A transformer-based multi-task deep learning model for simultaneous T-stage identification and segmentation of nasopharyngeal carcinoma.

Background: Accurate tumor target contouring and T staging are vital for precision radiation therapy in nasopharyngeal carcinoma (NPC). Identifying T-stage and contouring the Gross tumor volume (GTV) manually is a laborious and highly time-consuming process. Previous deep learning-based studies have mainly been focused on tumor segmentation, and few studies have specifically addressed the tumor staging of NPC. Objectives: To bridge this gap, we aim to devise a model that can simultaneously identify T-stage and perform accurate segmentation of GTV in NPC. Materials and methods: We have developed a transformer-based multi-task deep learning model that can perform two tasks simultaneously: delineating the tumor contour and identifying T-stage. Our retrospective study involved contrast-enhanced T1-weighted images (CE-T1WI) of 320 NPC patients (T-stage: T1-T4) collected between 2017 and 2020 at our institution, which were randomly allocated into three cohorts for three-fold cross-validations, and conducted the external validation using an independent test set. We evaluated the predictive performance using the area under the receiver operating characteristic curve (ROC-AUC) and accuracy (ACC), with a 95% confidence interval (CI), and the contouring performance using the Dice similarity coefficient (DSC) and average surface distance (ASD). Results: Our multi-task model exhibited sound performance in GTV contouring (median DSC: 0.74; ASD: 0.97 mm) and T staging (AUC: 0.85, 95% CI: 0.82-0.87) across 320 patients. In early T category tumors, the model achieved a median DSC of 0.74 and ASD of 0.98 mm, while in advanced T category tumors, it reached a median DSC of 0.74 and ASD of 0.96 mm. The accuracy of automated T staging was 76% (126 of 166) for early stages (T1-T2) and 64% (99 of 154) for advanced stages (T3-T4). Moreover, experimental results show that our multi-task model outperformed the other single-task models. Conclusions: This study emphasized the potential of multi-task model for simultaneously delineating the tumor contour and identifying T-stage. The multi-task model harnesses the synergy between these interrelated learning tasks, leading to improvements in the performance of both tasks. The performance demonstrates the potential of our work for delineating the tumor contour and identifying T-stage and suggests that it can be a practical tool for supporting clinical precision radiation therapy.

Cross-Domain Mutual-Assistance Learning Framework for Fully Automated Diagnosis of Primary Tumor in Nasopharyngeal Carcinoma.

Accurate T-staging of nasopharyngeal carcinoma (NPC) holds paramount importance in guiding treatment decisions and prognosticating outcomes for distinct risk groups. Regrettably, the landscape of deep learning-based techniques for T-staging in NPC remains sparse, and existing methodologies often exhibit suboptimal performance due to their neglect of crucial domain-specific knowledge pertinent to primary tumor diagnosis. To address these issues, we propose a new cross-domain mutual-assistance learning framework for fully automated diagnosis of primary tumor using H&N MR images. Specifically, we tackle primary tumor diagnosis task with the convolutional neural network consisting of a 3D cross-domain knowledge perception network (CKP net) for excavated cross-domain-invariant features emphasizing tumor intensity variations and internal tumor heterogeneity, and a multi-domain mutual-information sharing fusion network (M2SF net), comprising a dual-pathway domain-specific representation module and a mutual information fusion module, for intelligently gauging and amalgamating multi-domain, multi-scale T-stage diagnosis-oriented features. The proposed 3D cross-domain mutual-assistance learning framework not only embraces task-specific multi-domain diagnostic knowledge but also automates the entire process of primary tumor diagnosis. We evaluate our model on an internal and an external MR images dataset in a three-fold cross-validation paradigm. Exhaustive experimental results demonstrate that our method outperforms the state-of-the-art algorithms, and obtains promising performance for tumor segmentation and T-staging. These findings underscore its potential for clinical application, offering valuable assistance to clinicians in treatment decision-making and prognostication for various risk groups.

IGF-1R mediates crosstalk between nasopharyngeal carcinoma cells and osteoclasts and promotes tumor bone metastasis.

BACKGROUND: Nasopharyngeal carcinoma (NPC) poses a significant health burden in specific regions of Asia, and some of NPC patients have bone metastases at the time of initial diagnosis. Bone metastasis can cause pathologic fractures and pain, reducing patients' quality of life, and is associated with worse survival. This study aims to unravel the complex role of insulin-like growth factor 1 receptor (IGF-1R) in NPC bone metastasis, offering insights into potential therapeutic targets. METHODS: We assessed IGF-1R expression in NPC cells and explored its correlation with bone metastasis. Experiments investigated the impact of osteoclast-secreted IGF-1 on the IGF-1R/AKT/S6 pathway in promoting NPC cell proliferation within the bone marrow. Additionally, the reciprocal influence of tumor-secreted Granulocyte-macrophage colony-stimulating factor (GM-CSF) on osteoclast differentiation and bone resorption was examined. The effects of IGF-1 neutralizing antibody, IGF-1R specific inhibitor (NVP-AEW541) and mTORC inhibitor (rapamycin) on nasopharyngeal carcinoma bone metastasis were also explored in animal experiments. RESULTS: Elevated IGF-1R expression in NPC cells correlated with an increased tendency for bone metastasis. IGF-1, secreted by osteoclasts, activated the IGF-1R/AKT/S6 pathway, promoting NPC cell proliferation in the bone marrow. Tumor-secreted GM-CSF further stimulated osteoclast differentiation, exacerbating bone resorption. The IGF-1 neutralizing antibody, NVP-AEW541 and rapamycin were respectively effective in slowing down the rate of bone metastasis and reducing bone destruction. CONCLUSION: The intricate interplay among IGF-1R, IGF-1, and GM-CSF highlights potential therapeutic targets for precise control of NPC bone metastasis, providing valuable insights for developing targeted interventions.

Discovery of Taurocholic Acid Sodium Hydrate as a Novel Repurposing Drug for Intervertebral Disc Degeneration by Targeting MAPK3.

OBJECTIVE: Nowadays, more than 90% of people over 50 years suffer from intervertebral disc degeneration (IDD), but there are exist no ideal drugs. The aim of this study is to identify a new drug for IDD. METHODS: An approved small molecular drug library including 2040 small molecular compounds was used here. We found that taurocholic acid sodium hydrate (NAT) could induce chondrogenesis and osteogenesis in mesenchymal stem cells (MSCs). Then, an in vivo mouse model of IDD was established and the coccygeal discs transcriptome analysis and surface plasmon resonance analysis (SPR) integrated with liquid chromatography-tandem mass spectrometry assay (LC-MS) were performed in this study to study the therapy effect and target proteins of NAT for IDD. Micro-CT was used to evaluate the cancellous bone. The expression of osteogenic (OCN, RNX2), chondrogenic (COL2A1, SOX9), and the target related (ERK1/2, p-ERK1/2) proteins were detected. The alkaline phosphatase staining was performed to estimate osteogenic differentiation. Blood routine and blood biochemistry indexes were analyzed for the safety of NAT. RESULTS: The results showed that NAT could induce chondrogenesis and osteogenesis in MSCs. Further experiments confirmed NAT could ameliorate the secondary osteoporosis and delay the development of IDD in mice. Transcriptome analysis identified 128 common genes and eight Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways for NAT. SPR-LC-MS assay detected 57 target proteins for NAT, including MAPK3 (mitogen-activated protein kinase 3), also known as ERK1 (extracellular regulated protein kinase 1). Further verification experiment confirmed that NAT significantly reduced the expression of ERK1/2 phosphorylation. CONCLUSION: NAT would induce chondrogenesis and osteogenesis of MSCs, ameliorate the secondary osteoporosis and delay the progression of IDD in mice by targeting MAPK3.Furthermore, MAPK3, especially the phosphorylation of MAPK3, would be a potential therapeutic target for IDD treatment.

Chronic intermittent hypobaric hypoxia ameliorates vascular reactivity through upregulating adiponectin expression of PVAT in metabolic syndrome rats.

Cumulating evidence demonstrated that chronic intermittent hypobaric hypoxia (CIHH) had beneficial effects on the body. This study investigated the role of perivascular adipose tissue (PVAT) in ameliorating effect of CIHH on vascular reactivity by adiponectin in mesenteric artery of metabolic syndrome (MS) rats. Main methods: 6-week-old male Sprague-Dawley rats were randomly divided into four groups: control (CON), MS model, CIHH treatment, and MS + CIHH treatment group. The size of adipocytes in PVAT was measured by scanning electron microscopy. Serum adiponectin was measured. The microvessel recording technique was used to observe the effect of CIHH on contraction and relaxation in mesenteric artery rings. Also, the expressions of interleukin-1ß, tumor necrosis factor-α, adiponectin, AdipoR1, AdipoR2, APPL1, and endothelial nitric oxide synthase (eNOS) were assayed by Western blotting. Key findings: in MS rats, adipocyte size increased, serum adiponectin decreased, contraction reaction increased while relaxation reaction decreased, the expression of pro-inflammatory cytokines was upregulated, while adiponectin was downregulated in PVAT, and the expressions of AdipoR1, AdipoR2, APPL, and phosphorylated-eNOS were downregulated in mesenteric artery. All aforementioned abnormalities of MS were ameliorated in MS + CIHH rats. We concluded that CIHH treatment improves vascular reactivity through upregulating adiponectin expression and downregulating pro-inflammatory cytokine expression of PVAT in MS rats.

A multi-perspective information aggregation network for automatedT-staging detection of nasopharyngeal carcinoma.

AccurateT-staging is important when planning personalized radiotherapy. However,T-staging via manual slice-by-slice inspection is time-consuming while tumor sizes and shapes are heterogeneous, and junior physicians find such inspection challenging. With inspiration from oncological diagnostics, we developed a multi-perspective aggregation network that incorporated various diagnosis-oriented knowledge which allowed automated nasopharyngeal carcinomaT-staging detection (TSD Net). Specifically, our TSD Net was designed in multi-branch architecture, which can capture tumor size and shape information (basic knowledge), strongly correlated contextual features, and associations between the tumor and surrounding tissues. We defined the association between the tumor and surrounding tissues by a signed distance map which can embed points and tumor contours in higher-dimensional spaces, yielding valuable information regarding the locations of tissue associations. TSD Net finally outputs aT1-T4 stage prediction by aggregating data from the three branches. We evaluated TSD Net by using the T1-weighted contrast-enhanced magnetic resonance imaging database of 320 patients in a three-fold cross-validation manner. The results show that the proposed method achieves a mean area under the curve (AUC) as high as 87.95%. We also compared our method to traditional classifiers and a deep learning-based method. Our TSD Net is efficient and accurate and outperforms other methods.

EGF Contributes to Hypertrophy of Human Ligamentum Flavum via the TGF-ß1/Smad3 Signaling Pathway.

Background: The most common spinal disorder in elderly is lumbar spinal canal stenosis (LSCS). Previous studies showed that ligamentum flavum hypertrophy (LFH) with fibrosis as the main pathological change is one of the pathogenic factors leading to LSCS. Epidermal Growth Factor (EGF) is known to have an intimate relationship with fibrosis in various tissues. Nevertheless, currently, there are few studies regarding EGF in LFH. The effect of EGF on the development of LFH is unknown, and the underlying pathomechanism remains unclear. In this study, we investigated the role of EGF in LFH and its potential molecular mechanism. Methods: First, the expression levels of EGF, phosphorylation of EGF receptor (pEGFR), Transforming growth factor-ß1 (TGF-ß1), Phosphorylated Smad3 (pSmad3), collagen I and collagen III were examined via immunohistochemistry and Western blot in LF tissues from patients with LSCS or Non-LSCS. Second, primary LF cells were isolated from adults with normal LF thickness and were cultured with different concentrations of exogenous EGF with or without erlotinib/TGF-ß1-neutralizing antibody. Results: The results showed that EGF, pEGFR, TGF-ß1, pSmad3, collagen I and collagen III protein expression in the LSCS group was significantly higher than that in the Non-LSCS group. Meanwhile, pEGFR, TGF-ß1, pSmad3, collagen I and collagen III protein expression was significantly enhanced in LF cells after exogenous EGF exposure, which can be notably blocked by erlotinib. In addition, pSmad3, collagen I and collagen III protein expression was blocked by TGF-ß1-neutralizing antibody. Conclusions: EGF promotes the synthesis of collagen I and collagen III via the TGF-ß1/Smad3 signaling pathway, which eventually contributes to LFH.

Protein tyrosine phosphatase receptor type D gene promotes radiosensitivity via STAT3 dephosphorylation in nasopharyngeal carcinoma.

Radiotherapy is essential to the treatment of nasopharyngeal carcinoma (NPC) and acquired or innate resistance to this therapeutic modality is a major clinical problem. However, the underlying molecular mechanisms in the radiation resistance in NPC are not fully understood. Here, we reanalyzed the microarray data from public databases and identified the protein tyrosine phosphatase receptor type D (PTPRD) as a candidate gene. We found that PTPRD was downregulated in clinical NPC tissues and NPC cell lines with its promoter hypermethylated. Functional assays revealed that PTPRD overexpression sensitized NPC to radiation in vitro and in vivo. Importantly, miR-454-3p directly targets PTPRD to inhibit its expression and biological effect. Interestingly, mechanistic analyses indicate that PTPRD directly dephosphorylates STAT3 to enhance Autophagy-Related 5 (ATG5) transcription, resulting in triggering radiation-induced autophagy. The immunohistochemical staining of 107 NPC revealed that low PTPRD and high p-STAT3 levels predicted poor clinical outcome. Overall, we showed that PTPRD promotes radiosensitivity by triggering radiation-induced autophagy via the dephosphorylation of STAT3, thus providing a potentially useful predictive biomarker for NPC radiosensitivity and drug target for NPC radiosensitization.

CYLD Promotes Apoptosis of Nasopharyngeal Carcinoma Cells by Regulating NDRG1.

PURPOSE: Nasopharyngeal carcinoma (NPC) is among the most common malignancies derived from the epithelium of the nasopharynx. To date, the regulatory networks involved in NPC have not been fully identified. Previous studies revealed multiple loss-of-function mutations in NPC and specifically in cylindromatosis lysine 63 deubiquitinase (CYLD); however, the exact role of CYLD in NPC progression and its potential mechanism remains unclear. METHODS: We performed immunohistochemical (IHC) staining and real-time quantitative polymerase chain reaction (qPCR) to measure CYLD expression in NPC tissues, and Western blot was conducted to determine CYLD levels in NPC cell lines. Cell proliferation was detected by CCK8 assay and colony formation analysis, and apoptosis was determined by Annexin V/propidium iodide staining. Potential targets of CYLD were verified by co-immunoprecipitation and mass spectrometry. Xenograft assay was conducted to confirm the role of CYLD in vivo. RESULTS: We found that CYLD levels were significantly decreased in both NPC tissues and cell lines, and that CYLD overexpression inhibited NPC cell proliferation and promoted apoptosis. Additionally, we revealed that CYLD bound and upregulated N-Myc downstream regulated 1 (NDRG1), and that silencing NDRG1 abolished the tumor-suppressor effect of CYLD on NPC cells. Furthermore, CYLD suppressed tumor growth in xenograft mice models. CONCLUSION: These results suggest CYLD as a tumor suppressor, potential biomarker for diagnosing NPC, and therapeutic target.

Defactinib attenuates osteoarthritis by inhibiting positive feedback loop between H-type vessels and MSCs in subchondral bone.

BACKGROUND: Abnormal bone formation in subchondral bone resulting from uncoupled bone remodeling is considered a central feature in osteoarthritis (OA) pathogenesis. H-type vessels can couple angiogenesis and osteogenesis. We previously revealed that elevated H-type vessels in subchondral bone were correlated with OA and focal adhesion kinase (FAK) in MSCs is critical for H-type vessel formation in osteoporosis. The aim of this study was to explore the correlation between H-type vessels and MSCs in OA pathogenesis through regulation of H-type vessel formation using defactinib (an FAK inhibitor). METHODS: In vivo: 3-month-old male C57BL/6J (WT) mice were randomly divided into three groups: sham controls, vehicle-treated ACLT mice, and defactinib-treated ACLT mice (25 mg/kg, intraperitoneally weekly). In vitro: we explored the role of conditioned medium (CM) of MSCs from subchondral bone of different groups on the angiogenesis of endothelial cells (ECs). Flow cytometry, Western blotting, ELISA, real time (RT)-PCR, immunostaining, CT-based microangiography, and bone micro-CT (µCT) were used to detect changes in relative cells and tissues. RESULTS: This study demonstrated that inhibition of H-type vessels with defactinib alleviated OA by inhibiting H-type vessel-linked MSCs in subchondral bone. During OA pathogenesis, H-type vessels and MSCs formed a positive feedback loop contributing to abnormal bone formation in subchondral bone. Elevated H-type vessels provided indispensable MSCs for abnormal bone formation in subchondral bone. Flow cytometry and immunostaining results confirmed that the amount of MSCs in subchondral bone was obviously higher in vehicle-treated ACLT mice than that in sham controls and defactinib-treated ACLT mice. In vitro, p-FAK in MSCs from subchondral bone of vehicle-treated ALCT mice increased significantly relative to other groups. Further, the CM from MSCs of vehicle-treated ACLT mice enhanced angiogenesis of ECs through FAK-Grb2-MAPK-linked VEGF expression. CONCLUSIONS: Our results demonstrate that defactinib inhibits OA by suppressing the positive feedback loop between H-type vessels and MSCs in subchondral bone. THE TRANSLATIONAL POTENTIAL OF THIS ARTICLE: Our results provide a mechanistic rationale for the use of defactinib as an effective candidate for OA treatment.

Copper-Containing Alloy as Immunoregulatory Material in Bone Regeneration via Mitochondrial Oxidative Stress.

In the mammalian skeletal system, osteogenesis and angiogenesis are closely linked by type H vessels during bone regeneration and repair. Our previous studies confirmed the promotion of these processes by copper-containing metal (CCM) in vitro and in vivo. However, whether and how the coupling of angiogenesis and osteogenesis participates in the promotion of bone regeneration by CCM in vivo is unknown. In this study, M2a macrophages but not M2c macrophages were shown to be immunoregulated by CCM. A CCM, 316L-5Cu, was applied to drilling hole injuries of the tibia of C57/6 mice for comparison. We observed advanced formation of cortical bone and type H vessels beneath the new bone in the 316L-5Cu group 14 and 21 days postinjury. Moreover, the recruitment of CD206-positive M2a macrophages, which are regarded as the primary source of platelet-derived growth factor type BB (PDGF-BB), was significantly promoted at the injury site at days 14 and 21. Under the stimulation of CCM, mitochondria-derived reactive oxygen species were also found to be upregulated in CD206hi M2a macrophages in vitro, and this upregulation was correlated with the expression of PDGF-BB. In conclusion, our results indicate that CCM promotes the evolution of callus through the generation of type H vessels during the process of bone repair by upregulating the expression of PDGF-BB derived from M2a macrophages.

SDF-1/CXCR4 axis coordinates crosstalk between subchondral bone and articular cartilage in osteoarthritis pathogenesis.

Crosstalk between subchondral bone and articular cartilage is considered a central feature of osteoarthritis (OA) initiation and progression, but its underlying molecular mechanism remains elusive. Meanwhile, specific administration of drugs in subchondral bone is also a great challenge during investigation of the process. We here explore the role of stromal cell-derived factor 1 (SDF-1)/C-X-C chemokine receptor type 4 (CXCR4) axis in the crosstalk between subchondral bone and articular cartilage in OA pathogenesis, using osmotic infusion pumps implanted in tibial subchondral bone directly to ensure quantitative, continuous and steady drug delivery over the entire experiment. We found that increased SDF-1 in subchondral bone firstly induced subchondral bone deterioration by erroneous Mesenchymal Stem Cells (MSCs) recruitment and excessive bone resorption in anterior cruciate ligament transection (ACLT) mice. Deterioration of subchondral bone then led to the traverse of SDF-1 from subchondral bone to overlying cartilage. Finally, SDF-1 from underlying subchondral bone combined with CXCR4 in chondrocytes to induce articular cartilage degradation by promoting the shift of transforming growth factor-ß receptor type I (TßRI) in chondrocytes from activin receptor-like kinase 5 (ALK5) to activin receptor-like kinase 1 (ALK1). More importantly, specific inhibition of SDF-1/CXCR4 axis in ACLT rats attenuated OA by stabilizing subchondral bone microarchitecture, reducing SDF-1 in cartilage and abrogating the shift of TßRI in chondrocytes. Our data demonstrate that the SDF-1/CXCR4 axis may coordinate the crosstalk between subchondral bone and articular cartilage in OA pathogenesis. Therefore, specific inhibition of SDF-1/CXCR4 axis in subchondral bone or intervention in SDF-1 traverse may be therapeutic targets for OA.

Deep-learning-based detection and segmentation of organs at risk in nasopharyngeal carcinoma computed tomographic images for radiotherapy planning.

OBJECTIVE: Accurate detection and segmentation of organs at risks (OARs) in CT image is the key step for efficient planning of radiation therapy for nasopharyngeal carcinoma (NPC) treatment. We develop a fully automated deep-learning-based method (termed organs-at-risk detection and segmentation network (ODS net)) on CT images and investigate ODS net performance in automated detection and segmentation of OARs. METHODS: The ODS net consists of two convolutional neural networks (CNNs). The first CNN proposes organ bounding boxes along with their scores, and then a second CNN utilizes the proposed bounding boxes to predict segmentation masks for each organ. A total of 185 subjects were included in this study for statistical comparison. Sensitivity and specificity were performed to determine the performance of the detection and the Dice coefficient was used to quantitatively measure the overlap between automated segmentation results and manual segmentation. Paired samples t tests and analysis of variance were employed for statistical analysis. RESULTS: ODS net provides an accurate detection result with a sensitivity of 0.997 to 1 for most organs and a specificity of 0.983 to 0.999. Furthermore, segmentation results from ODS net correlated strongly with manual segmentation with a Dice coefficient of more than 0.85 in most organs. A significantly higher Dice coefficient for all organs together (p = 0.0003 < 0.01) was obtained in ODS net (0.861 ± 0.07) than in fully convolutional neural network (FCN) (0.8 ± 0.07). The Dice coefficients of each OAR did not differ significantly between different T-staging patients. CONCLUSION: The ODS net yielded accurate automated detection and segmentation of OARs in CT images and thereby may improve and facilitate radiotherapy planning for NPC. KEY POINTS: • A fully automated deep-learning method (ODS net) is developed to detect and segment OARs in clinical CT images. • This deep-learning-based framework produces reliable detection and segmentation results and thus can be useful in delineating OARs in NPC radiotherapy planning. • This deep-learning-based framework delineating a single image requires approximately 30 s, which is suitable for clinical workflows.