Spatial features of skip lesions in Crohn's disease.

Skip lesions are an enigmatic spatial feature characterizing Crohn's disease (CD). They comprise inflamed and adjacent non-inflamed tissue sections with a clear demarcation. Currently, spatial features of the human gastrointestinal (GI) system lack clarity regarding the organization of microbes, mucus, tissue, and host cells during inflammation. New technologies with multiplexing abilities and innovative approaches provide ways of examining the spatial organization of inflamed and non-inflamed tissues in CD, which may open new avenues for diagnosis, prognosis, and treatment. In this review, we present evidence of the relevance of spatial context in patients with CD and the methods and ideas recently published in studies of spatiality during inflammation. With this review, we aim to provide inspiration for further research to address existing gaps.

Effective multi-modal clustering method via skip aggregation network for parallel scRNA-seq and scATAC-seq data.

In recent years, there has been a growing trend in the realm of parallel clustering analysis for single-cell RNA-seq (scRNA) and single-cell Assay of Transposase Accessible Chromatin (scATAC) data. However, prevailing methods often treat these two data modalities as equals, neglecting the fact that the scRNA mode holds significantly richer information compared to the scATAC. This disregard hinders the model benefits from the insights derived from multiple modalities, compromising the overall clustering performance. To this end, we propose an effective multi-modal clustering model scEMC for parallel scRNA and Assay of Transposase Accessible Chromatin data. Concretely, we have devised a skip aggregation network to simultaneously learn global structural information among cells and integrate data from diverse modalities. To safeguard the quality of integrated cell representation against the influence stemming from sparse scATAC data, we connect the scRNA data with the aggregated representation via skip connection. Moreover, to effectively fit the real distribution of cells, we introduced a Zero Inflated Negative Binomial-based denoising autoencoder that accommodates corrupted data containing synthetic noise, concurrently integrating a joint optimization module that employs multiple losses. Extensive experiments serve to underscore the effectiveness of our model. This work contributes significantly to the ongoing exploration of cell subpopulations and tumor microenvironments, and the code of our work will be public at https://github.com/DayuHuu/scEMC.

SKIP-HOPS recruits TBC1D15 for a Rab7-to-Arl8b identity switch to control late endosome transport.

The endolysosomal system fulfils a myriad of cellular functions predicated on regulated membrane identity progressions, collectively termed maturation. Mature or "late" endosomes are designated by small membrane-bound GTPases Rab7 and Arl8b, which can either operate independently or collaborate to form a joint compartment. Whether, and how, Rab7 and Arl8b resolve this hybrid identity compartment to regain functional autonomy is unknown. Here, we report that Arl8b employs its effector SKIP to instigate inactivation and removal of Rab7 from select membranes. We find that SKIP interacts with Rab7 and functions as its negative effector, delivering the cognate GAP, TBC1D15. Recruitment of TBC1D15 to SKIP occurs via the HOPS complex, whose assembly is facilitated by contacts between Rab7 and the KMI motif of SKIP. Consequently, SKIP mediates reinstatement of single identity Arl8b sub-compartment through an ordered Rab7-to-Arl8b handover, and, together with Rab7's positive effector RILP, enforces spatial, temporal and morphological compartmentalization of endolysosomal organelles.

Predicting drug-target binding affinity through molecule representation block based on multi-head attention and skip connection.

Exiting computational models for drug-target binding affinity prediction have much room for improvement in prediction accuracy, robustness and generalization ability. Most deep learning models lack interpretability analysis and few studies provide application examples. Based on these observations, we presented a novel model named Molecule Representation Block-based Drug-Target binding Affinity prediction (MRBDTA). MRBDTA is composed of embedding and positional encoding, molecule representation block and interaction learning module. The advantages of MRBDTA are reflected in three aspects: (i) developing Trans block to extract molecule features through improving the encoder of transformer, (ii) introducing skip connection at encoder level in Trans block and (iii) enhancing the ability to capture interaction sites between proteins and drugs. The test results on two benchmark datasets manifest that MRBDTA achieves the best performance compared with 11 state-of-the-art models. Besides, through replacing Trans block with single Trans encoder and removing skip connection in Trans block, we verified that Trans block and skip connection could effectively improve the prediction accuracy and reliability of MRBDTA. Then, relying on multi-head attention mechanism, we performed interpretability analysis to illustrate that MRBDTA can correctly capture part of interaction sites between proteins and drugs. In case studies, we firstly employed MRBDTA to predict binding affinities between Food and Drug Administration-approved drugs and severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) replication-related proteins. Secondly, we compared true binding affinities between 3C-like proteinase and 185 drugs with those predicted by MRBDTA. The final results of case studies reveal reliable performance of MRBDTA in drug design for SARS-CoV-2.

Long-term outcomes of great saphenous vein harvest techniques for infrainguinal arterial bypass in a Medicare-matched registry database.

OBJECTIVE: Long-term outcomes for harvesting techniques for great saphenous vein (GSV) and its impact on the outcomes of infrainguinal arterial bypass remains largely unknown. Endoscopic GSV harvesting (EVH) has emerged as a less invasive alternative to conventional open techniques. Using the Vascular Quality initiative Vascular Implant Surveillance & Interventional Outcomes Network (VQI-VISION) database, we compared the long-term outcomes of infrainguinal arterial bypass using open and endoscopic GSV harvest techniques. METHODS: Patients who underwent infrainguinal GSV bypass between 2010 and 2019 were identified in the VQI-VISION Medicare linked database. Long-term outcomes of major/minor amputations, and reinterventions up to 5 years of follow-up were compared between continuous incisions, skip incision, and EVH, with continuous incisions being the reference group. Secondary outcomes included 30- and 90-day readmission, in addition to surgical site infections and patency rates at 6 months to 2 years postoperatively. Survival analysis using Kaplan-Meier curves and Cox regression hazard models were utilized to compare outcomes between groups. To adjust for multiple comparisons between the study groups, a P value of 2.5% was considered significant. RESULTS: Among the 8915 patients included in the study, continuous and skip vein harvest techniques were used in 44.4% and 43.4% of cases each, whereas 12.3% underwent EVH. The utilization of EVH remained relatively stable at around 12% throughout the study period. Compared with GSV harvest using continuous incisions, EVH was associated with higher rates of reintervention at 1 year (46.5% vs 41.3%; adjusted hazard ratio [aHR], 1.22; 95% confidence interval [CI], 1.06-1.41; P = .01]. However, no significant difference was observed between EVH and continuous incisions, and between skip and continuous incisions in terms of long-term reintervention or major and minor amputations on adjusted analysis. Compared with continuous incision vein harvest, both EVH and skip incisions were associated with lower surgical site infection rates within the first 6 months post-bypass (aHR, 0.53; 95% CI, 0.35-0.82 and aHR, 0.68; 95% CI, 0.53-0.87, respectively). Loss of primary, primary-assisted, and secondary patency was higher after EVH compared with continuous incision vein harvest. Among surgeons performing EVH, comparable long-term outcomes were observed regardless of low (<4 cases/year), medium (4-7 cases/year), or high procedural volumes (>7 cases/year). CONCLUSIONS: Despite higher 1-year reintervention rates, EVH for infrainguinal arterial bypass is not associated with a significant difference in long-term reintervention or amputation rates compared with other harvesting techniques. These outcomes are not influenced by procedural volumes for EVH technique.

The impact of genotype on age at loss of ambulation in individuals with Duchenne muscular dystrophy treated with corticosteroids: A single-center study of 555 patients.

INTRODUCTION/AIMS: Studies have demonstrated that certain genotypes in Duchenne muscular dystrophy (DMD) have milder or more severe phenotypes. These studies included individuals treated and not treated with corticosteroids and multiple sites with potentially varying standards of care. We aimed to assess genotype-phenotype correlations for age at loss of ambulation (LoA) in a large cohort of individuals with DMD treated with corticosteroids at one center. METHODS: In this retrospective review of medical records, encounters were included for individuals diagnosed with DMD if prescribed corticosteroids, defined as daily deflazacort or prednisone or high-dose weekend prednisone, for 12 consecutive months. Encounters were excluded if the participants were taking disease-modifying therapy. Data were analyzed using survival analysis for LoA and Fisher's exact tests to assess the percentage of late ambulatory (>14 years old) individuals for selected genotypes. RESULTS: Overall, 3948 encounters from 555 individuals were included. Survival analysis showed later age at LoA for exon 44 skip amenable (p = .004), deletion exons 3-7 (p < .001) and duplication exon 2 (p = .043) cohorts and earlier age at LoA for the exon 51 skip amenable cohort (p < .001) when compared with the rest of the cohort. Individuals with deletions of exons 3-7 had significantly more late ambulatory individuals than other cohorts (75%), while those with exon 51 skip amenable deletions had significantly fewer (11.9%) compared with other cohorts. DISCUSSION: This confirms previous observations of genotype-phenotype correlations in DMD and enhances information for trial design and clinical management.

Outcomes of non-contiguous two-level anterior cervical discectomy and fusion in patients with degenerative cervical myelopathy: a retrospective study.

BACKGROUND: Non-contiguous two-level Anterior Cervical Discectomy and Fusion (ACDF) may be a viable option for patients with degenerative cervical myelopathy and imaging-evident spine and radicular compression at two non-contiguous cervical levels. The risk of hastening degeneration and triggering Adjacent Segment Disease at the spine levels located between the fused levels is a putative adverse event, which was assessed in a few studies. The aim of this study is to investigate the clinical outcomes of patients undergoing non-contiguous two levels ACDF and to assess biomechanical modifications at non-fused segments. METHOD: We retrospectively reviewed all patients with noncontiguous two-level spine and radicular compression, who underwent simultaneous noncontiguous two-level ACDF at our center. We analyzed clinical and radiological outcomes and investigated the rate of adjacent segment disease. Radiographic parameters were calculated on pre- and postoperative images. RESULTS: Thirty-two patients underwent simultaneous noncontiguous two-level ACDF for cervical myelo-radiculopathy between 2015 and 2021 and were followed up for a mean period of 43.3 months. For all patients, the mJOA score significantly improved from 14.57 ± 2.3 to 16.5 ± 2.1 (p<0.01) and the NDI score significantly decreased from 21.45 ± 4.3 to 12.8 ± 2.3 (p<0.01) postoperatively. Cervical lordosis increased after surgery (from 9.65° ±9.47 to 15.12° ± 6.09); intermediate disc height decreased (5.68 mm ± 0.57 to 5.27 mm ±0.98); the ROMs of intermediate (from 12.45 ± 2.33 to 14.77 ± 1.98), cranial (from 14.63 ± 1.59 to 15.71 ± 1.02), and caudal (from 11.58 ± 2.32 to 13.33 ± 2.67) segments slightly increased. During follow-up assessment, in one patient the myelopathy worsened due to spine compression at the intermediate level. CONCLUSIONS: Simultaneous and non-contiguous two-level ACDF is a safe and effective procedure. The occurrence of postoperative adjacent and intermediate segment disease is rare.

Medical image segmentation of gastric adenocarcinoma based on dense connection of residuals.

BACKGROUND AND OBJECTIVE: Accurate segmentation of gastric cancer based on CT images of gastric adenocarcinoma is crucial for physicians to screen gastric diseases, clinical diagnosis, preoperative prediction, and postoperative evaluation plans. To address the issue of the inability of the segmentation algorithm to depict the correct boundaries due to unclear gastric contours in the lesion area and the visible irregular band-like dense shadow extending to the perigastric region, a 3D medical image segmentation model 3D UNet based on residual dense jumping method is proposed. METHODS: In the method we proposed, Residual Dense Block, which is applied to the image super-resolution module to remove CT artifacts, and Residual Block in ResNet are further fused. The quality of CT images is improved by Residual Dense Skip Block, which removes banded dense shadows, preserves image details and edge information, captures features, and improves the segmentation performance of gastric adenocarcinoma. The Instance Normalization layer position is modified to select the best result. Different loss functions are also combined in order to obtain the best gastric adenocarcinoma segmentation performance. RESULTS: We tested the model on a hospital-provided gastric adenocarcinoma dataset. The experimental results show that our model outperforms the existing methods in CT gastric adenocarcinoma segmentation, in which the method combining the hybrid loss function of Dice and CE obtains an average dice score of 82.3%, which is improved by 5.3% and 3.8% compared to TransUNet and Hiformer, respectively, and improves the cross-merge rate to 70.8%, compared to nnFormer, nnUNet by 1% and 0.9%, respectively. CONCLUSIONS: The residual jump connection structure indeed improves segmentation performance. The proposed method has the potential to be used as a screen for gastric diseases and to assist physicians in diagnosis.

Risk factors and impact of occult and skip metastasis in early-stage oral tongue squamous cell carcinoma.

OBJECTIVES: To investigate risk factors associated with occult lymph node metastases (ONM) and skip metastasis in early-stage oral tongue squamous cell carcinoma (OTSCC) patients. Meanwhile, to analyze the contribution of metastatic nodes to survival outcomes. MATERIALS AND METHODS: 544 OTSCC patients who were clinically staged T1-T2N0 with pathologic results from May 2018 to January 2024 were enrolled. Those with ONM were divided into subgroups with or without skip metastasis. Clinical, laboratorial, radiological and pathological factors between groups were analyzed by using univariate analysis and multivariate logistic analysis. The association of tumor growth behavior with the metastatic pattern of lymph nodes was summarized. Additionally, disease free survival (DFS) among different groups were compared using Kaplan-Meier analysis. RESULTS: Tumor growth behavior was associated with ONM. Tumor thickness with a threshold of 6.4 mm was not inferior to histological depth of invasion in predicting ONM. Only 1.3% of patients had nodal involvement of neck level IV or V. The DFS of patients with ONM were significantly reduced than those without ONM (P < 0.001). The DFS between patients with and without skip metastasis exhibited no statistical significance(P = 0.246). The 1-year, 2-year recurrence rates of patients with or without ONM were 31.9%, 37.5%, 10.1% and 14.0%, correspondingly. CONCLUSIONS: Tumor thickness with a threshold of 6.4 mm could be used as a preoperative predictor for ONM. Elective neck dissection of level I - III might be sufficient for early stage OTSCC patients. OTSCC patients with ONM should be closely observed during the first 2 years after surgery. CLINICAL RELEVANCE: The risk of ONM in early stage OTSCC patients might be predicted by tumor thickness calculated on MR imaging. Elective neck dissection of level I - III could remove micrometastases timely and effectively.

MASDF-Net: A Multi-Attention Codec Network with Selective and Dynamic Fusion for Skin Lesion Segmentation.

Automated segmentation algorithms for dermoscopic images serve as effective tools that assist dermatologists in clinical diagnosis. While existing deep learning-based skin lesion segmentation algorithms have achieved certain success, challenges remain in accurately delineating the boundaries of lesion regions in dermoscopic images with irregular shapes, blurry edges, and occlusions by artifacts. To address these issues, a multi-attention codec network with selective and dynamic fusion (MASDF-Net) is proposed for skin lesion segmentation in this study. In this network, we use the pyramid vision transformer as the encoder to model the long-range dependencies between features, and we innovatively designed three modules to further enhance the performance of the network. Specifically, the multi-attention fusion (MAF) module allows for attention to be focused on high-level features from various perspectives, thereby capturing more global contextual information. The selective information gathering (SIG) module improves the existing skip-connection structure by eliminating the redundant information in low-level features. The multi-scale cascade fusion (MSCF) module dynamically fuses features from different levels of the decoder part, further refining the segmentation boundaries. We conducted comprehensive experiments on the ISIC 2016, ISIC 2017, ISIC 2018, and PH2 datasets. The experimental results demonstrate the superiority of our approach over existing state-of-the-art methods.

ODNet: A High Real-Time Network Using Orthogonal Decomposition for Few-Shot Strip Steel Surface Defect Classification.

Strip steel plays a crucial role in modern industrial production, where enhancing the accuracy and real-time capabilities of surface defect classification is essential. However, acquiring and annotating defect samples for training deep learning models are challenging, further complicated by the presence of redundant information in these samples. These issues hinder the classification of strip steel surface defects. To address these challenges, this paper introduces a high real-time network, ODNet (Orthogonal Decomposition Network), designed for few-shot strip steel surface defect classification. ODNet utilizes ResNet as its backbone and incorporates orthogonal decomposition technology to reduce the feature redundancies. Furthermore, it integrates skip connection to preserve essential correlation information in the samples, preventing excessive elimination. The model optimizes the parameter efficiency by employing Euclidean distance as the classifier. The orthogonal decomposition not only helps reduce redundant image information but also ensures compatibility with the Euclidean distance requirement for orthogonal input. Extensive experiments conducted on the FSC-20 benchmark demonstrate that ODNet achieves superior real-time performance, accuracy, and generalization compared to alternative methods, effectively addressing the challenges of few-shot strip steel surface defect classification.

Predictors of Skip Laminotomy for Placement of Paddle Leads for Spinal Cord Stimulation.

OBJECTIVES: Placement of a standard paddle lead for spinal cord stimulation (SCS) requires a laminotomy for positioning of the lead within the epidural space. During initial placement, an additional laminotomy or laminectomy, termed a "skip" laminotomy, may be necessary at a higher level to pass the lead to the appropriate midline position. Patient and radiographic factors that predict the need for a skip laminotomy have yet to be identified. MATERIALS AND METHODS: Participants who underwent SCS paddle placement at Albany Medical Center between 2016 and 2017 were identified. Operative reports were reviewed to identify the paddle type, level of initial laminotomy, target level, and skip laminotomy level. Preoperative thoracic magnetic resonance images (MRIs) were reviewed, and spinal canal diameter, interpedicular distance, and dorsal cerebral spinal fluid thickness were measured for each participant when available. RESULTS: A total of 106 participants underwent thoracic SCS placement. Of these, 97 had thoracic MRIs available for review. Thirty-eight participants required a skip laminotomy for placement of the paddle compared with 68 participants who did not. There was no significant difference in demographic features including age, sex, body mass index, and surgical history. Univariate analyses that suggested trends were selected for further analysis using binary logistic regression. Level of initial laminotomy (odds ratio [OR] = 1.51, p = 0.028), spinal canal diameter (OR = 0.71, p = 0.015), and dorsal cerebrospinal fluid thickness (OR = 0.61, p = 0.011) were correlated with skip laminotomy. Target level (OR = 1.27, p = 0.138) and time from trial (1.01, p = 0.117) suggested potential association. The multivariate regression was statistically significant, X2(10) = 28.02, p = 0.002. The model explained 38.3% of the variance (Nagelkerke R2) and predicted skip laminectomy correctly in 73.3% of cases. However, for the multivariate regression, only a decrease in spinal canal diameter (OR = 0.59, p = 0.041) was associated with a greater odds of skip laminotomy. CONCLUSIONS: This study aims to characterize the patient and radiographic factors that may predict the need to perform a skip laminotomy during the initial placement of SCS paddles. Here, we show that radiographic and anatomic variables, primarily spinal canal diameter, play an important role in predicting the need for a skip laminotomy. Furthermore, we suggest that target level for placement and level of initial laminotomy also may contribute. Further investigation of the predictive factors for performing a skip laminotomy would help optimize surgical planning and preoperative patient selection and counseling.

A new deep learning technique reveals the exclusive functional contributions of individual cancer mutations.

Cancers are caused by genomic alterations that may be inherited, induced by environmental carcinogens, or caused due to random replication errors. Postinduction of carcinogenicity, mutations further propagate and drastically alter the cancer genomes. Although a subset of driver mutations has been identified and characterized to date, most cancer-related somatic mutations are indistinguishable from germline variants or other noncancerous somatic mutations. Thus, such overlap impedes appreciation of many deleterious but previously uncharacterized somatic mutations. The major bottleneck arises due to patient-to-patient variability in mutational profiles, making it difficult to associate specific mutations with a given disease outcome. Here, we describe a newly developed technique Continuous Representation of Codon Switches (CRCS), a deep learning-based method that allows us to generate numerical vector representations of mutations, thereby enabling numerous machine learning-based tasks. We demonstrate three major applications of CRCS; first, we show how CRCS can help detect cancer-related somatic mutations in the absence of matched normal samples, which has applications in cell-free DNA-based assessment of tumor mutation burden. Second, the proposed approach also enables identification and exploration of driver genes; our analyses implicate DMD, RSK4, OFD1, WDR44, and AFF2 as potential cancer drivers. Finally, we used CRCS to score individual mutations in a tumor sample, which was found to be predictive of patient survival in bladder urothelial carcinoma, hepatocellular carcinoma, and lung adenocarcinoma. Taken together, we propose CRCS as a valuable computational tool for analysis of the functional significance of individual cancer mutations.

Survival outcome of upfront surgery for clinical single-station N2 non-small cell lung cancer.

BACKGROUND: Pathological N2 (pN2) non-small cell lung cancer (NSCLC) is diverse; its treatment depends on the clinical N (cN) status. We aimed to determine the efficacy of upfront surgery for cN2pN2 NSCLC. METHODS: The study included 43 cN2pN2 NSCLC patients who underwent upfront surgery at the Shizuoka Cancer Center between 2002 and 2017. Survival outcome, focusing on cN2 status, was retrospectively investigated. Mediastinal lymph nodes were pre-operatively evaluated using computed tomography and positron emission tomography. Surgical eligibility criteria included single-station cN2. N2 with N1 and skip N2 were defined as N2 with and without ipsilateral hilar lymph node metastasis, respectively. A platinum-doublet regimen was used for adjuvant chemotherapy. Survival curves were analysed using the Kaplan-Meier method. Univariate and multivariate analyses were performed using the Cox proportional hazard regression model. RESULTS: Clinical-skip N2 and cN2 with N1 cases included 22 and 21 patients, respectively. Twenty-three patients received adjuvant chemotherapy. The median follow-up duration was 73 months. Clinical-skip N2 had a significantly better 5-year recurrence-free survival (RFS) than cN2 with N1 (58.3 vs 28.6%, P = 0.038) and was an independent favorable RFS predictor. Recurrence within 18 months occurred in 71% of cN2 with N1 cases. Five-year overall survival and RFS rates in patients receiving adjuvant chemotherapy vs those without adjuvant chemotherapy were 82.2 vs 41.9% (P = 0.019) and 56.5 vs 28.0% (P = 0.049), respectively. CONCLUSIONS: Clinical-skip N2 had an excellent prognosis, and upfront surgery was acceptable. Conversely, upfront surgery followed by chemotherapy is not recommended for cN2 with N1 patients because of early recurrence.

Influence of lymph node degeneration on metastases in prostate cancer: or why we must look for a needle in a haystack.

BACKGROUND: To evaluate the incidence of lymph node degeneration and its association with nodal metastatic pattern in prostate cancer. METHODS: A retrospective analysis of the submitted lymph node specimen of 390 prostatectomies in 2011 was performed. All lymph nodes were histologically re-evaluated and the degree of lymph node degeneration e.g. lipomatous atrophy, capsular and framework fibrosis, and calcifications as well as the lymph node size were recorded. Lymph node degeneration was compared in the anatomic regions of the pelvis as well as in lymph nodes with and without metastases of prostatic cancer. RESULTS: Eighty-one of 6026 lymph nodes demonstrated metastases. Complete histologic examination with analysis of a complete cross-section was possible in 5173 lymph nodes including all lymph nodes with metastases. The incidence of lymph node degeneration was different across the various landing sites. Lymph node metastases were primarily detected in less degenerative and therefore more functional lymph nodes. In metastatic versus non-metastatic lymph nodes low lipomatous atrophy was reported in 84.0% versus 66.7% (p = 0.004), capsular fibrosis in 14.8% versus 35.4% (p < 0.001), calcifications in 35.8% versus 46.1% (p = 0.072) and framework fibrosis in 69.8% versus 75.3% (p = 0.53). Metastases were also identified more frequently in larger than in smaller lymph nodes (63.0% vs. 47.5%; p = 0.007). CONCLUSIONS: Degenerative changes in pelvic lymph nodes are commonly detectable but occur with variable frequency in the various nodal landing sites in the pelvis. The degree of lymph node degeneration of single lymph nodes has a significant influence on whether a lymph node is infiltrated by tumor cells and may harbour metastases.

Role of skip N2 lymph node metastasis for patients with the stage III-N2 lung adenocarcinoma: a propensity score matching analysis.

PURPOSE: Recent studies have indicated some differences in the prognosis of patients with stage III-N2 lung adenocarcinoma, and the prognosis of patients with skip N2 lymph node metastasis (SKN2) is good. This study grouped patients with stage III-N2 lung adenocarcinoma by propensity score matching (PSM) to evaluate the impact of SKN2 on the prognosis of these patients. METHODS: The clinical data for patients who underwent radical lobectomy and had a postoperative pathological diagnosis of stage III-N2 lung adenocarcinoma at our centre from 2016 to 2018 were collected, and PSM was performed at a ratio of 1:1. RESULTS: A total of 456 patients were enrolled in this study. After PSM, 112 patients were included in the SKN2 group, and 112 patients were included in the non-SKN2 group. When comparing the SKN2 group with the non-SKN2 group, the 3-year OS rate was (71.4% vs. 12.5%, p < 0.001), and the 3-year DFS rate was (35.7% vs. 5.4%, p < 0.001). It is further divided into four groups:single-station SKN2 (N2a1),Multi-station SKN2 (N2a2),single-station non-SKN2 (N2b1) and Multi-station non-SKN2 (N2b2).The 3-year OS and DFS rates of skip lymph node metastasis were better than those of non-skip lymph node metastasis(OS:N2a1 vs. N2b1 68.4% vs. 23.5%,p < 0.001;N2a2 vs. N2b2 73.0% vs. 7.7%,p < 0.001)(DFS:N2a1 vs. N2b1 68.4% vs. 5.9%,p < 0.001;N2a2 vs. N2b2 62.2% vs. 5.1%,p < 0.001), regardless of the number of N2 station(OS:N2a1 vs. N2a2 68.4% vs. 73.0%,p = 0.584;N2b1 vs. N2b2 23.5% vs. 7.7%,p = 0.051). On multivariate analysis, sex (p = 0.008) ,Vascular tumour thrombus(p = 0.047),size(p = 0.002)and SKN2 (p < 0.001) were independent predictors of OS. CONCLUSION: For patients with stage III-N2 lung adenocarcinoma, the prognosis of SKN2 patients is better than non-SKN2 patients', and SKN2 may be used as an important factor in the N2 subgroup classification in future TNM staging.

Skip metastases in appendicular Ewing sarcoma: relationship to distant metastases at diagnosis, chemotherapy response and overall survival.

OBJECTIVE: To identify in appendicular Ewing sarcoma (ES), if skip metastases (SM) are associated with distant metastases at presentation, response to neoadjuvant chemotherapy and overall outcome. MATERIALS AND METHOD: Patients with appendicular ES from 2007 to 2021 who had whole-bone MRI to identify SM were included in the study. Patient demographics included age/gender, bone involved, the presence of SM, longitudinal tumour length, presence of extra-osseous disease and its axial depth if present from whole-bone MRI and lung metastases and distant bone metastases from staging studies. Response to neoadjuvant chemotherapy from resection specimens and overall survival were noted. Comparison of these factors between patients with and without SM was undertaken. RESULTS: Ninety-five patients (66 males; 29 females: mean age 15.8 years; range 5-48 years) were included. On whole-bone MRI, 80 (84.2%) patients had no SM and 15 (15.8%) patients had one or more SM. Of patients without a SM, lung metastases were present in 16 (21%), distant bone metastases in 7 (11%), while 51 (75%) had a good response to chemotherapy compared with 7 (50%), 3 (27%), and 10 (77%), respectively, in patients with a SM. SM were significantly associated with lung metastases (p = 0.02), but not with distant skeletal metastases (p = 0.13), chemotherapy response (p = 0.88), tumour length (p = 0.47), presence of (p = 0.15) or axial depth of extra-osseous disease (p = 0.4). SM were associated with a significantly poorer survival (p = 0.007) and three times greater risk of death during follow-up. CONCLUSIONS: In appendicular ES, the identification of a SM is associated with the presence of lung metastases at presentation and poorer survival.

"Skip" corpectomy technique in multilevel cervical spondylotic myelopathy and ossified posterior longitudinal ligament: Outcomes with over 10-years follow-up.

OBJECTIVE: Skip corpectomy is a surgical technique that includes C4 and C6 corpectomies and fusion via autografts and a cervical plate and is frequently performed in patients with CSM and OPLL. This study presents long-term clinical and radiological outcomes of 48 patients who underwent skip corpectomy with 10-year follow-up. METHODS: 48 patients who were diagnosed with CSM or OPLL were included. All patients underwent spinal canal decompression and fusion via skip corpectomy. Clinical assessment was performed using the JOA scoring system. The radiological assessment was performed using plain anteroposterior, lateral, and flexion-extension cervical spine radiographs; cervical spine MR imaging; and cervical spine CT scans. The spinal canal size, spinal cord occupation ratio, cervical lordosis, and T2 signal changes were evaluated preoperatively, and postoperatively. RESULTS: The mean follow-up period was 14.6 years (13-20 years). Preoperatively, the JOA score was 11.06 ± 3.09. The mean cervical lordosis was 2.08°±11.74 and the average SCOR was 62.1 ± 14.22. There was a significant improvement in SCOR in the early postoperative period. The average cervical lordosis increased to 13.81 ± 2.51 in the 2nd month and minimal loss of cervical lordosis was observed on the 10th year in two patients. Fusion was achieved in all patients, regardless of the graft type. None of the patients had implant failure and graft or hardware-related complications at the 10th-year follow-up. CONCLUSION: Skip corpectomy provides efficient decompression of the spinal cord and provides adequate sagittal alignment and fusion in patients with CSM and OPLL. Long-term radiological and clinical outcomes of the technique are favorable.

A Long Skip Connection for Enhanced Color Selectivity in CNN Architectures.

Some recent studies show that filters in convolutional neural networks (CNNs) have low color selectivity in datasets of natural scenes such as Imagenet. CNNs, bio-inspired by the visual cortex, are characterized by their hierarchical learning structure which appears to gradually transform the representation space. Inspired by the direct connection between the LGN and V4, which allows V4 to handle low-level information closer to the trichromatic input in addition to processed information that comes from V2/V3, we propose the addition of a long skip connection (LSC) between the first and last blocks of the feature extraction stage to allow deeper parts of the network to receive information from shallower layers. This type of connection improves classification accuracy by combining simple-visual and complex-abstract features to create more color-selective ones. We have applied this strategy to classic CNN architectures and quantitatively and qualitatively analyzed the improvement in accuracy while focusing on color selectivity. The results show that, in general, skip connections improve accuracy, but LSC improves it even more and enhances the color selectivity of the original CNN architectures. As a side result, we propose a new color representation procedure for organizing and filtering feature maps, making their visualization more manageable for qualitative color selectivity analysis.

Improved Skip-Gram Based on Graph Structure Information.

Applying the Skip-gram to graph representation learning has become a widely researched topic in recent years. Prior works usually focus on the migration application of the Skip-gram model, while Skip-gram in graph representation learning, initially applied to word embedding, is left insufficiently explored. To compensate for the shortcoming, we analyze the difference between word embedding and graph embedding and reveal the principle of graph representation learning through a case study to explain the essential idea of graph embedding intuitively. Through the case study and in-depth understanding of graph embeddings, we propose Graph Skip-gram, an extension of the Skip-gram model using graph structure information. Graph Skip-gram can be combined with a variety of algorithms for excellent adaptability. Inspired by word embeddings in natural language processing, we design a novel feature fusion algorithm to fuse node vectors based on node vector similarity. We fully articulate the ideas of our approach on a small network and provide extensive experimental comparisons, including multiple classification tasks and link prediction tasks, demonstrating that our proposed approach is more applicable to graph representation learning.