeq surpassed four CNNs (ResNet18, ResNet50, MobileNetV2, and EfficientNet) and two transformer (Vision Transformer and Swin Transformer) models, achieving the highest area under the receiver operating characteristic curve and area under the precision-recall curve values in predicting microsatellite instability status, BRAF mutation, and CpG island methylator phenotype status. Furthermore, DPSeq required less time for both training and prediction because of its simple architecture. Therefore, DPSeq appears to be the preferred choice over transformer and CNN models for predicting cancer biomarkers.

Colorectal Cancer Survival Prediction Using Deep Distribution Based Multiple-Instance Learning.

Most deep-learning algorithms that use Hematoxylin- and Eosin-stained whole slide images (WSIs) to predict cancer survival incorporate image patches either with the highest scores or a combination of both the highest and lowest scores. In this study, we hypothesize that incorporating wholistic patch information can predict colorectal cancer (CRC) cancer survival more accurately. As such, we developed a distribution-based multiple-instance survival learning algorithm (DeepDisMISL) to validate this hypothesis on two large international CRC WSIs datasets called MCO CRC and TCGA COAD-READ. Our results suggest that combining patches that are scored based on percentile distributions together with the patches that are scored as highest and lowest drastically improves the performance of CRC survival prediction. Including multiple neighborhood instances around each selected distribution location (e.g., percentiles) could further improve the prediction. DeepDisMISL demonstrated superior predictive ability compared to other recently published, state-of-the-art algorithms. Furthermore, DeepDisMISL is interpretable and can assist clinicians in understanding the relationship between cancer morphological phenotypes and a patient's cancer survival risk.

Adaptive reconstruction for coded aperture temporal compressive imaging.

This paper presents an adaptive reconstruction method for coded aperture temporal compressive imaging. A pixel-wise equal-exposure coding strategy is first implemented to induce speckle-like features in the moving area of the measured image. A moving area detection method is then proposed to reconstruct the adaptively segmented moving areas into a series of video frames, which are filled back into the stationary clear background. Both simulation and experimental results demonstrate that the proposed method significantly reduces the time consumption of video reconstruction without degradation of image quality.

Full monitoring for long-reach TWDM passive optical networks.

This paper presents a novel and simple fiber monitoring system based on multi-wavelength transmission-reflection analysis for long-reach time and wavelength division multiplexing passive optical networks. For the first time, the full localization functionality of long-reach passive optical networks is possible with the proposed monitoring scheme, including supporting fault detection, identification, and localization in both feeder and distribution fiber segments. By measuring the transmitted and reflected/backscattered optical powers launched by an unmodulated continuous-wave optical source, the proposed solution is able to supervise the network with good spatial accuracy, a high detection speed and a low impact on data traffic. Both the theoretical analysis and experimental validation show that the proposed scheme is capable of providing an accurate fault monitoring functionality for long-reach time and wavelength division multiplexing passive optical networks.

Localization and quantification of reflective events along an optical fiber using a bi-directional TRA technique.

We report on the theory and the implementation of a novel approach for the detection and localization of a reflective event along a fiber link. By launching a continuous-wave signal into both fiber ends and by analyzing the transmitted and reflected/backscattered optical powers, it is possible to localize an optical event and to quantify the induced insertion and return losses simultaneously. The novel idea of utilizing bi-directional measurement allows the localization of both reflective and non-reflective events. Theoretical and experimental studies show that for a 10 km-long single mode fiber, the localization accuracy can be in the range of 5.0 m.

Multi-wavelength Transmission-Reflection Analysis for fiber monitoring.

We propose and implement a novel approach based on multi-wavelength Transmission-Reflection Analysis (MW-TRA) technique for monitoring lossy events (e.g. disconnected connectors, fiber breaks and fiber bendings) along an optical fiber link. By launching un-modulated continuous-wave lights carried by different wavelengths into the fiber and measuring their transmitted and reflected/backscattered optical powers, our proposed MW-TRA scheme is able to localize any lossy event (including both reflective and non-reflective) and to quantify the corresponding insertion and return losses with high accuracy.

Dependences of water permeation through cyclic octa-peptide nanotubes on channel length and membrane thickness.

Effects of the channel length and membrane thickness on the water permeation through the transmembrane cyclic octa-peptide nanotubes (octa-PNTs) have been studied by molecular dynamics (MD) simulations. The water osmotic permeability (p(f)) through the PNTs of k × (WL)(4)/POPE (1-palmitoyl-2-oleoyl-glycerophosphoethanolamine; k = 6, 7, 8, 9, and 10) was found to decay with the channel length (L) along the axis (~L(-2.0)). Energetic analysis showed that a series of water binding sites exist in these transmembrane PNTs, with the barriers of ~3k(B)T, which elucidates the tendency of p(f) well. Water diffusion permeability (p(d)) exhibits a relationship of ~L(-1.8), which results from the novel 1-2-1-2 structure of water chain in such confined nanolumens. In the range of simulation accuracy, the ratio (p(f)/p(d)) of the water osmotic and diffusion permeability is approximately a constant. MD simulations of water permeation through the transmembrane PNTs of 8 × (WL)(4)/octane with the different octane membrane thickness revealed that the water osmotic and diffusion permeability (p(f) and p(d)) are both independent of the octane membrane thickness, confirmed by the weak and nearly same interactions between the channel water and octane membranes with the different thickness. The results may be helpful for revealing the permeation mechanisms of biological water channels and designing artificial nanochannels.

Improving feature extraction from histopathological images through a fine-tuning ImageNet model.

Background: Due to lack of annotated pathological images, transfer learning has been the predominant approach in the field of digital pathology. Pre-trained neural networks based on ImageNet database are often used to extract "off-the-shelf" features, achieving great success in predicting tissue types, molecular features, and clinical outcomes, etc. We hypothesize that fine-tuning the pre-trained models using histopathological images could further improve feature extraction, and downstream prediction performance. Methods: We used 100 000 annotated H&E image patches for colorectal cancer (CRC) to fine-tune a pre-trained Xception model via a 2-step approach. The features extracted from fine-tuned Xception (FTX-2048) model and Image-pretrained (IMGNET-2048) model were compared through: (1) tissue classification for H&E images from CRC, same image type that was used for fine-tuning; (2) prediction of immune-related gene expression, and (3) gene mutations for lung adenocarcinoma (LUAD). Five-fold cross validation was used for model performance evaluation. Each experiment was repeated 50 times. Findings: The extracted features from the fine-tuned FTX-2048 exhibited significantly higher accuracy (98.4%) for predicting tissue types of CRC compared to the "off-the-shelf" features directly from Xception based on ImageNet database (96.4%) (P value = 2.2 × 10-6). Particularly, FTX-2048 markedly improved the accuracy for stroma from 87% to 94%. Similarly, features from FTX-2048 boosted the prediction of transcriptomic expression of immune-related genes in LUAD. For the genes that had significant relationships with image features (P < 0.05, n = 171), the features from the fine-tuned model improved the prediction for the majority of the genes (139; 81%). In addition, features from FTX-2048 improved prediction of mutation for 5 out of 9 most frequently mutated genes (STK11, TP53, LRP1B, NF1, and FAT1) in LUAD. Conclusions: We proved the concept that fine-tuning the pretrained ImageNet neural networks with histopathology images can produce higher quality features and better prediction performance for not only the same-cancer tissue classification where similar images from the same cancer are used for fine-tuning, but also cross-cancer prediction for gene expression and mutation at patient level.

[Relationship between reference values of fibrinogen and geographical factors based on neural network analysis].

OBJECTIVE: To analyze the relationship between the reference values of fibrinogen (FIB) in healthy Chinese adults and geographical factors to provide scientific evidences for establishing the uniform standard. METHODS: The reference values of FIB of 10701 Chinese healthy adults from 103 cities were collected to investigate their relationship with 18 geographical factors including spatial index, terrain index, climate index, and soil index. Geographical factors that significantly correlated with the reference values were selected for constructing the BP neural network model. The spatial distribution map of the reference value of FIB of healthy Chinese adults was fitted by disjunctive kriging interpolation. We used the 5-layer neural network and selected 2000 times of training covering 11 hidden layers to build the simulation rule for simulating the relationship between FIB and geographical environmental factors using the MATLAB software. RESULTS: s The reference value of FIB in healthy Chinese adults was significantly correlated with the latitude, sunshine duration, annual average temperature, annual average relative humidity, annual precipitation, annual range of air temperature, average annual soil gravel content, and soil cation exchange capacity (silt). The artificial neural networks were created to analyze the simulation of the selected indicators of geographical factors. The spatial distribution map of the reference values of FIB in healthy Chinese adults showed a distribution pattern that FIB levels were higher in the South and lower in the North, and higher in the East and lower in the West. CONCLUSION: When the geographical factors of a certain area are known, the reference values of FIB in healthy Chinese adults can be obtained by establishing the neural network mode or plotting the spatial distribution map.

Cyclo-hexa-peptides at the water/cyclohexane interface: a molecular dynamics simulation.

Molecular dynamic (MD) simulations have been performed to study the behaviors of ten kinds of cyclo-hexa-peptides (CHPs) composed of amino acids with the diverse hydrophilic/hydrophobic side chains at the water/cyclohexane interface. All the CHPs take the "horse-saddle" conformations at the interface and the hydrophilicity/hydrophobicity of the side chains influences the backbones' structural deformations. The orientations and distributions of the CHPs at the interface and the differences of interaction energies (ΔΔE) between the CHPs and the two liquid phases have been determined. RDF analysis shows that the H-bonds were formed between the O(C) atoms of the CHPs' backbones and H(w) atoms of water molecules. N atoms of the CHPs' backbones formed the H-bonds or van der Waals interactions with the water solvent. It was found that there is a parallel relationship between ΔΔE and the lateral diffusion coefficients (D ( xy )) of the CHPs at the interface. The movements of water molecules close to the interface are confined to some extent, indicating that the dynamics of the CHPs and interfacial water molecules are strongly coupled.

Water diffusion behaviors and transportation properties in transmembrane cyclic hexa-, octa- and decapeptide nanotubes.

Molecular dynamics simulations have been performed on three transmembrane cyclic peptide nanotubes, i.e., 8 × (WL)(n=3,4,5)/POPE (with uniform lengths but various radii) to investigate the radial dependences of the water-chain structures, diffusions, and transportation properties. The diffusions of individual water molecules and collective coordinates of all the channel-water in the three systems are certified as unbiased Brownian motions. From the very good linear relationships between MSDs and time intervals, the diffusion coefficients and transportation permeabilities have been deduced efficiently. Under the hydrostatic pressure differences across the membrane, a net unidirectional water flow rose up, and the osmotic permeabilities were determined. The ratios of the osmotic and diffusion permeabilities (p(f)/p(d)) were examined for all the three channels.