Artificial intelligence-assisted system for the assessment of Forrest classification of peptic ulcer bleeding: a multicenter diagnostic study.

BACKGROUND: Inaccurate Forrest classification may significantly affect clinical outcomes, especially in high risk patients. Therefore, this study aimed to develop a real-time deep convolutional neural network (DCNN) system to assess the Forrest classification of peptic ulcer bleeding (PUB). METHODS: A training dataset (3868 endoscopic images) and an internal validation dataset (834 images) were retrospectively collected from the 900th Hospital, Fuzhou, China. In addition, 521 images collected from four other hospitals were used for external validation. Finally, 46 endoscopic videos were prospectively collected to assess the real-time diagnostic performance of the DCNN system, whose diagnostic performance was also prospectively compared with that of three senior and three junior endoscopists. RESULTS: The DCNN system had a satisfactory diagnostic performance in the assessment of Forrest classification, with an accuracy of 91.2% (95%CI 89.5%-92.6%) and a macro-average area under the receiver operating characteristic curve of 0.80 in the validation dataset. Moreover, the DCNN system could judge suspicious regions automatically using Forrest classification in real-time videos, with an accuracy of 92.0% (95%CI 80.8%-97.8%). The DCNN system showed more accurate and stable diagnostic performance than endoscopists in the prospective clinical comparison test. This system helped to slightly improve the diagnostic performance of senior endoscopists and considerably enhance that of junior endoscopists. CONCLUSION: The DCNN system for the assessment of the Forrest classification of PUB showed satisfactory diagnostic performance, which was slightly superior to that of senior endoscopists. It could therefore effectively assist junior endoscopists in making such diagnoses during gastroscopy.

Artificial intelligence-aided diagnostic imaging: A state-of-the-art technique in precancerous screening.

BACKGROUND AND AIM: Chromoendoscopy with the use of indigo carmine (IC) dye is a crucial endoscopic technique to identify gastrointestinal neoplasms. However, its performance is limited by the endoscopist's skill, and no standards are available for lesion identification. Thus, we developed an artificial intelligence (AI) model to replace chromoendoscopy. METHODS: This pilot study assessed the feasibility of our novel AI model in the conversion of white-light images (WLI) into virtual IC-dyed images based on a generative adversarial network. The predictions of our AI model were evaluated against the assessments of five endoscopic experts who were blinded to the purpose of this study with a staining quality rating from 1 (unacceptable) to 4 (excellent). RESULTS: The AI model successfully transformed the WLI of polyps with different morphologies and different types of lesions in the gastrointestinal tract into virtual IC-dyed images. The quality ratings of the real IC-dyed and AI images did not significantly differ concerning surface structure (AI vs IC: 3.08 vs 3.00), lesion border (3.04 vs 2.98), and overall contrast (3.14 vs 3.02) from 10 sets of images (10 AI images and 10 real IC-dyed images). Although the score depended significantly on the evaluator, the staining methods (AI or real IC) and evaluators had no significant interaction (P > 0.05) with each other. CONCLUSION: Our results demonstrated the feasibility of employing AI model's virtual IC staining, increasing the possibility of being employed in daily practice. This novel technology may facilitate gastrointestinal lesion identification in the future.

Artificial intelligence for diagnosing microvessels of precancerous lesions and superficial esophageal squamous cell carcinomas: a multicenter study.

BACKGROUND: Intrapapillary capillary loop (IPCL) is an important factor for predicting invasion depth of esophageal squamous cell carcinoma (ESCC). The invasion depth is closely related to the selection of treatment strategy. However, diagnosis of IPCLs is complicated and subject to interobserver variability. This study aimed to develop an artificial intelligence (AI) system to predict IPCLs subtypes of precancerous lesions and superficial ESCC. METHODS: Images of magnifying endoscopy with narrow band imaging from three hospitals were collected retrospectively. IPCLs subtypes were annotated on images by expert endoscopists according to Japanese Endoscopic Society classification. The performance of the AI system was evaluated using internal and external validation datasets (IVD and EVD) and compared with that of the 11 endoscopists. RESULTS: A total of 7094 images from 685 patients were used to train and validate the AI system. The combined accuracy of the AI system for diagnosing IPCLs subtypes in IVD and EVD was 91.3% and 89.8%, respectively. The AI system achieved better performance than endoscopists in predicting IPCLs subtypes and invasion depth. The ability of junior endoscopists to diagnose IPCLs subtypes (combined accuracy: 84.7% vs 78.2%, P < 0.0001) and invasion depth (combined accuracy: 74.4% vs 67.9%, P < 0.0001) were significantly improved with AI system assistance. Although there was no significant differences, the performance of senior endoscopists was slightly elevated. CONCLUSIONS: The proposed AI system could improve the diagnostic ability of endoscopists to predict IPCLs classification of precancerous lesions and superficial ESCC.

[Experimental Investigation of Infrared Spectral Emissivity of Copper at 300~1123 K].

In this study, a new reflective experimental apparatus, which can measure the spectral emissivity of opaque materials accurately and real timely, has been developed based on the Kirchhoff's law by using the GaAs semiconductor laser as the standard radiation source. The spectral emissivity of brass and red copper at wavelength λ=1.55 µm were investigated systematically with the temperatures range from 300 up to 1123 K by using this apparatus and the influence of temperature, oxidation and heating time on the spectral emissivity of two kinds of specimens were also discussed. The experimental data showed that the spectral emissivity increased with increase of temperature and appeared its peak value and valley value when the thickness of oxide film was at some degree. The spectral emissivity of red copper was always greater than that of brass. The formula for calculating the thickness of oxide film was derived from the reflection model composed of a metal and oxide film, then the peak and valley thickness of the red copper were estimated according to this model. The experimental data of constant temperature measurements showed that the spectral emissivity had a slight increase with heating time increasing. Two hours later, the spectral emissivity of two kinds of samples trended to be stable when the thickness of oxide film was at some degree. The values of spectral emissivity at high temperatures were always larger than that of low temperatures. The results of this study will further enrich spectral emissivity data of copper and provide experimental basis for its application.

[Effects of temperature and heating time on armco iron spectral emissivity in atmospheric environment].

The spectral emissivity of pure iron at 1.55 µm was investigated systematically by using our self-designed reflective experimental apparatus based on the Kirchhoff's law, and the influences of temperature and heating time on the spectral emissivity of pure iron were also discussed. The experimental data showed that the spectral emissivity of pure iron increased with temperature rising and its peak value and valley value appeared at certain temperatures. By analyzing the emissivity model of metal with oxidation layer, the variation of the spectral emissivity of pure iron was illustrated. The influence of heating time on the spectral emissivity was different at different temperature. The research results will further enrich pure iron spectral emissivity data, and provide the experimental basis for its application in atmospheric environment.

Construction synergetic adsorption and activation surface via confined Cu/Cu2O and Ag nanoparticles on TiO2 for effective conversion of CO2 to CH4.

High-performance photocatalysts for catalytic reduction of CO2 are largely impeded by inefficient charge separation and surface activity. Reasonable design and efficient collaboration of multiple active sites are important for attaining high reactivity and product selectivity. Herein, Cu-Cu2O and Ag nanoparticles are confined as dual sites for assisting CO2 photoreduction to CH4 on TiO2. The introduction of Cu-Cu2O leads to an all-solid-state Z-scheme heterostructure on the TiO2 surface, which achieves efficient electron transfer to Cu2O and adsorption and activation of CO2. The confined nanometallic Ag further enhances the carrier's separation efficiency, promoting the conversion of activated CO2 molecules to •COOH and further conversion to CH4. Particularly, this strategy is highlighted on the TiO2 system for a photocatalytic reduction reaction of CO2 and H2O with a CH4 generation rate of 62.5 µmol∙g-1∙h-1 and an impressive selectivity of 97.49 %. This work provides new insights into developing robust catalysts through the artful design of synergistic catalytic sites for efficient photocatalytic CO2 conversion.

The transcription factor, Lmx1b, promotes a neuronal glutamate phenotype and suppresses a GABA one in the embryonic trigeminal brainstem complex.

Achieving an appropriate balance between inhibitory and excitatory neuronal fate is critical for development of effective synaptic transmission. However, the molecular mechanisms dictating such phenotypic outcomes are not well understood, especially in the whisker-to-barrel cortex neuraxis, an oft-used model system for revealing developmental mechanisms. In trigeminal nucleus principalis (PrV), the brainstem link in the whisker-barrel pathway, the transcription factor Lmx1b marks glutamatergic cells. In PrV of Lmx1b knockout mice (-/-), initial specification of glutamatergic vs. GABAergic cell fate is normal until embryonic day 14.5. Subsequently, until the day of birth, glutamatergic markers (e.g., VGLUT2) stain significantly fewer PrV neurons, whereas, GABAergic markers (Pax2 and Gad1) stain significantly more PrV cells, notably in Lmx1b null PrV cells. These changes also occurred in Lmx1b/Bax double-/- mice, where PrV cells are rescued from Lmx1b-/- induced apoptosis; thus, effects upon excitatory/inhibitory cell ratios do not reflect a cell death confound. Electroporation-induced ectopic expression of Lmx1b in an array of sites decreases numbers of neurons that express GABAergic markers, but increases VGLUT2+ cell numbers or stain intensity. Thus, Lmx1b is not involved in the initial specification of glutamatergic cell fate, but is essential for maintaining a glutamatergic phenotype. Other experiments suggest that Lmx1b acts to suppress Pax2, a promoter of GABAergic cell fate, in a cell-autonomous manner, which may be a mechanism for maintaining a functional balance of glutamatergic and GABAergic cell types in development.

The transcription factor, Lmx1b, is necessary for the development of the principal trigeminal nucleus-based lemniscal pathway.

Little is known of transcriptional mechanisms underlying the development of the trigeminal (V) principal sensory nucleus (PrV), the brainstem nucleus responsible for the development of the whisker-to-barrel cortex pathway. Lmx1b, a LIM homeodomain transcription factor, is expressed in embryonic PrV. In Lmx1b knockout ((-)(/)(-)) mice, V primary afferent projections to PrV are normal, albeit reduced in number, whereas the PrV-thalamic lemniscal pathway is sparse and develops late. Excess cell death occurs in the embryonic Lmx1b(-)(/)(-) PrV, but not in Lmx1b/Bax double null mutants. Expression of Drg11, a downstream transcription factor essential for PrV development and pattern formation, is abolished in PrV, but not in the V ganglion. Consequently, whisker patterns fail to develop in PrV by birth. Rescued PrV cells in Lmx1b/Bax double (-)(/)(-)s failed to rescue whisker-related PrV pattern formation. Thus, Lmx1b and Drg11 may act in the same genetic signaling pathway that is essential for PrV pattern formation.

Mice lacking central serotonergic neurons show enhanced inflammatory pain and an impaired analgesic response to antidepressant drugs.

A large body of literature has implicated serotonin [5-hydroxytryptamine (5-HT)] in descending modulation of nociceptive transmission. Here, we have studied the pain behavior of Lmx1b conditional knock-out mice (Lmx1b(f/f/p)), which lack 5-HT neurons in the CNS. Lmx1b(f/f/p) mutant mice showed normal thermal and visceral pain responses but were less sensitive to mechanical stimuli and exhibited enhanced inflammatory pain compared with their littermate control mice. Importantly, the analgesic effect of several antidepressant drugs, including selective serotonin reuptake inhibitors (SSRIs), serotonin-norepinephrine reuptake inhibitors (SNRIs), and tricyclic antidepressants, was either abolished or greatly attenuated in Lmx1b(f/f/p) mice. Moreover, in the acute versus persistent pain settings, the analgesic actions of the SNRI duloxetine and the SSRI fluoxetine were differentially affected. Together, our results provide in vivo genetic evidence demonstrating that although the predominant role of the central 5-HT system in inflammatory pain is inhibitory, its role in acute mechanical pain is facilitatory. The findings that the analgesic effects of various antidepressant drugs are differentially dependent on the central 5-HT system should help us to understand the mechanism of the analgesic action of different classes of antidepressants in the management of persistent pain.

[Segmentation of the prostate ultrasound image based on an improved geodesic active contour model].

In this paper, we propose that, the need of the costly re-initialization procedure can be completely eliminated by using the variation formulation, thus increasing the speed of computing operations. The edge detecting function in the geodesic active contour model is improved by incorporating a prior knowledge. The accuracy of the segmentation algorithm can be enhanced by using the minimal variance. Experimental results show that the proposed algorithm can segment the prostate ultrasound images effectively and avoid the problem of contour leakage.

Cancer Research Advance in CKLF-like MARVEL Transmembrane Domain Containing Member Family (Review).

CKLF-like MARVEL transmembrane domain-containing family (CMTM) is a novel family of genes first reported at international level by Peking University Human Disease Gene Research Center. The gene products are between chemokines and the transmembrane-4 superfamily. Loaceted in several human chromosomes, CMTMs, which are unregulated in kinds of tumors, are potential tumor suppressor genes consisting of CKLF and CMTM1 to CMTM8. CMTMs play important roles in immune, male reproductive and hematopoietic systems. Also, it has been approved that CMTM family has strong connection with diseases of autoimmunity, haematopoietic system and haematopoietic system. The in-depth study in recent years found the close relation between CMTMs and umorigenesis, tumor development and metastasis. CMTM family has a significant clinical value in diagnosis and treatment to the diseases linking to tumor and immune system.

Clinical Characteristics, Radiological Features and Gene Mutation in 10 Chinese Families with Spinocerebellar Ataxias.

BACKGROUND: Spinocerebellar ataxias (SCAs) are a group of neurodegenerative disorders that primarily cause the degeneration in the cerebellum, spinal cord, and brainstem. We study the clinical characteristics, radiological features and gene mutation in Chinese families with SCAs. METHODS: In this study, we investigated 10 SCAs Chinese families with SCA1, SCA3/Machado-Joseph disease (MJD), SCA7, SCA8. There were 27 people who were genetically diagnosed as SCA, of which 21 people showed clinical symptoms, and 6 people had no clinical phenotype that we called them presymptomatic patients. In addition, 3 people with cerebellar ataxia and cataracts were diagnosed according to the Harding diagnostic criteria but failed to be recognized as SCAs on genetic testing. Clinical characteristic analyses of each type of SCAs and radiological examinations were performed. RESULTS: We found that SCA3/MJD was the most common subtype in Han population in China, and the ratio of the pontine tegmentum and the posterior fossa area was negatively correlated with the number of cytosine-adenine-guanine (CAG) repeats; the disease duration was positively correlated with the International Cooperative Ataxia Rating Scale score; and the CAG repeats number of abnormal alleles was negatively correlated with the age of onset. CONCLUSIONS: Collectively our study is a systematic research on SCAs in China, which may help for the clinical diagnosis and prenatal screening of this disease, and it may also aid toward better understanding of this disease.

Follistatin-like 1 suppresses sensory afferent transmission by activating Na+,K+-ATPase.

Excitatory synaptic transmission is modulated by inhibitory neurotransmitters and neuromodulators. We found that the synaptic transmission of somatic sensory afferents can be rapidly regulated by a presynaptically secreted protein, follistatin-like 1 (FSTL1), which serves as a direct activator of Na(+),K(+)-ATPase (NKA). The FSTL1 protein is highly expressed in small-diameter neurons of the dorsal root ganglion (DRG). It is transported to axon terminals via small translucent vesicles and secreted in both spontaneous and depolarization-induced manners. Biochemical assays showed that FSTL1 binds to the α1 subunit of NKA and elevates NKA activity. Extracellular FSTL1 induced membrane hyperpolarization in cultured cells and inhibited afferent synaptic transmission in spinal cord slices by activating NKA. Genetic deletion of FSTL1 in small DRG neurons of mice resulted in enhanced afferent synaptic transmission and sensory hypersensitivity, which could be reduced by intrathecally applied FSTL1 protein. Thus, FSTL1-dependent activation of NKA regulates the threshold of somatic sensation.

Molecular mapping of developing dorsal horn-enriched genes by microarray and dorsal/ventral subtractive screening.

The dorsal horn of the spinal cord consists of distinct laminae that serve as a pivotal region for relaying a variety of somatosensory signals such as temperature, pain, and touch. The molecular mechanisms underlying the development of the dorsal horn are poorly understood. To define a molecular map of the dorsal horn circuit, we have profiled dorsal horn-enriched (DHE) gene expression in dorsal spinal cords on embryonic day 15.5 (E15.5) by genome-wide microarray and smart subtractive screening based on polymerase chain reaction (PCR). High-throughput in situ hybridization (ISH) was carried out to validate the expression of 379 genes in the developing dorsal spinal cord. A total of 113 DHE genes were identified, of which 59% show lamina-specific expression patterns. Most lamina-specific genes were expressed across at least two laminae, however. About 32% of all DHE genes are transcription factors, which represent the largest percentage of the group of all DHE functional classifications. Importantly, several individual lamina-specific transcription factors such c-Maf, Rora, and Satb1 are identified for the first time. Epistasis studies revealed several putative effectors of known DHE transcription factors such as Drg11, Tlx3(Rnx), and Lmx1b. These effector genes, including Grp, Trpc3, Pcp4, and Enc1, have been implicated in synaptic transmission, calcium homeostasis, and structural function and thus may have similar roles in the dorsal horn. The identification of a large number of DHE genes, especially those that are lamina specific, lays a foundation for future studies on the molecular machinery that controls the development of the dorsal horn and on functional differences of these distinct laminae in the dorsal spinal cord.

Upregulation of S100A4 after spinal cord transection in adult rats.

AIM: To investigate whether spinal cord transection induces changes of gene expression of S100A4 protein. METHODS: In a spinal cord transection model, S100A4 expression and cellular localization were examined using cDNA microarray, Northern blot, immunohistochemistry, and immunofluorescence double-labeling methods. RESULTS: There was very limited S100A4 mRNA expression in the control spinal cord. However, S100A4 mRNA expression was increased significantly in both the rostral and caudal spinal cord segments adjacent to the injury site. Specifically, S100A4 gene expression was substantially increased at d 2, peaked at d 7 and d 14, and remained high up to 28 d post-injury. During its peak expression, S100A4 protein was localized in astrocytes of the spinal cord within 5 mm from the site of spinal transection. CONCLUSION: Spinal cord transection induces prolonged S100A4 expression at both mRNA and protein levels in areas close to the injury site. Increased expression of S100A4 in astrocytes after spinal cord transection may indicate that this molecule may play a role in astrocytic responses to injury.

Effects of annexins II and V on survival of neurons and astrocytes in vitro.

AIM: To study the effects of annexins II and V on the survival and neurite outgrowth of primary cultured neurons and the survival of astrocytes after peroxide and hypoxia insults in vitro. METHODS: Annexins II and V proteins and/or corresponding antibodies were added to the medium of primary neocortical cultures. H(2)O(2) and NaN(3) were used to induce neuron injury, respectively. Lactate dehydrogenase (LDH) release was measured. RESULTS: Addition of annexin II or V into the culture medium did not affect the normal survival and neurite outgrowth of cortical neurons. However, when an antibody against annexin II or V was added to the culture, the survival and neurite outgrowth of these neurons markedly declined. Further, addition of the two annexins into cortical cultures after peroxide and hypoxia insults markedly reduced the LDH release and cell death. CONCLUSION: Annexins II and V are essential for the survival and neurite outgrowth of developing cortical neurons, the survival of glial cells, and protect neurons and glial cells against peroxide and hypoxia injuries.

Upregulation of heparin-binding growth-associated molecule after spinal cord injury in adult rats.

AIM: To investigate whether traumatic spinal cord injury (SCI) induces changes of gene expression of heparin-binding growth-associated molecule (HB-GAM). METHODS: In a spinal cord transection model, HB-GAM expression and cellular localization were examined using Northern blot, RT-PCR, immunohistochemistry and immunofluorescence double-labeling methods. RESULTS: HB-GAM mRNA was significantly upregulated in spinal cord tissues rostral and caudal to the injury at 7 d after SCI. HB-GAM gene expression was markedly increased at 3 d, peaked at 7 d, and declined to the baseline level at 28 d post-injury. During its peak expression, HB-GAM was co-localized in astrocytes, oligodendrocytes, and neurons in spinal cord tissues within 7 mm from the site of spinal transection. CONCLUSION: SCI induces HB-GAM expression at both mRNA and protein levels in areas close to the injury. Both neurons and glial cells expressed HB-GAM implying that HB-GAM played a role in the process of injury and/or repair following SCI.

Differential gene expression in neural stem cells and oligodendrocyte precursor cells: a cDNA microarray analysis.

The use of neural stem cells (NSCs) or their progeny oligodendrocyte precursor cells (OPCs) represents a promising repair strategy for many neurological disorders. However, the molecular events and biological features during the transition from NSCs to OPCs remain unclear. In the present study, we isolated NSCs from the embryonic rat forebrain and induced them into OPCs by using B104 conditioned medium (B104CM) in vitro. We then employed cDNA array technology to compare changes in gene expression between the two cell populations. Among 1,176 genes examined, 40 were differentially expressed, and some of them may be involved in OPC differentiation from NSCs. Our findings thus provide new insights into the molecular basis of differentiation of OPCs from NSCs.