Development of high-quality artificial intelligence for computer-aided diagnosis in determining subtypes of colorectal cancer.

BACKGROUND AND AIM: There are no previous studies in which computer-aided diagnosis (CAD) diagnosed colorectal cancer (CRC) subtypes correctly. In this study, we developed an original CAD for the diagnosis of CRC subtypes. METHODS: Pretraining for the CAD based on ResNet was performed using ImageNet and five open histopathological pretraining image datasets (HiPreD) containing 3 million images. In addition, sparse attention was introduced to improve the CAD compared to other attention networks. One thousand and seventy-two histopathological images from 29 early CRC cases at Kyoto Prefectural University of Medicine from 2019 to 2022 were collected (857 images for training and validation, 215 images for test). All images were annotated by a qualified histopathologist for segmentation of normal mucosa, adenoma, pure well-differentiated adenocarcinoma (PWDA), and moderately/poorly differentiated adenocarcinoma (MPDA). Diagnostic ability including dice sufficient coefficient (DSC) and diagnostic accuracy were evaluated. RESULTS: Our original CAD, named Colon-seg, with the pretraining of both HiPreD and ImageNET showed a better DSC (88.4%) compared to CAD without both pretraining (76.8%). Regarding the attentional mechanism, Colon-seg with sparse attention showed a better DSC (88.4%) compared to other attentional mechanisms (dual: 79.7%, ECA: 80.7%, shuffle: 84.7%, SK: 86.9%). In addition, the DSC of Colon-seg (88.4%) was better than other types of CADs (TransUNet: 84.7%, MultiResUnet: 86.1%, Unet++: 86.7%). The diagnostic accuracy of Colon-seg for each histopathological type was 94.3% for adenoma, 91.8% for PWDA, and 92.8% for MPDA. CONCLUSION: A deep learning-based CAD for CRC subtype differentiation was developed with pretraining and fine-tuning of abundant histopathological images.

Effectiveness of FEES with artificial intelligence-assisted computer-aided diagnosis.

OBJECTIVES: FEES is a standard procedure for diagnosing dysphagia. However, appropriate evaluation of FEES findings is difficult for inexperienced evaluators. Recent progress in deep learning has highlighted the use of artificial intelligence-assisted computer-aided diagnosis (AI-assisted CAD) in medical applications. We investigated the detection accuracy of FEES findings evaluated by inexperienced evaluators with and without the use of CAD. METHODS: The algorithm for FEES-CAD was developed using 25,630 expert-annotated images. A total of 45 inexperienced evaluators from three groups of people (resident doctors, nurses, and medical students), evaluated 32 FEES videos from 32 patients. To confirm the effectiveness of FEES-CAD, first, 32 FEES videos were evaluated without the use of CAD. Second, one half was evaluated with, and one half without, the use of CAD. The detection accuracy of the FEES findings was investigated, and the evaluation results obtained with CAD were statistically compared with those obtained without CAD. RESULTS: In the first FEES evaluation, the total detection accuracy was 82.2 %. In the second evaluation, the total detection accuracy with CAD was 84.3 %, and that without CAD was 81.7 %. The detection accuracies by the resident doctors, nurses, and medical students with CAD were 90.1 %, 82.6 %, and 79.4 %, respectively, and those without CAD were 83.7 %, 80.9 % and 80.1 %, respectively. In the resident doctors, the detection accuracy was significantly better when CAD was used for evaluation, compared with the non-CAD evaluations. CONCLUSION: The present study demonstrated the effectiveness of FEES-CAD in improving the detection accuracy of resident doctors, however, the differences were small.

Evaluating the toxic mechanism of 1,2-diacetylbenzene in neural cells/tissues: The favorable impact of silibinin.

1,2-diacetylbenzene (1,2-DAB) is a neurotoxic component of aromatic solvents commonly used in industrial applications that induces neuropathological changes in animals. This study unraveled the toxic impact of 1,2-DAB in nerve tissues, explant cultures, and neuron-glial cultures, and explored whether herbal products can mitigate its toxicity. The effects of DAB on axonal transport were studied in retinal explant cultures grown in a micro-patterned dish. The mitochondrial movement in the axons was captured using time-lapse video recordings. The results showed that 1,2-DAB, but not 1,3-DAB inhibited axonal outgrowth and mitochondrial movement in a dose-dependent manner. The toxicity of 1,2-DAB was further studied in spinal cord tissues and cultures. 1,2-DAB selectively induced modifications of microtubules and neurofilaments in spinal cord tissues. 1,2-DAB also potently induced cell damage in both neuronal and glial cultures. Further, 1,2-DAB-induced cellular ATP depletion precedes cell damage in glial cells. Interestingly, treatment with the herbal products silibinin or silymarin effectively mitigated 1,2-DAB-induced toxicity in spinal cord tissues and neuronal/glial cultures. Collectively, the molecular toxicity of 1,2-DAB in neural tissues involves protein modification, ATP depletion, and axonal transport defects, leading to cell death. Silibinin and silymarin show promising neuroprotective effects against 2-DAB-induced toxicity.

Expert-level aspiration and penetration detection during flexible endoscopic evaluation of swallowing with artificial intelligence-assisted diagnosis.

Flexible endoscopic evaluation of swallowing (FEES) is considered the gold standard in diagnosing oropharyngeal dysphagia. Recent advances in deep learning have led to a resurgence of artificial intelligence-assisted computer-aided diagnosis (AI-assisted CAD) for a variety of applications. AI-assisted CAD would be a remarkable benefit in providing medical services to populations with inadequate access to dysphagia experts, especially in aging societies. This paper presents an AI-assisted CAD named FEES-CAD for aspiration and penetration detection on video recording during FEES. FEES-CAD segments the input FEES video and classifies penetration, aspiration, residue in the vallecula, and residue in the hypopharynx based on the segmented FEES video. We collected and annotated FEES videos from 199 patients to train the network and tested the performance of FEES-CAD using FEES videos from other 40 patients. These patients consecutively underwent FEES between December 2016 and August 2019 at Fukushima Medical University Hospital. FEES videos were deidentified, randomized, and rated by FEES-CAD and laryngologists with over 15 years of experience in performing FEES. FEES-CAD achieved an average Dice similarity coefficient of 98.6[Formula: see text]. FEES-CAD achieved expert-level accuracy performance on penetration (92.5[Formula: see text]), aspiration (92.5[Formula: see text]), residue in the vallecula (100[Formula: see text]), and residue in the hypopharynx (87.5[Formula: see text]) classification tasks. To the best of our knowledge, FEES-CAD is the first CNN-based system that achieves expert-level performance in detecting aspiration and penetration.

Automatic Measurement of Endometrial Thickness From Transvaginal Ultrasound Images.

Purpose: Endometrial thickness is one of the most important indicators in endometrial disease screening and diagnosis. Herein, we propose a method for automated measurement of endometrial thickness from transvaginal ultrasound images. Methods: Accurate automated measurement of endometrial thickness relies on endometrium segmentation from transvaginal ultrasound images that usually have ambiguous boundaries and heterogeneous textures. Therefore, a two-step method was developed for automated measurement of endometrial thickness. First, a semantic segmentation method was developed based on deep learning, to segment the endometrium from 2D transvaginal ultrasound images. Second, we estimated endometrial thickness from the segmented results, using a largest inscribed circle searching method. Overall, 8,119 images (size: 852 × 1136 pixels) from 467 cases were used to train and validate the proposed method. Results: We achieved an average Dice coefficient of 0.82 for endometrium segmentation using a validation dataset of 1,059 images from 71 cases. With validation using 3,210 images from 214 cases, 89.3% of endometrial thickness errors were within the clinically accepted range of ±2 mm. Conclusion: Endometrial thickness can be automatically and accurately estimated from transvaginal ultrasound images for clinical screening and diagnosis.

Niobium phosphates as new highly selective catalysts for the oxidative dehydrogenation of ethane.

Several niobium phosphate phases have been prepared, fully characterized and tested as catalysts for the selective oxidation of ethane to ethylene. Three distinct niobium phosphate catalysts were prepared, and each was comprised predominantly of a different bulk phase, namely Nb(2)P(4)O(15), NbOPO(4) and Nb(1.91)P(2.82)O(12). All of the niobium phosphate catalysts showed high selectivity towards ethylene, but the best catalyst was Nb(1.91)P(2.82)O(12), which was produced from the reduction of niobium oxide phosphate (NbOPO(4)) by hydrogen. It was particularly selective for ethylene, giving ca. 95% selectivity at 5% conversion, decreasing to ca. 90% at 15% conversion, and only produced low levels of carbon oxides. It was also determined that the only primary product from ethane oxidation over this catalyst was ethylene. Catalyst activity also increased with time-on-line, and this behaviour was ascribed to an increase of the concentration of the Nb(1.91)P(2.82)O(12) phase, as partially transformed NbOPO(4), formed during preparation, was converted to Nb(1.91)P(2.82)O(12) during use. Catalysts with predominant phases of Nb(2)P(4)O(15) and NbOPO(4) also showed appreciable activity and selectivities to ethylene with values around 75% and 85% respectively at 5% ethane conversion. The presence of phosphorous is required to achieve high ethylene selectivity, as orthorhombic and monoclinic Nb(2)O(5) catalysts showed similar activity, but displayed selectivities to ethylene that were <20% under the same reaction conditions. To the best of our knowledge, this is the first time that niobium phosphates have been shown to be highly selective catalysts for the oxidation of ethane to ethylene, and demonstrates that they are worthy candidates for further study.

Syringaldehyde promoting intestinal motility with suppressing α-amylase hinders starch digestion in diabetic mice.

The antihyperglycemic potential of syringaldehyde has been previously investigated; however, the underlying mechanism remains unclear. In this study, we performed a postprandial glucose test (in vivo) including oral glucose tolerance test (OGTT) and oral starch tolerance test (OSTT) in fructose-induced diabetic mice on a high-fat diet for mimicking type 2 diabetes to explore the hypoglycemic efficacy of syringaldehyde and the underlined molecular involvement of syringaldehyde in a glucose-lowering effect. The results revealed that syringaldehyde dose-dependently suppressed blood glucose in both the OSTT and OGTT when referenced to acarbose and metformin, respectively. Surprisingly, syringaldehyde triggered jejunum motility (ex vivo) via activation of the muscarinic-type acetylcholine receptor. By performing virtual screening with molecular docking, the data showed that syringaldehyde nicely interacted with glucagon-like peptide 1 receptor (GLP-1R), peroxisome proliferator-activated receptor (PPAR), dipeptidyl peptidase-IV (DPP-4), acetylcholine M2 receptor, and acetylcholinesterase. These results showed that syringaldehyde can potentiate intestinal contractility to abolish the α-amylase reaction when concurrently reducing retention time and glucose absorption to achieve a glucose-lowering effect in diabetic mice, suggesting its potential therapeutic benefits with improvement for use as a prophylactic and treatment.

High-precision deformation mapping in finFET transistors with two nanometre spatial resolution by precession electron diffraction.

Precession electron diffraction has been used to systematically measure the deformation in Si/SiGe blanket films and patterned finFET test structures grown on silicon-on-insulator type wafers. Deformation maps have been obtained with a spatial resolution of 2.0 nm and a precision of ±0.025%. The measured deformation by precession diffraction for the blanket films has been validated by comparison to energy dispersive x-ray spectrometry, X-Ray diffraction, and finite element simulations. We show that although the blanket films remain biaxially strained, the patterned fin structures are fully relaxed in the crystallographic planes that have been investigated. We demonstrate that precession diffraction is a viable deformation mapping technique that can be used to provide useful studies of state-of-the-art electronic devices.

Cross-sectional atom probe tomography sample preparation for improved analysis of fins on SOI.

Sample preparation for atom probe tomography of 3D semiconductor devices has proven to significantly affect field evaporation and the reliability of reconstructed data. A cross-sectional preparation method is applied to state-of-the-art Si finFET technology on SOI. This preparation approach advantageously provides a conductive path for voltage and heat, offers analysis of many fins within a single tip, and improves resolution across interfaces of particular interest. Measured B and Ge profiles exhibit good correlation with SIMS and EDX and show no signs of B clustering or pile-up near the Si/SiGe interface of the fin.

Assessment of a nanocrystal 3-D morphology by the analysis of single HAADF-HRSTEM images.

This work presents the morphological characterization of CeO2 nanocrystals by the analysis of single unfiltered high-angle annular dark-field (HAADF)-high-resolution scanning transmission electron microscopy (HRSTEM) images. The thickness of each individual atomic column is estimated by the classification of its HAADF integrated intensity using a Gaussian mixture model. The resulting thickness maps obtained from two example nanocrystals with distinct morphology were analyzed with aid of the symmetry from the CeO2 crystallographic structure, providing an approximation for their 3-D morphology with high spatial resolution. A confidence level of ±1 atom per atomic column along the viewing direction on the thickness estimation is indicated by the use of multislice image simulation. The described characterization procedure stands out as a simple approach for retrieving morphological parameters of individual nanocrystals, such as volume and specific surface areas for different crystalline planes. The procedure is an alternative to the tilt-series tomography technique for a number of nanocrystalline systems, since its application does not require the acquisition of multiple images from the same nanocrystal along different zone axes.

Oxidation of benzyl alcohol by using gold nanoparticles supported on ceria foam.

The efficacy of using cerium oxide foams as a support for Au nanoparticles and subsequent use as oxidation catalysts have been investigated. These were synthesized using L-asparagine to produce a cerium coordination polymer foam, which was calcined to give the oxide foam. Au nanoparticles were supported on the CeO(2) foams using a sol-immobilization method. The activity of the Au/foamCeO(2) for solvent-free benzyl alcohol oxidation was superior to standard Au/CeO(2) catalysts, and the activity was found to be dependent on the crystallization time of the precursor foam. A crystallization time of 4 h was found to produce the most active catalyst, which retained activity and a high selectivity to benzaldehyde (ca. 96 %) when re-used and this is related to the structure of the material. The high activity is attributed to the greater lability of surface oxygen in the support compared with commercial CeO(2) materials.