Application of convolutional neural networks in medical images: a bibliometric analysis.

Background: In the field of medical imaging, the rapid rise of convolutional neural networks (CNNs) has presented significant opportunities for conserving healthcare resources. However, with the wide spread application of CNNs, several challenges have emerged, such as enormous data annotation costs, difficulties in ensuring user privacy and security, weak model interpretability, and the consumption of substantial computational resources. The fundamental challenge lies in optimizing and seamlessly integrating CNN technology to enhance the precision and efficiency of medical diagnosis. Methods: This study sought to provide a comprehensive bibliometric overview of current research on the application of CNNs in medical imaging. Initially, bibliometric methods were used to calculate the frequency statistics, and perform the cluster analysis and the co-citation analysis of countries, institutions, authors, keywords, and references. Subsequently, the latent Dirichlet allocation (LDA) method was employed for the topic modeling of the literature. Next, an in-depth analysis of the topics was conducted, and the topics in the medical field, technical aspects, and trends in topic evolution were summarized. Finally, by integrating the bibliometrics and LDA results, the developmental trajectory, milestones, and future directions in this field were outlined. Results: A data set containing 6,310 articles in this field published from January 2013 to December 2023 was complied. With a total of 55,538 articles, the United States led in terms of the citation count, while in terms of the publication volume, China led with 2,385 articles. Harvard University emerged as the most influential institution, boasting an average of 69.92 citations per article. Within the realm of CNNs, residual neural network (ResNet) and U-Net stood out, receiving 1,602 and 1,419 citations, respectively, which highlights the significant attention these models have received. The impact of coronavirus disease 2019 (COVID-19) was unmistakable, as reflected by the publication of 597 articles, making it a focal point of research. Additionally, among various disease topics, with 290 articles, brain-related research was the most prevalent. Computed tomography (CT) imaging dominated the research landscape, representing 73% of the 30 different topics. Conclusions: Over the past 11 years, CNN-related research in medical imaging has grown exponentially. The findings of the present study provide insights into the field's status and research hotspots. In addition, this article meticulously chronicled the development of CNNs and highlighted key milestones, starting with LeNet in 1989, followed by a challenging 20-year exploration period, and culminating in the breakthrough moment with AlexNet in 2012. Finally, this article explored recent advancements in CNN technology, including semi-supervised learning, efficient learning, trustworthy artificial intelligence (AI), and federated learning methods, and also addressed challenges related to data annotation costs, diagnostic efficiency, model performance, and data privacy.

Transcriptome Analysis and QTL Mapping Identify Candidate Genes and Regulatory Mechanisms Related to Low-Temperature Germination Ability in Maize.

Low-temperature germination ability (LTGA) is an important characteristic for spring sowing maize. However, few maize genes related to LTGA were confirmed, and the regulatory mechanism is less clear. Here, maize-inbred lines Ye478 and Q1 with different LTGA were used to perform transcriptome analysis at multiple low-temperature germination stages, and a co-expression network was constructed by weighted gene co-expression network analysis (WGCNA). Data analysis showed that 7964 up- and 5010 down-regulated differentially expressed genes (DEGs) of Ye478 were identified at low-temperature germination stages, while 6060 up- and 2653 down-regulated DEGs of Q1 were identified. Gene ontology (GO) enrichment analysis revealed that ribosome synthesis and hydrogen peroxide metabolism were enhanced and mRNA metabolism was weakened under low-temperature stress for Ye478, while hydrogen peroxide metabolism was enhanced and mRNA metabolism was weakened for Q1. DEGs pairwise comparisons between the two genotypes found that Ye478 performed more ribosome synthesis at low temperatures compared with Q1. WGCNA analysis based on 24 transcriptomes identified 16 co-expressed modules. Of these, the MEbrown module was highly correlated with Ye478 at low-temperature stages and catalase and superoxide dismutase activity, and the MEred, MEgreen, and MEblack modules were highly correlated with Ye478 across low-temperature stages, which revealed a significant association between LTGA and these modules. GO enrichment analysis showed the MEbrown and MEred modules mainly functioned in ribosome synthesis and cell cycle, respectively. In addition, we conducted quantitative trait loci (QTL) analysis based on a doubled haploid (DH) population constructed by Ye478 and Q1 and identified a major QTL explanting 20.6% of phenotype variance on chromosome 1. In this QTL interval, we found three, four, and three hub genes in the MEbrown, MEred, and MEgreen modules, of which two hub genes (Zm00001d031951, Zm00001d031953) related to glutathione metabolism and one hub gene (Zm00001d031617) related to oxidoreductase activity could be the candidate genes for LTGA. These biological functions and candidate genes will be helpful in understanding the regulatory mechanism of LTGA and the directional improvement of maize varieties for LTGA.

Full-sized realistic 3D printed models of liver and tumour anatomy: a useful tool for the clinical medicine education of beginning trainees.

BACKGROUND: Simulation-based medical education (SBME) and three-dimensional printed (3DP) models are increasingly used in continuing medical education and clinical training. However, our understanding of their role and value in improving trainees' understanding of the anatomical and surgical procedures associated with liver surgery remains limited. Furthermore, gender bias is also a potential factor in the evaluation of medical education. Therefore, the aim of this study was to evaluate the educational benefits trainees receive from the use of novel 3DP liver models while considering trainees' experience and gender. METHODS: Full-sized 3DP liver models were developed and printed using transparent material based on anonymous CT scans. We used printed 3D models and conventional 2D CT scans of the liver to investigate thirty trainees with various levels of experience and different genders in the context of both small group teaching and formative assessment. We adopted a mixed methods approach involving both questionnaires and focus groups to collect the views of different trainees and monitors to assess trainees' educational benefits and perceptions after progressing through different training programs. We used Objective Structured Clinical Examination (OSCE) and Likert scales to support thematic analysis of the responses to the questionnaires by trainees and monitors, respectively. Descriptive analyses were conducted using SPSS statistical software version 21.0. RESULTS: Overall, a 3DP model of the liver is of great significance for improving trainees' understanding of surgical procedures and cooperation during operation. After viewing the personalized full-sized 3DP liver model, all trainees at the various levels exhibited significant improvements in their understanding of the key points of surgery (p < 0.05), especially regarding the planned surgical procedure and key details of the surgical procedures. More importantly, the trainees exhibited higher levels of satisfaction and self-confidence during the operation regardless of gender. However, with regard to gender, the results showed that the improvement of male trainees after training with the 3DP liver model was more significant than that of female trainees in understanding and cooperation during the surgical procedure, while no such trend was found with regard to their understanding of the base knowledge. CONCLUSION: Trainees and monitors agreed that the use of 3DP liver models was acceptable. The improvement of the learning effect for practical skills and theoretical understanding after training with the 3DP liver models was significant. This study also indicated that training with personalized 3DP liver models can improve all trainees' presurgical understanding of liver tumours and surgery and males show more advantage in understanding and cooperation during the surgical procedure as compared to females. Full-sized realistic 3DP models of the liver are an effective auxiliary teaching tool for SBME teaching in Chinese continuing medical education.

Dissecting the Genetic Basis of Flowering Time and Height Related-Traits Using Two Doubled Haploid Populations in Maize.

In the field, maize flowering time and height traits are closely linked with yield, planting density, lodging resistance, and grain fill. To explore the genetic basis of flowering time and height traits in maize, we investigated six related traits, namely, days to anthesis (AD), days to silking (SD), the anthesis-silking interval (ASI), plant height (PH), ear height (EH), and the EH/PH ratio (ER) in two locations for two years based on two doubled haploid (DH) populations. Based on the two high-density genetic linkage maps, 12 and 22 quantitative trait loci (QTL) were identified, respectively, for flowering time and height-related traits. Of these, ten QTLs had overlapping confidence intervals between the two populations and were integrated into three consensus QTLs (qFT_YZ1a, qHT_YZ5a, and qHT_YZ7a). Of these, qFT_YZ1a, conferring flowering time, is located at 221.1-277.0 Mb on chromosome 1 and explained 10.0-12.5% of the AD and SD variation, and qHT_YZ5a, conferring height traits, is located at 147.4-217.3 Mb on chromosome 5 and explained 11.6-15.3% of the PH and EH variation. These consensus QTLs, in addition to the other repeatedly detected QTLs, provide useful information for further genetic studies and variety improvements in flowering time and height-related traits.

First-in-class humanized FSH blocking antibody targets bone and fat.

Blocking the action of FSH genetically or pharmacologically in mice reduces body fat, lowers serum cholesterol, and increases bone mass, making an anti-FSH agent a potential therapeutic for three global epidemics: obesity, osteoporosis, and hypercholesterolemia. Here, we report the generation, structure, and function of a first-in-class, fully humanized, epitope-specific FSH blocking antibody with a KD of 7 nM. Protein thermal shift, molecular dynamics, and fine mapping of the FSH-FSH receptor interface confirm stable binding of the Fab domain to two of five receptor-interacting residues of the FSHß subunit, which is sufficient to block its interaction with the FSH receptor. In doing so, the humanized antibody profoundly inhibited FSH action in cell-based assays, a prelude to further preclinical and clinical testing.

Genetic mapping of quantitative trait loci and a major locus for resistance to grey leaf spot in maize.

KEY MESSAGE: The genetic basis of GLS resistance was dissected using two DH populations sharing a common resistant parent. A major QTL repeatedly detected in multiple developmental stages and environments was fine mapped in a backcross population. Grey leaf spot (GLS), caused by Cercospora zeae-maydis or Cercospora zeina, is a highly destructive foliar disease worldwide. However, the mechanism of resistance against GLS is not well understood. To study the inheritance of this resistance, we developed two doubled haploid (DH) populations sharing a common resistant parent. The two DH populations were grown in two locations representing the typical maize-growing mountain area in Southwest China for 2 years. GLS disease severity was investigated 2 or 3 times until maturity in the 2 years, and the area under the disease progress curve was calculated. Two high-density linkage maps were constructed by genotyping-by-sequencing. A total of 22 quantitative trait loci (QTLs) were detected for GLS resistance, with most QTLs being repeatedly detected in different stages, locations and years. The confidence intervals of two major QTLs (qGLS_Y2-2 and qGLS_Z2-1) on chromosome 2 from the two DH populations overlapped with each other and were integrated into one consensus QTL (qGLS_YZ2-1). Using highly resistant and highly susceptible plants from a BC3 population, we fine mapped this genetic locus to a genomic region of 2.4 Mb. Using a panel of 255 inbred lines from breeding programmes, we detected associations between markers in the qGLS_YZ2-1 region and GLS resistance. The peak marker (ID-B1) will be very useful for marker-assisted breeding for GLS resistance.

Defining optimal cutoff value of MGMT promoter methylation by ROC analysis for clinical setting in glioblastoma patients.

Resistance to temozolomide (TMZ) chemotherapy poses a significant challenge in the treatment of glioblastoma (GBM). Hypermethylation in O6-methylguanine-DNA methyltransferase (MGMT) promoter is thought to play a critical role in this resistance. Pyrosequencing (PSQ) has been shown to be accurate and robust for MGMT promoter methylation testing. The unresolved issue is the determination of a cut-off value for dichotomization of quantitative MGMT PSQ results into "MGMT methylated" and "MGMT unmethylated" patient subgroups as a basis for further treatment decisions. In this study, receiver operating characteristic (ROC) curve analysis was used to identify an optimal cutoff of MGMT promoter methylation by testing mean percentage of methylation of 4 CpG islands (76-79) within MGMT exon 1. The area under the ROC (AUC) as well as the best cutoff to classify the methylation were calculated. Positive likelihood ratio (LR+) was chosen as a diagnostic parameter for defining an optimal cut-off. Meanwhile, we also analyzed whether mean percentage of methylation at the investigated CpG islands could be regarded as a marker for evaluating prognostication. ROC analysis showed that the optimal threshold was 12.5% (sensitivity: 60.87%; specificity: 76%) in response to the largest LR+ 2.54. 12.5% was established to distinguish MGMT promoter methylation, which was confirmed using validation set. According to the cutoff value, the MGMT promoter methylation was found in 58.3% of GBM. Mean methylation level of the investigated CpG sites strong correlated with overall survival (OS), which means GBM patients with a high level of methylation survived longer than those with low level of methylation(log-rank test, P = 0.017). In conclusion, ROC curve analysis enables the best cutoff for discriminating MGMT promoter methylation status. LR+ can be used as a key factor that evaluates cutoff. The promoter methylation level of MGMT by PSQ in GBM patients had prognostic value.

Association of mtDNA D-loop polymorphisms with risk of gastric cancer in Chinese population.

The aim of present study was to evaluate the association of common polymorphisms detected in mitochondrial DNA (mtDNA) D-loop region (mononucleotide repetitive D310, single nucleotide polymorphism (SNP) D16521) with susceptibility to gastric cancer (GC) in northwestern Chinese population. A total of 180 GC patients and 218 healthy controls were investigated by using PCR- denaturing high performance liquid chromatography (DHPLC) assay. Genotype and allele distributions and haplotype construction were analyzed in case-control study. We found D310 and D16521 heteroplasmy were significantly different between GC cases and controls (p < 0.05), and D16521 homoplasmy showed association with histological grade of GC (p < 0.05). Haplotype 7C/T, 8C/C and 9C/C had significant association with GC risk implied from analysis of D310 and D16521. Taken together, these findings suggested that mtDNA D-Loop polymorphisms and haplotypes may contribute to genetic susceptibility to GC in Chinese population.

Chromosome elimination and in vivo haploid production induced by Stock 6-derived inducer line in maize (Zea mays L.).

In vivo haploid production induced by inducer lines derived from Stock 6 is widely used in breeding program of maize (Zea mays L.), but the mechanisms behind have not yet been fully understood. In this study, average frequency of haploid induction in four inbred lines by Stock 6-derived inducer line HZI1 was above 10%. About 0.2% kernels from the cross Hua24 x HZI1 had mosaic endosperm showing yellow shrunken parts from Hua24 to normal parts with purple aleurone from HZI1. Individual lagged chromosomes and micronuclei were observed in mitotic cells of ovules pollinated by HZI1. Above 56.4% of the radicles from the kernels with purple aleurone and colorless embryos were mixoploid (2n = 9-21), and more than 45.22% cells were haploid cells (2n = 10) in three crosses. More than 62.5% of the radicles from the kernels with purple aleurone and purple embryos were mixoploid (2n = 9-21) having 54.27% cells with 2n = 20. SSR analysis showed that all haploids from the cross Hua24 x HZI1 shared the same genomic compositions as Hua24 except for plants Nos. 862 and 857 with some polymorphic DNA bands. The results revealed that chromosome elimination after fertilization caused the haploid production in maize.

Sphincter of Oddi dysfunction in hypercholesterolemic rabbits.

BACKGROUND AND OBJECTIVES: The mechanisms that trigger gallbladder evacuation dysfunction, the key risk factor for gallstone formation, have not yet been fully elucidated. The sphincter of Oddi (SO) plays important roles in the regulation of gallbladder evacuation and maintenance of normal hydraulic pressure of the biliary tract. The aim of our study was to investigate the effects of hypercholesterolemia on the motility function of SO and the underlying mechanisms of SO dysfunction (SOD). METHODS: Forty New Zealand white rabbits were divided randomly into the control group fed with standard chow and the experimental (Ch) group fed with a high-cholesterol diet for 8 weeks. Changes in the maximal gallbladder emptying rate, gallbladder evacuation with cholecystokinin-octapeptide (CCK-8) stimulation and SO functions of both groups were measured in vivo; B ultrasound examination was used for dynamic observation of peristaltic movements in vivo; SO pressure was measured using manometry; morphological characteristics were observed by electronic microscope; laser scanning confocal fluorescence microscopy was used to identify changes in [Ca]i and Ca oscillation in primary SO smooth muscle cells (SMCs). RESULTS: Gallbladder cholestasis was observed during early stages of gallstone formation in Ch rabbits. CCK-8 could not improve the gallbladder cholestatic state in Ch group. Passive dilation of SO significantly improved the cholestatic state in Ch rabbits (P<0.05), although the maximal gallbladder emptying rate was still lower than that of the control group. Manometry data indicted a significant increase in the base pressure of the SO low-pressure ampulla segment and high-pressure segment (P<0.05) in Ch group. laser scanning confocal fluorescence microscopy assay data indicated that [Ca]i in SO cells of Ch group significantly increased and were in a state of overload (P<0.05); Ca oscillation signals in SO cells of Ch group were also abnormal. CONCLUSION: Hypercholesterolemia initially induced SOD, leading to increased gallbladder evacuation resistance and cholestasis. We suggested that [Ca]i overload and/or Ca oscillation abnormality potentially play important roles in the pathogenesis of SOD.

Molecular cloning, tissue distribution and bioinformatics analyses of the rabbit BK channel beta1 subunit gene.

Large-conductance, voltage-dependent and Ca(2+)-sensitive K(+) (BK) channels are composed of pore-forming alpha subunits and the modulatory beta subunits. In smooth muscle, the modulatory beta1 subunits are vital in rendering BK channels function as an important regulator of smooth muscle tone and excitability. In this study, we cloned and characterized the BK beta1 subunit gene from rabbits (New Zealand white) and observed its tissue distribution pattern. The full-length cDNA of the BK beta1 subunit, amplified by 5'-RACE and 3'-RACE, is 1,437 bp in nucleotide containing a 447 bp 5'-UTR, a 385 bp 3'-UTR and a 576 bp open reading frame (ORF) which encodes a peptide of 191 amino acids. Sequence analyses showed that the rabbit BK beta1 subunit cDNA is 90, 84 and 82% homologous with that of human, mouse and rat respectively. The similarity is 86, 83, and 83% at the deduced amino acids level with human, mouse and rat beta1 subunit gene, respectively. Northern blots indicated that the rabbit BK beta1 subunit gene is highly expressed in sphincter of Oddi (SO) and aortal smooth muscle tissues, whereas with relatively lower level of expression in heart and skeletal muscle tissues and with no expression found in tissues of liver, lung, kidney and brain. Bioinformatics analyses indicated that the encoded protein is a membrane protein with two transmembrane helical regions containing four functional domains, one possible PKA phosphorylation site (T14) at the N-terminal and two N-glycosylation sites (N80 and N142) at the extracellular loop. For the first time, we identified and characterized the full-length cDNA sequence of the rabbit BK channel beta1 subunit gene, which will set the basis for further investigation in the transcriptional regulation of this gene.

Down regulated expression of the beta1 subunit of the big-conductance Ca2+ sensitive K+ channel in sphincter of Oddi cells from rabbits fed with a high cholesterol diet.

Hypercholesterolemia, which is closely related to gallbladder bile stasis, can cause sphincter of Oddi dysfunction (SOD) by increasing the tension of sphincter of Oddi (SO). Intracellular calcium ion concentration ([Ca(2+)](i)) could influence the tension of SO. The beta1 subunit of the big-conductance Ca(2+) sensitive K(+) channel (BK(Ca)) can enhance the sensitivity of the BK(Ca) channel to [Ca(2+)](i). Absence and decline of the BKCa channel subunit beta1 could lead to many diseases. However, the relationship between hypercholesterolemia and the expression of beta1 subunit is not well understood. In this study, we successfully expressed and purified the rabbit BK(Ca) beta1 subunit protein and prepared its polyclonal antibody. The specificity of the prepared antibody was determined by Western blotting. A SOD rabbit model induced by a high cholesterol diet was established and the expression of the beta1 subunit of SO was determined by immunohistochemical staining and western blotting. Compared with the controls, our results demonstrated that hypercholesterolemia could decrease the expression of the beta1 subunit in the SO cells from rabbits. This indicates that lower expression of BKCa channel beta1 subunit might induce SOD.

[The influence on the BK channel beta1 subunit expression of SO caused by high cholesterol in rabbits].

AIM: To investigate the effects on protein expression of big-conductance Ca(2+)-sensitive K(+) channel (BKca) beta1 subunit caused by high cholesterol in Rabbit Oddi's sphincter (SO) cells. METHODS: The rat-anti-rabbit polyclonal antiserum against beta1 subunits of BKca channel of SO cell was prepared. And the protein expression of BKca channel beta1 subunit of SO tissue was detected by semi-quantitative immunohistochemical staining. RESULTS: The protein expression of BKca channels beta1 subunit of SO tissue in HC group was reduced, and there's statistically significant difference between the HC group and the control group. CONCLUSION: High cholesterol can reduce the protein expression of BK Channel's beta1 subunit in Rabbits' SO which suggests high cholesterol can affect the function of BKca channel.

[Preparation of polyclonal antiserum against beta subunit of rabbit BK channel].

AIM: To prepare polyclonal antiserum against beta subunit of rabbit BK channel in mice. METHODS: Gene encoding the intracellular fragment of rabbit BK channel's beta subunit was amplified by RT-PCR. The GST-beta fusion protein was expressed in E. coli. The fusion protein from PAGE gel was used to immunize BALB/c mice and prepare polyclonal antiserum. The specificity of antiserum was identified by ELISA and Western blot. RESULTS: A unique band about 300 bp was amplified by RT-PCR and was verified to be BK channel beta subunit by DNA sequencing. The SDS-PAGE analysis showed that the M(r) of the fusion protein was about 37,000. The purity of GST-beta fusion protein was over 95%. The polyclonal antiserum against GST-beta fusion protein could recognize both GST-beta fusion protein and beta protein in rabbit tissues. The highest titer of the antiserum was about 1:128,000, as shown by Western blot and ELISA, respectively. CONCLUSION: The gene encoding the intracellular fragment of rabbit BK channel's beta subunit has been cloned. The polyclonal antiserum against beta subunit of rabbit BK channel with high titer and specificity has been prepared successfully.