More Efficient Auto-Segmentation Framework Using Patient-Specific Information for CBCT-Guided Adaptive Radiotherapy

PURPOSE/OBJECTIVE(S)

Accurate and fast delineation of regions of interest (ROIs) is critical for online adaptive radiotherapy (ART). Due to the noise, artifacts and low contrast of soft tissue on cone-beam computed tomography (CBCT), the CBCT-guided ART (CBCTgART) needs an effective tool to assist the segmentation of ROIs. This study aims to develop an efficient auto-segmentation workflow with personalized modeling for CBCTgART based on deep learning (DL) model.

MATERIALS/METHODS

Five hundred fifty-two patients with nasopharyngeal carcinoma were enrolled in this study. At the beginning, a cycle consistent adversarial network (CycleGAN) model was trained on 172 patients' CBCT/CT paired images to improve the image quality of CBCT to CT level. A generalized segmentation model was trained using the planning CT (pCT) and contour data from 530 patients. For personalized modeling, the generalized segmentation model was fine-tuned on the specific patient's pCT and contour to get the personalized model. When CBCT images were available, the trained CycleGAN model transformed the CBCT to synthetic pCT. Then the personalized auto-segmentation model generated the contour of ROIs on the synthetic pCT. We randomly selected 22 patients for model test. The proposed method (DL-personalized) was compared with other DL methods based on the same architecture of network: regular deep learning method (DL-regular), which was trained on the CBCT and corresponding contours, and generalized model in our framework (DL-generalized). So, 22 personalized, 1 generalized and 1 regular DL models were tested. The paired t-test was performed to test the significance for mean dice similarity coefficient (DSC), mean distance to agreement (MDA), and Hausdorff distance (HD) between the alternative and proposed methods.

RESULTS

Two ROIs were included: the clinical target volume (CTV) and nasopharynx gross tumor volume (GTVnx). The proposed DL-personalized model achieved better results compared with others as shown in the table. The accuracy and robustness of our proposed framework was reliable. All of p values were under 0.01, which indicated the statistically significant difference.

CONCLUSION

The proposed framework utilizing patient-specific information can improve the segmentation accuracy of ART. This method has potential to be integrated into the ART workflow to improve efficiency.

Dose DIBH Really Reduce the Subclinical Cardiac Acute Injury? Analysis of Clinical Real World from Our Institute

PURPOSE/OBJECTIVE(S)

The study is aim to investigate whether Deep-inspirational breath-hold (DIBH), compared with free breathing (FB), could provide a short-term cardiac benefit in patients with early left breast cancer after breast-conserving surgery combined whole breast radiotherapy.

MATERIALS/METHODS

A total of 78 patients with early stage left breast cancer treated with radiotherapy between 2021-2022 after breast-conserving surgery were enrolled. Among them, 32 cases were treated with DIBH technique and 46 cases were treated with free breathing. Patients with previous cardiac disease such as coronary artery disease were excluded. We performed myocardial enzymes, ECG, and ECHO in all patients within 2 weeks before, during, and 6 months after radiotherapy. The results of the two groups were compared using nonparametric tests and chi-square tests, and P < 0.05 indicated statistical significance. Where subclinical acute cardiac injury was defined as new above-normal myocardial enzymes and/or electrocardiographic ST-T or T-wave changes and/or ECHO abnormalities after the start of radiotherapy.

RESULTS

The median follow-up of patients was 6 months and the mean age of patients was 52.3 years for FB and 44.9 years for DIBH. There were no significant differences in staging, molecular subtype, chemotherapy and endocrine therapy history. The proportion of subclinical acute cardiac injury was smaller in the DIBH group compared to the FB group (DIBH = 31/46 and FB = 28/32, p = 0.042). The most sensitive of the subclinical acute cardiac injury events were detected by myocardial enzymes rising, with cTnI (p = 0.034) and NT-proBNP (p = 0.023) appearing significantly lower in the DIBH patients during radiotherapy. The difference of cTnI between 2 groups at 6 months after radiotherapy became non-significant. In contrast, CK-MB was higher in DIBH compared with FB only 6 months after radiotherapy (p = 0.006). The differences in ECG and ECHO were not significant between the two groups.

CONCLUSION

After breast-conserving surgery combined with radiotherapy for left early breast cancer, DIBH compared to FB reduces the proportion of acute subclinical cardiac injury that occurs with the most sensitive changes in myocardial enzymes. Subsequent studies will explore the relationship between the short-term subclinical injury and irradiated dose, as well as long term cardiac injury.

A multiline fitting method for measuring ethylene concentration based on WMS-2f/1f

Coal spontaneous combustion risk assessment is a global technical challenge for the sustainable development of deep mining technology, and C2H4 is a key indicator for early warning of coal spontaneous combustion. Tunable diode laser absorption spectroscopy (TDLAS) has the advantages of high selectivity, high sensitivity, high accuracy and real-time on-line measurement, and it can detect multiple gases simultaneously, so it has significant advantages in the accurate detection of coal spontaneous combustion indicator gases. To address the problem of cross-interference between the near-infrared absorption lines of CH4 and C2H4, which are the indicator gases of spontaneous combustion in coal, a multi-line fitting method was proposed in this study to calibrate the concentration of C2H4. The high-precision Environics2000 automatic standard gas dispenser from the United States, which has a built-in CPU computer control and data control and processing system, was used. Its gas concentration accuracy: ± 1.0%, gas flow accuracy: ± 1.0%, gas repeatability accuracy: ± 1.0%, flow linearity accuracy: ± 0.5%, and inlet operating pressure: minimum 10 psig (0.67 bar) ~ 75 psig (5.04 bar). The measured and simulated WMS-2f/1f signals were multilinearly fitted using a multilinear fitting algorithm and wavelength modulation spectroscopy (WMS), and the measurement of C2H4 concentration was achieved based on the extracted spectral line information. The results show that the maximum relative error of C2H4 concentration measurement is 2.40%, which is 54% lower than that of the conventional 2f peak measurement method, thus demonstrating the effectiveness of the multilinear fitting algorithm in the inversion of C2H4 concentration under the interference of absorption lines. In addition, this study has far-reaching implications for the application of TDLAS technology in the accurate detection of coal spontaneous combustion indicator gases.

Exploring the mechanism of Semen Strychni in treating amyotrophic lateral sclerosis based on network pharmacology

Semen Strychni (SS), known as an agonist of central nervous system, is a traditional herb widely used in treating amyotrophic lateral sclerosis (ALS) in small doses to relieve muscle weakness and improve muscle strength. However, the potential mechanisms and the main components of SS in treating ALS remain unclear. To explore the underlying mechanism of SS in treating ALS based on network pharmacology and molecular docking. The active components of SS were obtained using TCMSP, Herb, ETCM, and BATMAN-TCM. The targets of SS were gained from PharmMapper. The targets of ALS were searched on Genecards, Drugbank, DisGeNET, OMIM, TTD and GEO database. After obtaining the coincidence targets, we submitted them to the STRING database to build a protein-protein interaction network. Gene ontology and Kyoto Encyclopedia of Genes and Genomes enrichment analysis were performed subsequently. The active components and targets were further investigated using molecular docking technology. 395 targets of SS and 1925 targets of ALS were obtained with 125 common targets. The protein-protein interaction analysis indicated that SRC, AKT1, MAPK1, EGFR, and HSP90AA1 received the higher degree value and were considered the central genes. The Ras, PI3K-Akt, and MAPK signaling pathway could be involved in the treatment of ALS. Brucine-N-oxide obtained the lowest binding energy in molecular docking. This study explored the mechanism of SS in the treatment of ALS and provides a new perspective for future study. However, further experimental studies are needed to validate the therapeutic effect.

Ginsenoside-Rg1 combined with a conditioned medium from induced neuron-like hUCMSCs alleviated the apoptosis in a cell model of ALS through regulating the NF-κB/Bcl-2 pathway

Amyotrophic lateral sclerosis (ALS) is a fatal neurodegenerative disease affecting both upper and lower motor neurons in the brain and spinal cord. One important aspect of ALS pathogenesis is superoxide dismutase 1 (SOD1) mutant-mediated mitochondrial toxicity, leading to apoptosis in neurons. This study aimed to evaluate the neural protective synergistic effects of ginsenosides Rg1 (G-Rg1) and conditioned medium (CM) on a mutational SOD1 cell model, and to explore the underlying mechanisms. We found that the contents of nerve growth factor, glial cell line-derived neurotrophic factor, and brain-derived neurotrophic factor significantly increased in CM after human umbilical cord mesenchymal stem cells (hUCMSCs) were exposed to neuron differentiation reagents for seven days. CM or G-Rg1 decreased the apoptotic rate of SOD1^G93A-NSC34 cells to a certain extent, but their combination brought about the least apoptosis, compared with CM or G-Rg1 alone. Further research showed that the anti-apoptotic protein Bcl-2 was upregulated in all the treatment groups. Proteins associated with mitochondrial apoptotic pathways, such as Bax, caspase 9 (Cas-9), and cytochrome c (Cyt c), were downregulated. Furthermore, CM or G-Rg1 also inhibited the activation of the nuclear factor-kappa B (NF-κB) signaling pathway by reducing the phosphorylation of p65 and IκBα. CM/G-Rg1 or their combination also reduced the apoptotic rate induced by betulinic acid (BetA), an agonist of the NF-κB signaling pathway. In summary, the combination of CM and G-Rg1 effectively reduced the apoptosis of SOD1^G93A-NSC34 cells through suppressing the NF-κB/Bcl-2 signaling pathway (Fig. 1 is a graphical representation of the abstract).

NRF2 activation suppresses motor neuron ferroptosis induced by the SOD1G93A mutation and exerts neuroprotection in amyotrophic lateral sclerosis

The progressive neurodegenerative disease amyotrophic lateral sclerosis (ALS) is caused by a decline in motor neuron function, resulting in worsened motor impairments, malnutrition, respiratory failure and mortality, and there is a lack of effective clinical treatments. The exact mechanism of motor neuronal degeneration remains unclear. Previously, we reported that ferroptosis, which is characterized by the accumulation of lipid peroxide and glutathione depletion in an iron-dependent manner, contributed to motor neuronal death in ALS cell models with the hSOD1^G93A (human Cu/Zn-superoxide dismutase) gene mutation. In this study, we further explored the role of ferroptosis in motor neurons and its regulation in mutant hSOD1^G93A cell and mouse models. Our results showed that ferroptosis was activated in hSOD1^G93A NSC-34 cells and mouse models, which was accompanied by decreased nuclear retention of nuclear factor erythroid 2-related factor 2 (NRF2) and downregulation of solute carrier family 7 member 11 (SLC7A11) and glutathione peroxidase 4 (GPX4) levels. Moreover, RTA-408, an NRF2 activator, inhibited ferroptosis in hSOD1^G93A NSC-34 cells by upregulating the protein expression of SLC7A11 and GPX4. Moreover, hSOD1^G93A mice treated with RTA-408 showed obvious improvements in body weight and motor function. Our study demonstrated that ferroptosis contributed to the toxicity of motor neurons and that activating NRF2 could alleviate neuronal degeneration in ALS with the hSOD1^G93A mutation.

Herbicide residues in soil decrease microbe-mediated plant protection

The residues of glyphosate are found to remain in soils longer than previously reported, affecting rhizosphere microbes. This may adversely affect crop and other non-target plants because the plant's resilience and resistance largely rely on plant-associated microbes. Ubiquitous glyphosate residues in soil and how they impact mutualistic microbes inhabiting the aboveground plant parts are largely unexplored. We studied the effects of herbicide residues in soil on Epichloë sp., which are common endophytic symbionts inhabiting aerial parts of cool-season grasses. In this symbiosis, the obligate symbiont subsists entirely on its host plant, and in exchange, it provides alkaloids conferring resistance to herbivores for the host grass that invests little in its own chemical defence. We first show decreased growth of Epichloë endophytes in vitro when directly exposed to two concentrations of glyphosate or glyphosate-based herbicides. Second, we provide evidence for a reduction of Epichloë-derived, insect-toxic loline alkaloids in endophyte-symbiotic meadow fescue (F. pratensis) plants growing in soil with a glyphosate history. Plants were grown for 2 years in an open field site, and natural herbivore infestation was correlated with the glyphosate-mediated reduction of loline alkaloid concentrations. Our findings indicate that herbicides residing in soil not only affect rhizosphere microbiota but also aerial plant endophyte functionality, which emphasizes the destructive effects of glyphosate on plant symbiotic microbes, here with cascading effects on plant-pest insect interactions.

[Outcome of surgical repair for aortic coarctation with coexisting descending aortic aneurysm in adult patients]

Objective: To evaluate the efficacy of surgical treatment of aortic coarctation combined with descending aortic aneurysm in adult patients. Methods: This is a retrospective cohort study. Adult patients with aortic coarctation who were hospitalized in Beijing Anzhen Hospital from January 2015 to April 2019 were enrolled. The aortic coarctation was diagnosed by aortic CT angiography, and the included patients were divided into the combined descending aortic aneurysm group and the uncomplicated descending aortic aneurysm group based on descending aortic diameter. General clinical data and surgery-related data were collected from the included patients, and death and complications were recorded at 30 days after surgery, and upper limb systolic blood pressure was measured in all patients at discharge. Patients were followed up after discharge by outpatient visit or telephone call for their survival and the occurrence of repeat interventions and adverse events, which included death, cerebrovascular events, transient ischemic attack, myocardial infarction, hypertension, postoperative restenosis, and other cardiovascular-related interventions. Results: A total of 107 patients with aortic coarctation aged (34.1±15.2) years were included, and 68 (63.6%) were males. There were 16 cases in the combined descending aortic aneurysm group and 91 cases in the uncomplicated descending aortic aneurysm group. In the combined descending aortic aneurysm group, 6 cases (6/16) underwent artificial vessel bypass, 4 cases (4/16) underwent thoracic aortic artificial vessel replacement, 4 cases (4/16) underwent aortic arch replacement+elephant trunk procedure, and 2 cases (2/16) underwent thoracic endovascular aneurysm repair. There was no statistically significant difference between the two groups in the choice of surgical approach (all P>0.05). In the combined descending aortic aneurysm group at 30 days after surgery, one case underwent re-thoracotomy surgery, one case developed incomplete paraplegia of the lower extremity, and one case died; and the differences in the incidence of endpoint events at 30 days after surgery were similar between the two groups (P>0.05). Systolic blood pressure in the upper extremity at discharge was significantly lower in both groups compared with the preoperative period (in the combined descending aortic aneurysm group: (127.3±16.3) mmHg vs. (140.9±16.3) mmHg, P=0.030, 1 mmHg=0.133 kPa; in the uncomplicated descending aortic aneurysm group: (120.7±13.2) mmHg vs. (151.8±26.3) mmHg, P=0.001). The follow-up time was 3.5 (3.1, 4.4) years. There were no new deaths in the combined descending aortic aneurysm group, no transient ischemic attack, myocardial infarction or re-thoracotomy surgery, and one patient (1/15) suffered cerebral infarction and 10 patients (10/15) were diagnosed with hypertension. The differences in the occurrence of endpoint events during postoperative follow-up were similar between the two groups (P>0.05). Conclusion: In experienced centers, long-term prognosis of patients with aortic coarctation combined with descending aortic aneurysm is satisfactory post surgical intervention.

Effect of ripening temperature on the chemical composition of lingonberries (Vaccinium vitis-idaea L.) of northern and southern origin

Lingonberries (Vaccinium vitis-idaea L.) from two locations, northern (69°N, 18°E) and southern (59°N, 10°E) Norway, were grown under controlled conditions in a phytotron at two temperatures (9 and 15 °C) to study the effects of the ripening temperature and origin on the chemical composition of the berries. The concentrations of phenolic compounds, sugars, and organic acids as well as the profile of volatile organic compounds (VOCs) were determined using chromatographic and mass spectrometric methods. Five anthocyanins, eleven flavonols, eight cinnamic acid derivatives, three flavan-3-ols, three sugars, three organic acids, and 77 VOCs were identified, of which 40 VOCs had not previously been reported in lingonberries. Berries from both locations, were found to have higher contents of anthocyanins and cinnamic acid derivatives when ripened at lower temperature (9 °C), compared to the higher temperature (15 °C). Lingonberries of northern origin had a different VOC profile and higher contents of anthocyanins and organic acids than berries originating from the south. Lingonberries from the northern location also had higher proportions of cyanidin-3-O-glucoside and cyanidin-3-O-arabinoside than lingonberries from the southern location. The results show that the composition of lingonberries is influenced by both the environment and the origin of the plants, with phenolic compounds mainly influenced by the growth temperature and VOCs mainly influenced by plant origin.

Clinical Relevance of Plaque Distribution for Basilar Artery Stenosis

BACKGROUND AND PURPOSE

There is no clear association between plaque distribution and postoperative complications in patients with basilar artery atherosclerotic stenosis. The aim of this study was to determine whether plaque distribution and postoperative complications after endovascular treatment for basilar artery stenosis are related.

MATERIALS AND METHODS

Our study enrolled patients with severe basilar artery stenosis who were scanned with high-resolution MR imaging and followed by DSA before the intervention. According to high-resolution MR imaging, plaques can be classified as ventral, lateral, dorsal, or involved in 2 quadrants. Plaques affecting the proximal, distal, or junctional segments of the basilar artery were classified according to DSA. An experienced independent team assessed ischemic events after the intervention using MR imaging. Further analysis was conducted to determine the relationship between plaque distribution and postoperative complications.

RESULTS

A total of 140 eligible patients were included in the study, with a postoperative complication rate of 11.4%. These patients were an average age of 61.9 (SD, 7.7) years. Dorsal wall plaques accounted for 34.3% of all plaques, and plaques distal to the anterior-inferior cerebellar artery accounted for 60.7%. Postoperative complications of endovascular treatment were associated with plaques located at the lateral wall (OR = 4.00; 95% CI, 1.21-13.23; P = .023), junctional segment (OR = 8.75; 95% CI, 1.16-66.22; P = .036), and plaque burden (OR = 1.03; 95% CI, 1.01-1.06; P = .042).

CONCLUSIONS

Plaques with a large burden located at the junctional segment and lateral wall of the basilar artery may increase the likelihood of postoperative complications following endovascular therapy. A larger sample size is needed for future studies.

Uncovering and engineering the mechanical properties of the adhesion GPCR ADGRG1 GAIN domain

Key cellular functions depend on the transduction of extracellular mechanical signals by specialized membrane receptors including adhesion G-protein coupled receptors (aGPCRs). While recently solved structures support aGPCR activation through shedding of the extracellular GAIN domain, the molecular mechanisms underpinning receptor mechanosensing remain poorly understood. When probed using single-molecule atomic force spectroscopy and molecular simulations, ADGRG1 GAIN dissociated from its tethered agonist at forces significantly higher than other reported signaling mechanoreceptors. Strong mechanical resistance was achieved through specific structural deformations and force propagation pathways under mechanical load. ADGRG1 GAIN variants computationally designed to lock the alpha and beta subdomains and rewire mechanically-induced structural deformations were found to modulate the GPS-Stachel rupture forces. Our study provides unprecedented insights into the molecular underpinnings of GAIN mechanical stability and paves the way for engineering mechanosensors, better understanding aGPCR function, and informing drug-discovery efforts targeting this important receptor class.

[Current practices in the harmonisation of autoantibodies test]

Standardisation and harmonisation of the detection of autoantibodies is important for the clinical application of autoantibodies. However, achieving complete standardisation is difficult and involves several challenges due to the complexity and particularity of autoantibody detection. Harmonisation is feasible and valued, but it involves all aspects and processes of autoantibody detection. Based on the consensus and practice of the clinical application of autoantibody detection in recent years, we discuss harmonisation in this review.

Gut microbiota and risk of five common cancers: A univariable and multivariable Mendelian randomization study

BACKGROUND

Previous studies have linked gut microbiota with cancer etiology, but the associations for specific gut microbiota are causal or owing to bias remain to be elucidated.

METHODS

We performed a two-sample Mendelian randomization (MR) analysis to assess the causal effect of gut microbiota on cancer risk. Five common cancers, including breast, endometrial, lung, ovarian, and prostate cancer as well as their subtypes (sample sizes ranging from 27,209 to 228,951) were included as the outcomes. Genetic information for gut microbiota was obtained from a genome-wide association study (GWAS) comprising 18,340 participants. In univariable MR (UVMR) analysis, the inverse variance weighted (IVW) method was conducted as the primary method, with the robust adjusted profile scores, weighted median, and MR Egger used as supplementary methods for causal inference. Sensitivity analyses including the Cochran Q test, Egger intercept test, and leave-one-out analysis were performed to verify the robustness of the MR results. Multivariable MR (MVMR) was performed to evaluate the direct causal effects of gut microbiota on the risk of cancers.

RESULTS

UVMR detected a higher abundance of genus Sellimonas predicted a higher risk of estrogen receptor-positive breast cancer (OR = 1.09, 95% CI 1.05-1.14, p = 2.01 × 10^-5 ), and a higher abundance of class Alphaproteobacteria was associated with a lower risk of prostate cancer (OR = 0.84, 95% CI 0.75-0.93, p = 1.11 × 10^-3 ). Sensitivity analysis found little evidence of bias in the current study. MVMR further confirmed that genus Sellimonas exerted a direct effect on breast cancer, while the effect of class Alphaproteobacteria on prostate cancer was driven by the common risk factors of prostate cancer.

CONCLUSION

Our study implies the involvement of gut microbiota in cancer development, which provides a novel potential target for cancer screening and prevention, and might have an implication for future functional analysis.

[Changes of serum calcium and bone turnover markers in maintenance hemodialysis patients treated with denosumab]

A total of 20 patients (8 males and 12 females) aged (64.8±13.9) years who underwent regular hemodialysis and had bone loss or osteoporosis in Peking University People's Hospital from July to December 2021 were recruited. Sixty milligrams of denosumab were given subcutaneously. The average serum calcium decreased by 0.31 mmol/L and 40% (8/20) of the patients had hypocalcemia one month after treatment. The markers of bone turnover began to decrease 3 days after treatment, and continued to decrease until 5 months after denosumab medication. Multivariate logistic regression analysis failed to detect any independent risk factor for hypocalcemia (R²=0.516, P=0.021). Therefore, denosumab can significantly inhibit bone turnover in hemodialysis patients, however, patients with denosumab medication are more prone to hypocalcemia.

[Exploration of pathogenesis and clinical diagnosis and treatment for portal vein thrombosis formation]

Portal vein thrombosis (PVT) refers to the thrombosis of the main portal vein and/or branches, with or without mesenteric vein and splenic vein thrombosis, and is the most common cause of extrahepatic portal vein obstruction. It is latent under chronic conditions and is often detected accidentally during physical examinations or liver cancer screenings. Notably, the understanding of PVT management is still limited, both at home and abroad. Therefore, the purpose of this article is to provide a reference for clinical diagnosis and treatment of PVT formation by summarizing the basis and standards for clinical diagnosis and treatment based on representative research with a large sample size and new perspectives with recent guidelines and consensus.

[Evaluation of diagnostic efficacy of digital liquid chip method for detection of specific antineutrophil cytoplasmic antibodies]

To explore the clinical diagnostic efficacy of antineutrophil cytoplasmic antibody associated vasculitis (AAV) by comparing the consistency and coincidence rate of serum anti-myeloperoxidase (MPO) antibody and anti-protease 3 (PR3) antibody detected by digital liquid chip method (DLCM) and enzyme-linked immunosorbent assay (ELISA). To provide reference for the selection of detection methods of anti-MPO antibody and anti-PR3 antibody in clinical laboratory. This study is a cross-sectional study, a total of 307 cases of antineutrophil cytoplasmic antibodies were detected in the Department of Clinical Immunology, West China Hospital of Sichuan University from January to March 2021. The serum samples and related clinical information were collected. At the same time, the levels of anti-MPO antibody and anti-PR3 antibody in serum samples were detected by ELISA and DLCM, indirect immunofluorescence (IIF) was used to re-test the differential samples between the two methods. SPSS 26.0 was used to analyze the test results, Cohen's kappa coefficient analysis was used to compare the consistency of the two methods, and paired chi-square test was used to compare the sensitivity and specificity of the two methods to AAV. The results showed that the positive cases of anti-MPO antibody detected by ELISA and DLCM were 63 and 44, and the negative cases were 244 and 263; the positive cases of anti-PR3 antibody detected by ELISA and DLCM were 34 and 28, and the negative cases were 273 and 279. The results of anti-MPO antibody and anti-PR3 antibody detected by the two methods had good consistency and coincidence rate, in which the total coincidence rate of anti-MPO antibody was 92.51%, the positive coincidence rate was 66.67%, and the negative coincidence rate was 99.18%. The results of consistency analysis showed that kappa=0.741 had well consistency. The total coincidence rate of anti-PR3 antibody is 96.74%, the positive coincidence rate is 76.47%, and the negative coincidence rate is 99.27%. The consistency analysis results show that kappa=0.821 had strong consistency. The results of IIF re-test of differential samples showed that the coincidence rate between DLCM and IIF was higher. The results of comparative analysis of anti-MPO antibody and anti-PR3 antibody showed that the specificity of DLCM was better than that of ELISA, and its sensitivity was lower than that of ELISA. In conclusion, the results of anti-MPO antibody and anti-PR3 antibody detected by DLCM were consistent with those of ELISA. In the combined detection of anti-MPO antibody and anti-PR3 antibody, the specificity of DLCM is better than that of ELISA.

A few-shot learning-based eye diseases screening method

OBJECTIVE

This study aims to construct a brand-new ophthalmic disease screening task and establish a practically valuable ophthalmic disease screening model in the case of insufficient data.

MATERIALS AND METHODS

The main methods are as follows: firstly, we mixed data from different sources (these data may come from different cameras, including different fundus diseases) to get a new dataset. Based on this dataset, we conducted subsequent experiments on fundus multi-disease screening. However, in the past public datasets, each dataset often only corresponded to the screening diagnosis of one disease. Secondly, we proposed a method to simulate the characteristics of different fundus cameras by using a method based on style transfer, and to augment the training data, so that the model could learn the features of ophthalmic diseases in a more comprehensive way. Finally, a robust disease screening model based on few-shot learning was constructed on the combined dataset, and compared with benchmark algorithms.

RESULTS

We focused on the study of eye disease screening methods based on the metric-based few-shot learning model, data augmentation methods, and focus on key technologies such as data augmentation based on style transfer. Experiments have shown that our method can significantly improve the generalization ability of the disease screening model.

CONCLUSIONS

By introducing few-shot learning theory and data augmentation based on style transfer into ophthalmic disease screening, the generalization ability of the model is greatly improved, and it has certain practical value.

Accelerated Fatigue Model for Predicting Composite Restoration Failure

An empirical method is proposed to predict the clinical performance of resin composite dental restorations by using laboratory data derived from simple specimens subjected to chemical degradation and accelerated cyclic fatigue. Three resin composites were used to fill dentin disks (2-mm inner diameter, 5-mm outer diameter, and 2 mm thick) made from bovine incisor roots. The specimens (n = 30 per group) were aged with different durations of a low-pH challenge (0, 24, and 48 h under pH 4.5) before being subjected to diametral compression with either a monotonically increasing load (fast fracture) or a cyclic load with a continuously increasing amplitude (accelerated fatigue). The data from 1 material were used to establish the relationship between laboratory time (number of cycles) and clinical time to failure (years) via the respective survival probability curves. The temporal relationship was then used to predict the clinical rates of failure for restorations made of the other 2 materials, and the predictions were compared with the clinical data to assess their accuracy. Although there were significant differences in the fast fracture strength among the groups of materials or durations of chemical challenge, fatigue testing was much better at separating the groups. Linear relationships were found between the laboratory and clinical times to failure for the first material (R2 = 0.90, 0.90, and 0.62 for the 0-, 24-, and 48-h low-pH groups, respectively). The clinical life of restorations made of the other 2 materials was best predicted with data from the 48-h low-pH groups. In conclusion, an accelerated fatigue model was successfully calibrated and applied to predict the clinical failure of resin composite restorations, and the predictions based on data obtained from chemically aged specimens provided the best agreement with clinical data.