Comprehensive multimodal deep learning survival prediction enabled by a transformer architecture: A multicenter study in glioblastoma.

Background: This research aims to improve glioblastoma survival prediction by integrating MR images, clinical, and molecular-pathologic data in a transformer-based deep learning model, addressing data heterogeneity and performance generalizability. Methods: We propose and evaluate a transformer-based nonlinear and nonproportional survival prediction model. The model employs self-supervised learning techniques to effectively encode the high-dimensional MRI input for integration with nonimaging data using cross-attention. To demonstrate model generalizability, the model is assessed with the time-dependent concordance index (Cdt) in 2 training setups using 3 independent public test sets: UPenn-GBM, UCSF-PDGM, and Rio Hortega University Hospital (RHUH)-GBM, each comprising 378, 366, and 36 cases, respectively. Results: The proposed transformer model achieved a promising performance for imaging as well as nonimaging data, effectively integrating both modalities for enhanced performance (UCSF-PDGM test-set, imaging Cdt 0.578, multimodal Cdt 0.672) while outperforming state-of-the-art late-fusion 3D-CNN-based models. Consistent performance was observed across the 3 independent multicenter test sets with Cdt values of 0.707 (UPenn-GBM, internal test set), 0.672 (UCSF-PDGM, first external test set), and 0.618 (RHUH-GBM, second external test set). The model achieved significant discrimination between patients with favorable and unfavorable survival for all 3 datasets (log-rank P 1.9 × 10-8, 9.7 × 10-3, and 1.2 × 10-2). Comparable results were obtained in the second setup using UCSF-PDGM for training/internal testing and UPenn-GBM and RHUH-GBM for external testing (Cdt 0.670, 0.638, and 0.621). Conclusions: The proposed transformer-based survival prediction model integrates complementary information from diverse input modalities, contributing to improved glioblastoma survival prediction compared to state-of-the-art methods. Consistent performance was observed across institutions supporting model generalizability.

Principles of artificial intelligence in radiooncology.

PURPOSE: In the rapidly expanding field of artificial intelligence (AI) there is a wealth of literature detailing the myriad applications of AI, particularly in the realm of deep learning. However, a review that elucidates the technical principles of deep learning as relevant to radiation oncology in an easily understandable manner is still notably lacking. This paper aims to fill this gap by providing a comprehensive guide to the principles of deep learning that is specifically tailored toward radiation oncology. METHODS: In light of the extensive variety of AI methodologies, this review selectively concentrates on the specific domain of deep learning. It emphasizes the principal categories of deep learning models and delineates the methodologies for training these models effectively. RESULTS: This review initially delineates the distinctions between AI and deep learning as well as between supervised and unsupervised learning. Subsequently, it elucidates the fundamental principles of major deep learning models, encompassing multilayer perceptrons (MLPs), convolutional neural networks (CNNs), recurrent neural networks (RNNs), transformers, generative adversarial networks (GANs), diffusion-based generative models, and reinforcement learning. For each category, it presents representative networks alongside their specific applications in radiation oncology. Moreover, the review outlines critical factors essential for training deep learning models, such as data preprocessing, loss functions, optimizers, and other pivotal training parameters including learning rate and batch size. CONCLUSION: This review provides a comprehensive overview of deep learning principles tailored toward radiation oncology. It aims to enhance the understanding of AI-based research and software applications, thereby bridging the gap between complex technological concepts and clinical practice in radiation oncology.

Multicenter privacy-preserving model training for deep learning brain metastases autosegmentation.

OBJECTIVES: This work aims to explore the impact of multicenter data heterogeneity on deep learning brain metastases (BM) autosegmentation performance, and assess the efficacy of an incremental transfer learning technique, namely learning without forgetting (LWF), to improve model generalizability without sharing raw data. MATERIALS AND METHODS: A total of six BM datasets from University Hospital Erlangen (UKER), University Hospital Zurich (USZ), Stanford, UCSF, New York University (NYU), and BraTS Challenge 2023 were used. First, the performance of the DeepMedic network for BM autosegmentation was established for exclusive single-center training and mixed multicenter training, respectively. Subsequently privacy-preserving bilateral collaboration was evaluated, where a pretrained model is shared to another center for further training using transfer learning (TL) either with or without LWF. RESULTS: For single-center training, average F1 scores of BM detection range from 0.625 (NYU) to 0.876 (UKER) on respective single-center test data. Mixed multicenter training notably improves F1 scores at Stanford and NYU, with negligible improvement at other centers. When the UKER pretrained model is applied to USZ, LWF achieves a higher average F1 score (0.839) than naive TL (0.570) and single-center training (0.688) on combined UKER and USZ test data. Naive TL improves sensitivity and contouring accuracy, but compromises precision. Conversely, LWF demonstrates commendable sensitivity, precision and contouring accuracy. When applied to Stanford, similar performance was observed. CONCLUSION: Data heterogeneity (e.g., variations in metastases density, spatial distribution, and image spatial resolution across centers) results in varying performance in BM autosegmentation, posing challenges to model generalizability. LWF is a promising approach to peer-to-peer privacy-preserving model training.

Generating synthetic computed tomography for radiotherapy: SynthRAD2023 challenge report.

Radiation therapy plays a crucial role in cancer treatment, necessitating precise delivery of radiation to tumors while sparing healthy tissues over multiple days. Computed tomography (CT) is integral for treatment planning, offering electron density data crucial for accurate dose calculations. However, accurately representing patient anatomy is challenging, especially in adaptive radiotherapy, where CT is not acquired daily. Magnetic resonance imaging (MRI) provides superior soft-tissue contrast. Still, it lacks electron density information, while cone beam CT (CBCT) lacks direct electron density calibration and is mainly used for patient positioning. Adopting MRI-only or CBCT-based adaptive radiotherapy eliminates the need for CT planning but presents challenges. Synthetic CT (sCT) generation techniques aim to address these challenges by using image synthesis to bridge the gap between MRI, CBCT, and CT. The SynthRAD2023 challenge was organized to compare synthetic CT generation methods using multi-center ground truth data from 1080 patients, divided into two tasks: (1) MRI-to-CT and (2) CBCT-to-CT. The evaluation included image similarity and dose-based metrics from proton and photon plans. The challenge attracted significant participation, with 617 registrations and 22/17 valid submissions for tasks 1/2. Top-performing teams achieved high structural similarity indices (≥0.87/0.90) and gamma pass rates for photon (≥98.1%/99.0%) and proton (≥97.3%/97.0%) plans. However, no significant correlation was found between image similarity metrics and dose accuracy, emphasizing the need for dose evaluation when assessing the clinical applicability of sCT. SynthRAD2023 facilitated the investigation and benchmarking of sCT generation techniques, providing insights for developing MRI-only and CBCT-based adaptive radiotherapy. It showcased the growing capacity of deep learning to produce high-quality sCT, reducing reliance on conventional CT for treatment planning.

PRL1 and PRL3 promote macropinocytosis via its lipid phosphatase activity.

PRL1 and PRL3, members of the protein tyrosine phosphatase family, have been associated with cancer metastasis and poor prognosis. Despite extensive research on their protein phosphatase activity, their potential role as lipid phosphatases remains elusive. Methods: We conducted comprehensive investigations to elucidate the lipid phosphatase activity of PRL1 and PRL3 using a combination of cellular assays, biochemical analyses, and protein interactome profiling. Functional studies were performed to delineate the impact of PRL1/3 on macropinocytosis and its implications in cancer biology. Results: Our study has identified PRL1 and PRL3 as lipid phosphatases that interact with phosphoinositide (PIP) lipids, converting PI(3,4)P2 and PI(3,5)P2 into PI(3)P on the cellular membranes. These enzymatic activities of PRLs promote the formation of membrane ruffles, membrane blebbing and subsequent macropinocytosis, facilitating nutrient extraction, cell migration, and invasion, thereby contributing to tumor development. These enzymatic activities of PRLs promote the formation of membrane ruffles, membrane blebbing and subsequent macropinocytosis. Additionally, we found a correlation between PRL1/3 expression and glioma development, suggesting their involvement in glioma progression. Conclusions: Combining with the knowledge that PRLs have been identified to be involved in mTOR, EGFR and autophagy, here we concluded the physiological role of PRL1/3 in orchestrating the nutrient sensing, absorbing and recycling via regulating macropinocytosis through its lipid phosphatase activity. This mechanism could be exploited by tumor cells facing a nutrient-depleted microenvironment, highlighting the potential therapeutic significance of targeting PRL1/3-mediated macropinocytosis in cancer treatment.

System analysis based on Anoikis-related genes identifies MAPK1 as a novel therapy target for osteosarcoma with neoadjuvant chemotherapy.

BACKGROUND: Osteosarcoma (OS) is the most common bone malignant tumor in children, and its prognosis is often poor. Anoikis is a unique mode of cell death.However, the effects of Anoikis in OS remain unexplored. METHOD: Differential analysis of Anoikis-related genes was performed based on the metastatic and non-metastatic groups. Then LASSO logistic regression and SVM-RFE algorithms were applied to screen out the characteristic genes. Later, Univariate and multivariate Cox regression was conducted to identify prognostic genes and further develop the Anoikis-based risk score. In addition, correlation analysis was performed to analyze the relationship between tumor microenvironment, drug sensitivity, and prognostic models. RESULTS: We established novel Anoikis-related subgroups and developed a prognostic model based on three Anoikis-related genes (MAPK1, MYC, and EDIL3). The survival and ROC analysis results showed that the prognostic model was reliable. Besides, the results of single-cell sequencing analysis suggested that the three prognostic genes were closely related to immune cell infiltration. Subsequently, aberrant expression of two prognostic genes was identified in osteosarcoma cells. Nilotinib can promote the apoptosis of osteosarcoma cells and down-regulate the expression of MAPK1. CONCLUSIONS: We developed a novel Anoikis-related risk score model, which can assist clinicians in evaluating the prognosis of osteosarcoma patients in clinical practice. Analysis of the tumor immune microenvironment and chemotherapeutic drug sensitivity can provide necessary insights into subsequent mechanisms. MAPK1 may be a valuable therapeutic target for neoadjuvant chemotherapy in osteosarcoma.

Two-view topogram-based anatomy-guided CT reconstruction for prospective risk minimization.

To facilitate a prospective estimation of the effective dose of an CT scan prior to the actual scanning in order to use sophisticated patient risk minimizing methods, a prospective spatial dose estimation and the known anatomical structures are required. To this end, a CT reconstruction method is required to reconstruct CT volumes from as few projections as possible, i.e. by using the topograms, with anatomical structures as correct as possible. In this work, an optimized CT reconstruction model based on a generative adversarial network (GAN) is proposed. The GAN is trained to reconstruct 3D volumes from an anterior-posterior and a lateral CT projection. To enhance anatomical structures, a pre-trained organ segmentation network and the 3D perceptual loss are applied during the training phase, so that the model can then generate both organ-enhanced CT volume and organ segmentation masks. The proposed method can reconstruct CT volumes with PSNR of 26.49, RMSE of 196.17, and SSIM of 0.64, compared to 26.21, 201.55 and 0.63 using the baseline method. In terms of the anatomical structure, the proposed method effectively enhances the organ shapes and boundaries and allows for a straight-forward identification of the relevant anatomical structures. We note that conventional reconstruction metrics fail to indicate the enhancement of anatomical structures. In addition to such metrics, the evaluation is expanded with assessing the organ segmentation performance. The average organ dice of the proposed method is 0.71 compared with 0.63 for the baseline model, indicating the enhancement of anatomical structures.

A real-world disproportionality analysis of the FDA adverse event reporting system events for ibuprofen.

INTRODUCTION: Ibuprofen is commonly used as an over-the-counter (OTC) antipyretic and analgesic. As the frequency of its use has increased, there has been a corresponding increase in reports of associated adverse events (AEs). However, these events have not been systematically reported in the literature. Meanwhile, the importance of effective pharmacovigilance in evaluating the benefits and risks of drugs is being recognized. METHODS: The data was obtained indirectly from FAERS using the OpenVigil 2 database, lexically mapped using software such as MySQL, Microsoft Excel, and the R language, and then subjected to four more rigorous algorithms to detect risk signals associated with ibuprofen AEs. RESULTS: By analyzing data from the past 18 years, 878 ibuprofen-related AEs were identified as primary AEs. Notably, unexpected reproductive system and breast diseases, etc., which were unexpected, were observed as important system organ classes (SOCs) associated with ibuprofen. Among the 651 preferred terms (PTs) that simultaneously satisfy the four arithmetic methods, renal tubular acidosis and lip oedema are proposed as new signals for ibuprofen AEs. CONCLUSION: This study explores the important and valuable potential AEs and ADRs of ibuprofen at the SOC and PT levels, respectively. To provide a reference on decision-making for ibuprofen to promote rational clinical dosing.

Discovery of novel macrocyclic derivatives as potent and selective cyclin-dependent kinase 2 inhibitors.

Cyclin-dependent kinase 2 (CDK2) is a member of CDK family of kinases (CDKs) that regulate the cell cycle. Its inopportune or over-activation leads to uncontrolled cell cycle progression and drives numerous types of cancers, especially ovarian, uterine, gastric cancer, as well as those associated with amplified CCNE1 gene. However, developing selective lead compound as CDK2 inhibitors remains challenging owing to similarities in the ATP pockets among different CDKs. Herein, we described the optimization of compound 1, a novel macrocyclic inhibitor targeting CDK2/5/7/9, aiming to discover more selective and metabolically stable lead compound as CDK2 inhibitor. Molecular dynamic (MD) simulations were performed for compound 1 and 9 to gain insights into the improved selectivity against CDK5. Further optimization efforts led to compound 22, exhibiting excellent CDK2 inhibitory activity, good selectivity over other CDKs and potent cellular effects. Based on these characterizations, we propose that compound 22 holds great promise as a potential lead candidate for drug development.

Metformin adverse event profile: a pharmacovigilance study based on the FDA Adverse Event Reporting System (FAERS) from 2004 to 2022.

BACKGROUND: Metformin has the potential for treating numerous diseases, but there are still many unrecognized and unreported adverse events (AEs). METHODS: We selected data from the United States FDA Adverse Event Reporting System (FAERS) database from the first quarter (Q1) of 2004 to the fourth quarter (Q4) of 2022 for disproportionality analysis to assess the association between metformin and related adverse events. RESULTS: In this study 10,500,295 case reports were collected from the FAERS database, of which 56,674 adverse events related to metformin were reported. A total of 643 preferred terms (PTs) and 27 system organ classes (SOCs) that were significant disproportionality conforming to the four algorithms simultaneously were included. The SOCs included metabolic and nutritional disorders (p = 0.00E + 00), gastrointestinal disorders (p = 0.00E + 00) and others. PT levels were screened for adverse drug reaction (ADR) signals such as acute pancreatitis (p = 0.00E + 00), melas syndrome, pemphigoid (p = 0.00E + 00), skin eruption (p = 0.00E + 00) and drug exposure during pregnancy (p = 0.00E + 00). CONCLUSION: Most of our results were consistent with the specification, but some new signals of adverse reactions such as acute pancreatitis were not included. Therefore, further studies are needed to validate unlabeled adverse reactions and provide important support for clinical monitoring and risk identification of metformin.

Toward a deep learning-based magnetic resonance imaging only workflow for postimplant dosimetry in I-125 seed brachytherapy for prostate cancer.

BACKGROUND AND PURPOSE: The current standard imaging-technique for creating postplans in seed prostate brachytherapy is computed tomography (CT), that is associated with additional radiation exposure and poor soft tissue contrast. To establish a magnetic resonance imaging (MRI) only workflow combining improved tissue contrast and high seed detectability, a deep learning-approach for automatic seed segmentation on MRI-scans was developed. MATERIAL AND METHODS: Patients treated with I-125 seed brachytherapy received a postplan-CT and a 1.5 T MRI-scan on nominal day 30 after implantation. For MRI-based seed visualization, DIXON-sequences were acquired and deep learning-based quantitative susceptibility maps (QSM) were generated from 3D-gradient-echo-sequences from 20 patients. Seed segmentations created on CT served as ground truth. For automatic seed segmentation on MRI, a 3D nnU-net model was trained using QSM and DIXON, both solely and combined. RESULTS: Of the implanted seeds 94.8 ± 2.4% were detected with deep learning automatic segmentation entrained on both QSM and DIXON data. Models trained on the individual sequence data-sets performed worse with detection rates of 87.5 ± 2.6% or 88.6 ± 7.5% for QSM and DIXON respectively. The seed centers identified on CT versus QSM and DIXON were on average 1.8 ± 1.3 mm apart. Postimplant dosimetry for evaluation of positioning inaccuracies revealed only small variations of up to 0.4 ± 4.26 Gy in D90 (dose 90% of the prostate receives) between the standard CT-approach and our MRI-only workflow. CONCLUSION: The proposed deep learning-based MRI-only workflow provided a promisingly accurate and robust seed localization and thus has the potential to compete with current state-of-the-art CT-based postimplant dosimetry in the future.

Benchmarking ChatGPT-4 on a radiation oncology in-training exam and Red Journal Gray Zone cases: potentials and challenges for ai-assisted medical education and decision making in radiation oncology.

Purpose: The potential of large language models in medicine for education and decision-making purposes has been demonstrated as they have achieved decent scores on medical exams such as the United States Medical Licensing Exam (USMLE) and the MedQA exam. This work aims to evaluate the performance of ChatGPT-4 in the specialized field of radiation oncology. Methods: The 38th American College of Radiology (ACR) radiation oncology in-training (TXIT) exam and the 2022 Red Journal Gray Zone cases are used to benchmark the performance of ChatGPT-4. The TXIT exam contains 300 questions covering various topics of radiation oncology. The 2022 Gray Zone collection contains 15 complex clinical cases. Results: For the TXIT exam, ChatGPT-3.5 and ChatGPT-4 have achieved the scores of 62.05% and 78.77%, respectively, highlighting the advantage of the latest ChatGPT-4 model. Based on the TXIT exam, ChatGPT-4's strong and weak areas in radiation oncology are identified to some extent. Specifically, ChatGPT-4 demonstrates better knowledge of statistics, CNS & eye, pediatrics, biology, and physics than knowledge of bone & soft tissue and gynecology, as per the ACR knowledge domain. Regarding clinical care paths, ChatGPT-4 performs better in diagnosis, prognosis, and toxicity than brachytherapy and dosimetry. It lacks proficiency in in-depth details of clinical trials. For the Gray Zone cases, ChatGPT-4 is able to suggest a personalized treatment approach to each case with high correctness and comprehensiveness. Importantly, it provides novel treatment aspects for many cases, which are not suggested by any human experts. Conclusion: Both evaluations demonstrate the potential of ChatGPT-4 in medical education for the general public and cancer patients, as well as the potential to aid clinical decision-making, while acknowledging its limitations in certain domains. Owing to the risk of hallucinations, it is essential to verify the content generated by models such as ChatGPT for accuracy.

Deep Learning and Registration-Based Mapping for Analyzing the Distribution of Nodal Metastases in Head and Neck Cancer Cohorts: Informing Optimal Radiotherapy Target Volume Design.

We introduce a deep-learning- and a registration-based method for automatically analyzing the spatial distribution of nodal metastases (LNs) in head and neck (H/N) cancer cohorts to inform radiotherapy (RT) target volume design. The two methods are evaluated in a cohort of 193 H/N patients/planning CTs with a total of 449 LNs. In the deep learning method, a previously developed nnU-Net 3D/2D ensemble model is used to autosegment 20 H/N levels, with each LN subsequently being algorithmically assigned to the closest-level autosegmentation. In the nonrigid-registration-based mapping method, LNs are mapped into a calculated template CT representing the cohort-average patient anatomy, and kernel density estimation is employed to estimate the underlying average 3D-LN probability distribution allowing for analysis and visualization without prespecified level definitions. Multireader assessment by three radio-oncologists with majority voting was used to evaluate the deep learning method and obtain the ground-truth distribution. For the mapping technique, the proportion of LNs predicted by the 3D probability distribution for each level was calculated and compared to the deep learning and ground-truth distributions. As determined by a multireader review with majority voting, the deep learning method correctly categorized all 449 LNs to their respective levels. Level 2 showed the highest LN involvement (59.0%). The level involvement predicted by the mapping technique was consistent with the ground-truth distribution (p for difference 0.915). Application of the proposed methods to multicenter cohorts with selected H/N tumor subtypes for informing optimal RT target volume design is promising.

Revealing the Mechanisms of Byu dMar 25 in the Treatment of Alzheimer's Disease through Network Pharmacology, Molecular Docking, and In Vivo Experiment.

BACKGROUND: Alzheimer's disease (AD) is the most common neurodegenerative disease, severely reducing the cognitive level and life quality of patients. Byu dMar 25 (BM25) has been proved to have a therapeutic effect on AD. However, the pharmacological mechanism is still unclear. Therefore, this study aims to reveal the potential mechanism of BM25 affecting AD from the perspective of network pharmacology and experimental validation. METHODS: The potential active ingredients of BM25 were obtained from the TCMSP database and literature. Possible targets were predicted using SwissTargetPrediction tools. AD-related genes were identified by using GeneCards, OMIM, DisGeNET, and Drugbank databases. The candidate genes were obtained by extraction of the intersection network. Additionally, the "drug-target-disease" network was constructed by Cytoscape 3.7.2 for visualization. The PPI network was constructed by the STRING database, and the core network modules were filtered by Cytoscape 3.7.2. Enrichment analysis of GO and KEGG was carried out in the Metascape platform. Ledock software was used to dock the critical components with the core target. Furthermore, protein levels were evaluated by immunohistochemistry. RESULTS: In this study, 112 active components, 1112 disease candidate genes, 3084 GO functions, and 277 KEGG pathways were obtained. Molecular docking showed that the effective components of BM25 in treating AD were ß-asarone and hydroxysafflor yellow A. The most important targets were APP, PIK3R1, and PIK3CA. Enrichment analysis indicated that the Golgi genetic regulation, peroxidase activity regulation, phosphatidylinositol 3-kinase complex IA, 5-hydroxytryptamine receptor complexes, cancer pathways, and neuroactive ligand-receptor interactions played vital roles against AD. The rat experiment verified that BM25 affected PI3K-Akt pathway activation in AD. CONCLUSIONS: This study reveals the mechanism of BM25 in treating AD with network pharmacology, which provides a foundation for further study on the molecular mechanism of AD treatment.

Research Progress on the Effect of Autophagy and Exosomes on Liver Fibrosis.

Chronic liver disease is a known risk factor for the development of liver cancer, and the development of microRNA (miRNA) liver therapies has been hampered by the difficulty of delivering miRNA to damaged tissues. In recent years, numerous studies have shown that hepatic stellate cell (HSC) autophagy and exosomes play an important role in maintaining liver homeostasis and ameliorating liver fibrosis. In addition, the interaction between HSC autophagy and exosomes also affects the progression of liver fibrosis. In this paper, we review the research progress of mesenchymal stem cell-derived exosomes (MSC-EVs) loaded with specific miRNA and autophagy, and their related signaling pathways in liver fibrosis, which will provide a more reliable basis for the use of MSC-EVs for therapeutic delivery of miRNAs targeting the chronic liver disease.

Deep learning for automatic head and neck lymph node level delineation provides expert-level accuracy.

Background: Deep learning-based head and neck lymph node level (HN_LNL) autodelineation is of high relevance to radiotherapy research and clinical treatment planning but still underinvestigated in academic literature. In particular, there is no publicly available open-source solution for large-scale autosegmentation of HN_LNL in the research setting. Methods: An expert-delineated cohort of 35 planning CTs was used for training of an nnU-net 3D-fullres/2D-ensemble model for autosegmentation of 20 different HN_LNL. A second cohort acquired at the same institution later in time served as the test set (n = 20). In a completely blinded evaluation, 3 clinical experts rated the quality of deep learning autosegmentations in a head-to-head comparison with expert-created contours. For a subgroup of 10 cases, intraobserver variability was compared to the average deep learning autosegmentation accuracy on the original and recontoured set of expert segmentations. A postprocessing step to adjust craniocaudal boundaries of level autosegmentations to the CT slice plane was introduced and the effect of autocontour consistency with CT slice plane orientation on geometric accuracy and expert rating was investigated. Results: Blinded expert ratings for deep learning segmentations and expert-created contours were not significantly different. Deep learning segmentations with slice plane adjustment were rated numerically higher (mean, 81.0 vs. 79.6, p = 0.185) and deep learning segmentations without slice plane adjustment were rated numerically lower (77.2 vs. 79.6, p = 0.167) than manually drawn contours. In a head-to-head comparison, deep learning segmentations with CT slice plane adjustment were rated significantly better than deep learning contours without slice plane adjustment (81.0 vs. 77.2, p = 0.004). Geometric accuracy of deep learning segmentations was not different from intraobserver variability (mean Dice per level, 0.76 vs. 0.77, p = 0.307). Clinical significance of contour consistency with CT slice plane orientation was not represented by geometric accuracy metrics (volumetric Dice, 0.78 vs. 0.78, p = 0.703). Conclusions: We show that a nnU-net 3D-fullres/2D-ensemble model can be used for highly accurate autodelineation of HN_LNL using only a limited training dataset that is ideally suited for large-scale standardized autodelineation of HN_LNL in the research setting. Geometric accuracy metrics are only an imperfect surrogate for blinded expert rating.

Fibroblast growth factor 9 reduces TBHP-induced oxidative stress in chondrocytes and diminishes mouse osteoarthritis by activating ERK/Nrf2 signaling pathway.

Osteoarthritis (OA) is a degenerative and progressive disease that affects joints. Pathologically, it is characterized by oxidative stress-mediated excessive chondrocyte apoptosis and mitochondrial dysfunction. Fibroblast growth factor 9 (FGF9) has been shown to exert antioxidant effects and prevent degenerative diseases by activating ERK-related signaling pathways. However, the mechanism of FGF9 in the pathogenesis of OA and its relationship with anti-oxidative stress and related pathways are unclear. In this study, mice with medial meniscus instability (DMM) were used as the in vivo model whereas TBHP-induced chondrocytes served as the in vitro model to explore the mechanism underlying the effects of FGF9 in OA and its association with anti-oxidative stress. Results showed that FGF9 reduced oxidative stress, apoptosis, and mitochondrial dysfunction in TBHP-treated chondrocytes and promoted the nuclear translocation of Nrf2 to activate the Nrf2/HO1 signaling pathway. Interestingly, silencing the Nrf2 gene or blocking the ERK signaling pathway abolished the antioxidant effects of FGF9. FGF9 treatment reduced joint space narrowing, cartilage ossification, and synovial thickening in the DMM model mice. In conclusion, the present findings demonstrate that FGF9 can inhibit TBHP-induced oxidative stress in chondrocytes through the ERK and Nrf2-HO1 signaling pathways and prevent the progression of OA in vivo.

Deep learning for terahertz image denoising in nondestructive historical document analysis.

Historical documents contain essential information about the past, including places, people, or events. Many of these valuable cultural artifacts cannot be further examined due to aging or external influences, as they are too fragile to be opened or turned over, so their rich contents remain hidden. Terahertz (THz) imaging is a nondestructive 3D imaging technique that can be used to reveal the hidden contents without damaging the documents. As noise or imaging artifacts are predominantly present in reconstructed images processed by standard THz reconstruction algorithms, this work intends to improve THz image quality with deep learning. To overcome the data scarcity problem in training a supervised deep learning model, an unsupervised deep learning network (CycleGAN) is first applied to generate paired noisy THz images from clean images (clean images are generated by a handwriting generator). With such synthetic noisy-to-clean paired images, a supervised deep learning model using Pix2pixGAN is trained, which is effective to enhance real noisy THz images. After Pix2pixGAN denoising, 99% characters written on one-side of the Xuan paper can be clearly recognized, while 61% characters written on one-side of the standard paper are sufficiently recognized. The average perceptual indices of Pix2pixGAN processed images are 16.83, which is very close to the average perceptual index 16.19 of clean handwriting images. Our work has important value for THz-imaging-based nondestructive historical document analysis.

Trainable joint bilateral filters for enhanced prediction stability in low-dose CT.

Low-dose computed tomography (CT) denoising algorithms aim to enable reduced patient dose in routine CT acquisitions while maintaining high image quality. Recently, deep learning (DL)-based methods were introduced, outperforming conventional denoising algorithms on this task due to their high model capacity. However, for the transition of DL-based denoising to clinical practice, these data-driven approaches must generalize robustly beyond the seen training data. We, therefore, propose a hybrid denoising approach consisting of a set of trainable joint bilateral filters (JBFs) combined with a convolutional DL-based denoising network to predict the guidance image. Our proposed denoising pipeline combines the high model capacity enabled by DL-based feature extraction with the reliability of the conventional JBF. The pipeline's ability to generalize is demonstrated by training on abdomen CT scans without metal implants and testing on abdomen scans with metal implants as well as on head CT data. When embedding RED-CNN/QAE, two well-established DL-based denoisers in our pipeline, the denoising performance is improved by 10%/82% (RMSE) and 3%/81% (PSNR) in regions containing metal and by 6%/78% (RMSE) and 2%/4% (PSNR) on head CT data, compared to the respective vanilla model. Concluding, the proposed trainable JBFs limit the error bound of deep neural networks to facilitate the applicability of DL-based denoisers in low-dose CT pipelines.