Region-based Analysis Research Articles

Diagnostic Accuracy of 18 F-FES PET/CT for the Detection of Recurrent and Metastatic Breast Cancer.

To evaluate the diagnostic value of 16α- 18 F-fluoro-17β-fluoroestradiol ( 18 F-FES) PET/CT for distant metastasis or recurrence in patients with estrogen receptor (ER)-positive breast cancer. Patients with ER-positive breast cancer and suspected of de novo metastasis or recurrence were retrospectively identified from a prospective cohort enrolled for a postmarketing surveillance study of 18 F-FES at our institution. Per-patient diagnostic accuracy was assessed using pathology or 2 or more standard-of-care imaging procedures with a minimum of 6 months of follow-up as the reference standard. The per-region detection rate of 18 F-FES PET/CT was evaluated and compared with that of standard-of-care imaging. Of the 162 included patients, 104 and 58 were suspected to have recurrence or de novo metastasis, respectively. The overall sensitivity and specificity of 18 F-FES PET/CT were 95% (95% confidence interval [CI], 89%-98%) and 89% (95% CI, 76%-96%), respectively. When stratified according to clinical settings, the sensitivity and specificity were 95% (95% CI, 88%-99%) and 96% (95% CI, 78%-100%), respectively, for detecting recurrence, and 94% (95% CI, 81%-99%) and 82% (95% CI, 60%-95%) for detecting distant metastasis. In region-based analysis, the overall detection rate of 18 F-FES PET/CT was significantly higher than that of standard-of-care imaging (92% [95% CI, 89%-94%] vs 83% [95% CI, 79%-87%], P < 0.001). 18 F-FES PET/CT showed excellent diagnostic performance in patients with ER-positive breast cancer suspected of de novo metastasis or recurrence.

Early detection of thermal image based T1 breast cancer using enhanced multiwavelet denoised convolution neural network with region based analysis

Medical Thermography image is used for the detection of breast cancer at an earlier stage. Thermography image shows the temperature change in the body due to cells. The metabolic rate of cancer cells is high compared to normal cells. A high metabolic rate increases the blood flow in cancer cells. High blood leads to changes in body temperature. The change in body temperature is used for cancer cell detection at an earlier stage. However, T1-stage cancer cells are smaller and have small temperature differences undetected with thermography. In this paper, a T1-stage cancer cell is heated by an external source; then, thermal images are acquired for earlier detection of small-size cancer cells. External heat source amplifies T1 stage cancer cell temperature. Amplified cancer cell images are analyzed using the proposed Multiwavelet-Deep Denoised Convolutional Neural Network (MWTDnCNN) algorithm for T1 cancer cell detection. Amplified T1 stage cancer cell has higher thermal conductivity (k) and heat capacity (Cp), which helps to detect T1 cancer cell tissue due to the enhanced pixel feature. The proposed MWTDnCNN algorithm has a T1-stage cancer cell detection accuracy of about 98% compared with traditional algorithms. The proposed MWTDnCNN algorithm detects T1-stage cancer of size 1.29 mm.

Region-based analysis with functional annotation identifies genes associated with cognitive function in South Asians from India.

The prevalence of dementia among South Asians across India is approximately 7.4% in those 60 years and older, yet little is known about genetic risk factors for dementia in this population. Most known risk loci for Alzheimer's disease (AD) have been identified from studies conducted in European Ancestry (EA) but are unknown in South Asians. Using whole-genome sequence data from 2680 participants from the Diagnostic Assessment of Dementia for the Longitudinal Aging Study of India (LASI-DAD), we performed a gene-based analysis of 84 genes previously associated with AD in EA. We investigated associations with the Hindi Mental State Examination (HMSE) score and factor scores for general cognitive function and five cognitive domains. For each gene, we examined missense/loss-of-function (LoF) variants and brain-specific promoter/enhancer variants, separately, both with and without incorporating additional annotation weights (e.g., deleteriousness, conservation scores) using the variant-Set Test for Association using Annotation infoRmation (STAAR). In the missense/LoF analysis without annotation weights and controlling for age, sex, state/territory, and genetic ancestry, three genes had an association with at least one measure of cognitive function (FDR q<0.1). APOE was associated with four measures of cognitive function, PICALM was associated with HMSE score, and TSPOAP1 was associated with executive function. The most strongly associated variants in each gene were rs429358 (APOE ε4), rs779406084 (PICALM), and rs9913145 (TSPOAP1). rs779406084 is a rare missense mutation that is more prevalent in LASI-DAD than in EA (minor allele frequency=0.075% vs. 0.0015%); the other two are common variants. No genes in the brain-specific promoter/enhancer analysis met criteria for significance. Results with and without annotation weights were similar. Missense/LoF variants in some genes previously associated with AD in EA are associated with measures of cognitive function in South Asians from India. Analyzing genome sequence data allows identification of potential novel causal variants enriched in South Asians.

SpatialCells: automated profiling of tumor microenvironments with spatially resolved multiplexed single-cell data.

Cancer is a complex cellular ecosystem where malignant cells coexist and interact with immune, stromal and other cells within the tumor microenvironment (TME). Recent technological advancements in spatially resolved multiplexed imaging at single-cell resolution have led to the generation of large-scale and high-dimensional datasets from biological specimens. This underscores the necessity for automated methodologies that can effectively characterize molecular, cellular and spatial properties of TMEs for various malignancies. This study introduces SpatialCells, an open-source software package designed for region-based exploratory analysis and comprehensive characterization of TMEs using multiplexed single-cell data. The source code and tutorials are available at https://semenovlab.github.io/SpatialCells. SpatialCells efficiently streamlines the automated extraction of features from multiplexed single-cell data and can process samples containing millions of cells. Thus, SpatialCells facilitates subsequent association analyses and machine learning predictions, making it an essential tool in advancing our understanding of tumor growth, invasion and metastasis.

Abstract 7409: SpatialCells: Automated profiling of tumor microenvironments with spatially resolved multiplexed single-cell data

Abstract Cancer is a complex cellular ecosystem where malignant cells coexist and interact with immune, stromal, and other cells within the tumor microenvironment. Recent technological advancements in spatially resolved multiplexed imaging at single-cell resolution have led to the generation of large-scale and high-dimensional datasets from biological specimens. This underscores the necessity for automated methodologies that can effectively characterize the molecular, cellular, and spatial properties of tumor microenvironments for various malignancies. This study introduces SpatialCells, an open-source software package designed for region-based exploratory analysis and comprehensive characterization of tumor microenvironments using spatially resolved multiplexed single-cell data. We have provided detailed documents and tutorials at https://github.com/SemenovLab/SpatialCells. SpatialCells is featured by its capability to computationally define regions of interest based on any group of cells and subsequently conduct region-based analyses. SpatialCells incorporates several modules to facilitate its functionality, such as the spatial module and the measurements module. Our workflow starts with developing the spatial module, including functions to establish regional boundaries and check the region in which a cell is located. In particular, SpatialCells supports macro-region and micro-region analyses. In the macro-region analysis, users can employ SpatialCells to divide the whole tissue into several subregions, such as tumor, tumor border, stroma border, and stroma. In the micro-region analysis, SpatialCells identifies small regions within an area, such as within the tumor area, enabling a finer-grained understanding of the cellular landscape. The measurements module contains functions to extract the properties of tumor cells, immune cells, and tumor-immune cell interactions, such as multivariate tumor proliferation index, immune infiltration score, and tumor-immune cell distances. These properties can be assessed for the whole tissue or local regions. Importantly, SpatialCells can efficiently process datasets containing millions of cells. In summary, SpatialCells is a novel software solution for spatially analyzing the tumor microenvironment and automatically extracting quantitative features of cells using multiplexed imaging data in a streamlined fashion with the capacity to process samples containing millions of cells. Thus, SpatialCells facilitates subsequent association analyses and machine learning predictions at scale, making it an essential tool in advancing our understanding of tumor growth, invasion, and metastasis. Citation Format: Guihong Wan. SpatialCells: Automated profiling of tumor microenvironments with spatially resolved multiplexed single-cell data [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 7409.

Patient-specific left atrium contraction quantification associated with atrial fibrillation: A region-based approach

Background and objectivesAtrial fibrillation (AF) is a widespread cardiac arrhythmia that significantly impacts heart function. AF disrupts atrial mechanical contraction, leading to irregular, uncoordinated, and slow blood flow inside the atria which favors the formation of clots, primarily within the left atrium (LA). A standardized region-based analysis of the LA is missing, and there is not even any consensus about how to define the LA regions. In this study we propose an automatic approach for regionalizing the LA into segments to provide a comprehensive 3D region-based LA contraction assessment. LA global and regional contraction were quantified in control subjects and in AF patients to describe mechanical abnormalities associated with AF. MethodsThe proposed automatic approach for LA regionalization was tested in thirteen control subjects and seventeen AF patients. After dividing LA into standard regions, we evaluated the global and regional mechanical function by measuring LA contraction parameters, such as regional volume, global and regional strains, regional wall motion and regional shortening fraction. ResultsLA regionalization was successful in all study subjects. In the AF group compared with control subjects, results showed: a global impairment of LA contraction which appeared more pronounced along radial and circumferential direction; a regional impairment of radial strain which was more pronounced in septal, inferior, and lateral regions suggesting a greater reduction in mechanical efficiency in these regions in comparison to the posterior and anterior ones. ConclusionAn automatic approach for LA regionalization was proposed. The regionalization method was proved to be robust with several LA anatomical variations and able to characterize contraction changes associated with AF.

Region-based cascading impact analysis in critical infrastructure systems

ABSTRACT Critical Infrastructure System (CIS) disruption cascades into all the dependent CISs, even the regions, over the internal system where the disruption occurs. This study develops a cascading impact analysis with a region-based path-flow estimation approach based on a complex network and provides a local-scale demonstration of the interdependent electric power-water supply system in Incheon, South Korea. The analytical framework facilitates the modeling of CISs considering three major structural relationships – CIS, CIS-CIS, and CIS-Region- with functional interdependency and quantifying cascading impacts on regions via a continuous service flow. The results demonstrate more practical cascading ranges and levels caused by each CIS disruption, highlighting that the potential impacts increase in an integrated network, which a single CIS network cannot capture. It emphasizes the need to consider functional interdependent CIS structures and estimate cascading impacts from a regional perspective. This study provides multiangle insights into how regions enhance resilience through infrastructure disaster management.

Oral Microbiome Stamp in Alzheimer's Disease.

Recent studies have suggested that periodontal disease and alterations in the oral microbiome may be associated with cognitive decline and Alzheimer's disease (AD) development. Here, we report a case-control study of oral microbiota diversity in AD patients compared to healthy seniors from Central Asia. We have characterized the bacterial taxonomic composition of the oral microbiome from AD patients (n = 64) compared to the healthy group (n = 71) using 16S ribosomal RNA sequencing. According to our results, the oral microbiome of AD has a higher microbial diversity, with an increase in Firmicutes and a decrease in Bacteroidetes in the AD group. LEfSe analysis showed specific differences at the genus level in both study groups. A region-based analysis of the oral microbiome compartment in AD was also performed, and specific differences were identified, along with the absence of differences in bacterial richness and on the functional side. Noteworthy findings demonstrated the decrease in periodontitis-associated bacteria in the AD group. Distinct differences were revealed in the distribution of metabolic pathways between the two study groups. Our study confirms that the oral microbiome is altered in AD. However, a comprehensive picture of the complete composition of the oral microbiome in patients with AD requires further investigation.

Volume and region-based analysis using risk-standardized mortality rates of outcomes in HPB surgery

Volume and region-based analysis using risk-standardized mortality rates of outcomes in HPB surgery

Small-Signal Modeling and Region-Based Stability Analysis Using Support Vector Machine (SVM) for Autonomous Hybrid AC and DC Microgrids

Small-Signal Modeling and Region-Based Stability Analysis Using Support Vector Machine (SVM) for Autonomous Hybrid AC and DC Microgrids

Evidence for inflammation in normal-appearing brain regions in patients with growing sporadic vestibular schwannoma: A PET study.

Nonauditory symptoms can be a prominent feature in patients with sporadic vestibular schwannoma (VS), but the cause of these symptoms is unknown. Inflammation is hypothesized to play a key role in the growth and symptomatic presentation of sporadic VS, and in this study, we investigated through translocator protein (TSPO) positron emission tomography (PET) whether inflammation occurred within the "normal appearing" brain of such patients and its association with tumor growth. Dynamic PET datasets from 15 patients with sporadic VS (8 static and 7 growing) who had been previously imaged using the TSPO tracer [11C](R)-PK11195 were included. Parametric images of [11C](R)-PK11195 binding potential (BPND) and the distribution volume ratio (DVR) were derived and compared across VS growth groups within both contralateral and ipsilateral gray (GM) and white matter (WM) regions. Voxel-wise cluster analysis was additionally performed to identify anatomical regions of increased [11C](R)-PK11195 binding. Compared with static tumors, growing VS demonstrated significantly higher cortical (GM, 1.070 vs. 1.031, P = .03) and whole brain (GM & WM, 1.045 vs. 1.006, P = .03) [11C](R)-PK11195 DVR values. The voxel-wise analysis supported the region-based analysis and revealed clusters of high TSPO binding within the precentral, postcentral, and prefrontal cortex in patients with growing VS. We present the first in vivo evidence of increased TSPO expression and inflammation within the brains of patients with growing sporadic VS. These results provide a potential mechanistic insight into the development of nonauditory symptoms in these patients and highlight the need for further studies interrogating the role of neuroinflammation in driving VS symptomatology.

Automatic non-destructive UAV-based structural health monitoring of steel container cranes

Container cranes are of key importance for maritime cargo transportation. The uninterrupted and all-day operation of these container cranes, which directly affects the efficiency of the port, necessitates the continuous inspection of these massive hoisting steel structures. Due to the large size of cranes, the current manual inspections performed by expert climbers are costly, risky, and time-consuming. This motivates further investigations on automated non-destructive approaches for the remote inspection of fatigue-prone parts of cranes. In this paper, we investigate the effectiveness of color space-based and deep learning-based approaches for separating the foreground crane parts from the whole image. Subsequently, three different ML-based algorithms (k-Nearest Neighbors, Random Forest, and Naive Bayes) are employed to detect the rust and repainting areas from detected foreground parts of the crane body. Qualitative and quantitative comparisons of the results of these approaches were conducted. While quantitative evaluation of pixel-based analysis reveals the superiority of the k-Nearest Neighbors algorithm in our experiments, the potential of Random Forest and Naive Bayes for region-based analysis of the defect is highlighted.

A region-based low-carbon operation analysis method for integrated electricity-hydrogen-gas systems

A region-based low-carbon operation analysis method for integrated electricity‑hydrogen-gas systems (IEHGSs) is proposed in this paper. The committed carbon emission operation region (CCE_OR) and the economic CCE_OR (ECCE_OR) are proposed to evaluate the low-carbon feasible space (LCFS) from the low-carbon, secure, and economic perspectives, which provides powerful tools for the low-carbon operation analyses of IEHGSs. Firstly, a coordinated mechanism of carbon flow and hydrogen flow is developed to excavate the low-carbon operation potential of IEHGSs. Then, based on the distributionally robust chance-constrained approach, the concepts and models of CCE_OR and ECCE_OR are defined and formulated considering uncertainties and hydrogen merits. Furthermore, a region model reformulation method is proposed to reformulate the uncertain and non-convex CCE_OR and ECCE_OR models into the deterministic and mixed-integer linear region models. The geometric properties of CCE_OR and ECCE_OR are both theoretically proven, which are the union set of finite convex and bounded polytopes. The united region solution methodology considering the geometric properties of CCE_OR and ECCE_OR is presented to effectively characterize CCE_OR and ECCE_OR. Finally, the numerical results on two test systems verify the effectiveness and applicability of the proposed methods. Simulation results indicate that the proposed methods can improve the LCFS, characterize visual coupling relationships between LCFS and observation variables, and provide comprehensive low-carbon operation information of operation points.

Quantification of extracellular volume with cardiac computed tomography in patients with left ventricle dysfunction

Abstract Background Myocardial fibrosis (MF) is a common endpoint in the majority of pathological mechanisms affecting cardiac muscle, ultimately resulting in cardiac failure. Cardiovascular magnetic resonance (CMR) with gadolinium-based contrast agent and T1 mapping technique is able to deliver a non-invasive quantification of the extracellular volume (ECV) expansion. Recently, ECV estimation with cardiac computed tomography (ECV-CCT) has been validated in the setting of cardiac amyloidosis, showing good agreement with ECV-CMR. However, no evidence is available with last generation single source single energy CT scanner in the clinical context of newly diagnosed left ventricular dysfunction. Therefore, the aim of this study is to test the diagnostic accuracy of ECV-CCT for the detection of MF in patients with recent diagnosis of reduced left ventricle ejection fraction (LVEF), having ECV-CMR as reference technique. Methods A consecutive cohort of 41 patients with newly diagnosed left ventricle dysfunction (LVEF &amp;lt; 50%) was enrolled. Time interval of maximum 30 days was present between CCT and CMR. Myocardial segment evaluability with each technique, agreement between ECV-CMR and ECV-CCT, regression analysis and Bland-Altman analysis were performed. Results Age of enrolled patients was 62 ± 11 years, and mean LVEF at CMR was 35.2 ± 10.7%. Mean total amount of administrated iodinated contrast was 112.6 ± 22.2 ml, while overall radiation exposure for ECV estimation was2.1 ± 1.1 mSv. Out of 656 myocardial segments available, 656 (100%) segments were evaluable at CCT while 640 (97.6%) were evaluable at CMR. ECV-CCT demonstrated slightly lower values compared to ECV-CMR both at global and region-based comparison (all segments, 33.1 ± 8.5% vs 35.5 ± 10.7%, p &amp;lt; 0.001). At regression analysis, strong correlations were described both at global and region-based analysis (all segments, r = 0.877, 95% CI: 0.840 to 0.914). On Bland-Altman analysis, bias between ECV-CMR and ECV-CCT for global analysis was 2.4 (95% CI: -7.8 to 12.7). Conclusions ECV estimation with last-generation single source single energy CT scanner is feasible and accurate. Integration of ECV measurement in comprehensive CCT evaluation of patients with newly diagnosed left ventricular dysfunction can be performed with a small increase in overall radiation exposure.

Genetic Association Analysis of Copy Number Variations for Meat Quality in Beef Cattle.

Meat quality is an economically important trait for global food production. Copy number variations (CNVs) have been previously implicated in elucidating the genetic basis of complex traits. In this article, we detected a total of 112,198 CNVs and 10,102 CNV regions (CNVRs) based on the Bovine HD SNP array. Next, we performed a CNV-based genome-wide association analysis (GWAS) of six meat quality traits and identified 12 significant CNV segments corresponding to eight candidate genes, including PCDH15, CSMD3, etc. Using region-based association analysis, we further identified six CNV segments relevant to meat quality in beef cattle. Among these, TRIM77 and TRIM64 within CNVR4 on BTA29 were detected as candidate genes for backfat thickness (BFT). Notably, we identified a 34 kb duplication for meat color (MC) which was supported by read-depth signals, and this duplication was embedded within the keratin gene family including KRT4, KRT78, and KRT79. Our findings will help to dissect the genetic architecture of meat quality traits from the aspects of CNVs, and subsequently improve the selection process in breeding programs.

Detection of resting-state functional connectivity in the lumbar spinal cord with 3T MRI

Functional MRI (fMRI) of the spinal cord is an expanding area of research with potential to investigate neuronal activity in the central nervous system. We aimed to characterize the functional connectivity features of the human lumbar spinal cord using resting-state fMRI (rs-fMRI) at 3T, using region-based and data-driven analysis approaches. A 3D multi-shot gradient echo resting-state blood oxygenation level dependent-sensitive rs-fMRI protocol was implemented in 26 healthy participants. Average temporal signal-to-noise ratio in the gray matter was 16.35 ± 4.79 after denoising. Evidence of synchronous signal fluctuations in the ventral and dorsal horns with their contralateral counterparts was observed in representative participants using interactive, exploratory seed-based correlations. Group-wise average in-slice Pearson’s correlations were 0.43 ± 0.17 between ventral horns, and 0.48 ± 0.16 between dorsal horns. Group spatial independent component analysis (ICA) was used to identify areas of coherent activity¸ and revealed components within the gray matter corresponding to anatomical regions. Lower-dimensionality ICA revealed bilateral components corresponding to ventral and dorsal networks. Additional separate ICAs were run on two subsets of the participant group, yielding two sets of components that showed visual consistency and moderate spatial overlap. This work shows feasibility of rs-fMRI to probe the functional features and organization of the lumbar spinal cord.

A survey on brain tumor image analysis.

Medical imaging, also known as radiology, is the field of medicine in which medical professionals recreate various images of parts of the body for diagnostic or treatment purposes. Medical imaging procedures include non-invasive tests that allow doctors to diagnose injuries and diseases without being intrusive TechTarget (n.d.). A number of tools and techniques are used to automate the analysis of medical images acquired with various image processing methods. The brain is one of the largest and most complex organs of the human body and anomaly detection from brain images (i.e., MRI, CT, PET, etc.) is one of the major research areas of medical image analysis. Image processing methods such as filtering and thresholding models, geometry models, graph models, region-based analysis, connected component analysis, machine learning (ML) models, the recent deep learning (DL) models, and various hybrid models are used in brain image analysis. Brain tumors are one of the most common brain diseases with a high mortality rate, and it is difficult to analyze from brain images for the versatility of the shape, location, size, texture, and other characteristics. In this paper, a comprehensive review on brain tumor image analysis is presented with basic ideas of brain tumor, brain imaging, brain image analysis tasks, brain image analysis models, brain tumor image features, performance metrics used for evaluating the models, and some available datasets on brain tumor/medical images. Some challenges of brain tumor analysis are also discussed including suggestions for future research directions. The graphical abstract summarizes the contributions of this paper.

A Brain Morphometry Study with Across-Site Harmonization Using a ComBat-Generalized Additive Model in Children and Adolescents.

Regional anatomical structures of the brain are intimately connected to functions corresponding to specific regions and the temporospatial pattern of genetic expression and their functions from the fetal period to old age. Therefore, quantitative brain morphometry has often been employed in neuroscience investigations, while controlling for the scanner effect of the scanner is a critical issue for ensuring accuracy in brain morphometric studies of rare orphan diseases due to the lack of normal reference values available for multicenter studies. This study aimed to provide across-site normal reference values of global and regional brain volumes for each sex and age group in children and adolescents. We collected magnetic resonance imaging (MRI) examinations of 846 neurotypical participants aged 6.0-17.9 years (339 male and 507 female participants) from 5 institutions comprising healthy volunteers or neurotypical patients without neurological disorders, neuropsychological disorders, or epilepsy. Regional-based analysis using the CIVET 2.1.0. pipeline provided regional brain volumes, and the measurements were across-site combined using ComBat-GAM harmonization. The normal reference values of global and regional brain volumes and lateral indices in our study could be helpful for evaluating the characteristics of the brain morphology of each individual in a clinical setting and investigating the brain morphology of ultra-rare diseases.

Detection and Molecular Characterization of Animal Adenovirus and Astrovirus from Western Maharashtra, India.

Astroviruses (AstV) and adenoviruses (AdV) are associated with diarrhoea in young animals. However, the epidemiology and genetic diversity of AstVs and AdVs in animals is not well studied. Hence, the present study was conducted to detect and characterize AstVs and AdVs in calves, piglets and puppies from Western Maharashtra, India. Out of the processed porcine (48), canine (80), and bovine (65) faecal samples, the porcine AstV (PAstV), bovine AstV (BAstV), canine AstV (CAstV), and porcine AdV (PAdV) were detected in 12.5%, 7.69%, 3.75% and 4.1% of samples, respectively. In the RNA-dependent RNA polymerase region-based phylogenetic analysis, the detected BAstV strains grouped with MAstV-28, MAstV-33, and MAstV-35, CAstV strains belonged to MAstV-5; PAstV strains belonged to MAstV-24, MAstV-26, and MAstV-31. However, in hexon gene-based phylogeny, both the detected PAdV were of genotype 3, exhibiting 91.9-92.5% nucleotide identity with Ivoirian and Chinese strains. The study reports first-time BAstVs from calves and PAdV-3 from piglets in India. The study revealed diversity in the circulation of AstVs in tested animals and AdVs in pigs, and suggested that they alone might be associated with other diarrhoea or in combination with other enteric pathogens, thus highlighting the necessity of extensive epidemiological investigations to develop diagnostic tools and control measures.

Performance of deep learning-based autodetection of arterial stenosis on head and neck CT angiography: an independent external validation study.

To externally validate the performance of automated stenosis detection on head and neck CT angiography (CTA) and investigate the impact factors using an independent bi-center dataset with digital subtraction angiography (DSA) as the ground truth. Patients who underwent head and neck CTA and DSA between January 2019 and December 2021 were retrospectively included. The degree of stenosis was automatically evaluated using CerebralDoc based on CTA. The performance of CerebralDoc across levels (per-patient, per-region, per-vessel, and per-segment) and thresholds (≥ 50%, ≥ 70%, and = 100%) was evaluated. Logistic regression was performed to identify independent factors associated with false negative results. 296 patients were analyzed. Specificity across levels and thresholds was high, exceeding 92%. The area under the curve ranged from poor (0.615, 95% CI: 0.544, 0.686; at the region-based analysis for stenosis ≥ 70%) to excellent (0.945, 95% CI: 0.905, 0.985; at the patient-based analysis for stenosis ≥ 50%). Sensitivity ranged from 0.714 (95% CI: 0.675, 0.750) at the segment-based analysis for stenosis ≥ 70% to 0.895 (95% CI: 0.849, 0.919) at the patient-based analysis for stenosis ≥ 50%. The multiple logistic regression analysis revealed that false negative results were primarily more likely to specific stenosis locations (particularly the M2 segment and skull base segment of the internal carotid artery) and occlusion. CerebralDoc has the potential to automated stenosis detection on head and neck CTA, but further efforts are needed to optimize its performance.