Innervate Commercial Fabrics with Spirally-Layered Iontronic Fibrous Sensors Toward Dual-Functional Smart Garments.

Electronic fabrics exhibit desirable breathability, wearing comfort, and easy integration with garments. However, surficial deposition of electronically functional materials/compounds onto fabric substrates would consequentially alter their intrinsic properties (e.g., softness, permeability, biocompatibility, etc.). To address this issue, here, a strategy to innervate arbitrary commercial fabrics with unique spirally-layered iontronic fibrous (SLIF) sensors is presented to realize both mechanical and thermal sensing functionalities without sacrificing the intrinsic fabric properties. The mechanical sensing function is realized via mechanically regulating the interfacial ionic supercapacitance between two perpendicular SLIF sensors, while the thermal sensing function is achieved based on thermally modulating the intrinsic ionic impedance in a single SLIF sensor. The resultant SLIF sensor-innervated electronic fabrics exhibit high mechanical sensitivity of 81 N-1, superior thermal sensitivity of 34,400 Ω °C-1, and more importantly, greatly minimized mutual interference between the two sensing functions. As demonstrations, various smart garments are developed for the precise monitoring of diverse human physiological signals. Moreover, artificial intelligence-assisted object recognition with high-accuracy (97.8%) is demonstrated with a SLIF sensor-innervated smart glove. This work opens up a new path toward the facile construction of versatile smart garments for wearable healthcare, human-machine interfaces, and the Internet of Things.

Targeting apoptotic pathways for cancer therapy.

Apoptosis is a form of programmed cell death that is mediated by intrinsic and extrinsic pathways. Dysregulation of and resistance to cell death are hallmarks of cancer. For over three decades, the development of therapies to promote treatment of cancer by inducing various cell death modalities, including apoptosis, has been a main goal of clinical oncology. Apoptosis pathways also interact with other signaling mechanisms, such as the p53 signaling pathway and the integrated stress response (ISR) pathway. In addition to agents directly targeting the intrinsic and extrinsic pathway components, anticancer drugs that target the p53 and ISR signaling pathways are actively being developed. In this Review, we discuss selected and promising anticancer therapies in various stages of development, including drug targets, mechanisms, and resistance to related treatments, focusing especially on B cell lymphoma 2 (BCL-2) inhibitors, TRAIL analogues, DR5 antibodies, and strategies that target p53, mutant p53, and the ISR.

Multi-line-of-sight Hartmann-Shack wavefront sensing based on image segmentation and K-means sorting.

Multi-line-of-sight wavefront sensing, crucial for next-generation astronomy and laser applications, often increases system complexity by adding sensors. This research introduces, to the best of our knowledge, a novel method for multi-line-of-sight Hartmann-Shack wavefront sensing by using a single sensor, addressing challenges in centroid estimation and classification under atmospheric turbulence. This method contrasts with existing techniques that rely on multiple sensors, thereby reducing system complexity. Innovations include combining edge detection and peak extraction for precise centroid calculation, improved k-means clustering for robust centroid classification, and a centroid filling algorithm for subapertures with light loss. The method's effectiveness was confirmed through simulations for a five-line-of-sight system and experimental setup for two-line and three-line-of-sight systems, demonstrating its potential in real atmospheric aberration correction conditions. Experimental findings indicate that, when implemented in a closed-loop configuration, the method significantly reduces wavefront residuals from 1 λ to 0.1 λ under authentic atmospheric turbulence conditions. Correspondingly, the quality of the far-field spot is enhanced by a factor of 2 to 4. These outcomes collectively highlight the method's robust capability in enhancing optical system performance in environments characterized by genuine atmospheric turbulence.

Nonvolatile multilevel adjustable optical switch based on plasmonic slot waveguide and GST segmented structure.

Optical computing has gradually demonstrated its efficiency in handling increasingly complex computational demands, attracting widespread attention. Optical switches can effectively control and modulate optical signals, providing flexibility and efficiency for optical computing systems. However, traditional optical switches face performance issues such as power consumption, switching speed, and compactness, severely limiting the implementation of large-scale photonic integrated circuits and optical neural networks. This paper proposes an innovative design structure for a non-volatile multi-level adjustable optical switch by combining a plasmonic slot waveguide with segmented phase-change materials. Modulation of waveguide light transmission is achieved by adjusting the phase state of Ge2Sb2Te5(GST). At a wavelength of 1550 nm, a low insertion loss of 0.5dB has been achieved, with approximately an 85% difference in optical transmittance between amorphous state (aGST) and crystalline state (cGST). The high transmittance difference contributes to achieving a wide range of weight variations and supports precise weight updates. Based on this design, we successfully implemented a handwritten digit recognition task with an accuracy of 95%, laying the foundation for future more efficient memory computing neural morphic networks.

In situ hydrogel based on Cu-Fe3O4 nanoclusters exploits oxidative stress and the ferroptosis/cuproptosis pathway for chemodynamic therapy.

Chemodynamic therapy (CDT) involving the use of metal nanozymes presents new opportunities for the treatment of deep-seated tumors. However, the lower ROS catalytic rate and dependence on high H2O2 concentrations affect therapeutic efficacy. To address this issue, a hydrogel was constructed for the treatment of osteosarcoma by combining Cu-Fe3O4 nanozymes (NCs) and artemisinin (AS) coencapsulated in situ with sodium alginate (ALG) and calcium ions. This hydrogel can release nanoparticles and AS within tumor tissue for an extended period of time, utilizing the multienzyme activity of NCs to achieve ROS accumulation. The carbon radicals (•C) generated from the interaction of Fe2+/Cu2+ with AS amplify oxidative stress, leading to tumor cell damage. Simultaneously, the NCs activate ferroptosis via the GPX4 pathway by depleting GSH and activate cuproptosis via the DLAT pathway by causing intracellular copper overload, enhancing therapeutic efficacy. In vitro experiments confirmed that the NCs-AS-ALG hydrogel has an excellent tumor cell killing effect, while in vivo experimental results demonstrated that it can effectively eliminate tumors with excellent biocompatibility, providing a new approach for osteosarcoma treatment.

Dopamine pre-treatment impairs the anti-cancer effect of integrated stress response- and TRAIL pathway-inducing ONC201, ONC206 and ONC212 imipridones in pancreatic, colorectal cancer but not DMG cells.

ONC201 (originally discovered as TRAIL-Inducing Compound #10 or TIC10) and analogue ONC206 have been found to induce an integrated stress response with suggested primary targets and mechanisms involving targeting mitochondrial protein ClpP and antagonism of dopamine receptors D2/3 (DRD2/3). We hypothesized that dopamine, the agonist of DRD2, may counteract ONC201 or ONC206 for DRD2/3 and impair the anti-cancer effect of ONC201 or ONC206, thus protect the tumor cells from the cytotoxic effect of ONC201 or ONC206. We therefore pre-treated cancer cells from different tissue origins including breast cancer, pancreatic cancer, colorectal cancer, and diffuse midline glioma (DMG) with dopamine, followed by treatment of ONC201, ONC206 or ONC212. We observed that 48 hours of pre-treatment with dopamine impaired the cell viability suppression effect of ONC201, ONC206 and ONC212 in pancreatic cancer cells and colorectal cancer cells. We pre-treated multiple cancer cell lines with dopamine for one week followed by ONC201, ONC206, or ONC212 treatment and performed colony assays. Pre-treatment with dopamine impaired the anti-cancer effect of ONC201 or ONC206 in pancreatic cancer and colorectal cancer. Impairment of ONC212 effect by pre-treatment with dopamine was also seen in colony assay for colorectal cancer, but not in pancreatic cancer cells by colony assay. No protection from killing by imipridones was observed with DRD2 agonist sumanirole in tumor cells, or with brain tumor cell lines pretreated with dopamine. Immunoblotting was conducted to investigate whether dopamine pre-treatment impacts signaling pathways reported to be affected by ONC201. The dopamine pre-treatment did not impact changes in ATF4, CHOP, DR5 and ClpX which were reported to be affected by ONC201. The mechanism of impairment of ONC201/206/212 effect caused by dopamine pre-treatment appears to involve upregulation of anti-apoptotic p-Bad, XIAP, FLIP and pAkt. Our results shed light on mechanisms of cancer cell protection by dopamine after imipridone treatment, heterogeneity among different tumor cell types, and suggest that effects of dopamine adaptation on tumor cells may impact on cell survival pathways in ways that may or may not depend on expression of dopamine receptors.

An essential gene signature of breast cancer metastasis reveals targetable pathways.

BACKGROUND: The differential gene expression profile of metastatic versus primary breast tumors represents an avenue for discovering new or underappreciated pathways underscoring processes of metastasis. However, as tumor biopsy samples are a mixture of cancer and non-cancer cells, most differentially expressed genes in metastases would represent confounders involving sample biopsy site rather than cancer cell biology. METHODS: By paired analysis, we defined a top set of differentially expressed genes in breast cancer metastasis versus primary tumors using an RNA-sequencing dataset of 152 patients from The Breast International Group Aiming to Understand the Molecular Aberrations dataset (BIG-AURORA). To filter the genes higher in metastasis for genes essential for breast cancer proliferation, we incorporated CRISPR-based data from breast cancer cell lines. RESULTS: A significant fraction of genes with higher expression in metastasis versus paired primary were essential by CRISPR. These 264 genes represented an essential signature of breast cancer metastasis. In contrast, nonessential metastasis genes largely involved tumor biopsy site. The essential signature predicted breast cancer patient outcome based on primary tumor expression patterns. Pathways underlying the essential signature included proteasome degradation, the electron transport chain, oxidative phosphorylation, and cancer metabolic reprogramming. Transcription factors MYC, MAX, HDAC3, and HCFC1 each bound significant fractions of essential genes. CONCLUSIONS: Associations involving the essential gene signature of breast cancer metastasis indicate true biological changes intrinsic to cancer cells, with important implications for applying existing therapies or developing alternate therapeutic approaches.

Phase unwrapping algorithm based on phase diversity wavefront reconstruction and virtual Hartmann-Shack technology.

Phase unwrapping (PU) algorithms play a crucial role in various phase measurement techniques. Traditional algorithms cannot work well in strong noise environments, which makes it very difficult to obtain the accurate absolute phase from the noisy wrapped phase. In this Letter, we introduce a novel, to the best of our knowledge, phase unwrapping algorithm named PD-VHS. This algorithm innovatively employs point spread function (PSF) filtering to eliminate noise from the wrapped phase. Furthermore, it combines a phase diversity (PD) wavefront reconstruction technology with a virtual Hartmann-Shack (VHS) technology for phase reconstruction and phase unwrapping of the filtered PSFs. In simulations, hundreds of random noise wrapped phases, containing the first 45 Zernike polynomials (excluding piston and the two tilt terms) and the wavefront RMS = 0.5λ and 1λ, are used to compare the classical quality-map guided algorithm, the VHS algorithm with decent noise immunity, with our PD-VHS algorithm. When signal-to-noise ratio (SNR) drops to just 2 dB, the mean root mean square errors (RMSEs) of the residual wavefront between the unwrapped result and the absolute phase of the quality-map guided algorithm and the VHS algorithm are up to 3.99λ, 0.44λ, 4.29λ, and 0.85λ, respectively; however, our algorithm RMSEs are low: 0.11λ and 0.17λ. Simulation results demonstrated that the PD-VHS algorithm significantly outperforms the quality-map guided algorithm and the VHS algorithm under large-scale noise conditions.

Chondroitin polymerizing factor promotes development and progression of colorectal cancer via facilitating transcription of VEGFB.

Colorectal cancer (CRC) is a highly prevalent malignancy affecting the digestive system on a global scale. This study aimed to explore the previously unexplored role of CHPF in the progression of CRC. Our results revealed a significant upregulation of CHPF expression in CRC tumour tissues compared to normal tissues, with its levels correlating with tumour malignancy. In vitro experiments using CRC cell lines demonstrated that inhibiting CHPF expression suppressed cell proliferation, colony formation and cell migration, while promoting apoptosis. Conversely, overexpressing CHPF had the opposite effect. Additionally, our xenograft models in mice confirmed the inhibitory impact of CHPF knockdown on CRC progression using various cell models. Mechanistic investigations unveiled that CHPF may enhance VEGFB expression through E2F1-mediated transcription. Functionally, suppressing VEGFB expression successfully mitigated the oncogenic effects induced by CHPF overexpression. Collectively, these findings suggest that CHPF may act as a tumour promoter in CRC, operating in a VEGFB-dependent manner and could be a potential target for therapeutic interventions in CRC treatment.

An in vitro feasibility study of 355 nm laser atherectomy for the treatment of peripheral atherosclerotic lesions.

In this study, we utilized a novel 355 nm laser to ablate porcine aortas in the presence of physiological saline and contrast agent. Subsequently, we investigated the shape and depth of the resulting injuries. After ablating bovine tendons and aortas with the laser, we analyzed the size and quantity of particles postablation. Finally, we conducted ablation experiments using human ex vivo plaques. The analysis revealed minimal damage to porcine aortas within 2 s of exposure to the 355 nm laser. The degree of injury in the presence of contrast agent was higher than that in the presence of physiological saline but significantly lower than the damage caused by 308 nm laser. Regardless of whether it was bovine tendon or porcine aorta tissue, the proportion of particles <25 µm postlaser ablation exceeded 99%. Lastly, the 355 nm laser successfully opened three types of plaques: chronically occluded, stent restenosis, and stale thrombosis.

All-optical neuromorphic XOR and XNOR operation utilizing a photonic spiking neuron based on a passive add-drop microring resonator.

We propose a concise hardware architecture supporting efficient exclusive OR (XOR) and exclusive NOR (XNOR) operations, by employing a single photonic spiking neuron based on a passive add-drop microring resonator (ADMRR). The threshold mechanism and inhibitory dynamics of the ADMRR-based spiking neuron are numerically discussed on the basis of the coupled mode theory. It is shown that a precise XOR operation in the ADMRR-based spiking neuron can be implemented by adjusting temporal differences within the inhibitory window. Additionally, within the same framework, the XNOR function can also be carried out by accumulating the input power over time to trigger an excitatory behavior. This work presents a novel, to the best of our knowledge, and pragmatic technique for optical neuromorphic computing and information processing utilizing passive devices.

Identification and validation of UBE2B as a prognostic biomarker promoting the development of esophageal carcinomas.

Introduction: Ubiquitination is a crucial biological mechanism in humans, essential for regulating vital biological processes, and has been recognized as a promising focus for cancer therapy. Our objective in this research was to discover potential enzymes associated with ubiquitination that may serve as therapeutic targets for individuals with esophageal carcinoma (ESCA). Methods: To identify genes linked to the prognosis of ESCA, we examined mRNA sequencing data from patients with ESCA in the TCGA database. Further investigation into the role of the candidate gene in ESCA was conducted through bioinformatic analyses. Subsequently, we carried out biological assays to assess its impact on ESCA development. Results: Through univariate Cox regression analysis, we identified Ubiquitin Conjugating Enzyme E2 B (UBE2B) as a potential gene associated with the prognosis of ESCA. UBE2B exhibited significant upregulation and was found to be correlated with survival outcomes in ESCA as well as other cancer types. Additionally, UBE2B was observed to be involved in various biological pathways linked to the development of ESCA, including TNF-a signaling via NF-κB, epithelial-mesenchymal transition, inflammatory response, and hypoxia. Moreover, immune-related pathways like B cell activation (GO: 0042113), B cell receptor signaling pathway (GO: 0050853) and B cell mediated immunity (GO:0019724) were also involved. It was found that high expression of UBE2B was correlated with the increase of several kinds of T cells (CD8 T cells, Th1 cells) and macrophages, while effector memory T cell (Tem) and Th17 cells decreased. Furthermore, UBE2B showed potential as a prognostic biomarker for ESCA, displaying high sensitivity and specificity. Notably, proliferation and migration in ESCA cells were effectively suppressed when the expression of UBE2B was knocked down. Conclusions: To summarize, this study has made a discovery regarding the importance of gaining new insights into the role of UBE2B in ESCA. UBE2B might be an oncogene with good ability in predicting and diagnosing ESCA. Consequently, this discovery highlights the feasibility of targeting UBE2B as a viable approach for treating patients with ESCA.

Long-term outcomes of endoscopic resection for well-differentiated nonampullary duodenal neuroendocrine tumors.

BACKGROUND & AIMS: Nonampullary duodenal neuroendocrine tumors (NAD-NETs) are rare with limited evidence regarding endoscopic treatment. The study aimed to investigate the efficacy and safety of endoscopic resection of well-differentiated NAD-NETs and evaluate long-term outcomes, including local recurrence and metastasis. METHODS: A total of 78 patients with NAD-NETs who underwent endoscopic resection between January 2011 and August 2022 were included. The clinicopathologic characteristics and treatment outcomes were collected and analyzed. RESULTS: En bloc resection was achieved for 74 of the tumors (94.9%) and R0 resection was obtained in 68 of the tumors (87.2%). Univariate analysis identified tumors in the second part of the duodenum, tumor size ≥ 10 mm and muscularis propria invasion as risk factors for non-curative resection. Two patients with R1 resection (vertical margin involvement) and two patients with lymphovascular invasion underwent additional surgery. Four patients experienced adverse events (5.1%), including two cases of delayed bleeding and two cases of perforation, all successfully managed conservatively. During a median follow-up period of 62.6 months, recurrence and lymph node metastasis were only detected in one patient with R1 resection 3 months after the original procedure. CONCLUSION: Endoscopic resection is safe and effective and provides a favorable long-term outcome for patients with well-differentiated NAD-NETs without regional lymph node or distant metastasis.

Germline structural variation globally impacts the cancer transcriptome including disease-relevant genes.

Germline variation and somatic alterations contribute to the molecular profile of cancers. We combine RNA with whole genome sequencing across 1,218 cancer patients to determine the extent germline structural variants (SVs) impact expression of nearby genes. For hundreds of genes, recurrent and common germline SV breakpoints within 100 kb associate with increased or decreased expression in tumors spanning various tissues of origin. A significant fraction of germline SV expression associations involves duplication of intergenic enhancers or 3' UTR disruption. Genes altered by both somatic and germline SVs include ATRX and CEBPA. Genes essential in cancer cell lines include BARD1 and IRS2. Genes with both expression and germline SV breakpoint patterns associated with patient survival include GCLM. Our results capture a class of phenotypic variation at work in the disease setting, including genes with cancer roles. Specific germline SVs represent potential cancer risk variants for genetic testing, including those involving genes with targeting implications.

The violent collapse of vapor bubbles in cryogenic liquids.

Vapor bubbles in cryogenic fluids may collapse violently under subcooled and pressurized conditions. Despite important implications for engineering applications such as cavitation erosion in liquid propellant rocket engines, these intense phenomena are still largely unexplored. In this paper, we systematically investigate the ambient conditions leading to the occurrence of violent collapses in liquid nitrogen and analyze their thermodynamic characteristics. Using Brenner's time ratio χ, the regime of violent collapse is identified in the ambient pressure-temperature parameter space. Complete numerical simulations further refine the prediction and illustrate two classes of collapses. At 1 < χ < 10, the collapse is impacted by significant thermal effects and attains only moderate wall velocity. Only when χ > 10 does the collapse show more inertial features. A mechanism analysis pinpoints a critical time when the surrounding liquid enters supercritical state. The ultimate collapse intensity is shown to be closely associated with the dynamics at this moment. Our study provides a fresh perspective to the treatment of cavitation in cryogenic fluids. The findings can be instrumental in engineering design to mitigate adverse effects arising from intense cavitational activities.

Position-Sensing Graph Neural Networks: Proactively Learning Nodes Relative Positions.

Most existing graph neural networks (GNNs) learn node embeddings using the framework of message passing and aggregation. Such GNNs are incapable of learning relative positions between graph nodes within a graph. To empower GNNs with the awareness of node positions, some nodes are set as anchors. Then, using the distances from a node to the anchors, GNNs can infer relative positions between nodes. However, position-aware GNNs (P-GNNs) arbitrarily select anchors, leading to compromising position awareness and feature extraction. To eliminate this compromise, we demonstrate that selecting evenly distributed and asymmetric anchors is essential. On the other hand, we show that choosing anchors that can aggregate embeddings of all the nodes within a graph is NP-complete. Therefore, devising efficient optimal algorithms in a deterministic approach is practically not feasible. To ensure position awareness and bypass NP-completeness, we propose position-sensing GNNs (PSGNNs), learning how to choose anchors in a backpropagatable fashion. Experiments verify the effectiveness of PSGNNs against state-of-the-art GNNs, substantially improving performance on various synthetic and real-world graph datasets while enjoying stable scalability. Specifically, PSGNNs on average boost area under the curve (AUC) more than 14% for pairwise node classification and 18% for link prediction over the existing state-of-the-art position-aware methods. Our source code is publicly available at: https://github.com/ZhenyueQin/PSGNN.

Endoscopic resection for calcifying fibrous tumors of the gastrointestinal tract.

BACKGROUND: Calcifying fibrous tumors (CFTs) are rare mesenchymal lesions that can occur in various sites throughout the body, including the tubular gastrointestinal (GI) tract. AIM: To analyze the clinical findings of 36 patients with GI tract CFTs to provide guidance for diagnosis and treatment. METHODS: This retrospective study included 36 patients diagnosed with CFTs of the GI tract. We collected demographic and clinical information and conducted regular follow-ups to assess for local recurrence. RESULTS: The stomach was the most commonly involved site, accounting for 72.2% of the 36 CFTs. Endoscopic mucosal resection (n = 1, 2.8%), endoscopic submucosal dissection (n = 14, 38.9%), endoscopic full-thickness resection (n = 16, 44.4%), and submucosal tunneling endoscopic resection (n = 5, 13.9%) were used to resect calcifying fibrous tumors. Overall, 34 (94.4%) CFTs underwent complete endoscopic resections with a mean procedure time of 39.8 ± 29.8 min. The average maximum diameter of the tumors was 10.6 ± 4.3 cm. No complications, such as bleeding or perforation, occurred during an average hospital stay of 2.9 ± 1.2 d. In addition, two patients developed new growth of CFTs near the primary tumor sites, and none of the patients developed distant metastases during the follow-up period. CONCLUSION: GI tract CFTs are rare and typically benign tumors that can be effectively managed with endoscopic procedures.

Numerical Simulation on Erosion and Stress on a Small-Diameter Hydrajet Fracturing Tool.

Conventional hydrajet fracturing techniques are often frustrated when they are applied to some specific well types, such as casing-damaged and small-diameter wells. It is of great significance to investigate the erosion and stress on a small-diameter hydrajet fracturing tool during its service and clarify the relevant influencing factors. Based on the solid-liquid two-phase flow theory and erosion model, a numerical simulation was conducted on the erosion and stress on a small-diameter hydrajet fracturing tool by using the computational fluid dynamics approach in order to understand how the inlet flow rate, particle size, and particle mass concentration affect the erosion and stress on the tool. The results show that the erosion on the small-diameter hydrajet fracturing tool is generally a cutting erosion of proppant particles on the tool body. Such erosion occurs on the lower wall of the nozzle, and the erosion at the upper-0° nozzle is higher in rate and smaller in area than that at the lower-180° nozzle. The maximum stress of the small-diameter hydrajet fracturing tool is concentrated on the upper and lower walls of the upper and lower nozzles, especially the lower part inside the upper nozzle. The maximum erosion rate, average erosion rate, and maximum stress on the wall near the nozzle during fracturing increase as the inlet flow rate and particle mass concentration increase and decrease as the proppant particle size increases.

Expression of SOX4 Significantly Predicts the Risk of Lymph Node Metastasis for Patients With Early-Stage Esophageal Squamous Cell Carcinoma.

Esophageal squamous cell carcinoma stands as a notably aggressive malignancy within the digestive system. In cases of early esophageal cancer without lymph node metastasis, endoscopic surgical resection offers a viable alternative, often resulting in improved patient quality of life. However, the paucity of methods to preoperatively ascertain lymph node involvement complicates surgical planning. SOX4 gene was previously found to be highly associated with invasive metastasis in our work through single-cell RNA sequencing on 5 paired tumor/peritumor tissues. This research included the collection of 124 tissue samples from 106 patients (106 tumor and 18 lymph node specimens). Samples were methodically arranged into a tissue microarray and treated with immunohistochemical staining. Statistical analysis was conducted to assess the relationship between them. In the univariate analysis, 3 factors were identified as statistically significant in relation to lymph node metastasis: T category (P = .014), vascular invasion (P < .001), and SOX4 intensity (P = .001). Additionally, when evaluating SOX4 intensity alongside other clinical indicators, SOX4 was shown to independently influence lymph node metastasis. Further, the multivariate analysis revealed that vascular invasion (P < .001) and SOX4 intensity (P = .003) were significantly associated with lymph node metastasis, exhibiting hazard ratios of 10.174 and 7.142, respectively. The results of our study indicate that both SOX4 expression and vascular invasion serve as predictors of lymph node metastasis in patients diagnosed with category T1 esophageal squamous cell carcinoma, underscoring the potential utility of SOX4 in prognostic evaluations.

MAD-Former: A Traceable Interpretability Model for Alzheimer's Disease Recognition Based on Multi-Patch Attention.

The integration of structural magnetic resonance imaging (sMRI) and deep learning techniques is one of the important research directions for the automatic diagnosis of Alzheimer's disease (AD). Despite the satisfactory performance achieved by existing voxel-based models based on convolutional neural networks (CNNs), such models only handle AD-related brain atrophy at a single spatial scale and lack spatial localization of abnormal brain regions based on model interpretability. To address the above limitations, we propose a traceable interpretability model for AD recognition based on multi-patch attention (MAD-Former). MAD-Former consists of two parts: recognition and interpretability. In the recognition part, we design a 3D brain feature extraction network to extract local features, followed by constructing a dual-branch attention structure with different patch sizes to achieve global feature extraction, forming a multi-scale spatial feature extraction framework. Meanwhile, we propose an important attention similarity position loss function to assist in model decision-making. The interpretability part proposes a traceable method that can obtain a 3D ROI space through attention-based selection and receptive field tracing. This space encompasses key brain tissues that influence model decisions. Experimental results reveal the significant role of brain tissues such as the Fusiform Gyrus (FuG) in AD recognition. MAD-Former achieves outstanding performance in different tasks on ADNI and OASIS datasets, demonstrating reliable model interpretability.