Minimal Detectable Bone Fracture Gaps in CT Images and Digital Three-Dimensional (3D) Radii Models.

Knowledge of the minimal detectable bone fracture gap is essential in three-dimensional (3D) models, particularly in pre-operative planning of osteosynthesis to avoid overlooking gaps. In this study, defined incisions and bony displacements ranging from 100 to 400 µm were created in diaphyseal radii in 20 paired forearm specimens and verified with light microscopy. The specimens were scanned utilizing different computed tomography (CT) technologies/scanners, specimen positionings, scan protocols, image segmentations, and processing protocols. Inter- and intra-operator variabilities were reported as coefficient kappa. In CT images, fracture gaps of 100 µm and bone lamellae of 300 µm and 400 µm width were identified at a rate of 80 to 100%, respectively, independent of the investigated settings. In contrast, only 400µm incisions and bony displacements were visible in digital 3D models, with detection rates dependent on CT technology, image segmentation, and post-processing algorithm. 3D bone models based on state-of-the-art CT imaging can reliably visualize clinically relevant bone fracture gap sizes. However, verification of fractures to be surgically addressed should be verified with the original CT image series.

Effect of hot water maceration, rehydration, and soft tissue presence on 3D geometry of bone.

PURPOSE: In forensic medicine, maceration is often essential for examining bone surfaces, serving purposes such as identifying cut marks, making geometric measurements, and determining the victim's age. While hot water maceration removes soft tissue effectively, it is known to cause bone surface shrinkage. This raises the question of whether this effect is permanent or if it can be partially reversed through rehydration, considering the presence of soft tissue. METHODS: Computed tomography (CT) scans were conducted on the radii of 20 paired human anatomic forearm specimens. Subsequently, the radii were extracted, macerated in 60 °C water, CT-scanned in an air environment, rehydrated, re-implanted into the forearms, and CT-scanned again. RESULTS: Maceration resulted in a mean shrinkage of 0.12 mm on the outer bone surface. This shrinkage was nearly fully recoverable for the diaphysis after rehydration and accounting for soft tissue surrounding the bone. In contrast, the epiphysis showed permanent shrinkage, likely due to the loss of small bone fragments. Analysis of the inner bone surface indicated a smaller effect, but with significant standard deviations, especially for the epiphysis, possibly related to the less well-defined nature of the inner bone surface. CONCLUSION: The epiphyseal surface of hot water-macerated bone will, on average, be approximately 0.15 mm deflated and cannot retain the original surface. On the other hand, the diaphyseal surface is less affected and can be nearly completely restored after rehydration and accounting for soft tissue surrounding the bone.

Accuracy Analysis of 3D Bone Fracture Models: Effects of Computed Tomography (CT) Imaging and Image Segmentation.

The introduction of three-dimensional (3D) printed anatomical models has garnered interest in pre-operative planning, especially in orthopedic and trauma surgery. Identifying potential error sources and quantifying their effect on the model dimensional accuracy are crucial for the applicability and reliability of such models. In this study, twenty radii were extracted from anatomic forearm specimens and subjected to osteotomy to simulate a defined fracture of the distal radius (Colles' fracture). Various factors, including two different computed tomography (CT) technologies (energy-integrating detector (EID) and photon-counting detector (PCD)), four different CT scanners, two scan protocols (i.e., routine and high dosage), two different scan orientations, as well as two segmentation algorithms were considered to determine their effect on 3D model accuracy. Ground truth was established using 3D reconstructions of surface scans of the physical specimens. Results indicated that all investigated variables significantly impacted the 3D model accuracy (p < 0.001). However, the mean absolute deviation fell within the range of 0.03 ± 0.20 to 0.32 ± 0.23 mm, well below the 0.5 mm threshold necessary for pre-operative planning. Intra- and inter-operator variability demonstrated fair to excellent agreement for 3D model accuracy, with an intra-class correlation (ICC) of 0.43 to 0.92. This systematic investigation displayed dimensional deviations in the magnitude of sub-voxel imaging resolution for all variables. Major pitfalls included missed or overestimated bone regions during the segmentation process, necessitating additional manual editing of 3D models. In conclusion, this study demonstrates that 3D bone fracture models can be obtained with clinical routine scanners and scan protocols, utilizing a simple global segmentation threshold, thereby providing an accurate and reliable tool for pre-operative planning.

Dimensional accuracy and precision and surgeon perception of additively manufactured bone models: effect of manufacturing technology and part orientation.

BACKGROUND: Additively manufactured (AM) anatomical bone models are primarily utilized for training and preoperative planning purposes. As such, they must meet stringent requirements, with dimensional accuracy being of utmost importance. This study aimed to evaluate the precision and accuracy of anatomical bone models manufactured using three different AM technologies: digital light processing (DLP), fused deposition modeling (FDM), and PolyJetting (PJ), built in three different part orientations. Additionally, the study sought to assess surgeons' perceptions of how well these models mimic real bones in simulated osteosynthesis. METHODS: Computer-aided design (CAD) models of six human radii were generated from computed tomography (CT) imaging data. Anatomical models were then manufactured using the three aforementioned technologies and in three different part orientations. The surfaces of all models were 3D-scanned and compared with the original CAD models. Furthermore, an anatomical model of a proximal femur including a metastatic lesion was manufactured using the three technologies, followed by (mock) osteosynthesis performed by six surgeons on each type of model. The surgeons' perceptions of the quality and haptic properties of each model were assessed using a questionnaire. RESULTS: The mean dimensional deviations from the original CAD model ranged between 0.00 and 0.13 mm with maximal inaccuracies < 1 mm for all models. In surgical simulation, PJ models achieved the highest total score on a 5-point Likert scale ranging from 1 to 5 (with 1 and 5 representing the lowest and highest level of agreement, respectively), (3.74 ± 0.99) in the surgeons' perception assessment, followed by DLP (3.41 ± 0.99) and FDM (2.43 ± 1.02). Notably, FDM was perceived as unsuitable for surgical simulation, as the material melted during drilling and sawing. CONCLUSIONS: In conclusion, the choice of technology and part orientation significantly influenced the accuracy and precision of additively manufactured bone models. However, all anatomical models showed satisfying accuracies and precisions, independent of the AM technology or part orientation. The anatomical and functional performance of FDM models was rated by surgeons as poor.

Proteomic Atlas of Atherosclerosis: The Contribution of Proteoglycans to Sex Differences, Plaque Phenotypes, and Outcomes.

BACKGROUND: Using proteomics, we aimed to reveal molecular types of human atherosclerotic lesions and study their associations with histology, imaging, and cardiovascular outcomes. METHODS: Two hundred nineteen carotid endarterectomy samples were procured from 120 patients. A sequential protein extraction protocol was employed in conjunction with multiplexed, discovery proteomics. To focus on extracellular proteins, parallel reaction monitoring was employed for targeted proteomics. Proteomic signatures were integrated with bulk, single-cell, and spatial RNA-sequencing data, and validated in 200 patients from the Athero-Express Biobank study. RESULTS: This extensive proteomics analysis identified plaque inflammation and calcification signatures, which were inversely correlated and validated using targeted proteomics. The inflammation signature was characterized by the presence of neutrophil-derived proteins, such as S100A8/9 (calprotectin) and myeloperoxidase, whereas the calcification signature included fetuin-A, osteopontin, and gamma-carboxylated proteins. The proteomics data also revealed sex differences in atherosclerosis, with large-aggregating proteoglycans versican and aggrecan being more abundant in females and exhibiting an inverse correlation with estradiol levels. The integration of RNA-sequencing data attributed the inflammation signature predominantly to neutrophils and macrophages, and the calcification and sex signatures to smooth muscle cells, except for certain plasma proteins that were not expressed but retained in plaques, such as fetuin-A. Dimensionality reduction and machine learning techniques were applied to identify 4 distinct plaque phenotypes based on proteomics data. A protein signature of 4 key proteins (calponin, protein C, serpin H1, and versican) predicted future cardiovascular mortality with an area under the curve of 75% and 67.5% in the discovery and validation cohort, respectively, surpassing the prognostic performance of imaging and histology. CONCLUSIONS: Plaque proteomics redefined clinically relevant patient groups with distinct outcomes, identifying subgroups of male and female patients with elevated risk of future cardiovascular events.

Evaluation of perivascular fat attenuation with coronary CT angiography in cardiac transplantation patients: an imaging biomarker candidate for prediction of cardiac mortality and re-transplantation.

OBJECTIVES: In cardiac transplant recipients, non-invasive allograft surveillance for identifying patients at risk for graft failure remains challenging. The fat attenuation index (FAI) of the perivascular adipose tissue in coronary computed tomography angiography (CCTA) predicts outcomes in coronary artery disease in non-transplanted hearts; however, it has not been evaluated in cardiac transplant patients. METHODS: We followed 39 cardiac transplant patients with two or more CCTAs obtained between 2010 and 2021. We performed FAI measurements around the proximal 4 cm segments of the left anterior descending (LAD), right coronary artery (RCA), and left circumflex artery (LCx) using a previously validated methodology. The FAI was analyzed at a threshold of - 30 to - 190 Hounsfield units. RESULTS: FAI measurements were completed in 113 CCTAs, obtained on two same-vendor CT models. Within each CCTA, the FAI values between coronary vessels were strongly correlated (RCA and LAD R = 0.67 (p < 0.0001), RCA and LCx R = 0.58 (p < 0.0001), LAD and LCx R = 0.67 (p < 0.0001)). The FAIs of each coronary vessel between the patient's first and last CCTA completed at 120 kV were also correlated (RCA R = 0.73 (p < 0.0001), LAD R = 0.81 (p < 0.0001), LCx R = 0.55 (p = 0.0069). Finally, a high mean FAI value of all three coronary vessels at baseline (mean ≥ - 71 HU) was predictive of cardiac mortality or re-transplantation, however, not predictive of all cause-mortality. CONCLUSION: High baseline FAI values may identify a higher-risk cardiac transplant population; thus, FAI may support the implementation of CCTA in post-transplant surveillance. KEY POINT: • Perivascular fat attenuation measured with coronary CT is feasible in cardiac transplant patients and may predict cardiac mortality or need for re-transplantation.

Patient-Specific Guides for Accurate and Precise Positioning of Osseointegrated Implants in Transfemoral Amputations: A Proof-of-Concept In Vitro Study.

Background and Objectives: The treatment of transfemoral amputees using osseointegrated implants for prosthetic anchorage requires accurate implant positioning when using threaded bone-anchoring implants due to the curvature of the femur and the risk of cortical penetration in misaligned implants. This study investigated the accuracy and precision in implant positioning using additively manufactured case-specific positioning guides. Materials and Methods: The geometry and density distribution of twenty anatomic specimens of human femora were assessed in quantitative computed tomography (QCT) scanning. The imaging series were used to create digital 3D specimen models, preoperatively plan the optimal implant position and manufacture specimen-specific positioning guides. Following the surgical bone preparation and insertion of the fixture (threaded bone-anchoring element) (OPRA; Integrum AB, Mölndal, Sweden), a second QCT imaging series and 3D model design were conducted to assess the operatively achieved implant position. The 3D models were registered and the deviations of the intraoperatively achieved implant position from the preoperatively planned implant position were analyzed as follows. The achieved, compared to the planned implant position, was presented as resulting mean hip abduction or adduction (A/A) and extension or flexion (E/F) and mean implant axis offset in medial or lateral (M/L) and anterior or posterior (A/P) direction measured at the most distal implant axis point. Results: The achieved implant position deviated from the preoperative plan by 0.33 ± 0.33° (A/A) and 0.68 ± 0.66° (E/F) and 0.62 ± 0.55 mm (M/L) and 0.68 ± 0.56 mm (A/P), respectively. Conclusions: Using case-specific guides, it was feasible to achieve not only accurate but also precise positioning of the implants compared to the preoperative plan. Thus, their design and application in the clinical routine should be considered, especially in absence of viable alternatives.

Realistic 3D printed CT imaging tumor phantoms for validation of image processing algorithms.

Medical imaging phantoms are widely used for validation and verification of imaging systems and algorithms in surgical guidance and radiation oncology procedures. Especially, for the performance evaluation of new algorithms in the field of medical imaging, manufactured phantoms need to replicate specific properties of the human body, e.g., tissue morphology and radiological properties. Additive manufacturing (AM) technology provides an inexpensive opportunity for accurate anatomical replication with customization capabilities. In this study, we proposed a simple and cheap protocol using Fused Deposition Modeling (FDM) technology to manufacture realistic tumor phantoms based on the filament 3D printing technology. Tumor phantoms with both homogenous and heterogeneous radiodensity were fabricated. The radiodensity similarity between the printed tumor models and real tumor data from CT images of lung cancer patients was evaluated. Additionally, it was investigated whether a heterogeneity in the 3D printed tumor phantoms as observed in the tumor patient data had an influence on the validation of image registration algorithms. A radiodensity range between -217 to 226 HUs was achieved for 3D printed phantoms using different filament materials; this range of radiation attenuation is also observed in the human lung tumor tissue. The resulted HU range could serve as a lookup-table for researchers and phantom manufactures to create realistic CT tumor phantoms with the desired range of radiodensities. The 3D printed tumor phantoms also precisely replicated real lung tumor patient data regarding morphology and could also include life-like heterogeneity of the radiodensity inside the tumor models. An influence of the heterogeneity on accuracy and robustness of the image registration algorithms was not found.

Detection of Post-COVID-19 Lung Abnormalities: Photon-counting CT versus Same-Day Energy-integrating Detector CT.

Background Photon-counting detector (PCD) CT enables ultra-high-resolution lung imaging and may shed light on morphologic correlates of persistent symptoms after COVID-19. Purpose To compare PCD CT with energy-integrating detector (EID) CT for noninvasive assessment of post-COVID-19 lung abnormalities. Materials and Methods For this prospective study, adult participants with one or more COVID-19-related persisting symptoms (resting or exertional dyspnea, cough, fatigue) underwent same-day EID and PCD CT between April 2022 and June 2022. The 1.0-mm EID CT images and, subsequently, 1.0-, 0.4-, and 0.2-mm PCD CT images were reviewed for the presence of lung abnormalities. Subjective and objective EID and PCD CT image quality were evaluated using a five-point Likert scale (-2 to 2) and lung signal-to-noise ratios (SNRs). Results Twenty participants (mean age, 54 years ± 16 [SD]; 10 men) were included. EID CT showed post-COVID-19 lung abnormalities in 15 of 20 (75%) participants, with a median involvement of 10% of lung volume [IQR, 0%-45%] and 3.5 lobes [IQR, 0-5]. Ground-glass opacities and linear bands (10 of 20 participants [50%] for both) were the most frequent findings at EID CT. PCD CT revealed additional lung abnormalities in 10 of 20 (50%) participants, with the most common being bronchiectasis (10 of 20 [50%]). Subjective image quality was improved for 1.0-mm PCD versus 1.0-mm EID CT images (median, 1; IQR, 1-2; P < .001) and 0.4-mm versus 1.0-mm PCD CT images (median, 1; IQR, 1-1; P < .001) but not for 0.4-mm versus 0.2-mm PCD CT images (median, 0; IQR, 0-0.5; P = .26). PCD CT delivered higher lung SNR versus EID CT for 1.0-mm images (mean difference, 0.53 ± 0.96; P = .03) but lower SNR for 0.4-mm versus 1.0-mm images and 0.2-mm vs 0.4-mm images (-1.52 ± 0.68 [P < .001] and -1.15 ± 0.43 [P < .001], respectively). Conclusion Photon-counting detector CT outperformed energy-integrating detector CT in the visualization of subtle post-COVID-19 lung abnormalities and image quality. © RSNA, 2023 Supplemental material is available for this article.

Ultrahigh-Resolution Photon-Counting Detector CT of the Lungs: Association of Reconstruction Kernel and Slice Thickness With Image Quality.

BACKGROUND. Prior work has shown improved image quality for photon-counting detector (PCD) CT of the lungs compared with energy-integrating detector CT. A paucity of the literature has compared PCD CT of the lungs using different reconstruction parameters. OBJECTIVE. The purpose of this study is to the compare the image quality of ultra-high-resolution (UHR) PCD CT image sets of the lungs that were reconstructed using different kernels and slice thicknesses. METHODS. This retrospective study included 29 patients (17 women and 12 men; median age, 56 years) who underwent noncontrast chest CT from February 15, 2022, to March 15, 2022, by use of a commercially available PCD CT scanner. All acquisitions used UHR mode (1024 × 1024 matrix). Nine image sets were reconstructed for all combinations of three sharp kernels (BI56, BI60, and BI64) and three slice thicknesses (0.2, 0.4, and 1.0 mm). Three radiologists independently reviewed reconstructions for measures of visualization of pulmonary anatomic structures and pathologies; reader assessments were pooled. Reconstructions were compared with the clinical reference reconstruction (obtained using the BI64 kernel and a 1.0-mm slice thickness [BI641.0-mm]). RESULTS. The median difference in the number of bronchial divisions identified versus the clinical reference reconstruction was higher for reconstructions with BI640.4-mm (0.5), BI600.4-mm (0.3), BI640.2-mm (0.5), and BI600.2-mm (0.2) (all p < .05). The median bronchial wall sharpness versus the clinical reference reconstruction was higher for reconstructions with BI640.4-mm (0.3) and BI640.2-mm (0.3) and was lower for BI561.0-mm (-0.7) and BI560.4-mm (-0.3) (all p < .05). Median pulmonary fissure sharpness versus the clinical reference reconstruction was higher for reconstructions with BI640.4-mm (0.3), BI600.4-mm (0.3), BI560.4-mm (0.5), BI640.2-mm (0.5), BI600.2-mm (0.5), and BI560.2-mm (0.3) (all p < .05). Median pulmonary vessel sharpness versus the clinical reference reconstruction was lower for reconstructions with BI561.0-mm (-0.3), BI600.4-mm (-0.3), BI560.4-mm (-0.7), BI640.2-mm (-0.7), BI600.2-mm (-0.7), and BI560.2-mm (-0.7). Median lung nodule conspicuity versus the clinical reference reconstruction was lower for reconstructions with BI561.0-mm (-0.3) and BI560.4-mm (-0.3) (both p < .05). Median conspicuity of all other pathologies versus the clinical reference reconstruction was lower for reconstructions with BI561.0 mm (-0.3), BI560.4-mm (-0.3), BI640.2-mm (-0.3), BI600.2-mm (-0.3), and BI560.2-mm (-0.3). Other comparisons among reconstructions were not significant (all p > .05). CONCLUSION. Only the reconstruction using BI640.4-mm yielded improved bronchial division identification and bronchial wall and pulmonary fissure sharpness without a loss in pulmonary vessel sharpness or conspicuity of nodules or other pathologies. CLINICAL IMPACT. The findings of this study may guide protocol optimization for UHR PCD CT of the lungs.

Biomechanical evaluation of an allograft fixation system for ACL reconstruction.

The purpose of this study was to compare the biomechanical stability, especially graft slippage of an allograft screw and a conventional interference screw for tibial implant fixation in ACL reconstruction. Twenty-four paired human proximal tibia specimens underwent ACL reconstruction, with the graft in one specimen of each pair fixed using the allograft screw and the other using the conventional interference screw. Specimens were subjected to cyclic tensile loading until failure. The two fixation methods did not show any statistical difference in load at graft slippage (p = 0.241) or estimated mean survival until slippage onset (p = 0.061). The ultimate load and the estimated mean survival until failure were higher for the interference screw (p = 0.04, and p = 0.018, respectively). Graft displacement at ultimate load reached values of up to 7.2 (interference screw) and 11.3 mm (allograft screw). The allograft screw for implant fixation in ACL reconstruction demonstrated comparable behavior in terms of graft slippage to the interference screw but underperformed in terms of ultimate load. However, the ultimate load, occurring at progressive graft slippage, may not be considered a direct indicator of clinical failure.

Dynamic electrophysiological mechanism in patients with long-standing persistent atrial fibrillation.

Background: Improved understanding of the mechanisms that sustain persistent and long-standing persistent atrial fibrillation (LSpAF) is essential for providing better ablation solutions. The findings of traditional catheter-based electrophysiological studies can be impacted by the sedation required for these procedures. This is not required in non-invasive body-surface mapping (ECGI). ECGI allows for multiple mappings in the same patient at different times. This would expose potential electrophysiological changes over time, such as the location and stability of extra-pulmonary vein drivers and activation patterns in sustained AF. Materials and methods: In this electrophysiological study, 10 open-heart surgery candidates with LSpAF, without previous ablation procedures (6 male, median age 73 years), were mapped on two occasions with a median interval of 11 days (IQR: 8-19) between mappings. Bi-atrial epicardial activation sequences were acquired using ECGI (CardioInsight™, Minneapolis, MN, United States). Results: Bi-atrial electrophysiological abnormalities were documented in all 20 mappings. Interestingly, the anatomic location of focal and rotor activities changed between the mappings in all patients [100% showed changes, 95%CI (69.2-100%), p < 0.001]. Neither AF driver type nor their number varied significantly between the mappings in any patient (median total number of focal activities 8 (IQR: 1-16) versus 6 (IQR: 2-12), p = 0.68; median total number of rotor activities 48 (IQR: 44-67) versus 55 (IQR: 44-61), p = 0.30). However, individual zones showed a high number of quantitative changes (increase/decrease) of driver activity. Most changes of focal activity were found in the left atrial appendage, the region of the left lower pulmonary vein and the right atrial appendage. Most changes in rotor activity were found also at the left lower pulmonary vein region, the upper half of the right atrium and the right atrial appendage. Conclusion: This clinical study documented that driver location and activation patterns in patients with LSpAF changes constantly. Furthermore, bi-atrial pathophysiology was demonstrated, which underscores the importance of treating both atria in LSpAF and the significant role that arrhythmogenic drivers outside the pulmonary veins seem to have in maintaining this complex arrhythmia.

Biomechanical Assessment of Fracture Loads and Patterns of the Odontoid Process.

STUDY DESIGN: Laboratory study. OBJECTIVE: This study aimed to investigate the biomechanical competence and fracture characteristics of the odontoid process. SUMMARY OF BACKGROUND DATA: Odontoid fractures of the second cervical vertebra (C2) represent the most common spine fracture type in the elderly. However, very little is known about the underlying biomechanical fracture mechanisms. MATERIALS AND METHODS: A total of 42 C2 human anatomic specimens were scanned via computed tomography, divided in six groups, and subjected to combined quasistatic loading at -15°, 0°, and 15° in sagittal plane and -50° and 0° in transverse plane until fracturing. Bone mineral density (BMD), height, fusion state of the ossification centers, stiffness, yield load, and ultimate load were assessed. RESULTS: While lowest values for stiffness, yield load, and ultimate load were observed at load inclination of 15° in sagittal plane, no statistically significant differences were observed between the study groups ( P ≥0.235). BMD correlated positively with yield load ( r2 =0.350, P <0.001) and ultimate load ( r2 =0.955, P <0.001) but not with stiffness ( r2 =0.082, P =0.07). The specimens with clearly distinguishable fusion of the ossification centers revealed less data scattering of the biomechanical outcomes. CONCLUSION: Load direction plays a subordinate role in traumatic fractures of the odontoid process. BMD was associated with significant correlation to the biomechanical outcomes. Thus, odontoid fractures appear to result from of an interaction between the load magnitude and bone quality.

Nose-shaped mass in the ascending aorta.

A 79-year-old man was referred for urgent coronary artery bypass grafting. Contrast-enhanced computed tomography revealed an atypically nose-shaped contrast defect, which intraoperatively turned out to be an atheromatous plaque. Its preoperative detection allowed us to prevent an adverse cerebral event. This case highlights that a thorough preoperative work-up should ideally include a CT angiography, in patients where atherosclerotic changes are to be expected.

The HeartMate 6 and CardioMEMS for Fixed Pulmonary Hypertension.

Fixed pulmonary hypertension (FPH) is a contraindication for heart transplantation (HTX). However, this condition might be reversed by continuous left-ventricular unloading with a left-ventricular assist device. We present a case of apical hypertrophic cardiomyopathy with extensive left-ventricular endocardial calcification and severe FPH (systolic pulmonary artery pressure, 102 mm Hg). To bridge the patient to candidacy for HTX, two Abbott HeartMate 3 ventricular assist devices were implanted in a total artificial heart (TAH) configuration ("HeartMate 6"). Before TAH implantation, an Abbott CardioMEMS pressure sensor was implanted to assess reversal of FPH before listing for HTX.

Semi-automatic vessel detection for challenging cases of peripheral arterial disease.

OBJECTIVES: Manual or semi-automated segmentation of the lower extremity arterial tree in patients with Peripheral arterial disease (PAD) remains a notoriously difficult and time-consuming task. The complex manifestations of the disease, including discontinuities of the vascular flow channels, the presence of calcified atherosclerotic plaque in close vicinity to adjacent bone, and the presence of metal or other imaging artifacts currently preclude fully automated vessel identification. New machine learning techniques may alleviate this challenge, but require large and reasonably well segmented training data. METHODS: We propose a novel semi-automatic vessel tracking approach for peripheral arteries to facilitate and accelerate the creation of annotated training data by expert cardiovascular radiologists or technologists, while limiting the number of necessary manual interactions, and reducing processing time. After automatically classifying blood vessels, bones, and other tissue, the relevant vessels are tracked and organized in a tree-like structure for further visualization. RESULTS: We conducted a pilot (N = 9) and a clinical study (N = 24) in which we assess the accuracy and required time for our approach to achieve sufficient quality for clinical application, with our current clinically established workflow as the standard of reference. Our approach enabled expert physicians to readily identify all clinically relevant lower extremity arteries, even in problematic cases, with an average sensitivity of 92.9%, and an average specificity and overall accuracy of 99.9%. CONCLUSIONS: Compared to the clinical workflow in our collaborating hospitals (28:40 ± 7:45 [mm:ss]), our approach (17:24 ± 6:44 [mm:ss]) is on average 11:16 [mm:ss] (39%) faster.