Synergistic Polymer Blending Informs Efficient Terpolymer Design and Machine Learning Discerns Performance Trends for pDNA Delivery.

Cationic polymers offer an alternative to viral vectors in nucleic acid delivery. However, the development of polymer vehicles capable of high transfection efficiency and minimal toxicity has remained elusive, and continued exploration of the vast design space is required. Traditional single polymer syntheses with large monomer bases are very time-intensive, limiting the speed at which new formulations are identified. In this work, we present an experimental method for the quick probing of the design space, utilizing a combinatorial set of 90 polymer blends, derived from 6 statistical copolymers, to deliver pDNA. This workflow facilitated rapid screening of polyplex compositions, successfully tailoring polyplex hydrophobicity, particle size, and payload binding affinity. This workflow identified blended polyplexes with high levels of transfection efficiency and cell viability relative to single copolymer controls and commercial JetPEI, indicating synergistic benefits from copolymer blending. Polyplex composition was coupled with biological outputs to guide the synthesis of single terpolymer vehicles, with high-performing polymers P10 and M20, providing superior transfection of HEK293T cells in serum-free and serum-containing media, respectively. Machine learning coupled with SHapley Additive exPlanations (SHAP) was used to identify polymer/polyplex attributes that most impact transfection efficiency, viability, and overall effective efficiency. Subsequent transfections on ARPE-19 and HDFn cells found that P10 and M20 were surpassed in performance by M10, contrasting with results in HEK293T cells. This cell type dependency reinforced the need to evaluate transfection conditions with multiple cell models to potentially identify moieties more beneficial to delivery in certain tissues. Overall, the workflow employed can be used to expedite the exploration of the polymer design space, bypassing extensive synthesis, and to develop improved polymer delivery vehicles more readily for nucleic acid therapies.

Development and Characterization of Quercetin-Loaded Polymeric Liposomes with Gelatin-Poly(ethylene glycol)-Folic Acid Coating to Increase Their Long-Circulating and Anticancer Activity.

Liposomes as drug-delivery systems have been researched and applied in multiple scientific reports and introduced as patented products with interesting therapeutic properties. Despite various advantages, this drug carrier faces major difficulties in its innate stability, cancer cell specificity, and control over the release of hydrophobic drugs, particularly quercetin, a naturally derived drug that carries many desirable characteristics for anticancer treatment. To improve the effectiveness of liposomes to deliver quercetin by tackling and mitigating the mentioned hurdles, we developed a strategy to establish the ability to passively target cancerous cells, as well as to increase the bioavailability of loaded drugs by incorporating poly(ethylene glycol), gelatin, and folic acid moieties to modify the liposomal system's surface. This research developed a chemically synthesized gelatin, poly(ethylene glycol), and folic acid as a single polymer to coat drug-loaded liposome systems. Liposomes were coated with gelatin-poly(ethylene glycol)-folic acid by electrostatic interaction, characterized by their size, morphology, ζ potential, drug loading efficiency, infrared structures, differential scanning calorimetry spectra, and drug-releasing profiles, and then evaluated for their cytotoxicity to MCF-7 breast cancer cells, as well as cellular uptake, analyzed by confocal imaging to further elaborate on the in vitro behavior of the coated liposome. The results indicated an unusual change in size with increased coating materials, followed by increased colloidal stability, ζ potential, and improved cytotoxicity to cancer cells, as shown by the cellular viability test with MCF-7. Cellular uptake also confirmed these results, providing data for the effects of biopolymer coating, while confirming that folic acid can increase the uptake of liposome by cancer cells. In consideration of such results, the modified gelatin-poly(ethylene glycol)-folic acid-coated liposome can be a potential system in delivering the assigned anticancer compound. This modified biopolymer showed excellent properties as a coating material and should be considered for further practical applications in the future.

Tuberous sclerosis complex associated lymphangioleiomyomatosis caused by de novo mutation of TSC2 gene in Vietnam: A case report.

Lymphangioleiomyomatosis (LAM) represents a rare, insidiously progressive disease of the pulmonary system, marked by cystic degradation of lung tissues leading to respiratory compromise. Pulmonary LAM has been identified as being associated with tuberous sclerosis complex (TSC) in its pulmonary manifestation (TSC-LAM), a multisystem genetic disorder resulting from mutations in either the TSC1 or TSC2 genes. Herein, we describe an early 20s female admitted to the hospital with dyspnea, chest pain, hypopigmented macules, and facial fibroadenomas. She has a medical history of renal angiomyolipomas (ALMs) and pneumothoraces. Diagnosis with LAM was confirmed through high-resolution computed tomography (HRCT) scan and histopathology of lung biopsy. Whole exome sequencing analysis identified a frameshift mutation c.4504del (p.L1502Cfs*74) in the patient's TSC2 gene. This variant was de novo due to its absence in the patient's parents. This is the first report on the clinical and genetic etiology of TSC-LAM in Vietnam.

Antiangiogenic Therapeutic mRNA Delivery Using Lung-Selective Polymeric Nanomedicine for Lung Cancer Treatment.

Therapeutic antibodies that block vascular endothelial growth factor (VEGF) show clinical benefits in treating nonsmall cell lung cancers (NSCLCs) by inhibiting tumor angiogenesis. Nonetheless, the therapeutic effects of systemically administered anti-VEGF antibodies are often hindered in NSCLCs because of their limited distribution in the lungs and their adverse effects on normal tissues. These challenges can be overcome by delivering therapeutic antibodies in their mRNA form to lung endothelial cells, a primary target of VEGF-mediated pulmonary angiogenesis, to suppress the NSCLCs. In this study, we synthesized derivatives of poly(ß-amino esters) (PBAEs) and prepared nanoparticles to encapsulate the synthetic mRNA encoding bevacizumab, an anti-VEGF antibody used in the clinic. Optimization of nanoparticle formulations resulted in a selective lung transfection after intravenous administration. Notably, the optimized PBAE nanoparticles were distributed in lung endothelial cells, resulting in the secretion of bevacizumab. We analyzed the protein corona on the lung- and spleen-targeting nanoparticles using proteomics and found distinctive features potentially contributing to their organ-selectivity. Lastly, bevacizumab mRNA delivered by the lung-targeting PBAE nanoparticles more significantly inhibited tumor proliferation and angiogenesis than recombinant bevacizumab protein in orthotopic NSCLC mouse models, supporting the therapeutic potential of bevacizumab mRNA therapy and its selective delivery through lung-targeting nanoparticles. Our proof-of-principle results highlight the clinical benefits of nanoparticle-mediated mRNA therapy in anticancer antibody treatment in preclinical models.

Successful Endoscopic Submucosal Dissection Using Open Peroral Endoscopic Myotomy for Duodenal Neuroendocrine Tumor.

Duodenal neuroendocrine tumors (NETs) are subepithelial tumors that are difficult to remove endoscopically, particularly when located just beyond the pylorus. This paper reports a case of a successful endoscopic submucosal dissection (ESD) using open gastric peroral endoscopic myotomy (POEM) for a remnant duodenal NET detected after endoscopic mucosal resection (EMR). A 67-year-old male presented with a 5 mm remnant duodenal NET close to the pylorus after EMR for a duodenal polypoid lesion performed four months earlier. Duodenal ESD was performed under conscious sedation using I-type and IT II knives. The tumor adhered to the fibrotic tissue, and the submucosal cushion was insufficient. Open gastric POEM was performed concurrently during ESD, resulting in the complete resection of the NET. This case suggests that while challenging, open gastric POEM can serve as a valuable technique for endoscopic resection in cases of early gastric cancer or duodenal masses located around the pylorus.

A genetic association study of circulating coagulation factor VIII and von Willebrand factor levels.

ABSTRACT: Coagulation factor VIII (FVIII) and its carrier protein von Willebrand factor (VWF) are critical to coagulation and platelet aggregation. We leveraged whole-genome sequence data from the Trans-Omics for Precision Medicine (TOPMed) program along with TOPMed-based imputation of genotypes in additional samples to identify genetic associations with circulating FVIII and VWF levels in a single-variant meta-analysis, including up to 45 289 participants. Gene-based aggregate tests were implemented in TOPMed. We identified 3 candidate causal genes and tested their functional effect on FVIII release from human liver endothelial cells (HLECs) and VWF release from human umbilical vein endothelial cells. Mendelian randomization was also performed to provide evidence for causal associations of FVIII and VWF with thrombotic outcomes. We identified associations (P < 5 × 10-9) at 7 new loci for FVIII (ST3GAL4, CLEC4M, B3GNT2, ASGR1, F12, KNG1, and TREM1/NCR2) and 1 for VWF (B3GNT2). VWF, ABO, and STAB2 were associated with FVIII and VWF in gene-based analyses. Multiphenotype analysis of FVIII and VWF identified another 3 new loci, including PDIA3. Silencing of B3GNT2 and the previously reported CD36 gene decreased release of FVIII by HLECs, whereas silencing of B3GNT2, CD36, and PDIA3 decreased release of VWF by HVECs. Mendelian randomization supports causal association of higher FVIII and VWF with increased risk of thrombotic outcomes. Seven new loci were identified for FVIII and 1 for VWF, with evidence supporting causal associations of FVIII and VWF with thrombotic outcomes. B3GNT2, CD36, and PDIA3 modulate the release of FVIII and/or VWF in vitro.

Expression of virus-like particles (VLPs) of foot-and-mouth disease virus (FMDV) using Saccharomyces cerevisiae.

We engineered Saccharomyces cerevisiae to express structural proteins of foot-and-mouth disease virus (FMDV) and produce virus-like particles (VLPs). The gene, which encodes four structural capsid proteins (VP0 (VP4 and VP2), VP3, and VP1), followed by a translational "ribosomal skipping" sequence consisting of 2A and protease 3C, was codon-optimized and chemically synthesized. The cloned gene was used to transform S. cerevisiae 2805 strain. Western blot analysis revealed that the polyprotein consisting of VP0, VP3, and VP1 was processed into the discrete capsid proteins. Western blot analysis of 3C confirmed the presence of discrete 3C protein, suggesting that the 2A sequence functioned as a "ribosomal skipping" signal in the yeast for an internal re-initiation of 3C translation from a monocistronic transcript, thereby indicating polyprotein processing by the discrete 3C protease. Moreover, a band corresponding to only VP2, which was known to be non-enzymatically processed from VP0 to both VP4 and VP2 during viral assembly, further validated the assembly of processed capsid proteins into VLPs. Electron microscopy showed the presence of the characteristic icosahedral VLPs. Our results clearly demonstrate that S. cerevisiae processes the viral structural polyprotein using a viral 3C protease and the resulting viral capsid subunits are assembled into virion particles. KEY POINTS: • Ribosomal skipping by self-cleaving FMDV peptide in S. cerevisiae. • Proteolytic processing of a structural polyprotein from a monocistronic transcript. • Assembly of the processed viral capsid proteins into a virus-like particle.

The Prognostic Value of Sequential 18 F-FDG PET/CT Metabolic Parameters in Outcomes of Upper-Third Esophageal Squamous Cell Carcinoma Patients Treated with Definitive Chemoradiotherapy.

Objective The aim of this study is to determine prognostic values of sequential 18 F-FDG PET/CT metabolic parameters in locally advanced esophageal squamous cell carcinoma (ESCC) patients treated with definitive chemoradiotherapy. Materials and Methods Forty locally advanced ESCC patients treated with definitive chemoradiotherapy (dCRT) who received pre-treatment 18 F-FDG PET/CT (PET1) and 3-months post-treatment 18 F-FDG PET/CT (PET2) were enrolled in the prospective study. 18 F-FDG PET parameters of the primary tumor including maximum and mean standardized uptake values (SUVmax, SUVmean), metabolic tumor volume (MTV), and total lesion glycolysis (TLG) were calculated on PET delineated primary tumor. Using Kaplan-Meier curves to estimated overall survival (OS), progression-free survival (PFS), and local-regional control (LRC). Cox regression analysis was performed to find significant prognostic factors for survival. Results With a median follow-up of 13.5 months, the 4-year OS, PFS, and LRC rates were 67.3%, 52.6%, and 53.4% respectively. Patients with MTV 2 > 5.7 had lower OS, PFS, and LRC rates than the lower MTV 2 group (p < 0.05). Univariate Cox regression analysis showed that MTV2 was a significant prognostic factor for OS, PFS, and LRC (p < 0.05). Conclusion MTV parameter of sequential 18 F-FDG PET/CT could be used as a prognostic factor for OS, PFS, and LRC in locally advanced ESCC patients treated with dCRT.

Nonribosomal Peptide Synthetases in Animals.

Nonribosomal peptide synthetases (NRPSs) are a class of cytosolic enzymes that synthesize a range of bio-active secondary metabolites including antibiotics and siderophores. They are widespread among both prokaryotes and eukaryotes but are considered rare among animals. Recently, several novel NRPS genes have been described in nematodes, schistosomes, and arthropods, which led us to investigate how prevalent NRPS genes are in the animal kingdom. We screened 1059 sequenced animal genomes and showed that NRPSs were present in 7 out of the 19 phyla analyzed. A phylogenetic analysis showed that the identified NRPSs form clades distinct from other adenylate-forming enzymes that contain similar domains such as fatty acid synthases. NRPSs show a remarkably scattered distribution over the animal kingdom. They are especially abundant in rotifers and nematodes. In rotifers, we found a large variety of domain architectures and predicted substrates. In the nematode Plectus sambesii, we identified the beta-lactam biosynthesis genes L-δ-(α-aminoadipoyl)-L-cysteinyl-D-valine synthetase, isopenicillin N synthase, and deacetoxycephalosporin C synthase that catalyze the formation of beta-lactam antibiotics in fungi and bacteria. These genes are also present in several species of Collembola, but not in other hexapods analyzed so far. In conclusion, our survey showed that NRPS genes are more abundant and widespread in animals than previously known.

Liposomal co-delivery of toll-like receptors 3 and 7 agonists induce a hot triple-negative breast cancer immune environment.

Triple-negative breast cancer (TNBC) is highly aggressive and has no standard treatment. Although being considered as an alternative to conventional treatments for TNBC, immunotherapy has to deal with many challenges that hinder its efficacy, particularly the poor immunogenic condition of the tumor microenvironment (TME). Herein, we designed a liposomal nanoparticle (LN) platform that delivers simultaneously toll-like receptor 7 (imiquimod, IQ) and toll-like receptor 3 (poly(I:C), IC) agonists to take advantage of the different toll-like receptor (TLR) signaling pathways, which enhances the condition of TME from a "cold" to a "hot" immunogenic state. The optimized IQ/IC-loaded LN (IQ/IC-LN) was effectively internalized by cancer cells, macrophages, and dendritic cells, followed by the release of the delivered drugs and subsequent stimulation of the TLR3 and TLR7 signaling pathways. This stimulation encouraged the secretion of type I interferon (IFN-α, IFN-ß) and CXCLl0, a T-cell and antigen-presenting cells (APCs) recruitment chemokine, from both cancer cells and macrophages and polarized macrophages to the M1 subtype in in vitro studies. Notably, systemic administration of IQ/IC-LN allowed for the high accumulation of drug content in the tumor, followed by the effective uptake by immune cells in the TME. IQ/IC-LN treatment comprehensively enhanced the immunogenic condition in the TME, which robustly inhibited tumor growth in tumor-bearing mice. Furthermore, synergistic antitumor efficacy was obtained when the IQ/IC-LN-induced immunogenic state in TME was combined with anti-PD1 antibody therapy. Thus, our results suggest the potential of combining 2 TLR agonists to reform the TME from a "cold" to a "hot" state, supporting the therapeutic efficacy of immune checkpoint inhibitors.

Polygenic risk of major depressive disorder as a risk factor for venous thromboembolism.

Major depressive disorder (MDD), bipolar disorder (BD), and schizophrenia (SCZ) are associated with an increased risk of cardiovascular diseases, including venous thromboembolism (VTE). The reasons for this are complex and include obesity, smoking, and use of hormones and psychotropic medications. Genetic studies have increasingly provided evidence of the shared genetic risk of psychiatric and cardiometabolic illnesses. This study aimed to determine whether a genetic predisposition to MDD, BD, or SCZ is associated with an increased risk of VTE. Genetic correlations using the largest genome-wide genetic meta-analyses summary statistics for MDD, BD, and SCZ (Psychiatric Genetics Consortium) and a recent genome-wide genetic meta-analysis of VTE (INVENT Consortium) demonstrated a positive association between VTE and MDD but not BD or SCZ. The same summary statistics were used to construct polygenic risk scores for MDD, BD, and SCZ in UK Biobank participants of self-reported White British ancestry. These were assessed for impact on self-reported VTE risk (10 786 cases, 285 124 controls), using logistic regression, in sex-specific and sex-combined analyses. We identified significant positive associations between polygenic risk for MDD and the risk of VTE in men, women, and sex-combined analyses, independent of the known risk factors. Secondary analyses demonstrated that this association was not driven by those with lifetime experience of mental illness. Meta-analyses of individual data from 6 additional independent cohorts replicated the sex-combined association. This report provides evidence for shared biological mechanisms leading to MDD and VTE and suggests that, in the absence of genetic data, a family history of MDD might be considered when assessing the risk of VTE.

Quantification of liver-Lung shunt fraction on 3D SPECT/CT images for selective internal radiation therapy of liver cancer using CNN-based segmentations and non-rigid registration.

PURPOSE: Selective internal radiation therapy (SIRT) has been proven to be an effective treatment for hepatocellular carcinoma (HCC) patients. In clinical practice, the treatment planning for SIRT using 90Y microspheres requires estimation of the liver-lung shunt fraction (LSF) to avoid radiation pneumonitis. Currently, the manual segmentation method to draw a region of interest (ROI) of the liver and lung in 2D planar imaging of 99mTc-MAA and 3D SPECT/CT images is inconvenient, time-consuming and observer-dependent. In this study, we propose and evaluate a nearly automatic method for LSF quantification using 3D SPECT/CT images, offering improved performance compared with the current manual segmentation method. METHODS: We retrospectively acquired 3D SPECT with non-contrast-enhanced CT images (nCECT) of 60 HCC patients from a SPECT/CT scanning machine, along with the corresponding diagnostic contrast-enhanced CT images (CECT). Our approach for LSF quantification is to use CNN-based methods for liver and lung segmentations in the nCECT image. We first apply 3D ResUnet to coarsely segment the liver. If the liver segmentation contains a large error, we dilate the coarse liver segmentation into the liver mask as a ROI in the nCECT image. Subsequently, non-rigid registration is applied to deform the liver in the CECT image to fit that obtained in the nCECT image. The final liver segmentation is obtained by segmenting the liver in the deformed CECT image using nnU-Net. In addition, the lung segmentations are obtained using 2D ResUnet. Finally, LSF quantitation is performed based on the number of counts in the SPECT image inside the segmentations. Evaluations and Results: To evaluate the liver segmentation accuracy, we used Dice similarity coefficient (DSC), asymmetric surface distance (ASSD), and max surface distance (MSD) and compared the proposed method to five well-known CNN-based methods for liver segmentation. Furthermore, the LSF error obtained by the proposed method was compared to a state-of-the-art method, modified Deepmedic, and the LSF quantifications obtained by manual segmentation. The results show that the proposed method achieved a DSC score for the liver segmentation that is comparable to other state-of-the-art methods, with an average of 0.93, and the highest consistency in segmentation accuracy, yielding a standard deviation of the DSC score of 0.01. The proposed method also obtains the lowest ASSD and MSD scores on average (2.6 mm and 31.5 mm, respectively). Moreover, for the proposed method, a median LSF error of 0.14% is obtained, which is a statically significant improvement to the state-of-the-art-method (p=0.004), and is much smaller than the median error in LSF manual determination by the medical experts using 2D planar image (1.74% and p<0.001). CONCLUSIONS: A method for LSF quantification using 3D SPECT/CT images based on CNNs and non-rigid registration was proposed, evaluated and compared to state-of-the-art techniques. The proposed method can quantitatively determine the LSF with high accuracy and has the potential to be applied in clinical practice.

Sooty Tern Optimization Algorithm-Based Deep Learning Model for Diagnosing NSCLC Tumours.

Lung cancer is one of the most common causes of cancer deaths in the modern world. Screening of lung nodules is essential for early recognition to facilitate treatment that improves the rate of patient rehabilitation. An increase in accuracy during lung cancer detection is vital for sustaining the rate of patient persistence, even though several research works have been conducted in this research domain. Moreover, the classical system fails to segment cancer cells of different sizes accurately and with excellent reliability. This paper proposes a sooty tern optimization algorithm-based deep learning (DL) model for diagnosing non-small cell lung cancer (NSCLC) tumours with increased accuracy. We discuss various algorithms for diagnosing models that adopt the Otsu segmentation method to perfectly isolate the lung nodules. Then, the sooty tern optimization algorithm (SHOA) is adopted for partitioning the cancer nodules by defining the best characteristics, which aids in improving diagnostic accuracy. It further utilizes a local binary pattern (LBP) for determining appropriate feature retrieval from the lung nodules. In addition, it adopts CNN and GRU-based classifiers for identifying whether the lung nodules are malignant or non-malignant depending on the features retrieved during the diagnosing process. The experimental results of this SHOA-optimized DNN model achieved an accuracy of 98.32%, better than the baseline schemes used for comparison.

A design of experiments screen reveals that Clostridium novyi-NT spore germinant sensing is stereoflexible for valine and its analogs.

Although Clostridium novyi-NT is an anti-cancer bacterial therapeutic which germinates within hypoxic tumors to kill cancer cells, the actual germination triggers for C. novyi-NT are still unknown. In this study, we screen candidate germinants using combinatorial experimental designs and discover by serendipity that D-valine is a potent germinant, inducing 50% spore germination at 4.2 mM concentration. Further investigation revealed that five D-valine analogs are also germinants and four of these analogs are enantiomeric pairs. This stereoflexible effect of L- and D-amino acids shows that spore germination is a complex process where enantiomeric interactions can be confounders. This study also identifies L-cysteine as a germinant, and hypoxanthine and inosine as co-germinants. Several other amino acids promote (L-valine, L-histidine, L-threonine and L-alanine) or inhibit (L-arginine, L-glycine, L-lysine, L-tryptophan) germination in an interaction-dependent manner. D-alanine inhibits all germination, even in complex growth media. This work lays the foundation for improving the germination efficacy of C. novyi-NT spores in tumors.

Screening and treatment of brain metastasis from papillary thyroid carcinoma: a case series.

BACKGROUND: The brain metastasis from differentiated thyroid carcinoma (DTC) is a rare condition and its prognosis is poor. The standard protocol for screening and treatment of patients with brain metastases from papillary thyroid cancer (PTC) remains controversial. This report aims to share the experience of a single center in the management of brain metastases from DTC. MATERIAL AND METHODS: Five patients with brain metastases were identified from 5000 patients with DTC attending the department of nuclear medicine, Hospital 108 between 2016 to 2022. The statistical software Statistical Package for Social Sciences (SPSS) 20.0 (SPSS Inc., Chicago, IL, USA) was used to analyze the data. RESULTS: Five patients with brain metastases from DTC were revealed by MRI, 18F-FDG PET/CT with contrast enhancement, and 131I-SPECT/CT. The median time of overall survival (OS) was 15 months, ranging from 10 to 65 months. Two out of the five patients underwent surgery, and futher 2 patients were treated with stereotactic surgery (SRS). All patients are still alive. CONCLUSIONS: Brain metastases from DTC are rare. MRI is the preferred imaging mobility to screen brain lesions in DTC. The primary treatment modalities are surgery and SRS.

Measurement of Medial Tibial Eminence Dimensions for the Clinical Evaluation of ACL-Injured Knees: A Comparison between CT and MRI.

Anterior cruciate ligament (ACL) injuries commonly lead to translational and rotational tibiofemoral instability. The morphology of the medial tibial eminence (MTE) has received increased attention regarding its role in tibiofemoral stability in ACL-injured knees. Therefore, quantification of MTE dimensions on clinical imaging may help clinicians predict knee stability after ACL injury. Although magnetic resonance imaging (MRI) is routinely obtained in patients with ACL injuries, whether the dimensions of the MTE can be accurate quantified on MRI is unknown. The purpose of this study was to assess the degree of correlation between measurements of MTE height and width on computed tomography (CT) versus MRI. An institutional picture archiving and communication system imaging database was used to identify patients aged between 15 and 60 years who received concurrent MRI and CT of the same knee within a 1-year interval. Knees with significant arthrosis, deformity, intraarticular fracture, or hardware-related artifact that obscured visualization of the MTE were excluded. Mean differences and interstudy agreement between CT and MRI MTE measurements were compared using concordance correlation coefficient (r c) and Bland-Altman analysis. A total of 41 knees in 38 patients (mean age, 37 years; 82% male) were analyzed. Interrater reliability for CT and MRI measurements was high (intraclass correlation coefficient = 0.740-0.954). On coronal CT and MRI, mean MTE height measurements were 10.4 ± 1.9 and 10.4 ± 1.8 mm, respectively; mean MTE width measurements were 14.6 ± 3.6 and 14.2 ± 3.0 mm, respectively. On sagittal CT and MRI, mean MTE height measurements were 11.6 ± 1.7 and 11.7 ± 1.7 mm, respectively; mean MTE width measurements were 36.5 ± 4.8 and 36.2 ± 5.0 mm, respectively. Good agreement was observed between CT and MRI measurements of MTE height and width on coronal and sagittal planes (r c = 0.947-0.969). Measurements of MTE height and width were similar on MRI relative to CT on both coronal and sagittal planes. MRI may be suitable for characterizing the dimensions of the MTE when clinically evaluating patients with ACL injuries, potentially allowing for individualized patient care.

Automatic Detection of Meniscus Tears Using Backbone Convolutional Neural Networks on Knee MRI.

BACKGROUND: Timely diagnosis of meniscus injuries is key for preventing knee joint dysfunction and improving patient outcomes because it decreases morbidity and facilitates treatment planning. PURPOSE: To train and evaluate a deep learning model for automated detection of meniscus tears on knee magnetic resonance imaging (MRI). STUDY TYPE: Bicentric retrospective study. SUBJECTS: In total, 584 knee MRI studies, divided among training (n = 234), testing (n = 200), and external validation (n = 150) data sets, were used in this study. The public data set MRNet was used as a second external validation data set to evaluate the performance of the model. SEQUENCE: A 3 T, coronal, and sagittal images from T1-weighted proton density (PD) fast spin-echo (FSE) with fat saturation and T2-weighted FSE with fat saturation sequences. ASSESSMENT: The detection system for meniscus tear was based on the improved YOLOv4 model with Darknet-53 as the backbone. The performance of the model was also compared with that of three radiologists of varying levels of experience. The determination of the presence of a meniscus tear from surgery reports was used as the ground truth for the images. STATISTICAL TESTS: Sensitivity, specificity, prevalence, positive predictive value, negative predictive value, accuracy, and receiver operating characteristic curve were used to evaluate the performance of the detection model. Two-way analysis of variance, Wilcoxon signed-rank test, and Tukey's multiple tests were used to evaluate differences in performance between the model and radiologists. RESULTS: The overall accuracies for detecting meniscus tears using our model on the internal testing, internal validation, and external validation data sets were 95.4%, 95.8%, and 78.8%, respectively. One radiologist had significantly lower performance than our model in detecting meniscal tears (accuracy: 0.9025 ± 0.093 vs. 0.9580 ± 0.025). DATA CONCLUSION: The proposed model had high sensitivity, specificity, and accuracy for detecting meniscus tears on knee MRIs. EVIDENCE LEVEL: 3 TECHNICAL EFFICACY: Stage 2.

End-to-end deep learning model for segmentation and severity staging of anterior cruciate ligament injuries from MRI.

PURPOSE: The purpose of this study was to develop a semi-supervised segmentation and classification deep learning model for the diagnosis of anterior cruciate ligament (ACL) tears on MRI based on a semi-supervised framework, double-linear layers U-Net (DCLU-Net). MATERIALS AND METHODS: A total of 297 participants who underwent of total of 303 MRI examination of the knee with fat-saturated proton density (PD) fast spin-echo (FSE) sequence in the sagittal plane were included. There were 214 men and 83 women, with a mean age of 37.46 ± 1.40 (standard deviation) years (range: 29-44 years). Of these, 107 participants had intact ACL (36%), 98 had partially torn ACL (33%), and 92 had fully ruptured ACL (31%). The DCLU-Net was combined with radiomic features for enhancing performances in the diagnosis of ACL tear. The different evaluation metrics for both classification (accuracy, sensitivity, accuracy) and segmentation (mean Dice similarity coefficient and root mean square error) were compared individually for each image class across the three phases of the model, with each value being compared to its respective value from the previous phase. Findings at arthroscopic knee surgery were used as the standard of reference. RESULTS: With the addition of radiomic features, the final model yielded accuracies of 90% (95% CI: 83-92), 82% (95% CI: 73-86), and 92% (95% CI: 87-94) for classifying ACL as intact, partially torn and fully ruptured, respectively. The DCLU-Net achieved mean Dice similarity coefficient and root mean square error of 0.78 (95% CI: 0.71-0.80) and 0.05 (95% CI: 0.06-0.07), respectively, when segmenting the three ACL conditions with pseudo data (P < 0.001). CONCLUSION: A dual-modules deep learning model with segmentation and classification capabilities was successfully developed. In addition, the use of semi-supervised techniques significantly reduced the amount of manual segmentation data without compromising performance.