Automatic Identification of Coronary Stent in Coronary Calcium Scoring CT using Deep Learning

Automatic Identification of Coronary Stent in Coronary Calcium Scoring CT using Deep Learning | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Article Automatic Identification of Coronary Stent in Coronary Calcium Scoring CT using Deep Learning Yura Ahn, Gyu-Jun Jeong, Dabee Lee, Cherry Kim, June-Goo Lee, and 1 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-4543450/v1 This work is licensed under a CC BY 4.0 License Status: Published Journal Publication published 28 Oct, 2024 Read the published version in Scientific Reports → Version 1 posted 11 You are reading this latest preprint version Abstract Automatic pre-screening of pre-existing stents, whose prognostic value remains uncertain, could potentially reduce workload and enhance efficiency. However, such a solution has not yet been developed and validated. We aimed to develop and evaluate a deep learning-based coronary stent filtering algorithm (Stent_filter) in CAC scoring CT scans using a multicenter CAC dataset. We developed Stent_filter comprising two main processes: stent identification and false-positive reduction. Development utilized 108 non-enhanced echocardiography-gated CAC scans (including 74 with manually labeled stents), and for false positive reduction, 2063 CAC scans with significant coronary calcium (average Agatston score: 523.8) but no stents were utilized. Stent_filter’s performance was evaluated on two independent internal test sets (n = 355 and 396; one without coronary stents) and two external test sets from different institutions (n = 105 and 62), each with manually labeled stents. We calculated the per-patient sensitivity, specificity, and false-positive rate of Stent_filter. Stent_filter demonstrated a high overall per-patient sensitivity of 98.8% (511/517 cases with stents) and a false-positive rate of 0.022 (20/918). Notably, the false-positive ratio was significantly lower in the dataset containing stents (Internal-1; 0.008 [3/355]) compared with the dataset without stents (Internal-2; 0.043 [17/396], p = 0.008). All false-positive identifications were attributed to dense coronary calcifications, with no false positives identified in extracoronary locations. The automated Stent_filter accurately distinguished coronary stents from preexisting coronary calcifications. This approach holds potential as a filter within a fully automated CAC scoring workflow, streamlining the process efficiently. Coronary artery calcium score Computed tomography Coronary stent Artificial intelligence Accuracy Figures Figure 1 Figure 2 Figure 3 Introduction Coronary artery calcium (CAC) scoring has become a common method for assessing cardiovascular risk in clinical practice 1 , 2 . Traditionally, this scoring has been performed using non-enhanced electrocardiography-synchronized cardiac computed tomography (CT), requiring skilled professionals for manual interventions, which can be time-consuming 3 , 4 . However, recent advances in artificial intelligence (AI) have facilitated the development of fully automatic calcium scoring system using non-enhanced CT 5 , including non-gated low-dose chest CT scans or radiation therapy treatment planning CT 6 , 7 . Despite these technological advancements promising to streamline the CAC scoring workflow, one crucial manual step remains: identifying patients with coronary stents. The utility of CAC scoring in patients who have undergone stent or bypass surgery is known to be uncertain. In addition, stents themselves can mimic coronary calcium, leading to an overestimation of calcification burden. As a result, radiologic professionals manually exclude these patients from CAC scoring. Consequently, if it becomes feasible to automatically identify coronary stents on CAC CT scans and triage such patients, it could significantly improve the efficiency of both manual and automated CAC scoring workflows. Despite the potential benefits, no automated algorithm for stent identification from CAC scans exists, to our knowledge. Moreover, the technical challenge arises not only from identifying the stent but also from distinguishing it from the heavy coronary artery calcification that often coexists in patients who have underwent stent insertion 8 . Therefore, in this study, we aimed to develop and evaluate a deep learning-based coronary stent filtering algorithm (Stent_filter), along with a false-positive reduction algorithm, for use in CAC scoring CT scans, using a multicenter CAC dataset. Methods This retrospective study was approved by the institutional review board committee of each participating institution (institutional review board committee of the Asan Medical Center, institutional review board committee of the Dankook University Hospital, and institutional review board committee of the Korea University Ansan Hospital) and informed consent was waived by the institutional review board from all of three institutions due to the retrospective nature of observational study. This study was performed in accordance with the Helsinki Declaration. Development of Stent_filter The development and validation dataset comprised non-enhanced CAC scans obtained from 108 patients (mean age, 60.7 ± 11.0 years; range, 37–84; 84 men [77.8%]) who underwent CT scans for chest pain between January 2016 to December 2020 (Table 1 ). All scans were acquired from the three different scanners from a single vendor, with a slice thickness of 3 mm. Initially, segmentation for the labeled dataset was conducted using a research prototype (AVIEW CAC, Coreline Soft, Co. Ltd.). Subsequently, a cardiothoracic radiologist (Y.A., with 2 years of experience in cardiothoracic radiology) meticulously reviewed the entire set of generated masks and manually differentiated them for stents against those for coronary calcification. In cases when stents and coronary calcification were contiguous, manual adjustments were made to separate the stent from the calcification. In cases where discrimination between stent and calcification was ambiguous, medical records, including invasive coronary angiography records, were reviewed. Table 1 Characteristics of development and test datasets Characteristics Development Dataset Test Dataset Development Validation Internal-1 Internal-2 External-1 External-2 No. of patients 77 31 355 396 105 62 Age (y) 61.0 ± 10.7 60.0 ± 11.9 62.1 ± 9.6 62.5 ± 9.2 58.3 ± 10.7 65.3 ± 9.3 No. of male patients 63 (81.8) 21 (67.7) 277 (78.0) 310 (78.3) 91 (86.7) 50 (80.6) Tube voltage (kVp) 100 0 (0.0) 0 (0.0) 18 (5.1) 5 (1.3) 0 (0.0) 0 (0.0) 120 77 (100.0) 31 (100.0) 337 (94.9) 391 (98.7) 105 (100.0) 62 (100.0) Tube current (mAs) 83.0 ± 20.4 81.5 ± 19.6 72.5 ± 28.8 32.3 ± 42.2 43.6 ± 4.2 30.2 ± 2.6 Slice thickness 2.5 mm 0 (0.0) 0 (0.0) 1 (0.3) 252 (63.6) 102 (97.1) 62 (100.0) 3 mm 77 (100.0) 31 (100.0) 354 (99.7) 144 (36.4) 0 (0.0) 0 (0.0) 4 mm 0 (0.0) 0 (0.0) 0 (0.0) 0 (0.0) 3 (2.9) 0 (0.0) Scanners SOMATOM Definition, SOMATOM Definition Flash, SOMATOM Force SOMATOM Definition, SOMATOM Definition Flash, SOMATOM Force SOMATOM Definition, SOMATOM Definition Flash, LightSpeed VCT SOMATOM Definition, SOMATOM Definition Flash, LightSpeed VCT, Discovery CT750 HD Ingenuity Core 128, IQon - Spectral CT iCT 256, Ingenuity CT Mask Calcification in LAD 56 26 260 391 83 49 Calcification in LCX 46 23 266 289 67 42 Calcification in LM 38 16 193 218 35 0 Calcification in RCA 51 27 295 356 80 57 Stent 62 12 352 0 103 62 Note.—Data are presented as numbers with percentage in parenthesis or as means ± standard deviations. LAD = left anterior descending artery, LCX = left circumflex coronary artery, LM = left main coronary artery, RCA = right coronary artery. To develop the false-positive reduction algorithm in the final architecture, 2036 CAC scans with substantial coronary calcium without stents (average Agatston calcium score: 523.8) were utilized. These scans constituted a random subset of asymptomatic individuals who underwent health screening 9 . Test dataset The portion of the internal test dataset (Internal-2) was derived from a previous study 5 , which was utilized for the development and validation of a deep learning system for automatic coronary calcium scoring. This dataset comprised 396 patients and exclusively contained coronary calcifications without stents, serving as a negative control. The remaining three test datasets, obtained from different institutions, included both stents and coronary calcifications. Similar to the development dataset, masks for the labeled dataset were initially generated via software, and all generated masks were reviewed by the same cardiothoracic radiologist (Y.A.). Ultimately, a total of four datasets comprising 918 non-enhanced CAC images, including 517 stents, were employed to evaluate the performance of Stent_filter. Detailed information regarding the test dataset is presented in Table 1 . Automatic coronary stent identification: Stent_filter We developed the Stent_filter methodology to detect coronary stents and calcium separately. This methodology involves two main pipelines: 1) For stent and calcium detection, we used an advanced technique called Mask RCNN, based on the ResNet101 architecture, to accurately identify stents and calcium deposits 10 . This model was initialized with pre-trained weights from a large dataset called Microsoft Common Objects in Context (MSCOCO), which helped improve its accuracy. The ResNet101 architecture has special features called residual connections, which effectively capture intricate features while mitigating the issue of vanishing gradients in deep networks. The Mask RCNN model generates potential object candidates by use of a region proposal network. Then, it utilizes the feature pyramid network to detect objects of varying sizes efficiently. To ensure accurate and robust performance, several implementation details were carefully considered. Image preprocessing involved windowing with a window width of 1500 HU and a window level of 300 HU. The input tensor for the model was constructed in 2.5D, incorporating frames from both above and below. Input frames were standardized to RGB channels of size 512 x 512. During training, model parameters were optimized using a stochastic gradient descent optimizer with a mini-batch size of 2 images. Training spanned 100 epochs with an initial learning rate of 1e-3 and included weight decay. Learning rate adjustments were made, reducing by a factor of 0.1 every 25 epochs. The multi-task loss function employed for detection encompassed regression and cross-entropy components. Augmentation techniques, including sharpening, Gaussian blur, additive Gaussian noise, multiplication, and affine transformations (scale, shift, rotate), were also applied. During testing, we set a minimum confidence score of 0.5 for stent detection. 2) For false positive reduction, we applied a rule-based process to further refine the results. We performed 3D connective component analysis to make multiple consecutive detections into one stent candidate lesion. The candidate lesion which is smaller than 100 mm 3 in size was discarded. The overall procedure is illustrated in Fig. 1 . We evaluated the performance of both pipelines using a computer with specifications: an Intel core i7-6700 processor, 32 GB of RAM, and an NVIDIA GeForce GTX 1060 graphics card with 3 GB memory. The time it took for the model to analyze a CT scan with 45 to 60 frames ranged from approximately 20 to 30 seconds. Statistical analysis Analyses were performed on the per-patient and per-lesion diagnostic performance of Stent_filter. In the per-patient analysis, lesions identified by both Stent_filter and by the radiologist were considered stents, while patients with any overlapping lesion were classified true positives for having stents. Mismatched lesions were classified into false-positive (i.e., coronary calcifications determined as stent by Stent_filter) and false-negative results (i.e., missed stent by Stent_filter). For per-lesion analysis, performance metrics were calculated based on the target coronary artery (i.e., left anterior descending 11 and left main coronary artery [LM], left circumflex coronary artery [LCX], and right coronary artery [RCA]). The coronary artery containing the greater portion of the stent mask was designated as the target coronary artery. Per-lesion sensitivity was calculated as (number of true positives /total number of stents in the corresponding vessel) × 100. The false-positive ratio was calculated as (number of false-positive lesions/total number of patients). A cardiothoracic radiologist reviewed all mismatched lesions to determine the cause of the mismatch. Results Stent_filter for coronary stent identification In the per-patient analysis, Stent_filter demonstrated high accuracy in distinguishing stents from coronary calcifications, achieving an overall sensitivity of 98.8% (511/517) and a specificity of 95.0% (381/401) (Fig. 2 ). The overall false-positive ratio was 0.022 (20/918) (Table 2 ). Notably, review of the mismatched lesions revealed that all false-positive results were tubular calcifications mimicking stents without extracoronary lesions (Fig. 3 ) Table 2 Performance of Stent_filter: Analysis per patient Test Dataset Performance measure Internal-1 Internal-2 External-1 External-2 Overall No. of patients 355 396 105 62 918 True positive 349 NA 101 61 511 False negative 3 NA 2 1 6 False positive 3 17 0 0 20 True negative 0 379 2 0 381 Sensitivity 0.991 NA 0.980 0.984 0.988 Specificity 0.000 0.957 1.000 0.000 0.950 FP ratio 0.008 0.043 0.000 0.000 0.022 FP = false positive, NA = not applicable. Across the three test datasets from different institutions, Stent_filter consistently demonstrated high sensitivity, ranging from 98.0% (External-1; 101/103) to 99.1% (Internal-1; 349/352). It also exhibited a specificity of 95.7% (379/396) in the Internal-2 dataset, which served as a negative control. Notably, the false-positive ratio was significantly lower in the dataset containing stents (Internal-1; 0.008 [3/355]) compared with the dataset without stents (Internal-2; 0.043 [17/396], p = 0.008). Furthermore, in the two external test datasets, there were no instances of completely false-positive results for a single patient without stent. Per-lesion analysis Overall, per-lesion sensitivity ranged from 94.5% (449/475) for the LAD and LM to 98.4% (179/182) for the RCA, as detailed in Table 3 . Among all test datasets, the LAD and LM regions, which had the highest number of stents, also exhibited the highest false-positive ratio, with an overall false-positive ratio for LAD and LM at 0.038 (35/918). Notably, in the external-2 dataset, the false-positive ratio was highest across all coronary arteries: 0.113 for LAD and LM (7/62), 0.065 for the LCX (4/62), and 0.081 for the RCA (5/62), compared with the other three datasets. Table 3 Performance of Stent_filter: Analysis per lesion Test Dataset Performance measure Internal-1 Internal-2 External-1 External-2 Overall No. of patients 355 396 105 62 918 Total lesions 548 NA 144 87 779 Vessel LAD, LM LCX RCA LAD, LM LCX RCA LAD, LM LCX RCA LAD, LM LCX RCA LAD, LM LCX RCA No. of lesions 368 81 129 12 1 5 84 30 46 58 20 25 510 131 200 No. of stents 348 79 121 NA NA NA 76 27 41 51 16 20 475 122 182 True positive 327 76 119 NA NA NA 72 27 40 50 15 20 449 118 179 False negative 21 3 2 NA NA NA 4 0 1 1 1 0 26 4 3 False positive 20 2 8 12 1 5 8 3 5 7 4 5 35 9 18 Sensitivity 0.940 0.962 0.983 NA NA NA 0.947 1.000 0.976 0.980 0.938 1.000 0.945 0.967 0.984 FP ratio 0.056 0.006 0.023 0.030 0.003 0.013 0.076 0.029 0.048 0.113 0.065 0.081 0.038 0.010 0.020 FP = false positive, NA = not applicable, LAD = left anterior descending artery, LCX = left circumflex coronary artery, LM = left main coronary artery, RCA = right coronary artery. Stent_filter according to the stent length Across the vessels, sensitivity was highest for medium-length stents (319/334) and lowest for short stents (36/40) (Table 4 ). Notably, Stent_filter exhibited the lowest sensitivity for short (≤ 15 mm) stents in the RCA (9/11), while it showed the highest sensitivity for short stents in the LCX (6/6) and medium-length stents (15 mm < stent length ≤ 30 mm) in the RCA (91/91). Table 4 Performance of Stent_filter: according to the stent length Test Dataset Internal-1 External-1 External-2 Overall Stent length, mm x < = 15 15 < x < = 30 30 < x x < = 15 15 < x < = 30 30 < x x < = 15 15 < x < = 30 30 < x x < = 15 15 < x < = 30 30 < x LAD, LM 19 113 170 2 44 24 2 20 20 23 177 214 No. of FN 1 5 14 1 2 1 0 1 0 2 10 22 Sensitivity 0.947 0.956 0.918 0.500 0.955 0.958 1.000 0.950 1.000 0.913 0.944 0.897 LCX 5 35 27 1 19 6 0 12 4 6 66 37 No. of FN 0 4 0 0 0 0 0 1 0 0 5 1 Sensitivity 1.000 0.886 1.000 1.000 1.000 1.000 NA 0.917 1.000 1.000 0.924 0.973 RCA 7 55 69 3 27 16 1 9 13 11 91 98 No. of FN 1 0 1 1 0 0 0 0 0 2 0 1 Sensitivity 0.857 1.000 0.986 0.667 1.000 1.000 1.000 1.000 1.000 0.818 1.000 0.90 Total of no. of lesions 31 203 266 6 90 46 3 41 37 40 334 349 No. of FN 2 9 15 2 2 1 0 2 0 4 15 24 Sensitivity 0.935 0.956 0.944 0.667 0.978 0.978 1.000 0.951 1.000 0.900 0.955 0.931 Total 500 142 81 723 No. of FN 26 5 2 43 Sensitivity 0..948 0.965 0.975 0.941 FN = false negative, NA = not applicable, LAD = left anterior descending artery, LCX = left circumflex coronary artery, LM = left main coronary artery, RCA = right coronary artery Discussion Automatic pre-screening of pre-existing stents, whose prognostic value remains uncertain, could potentially reduce workload and enhance efficiency. However, such a solution has not yet been developed and validated. In this study, we developed and tested the performance of an automatic identification algorithm for stents. Our results demonstrate highly accurate stent identification (sensitivity of 98.8% [511/517] and specificity of 95.0% [381/401]), along with a false-positive rate of 0.022 across both the internal and external test datasets. With recent advancements in AI, automated CAC scoring has reached a state-of-the-art level, demonstrating sensitivities from 93–98% and specificities from 93–100%, along with a high interclass correlation coefficient of 1.00 across large multicenter and multivendor cohorts 5 , 12 , 13 . As a result, the primary challenge remaining for automated CAC scoring, to minimize additional work for radiologists, lies in the ability of the AI to differentiate non-coronary high-density lesions from vascular calcifications. In particular, the automated identification of coronary stents holds significance beyond simply reducing false positives; coronary stents introduce uncertainty into the prognostic value of CAC scoring, suggesting that their automatic exclusion could substantially decrease workload. However, this area remains under-investigated, and even in studies focused on automated CAC scoring, the presence of a coronary stent has often been a criterion for exclusion 5 , 12 , 13 . In our multicenter and multivendor cohort study involving patients with preexisting coronary calcifications, we evaluated the performance of our Stent_filter. The filter demonstrated a sensitivity exceeding 98% and a specificity ranging from 95.7–100% across all internal and external datasets, which is considered clinically acceptable compared with the reported performance of automated CAC scoring methods. Specifically, the Internal-2 dataset, characterized by the presence of only coronary artery calcifications, yielded a specificity of 95.7% (379/396), underscoring the effectiveness of the algorithm in discrimination against calcifications. Moreover, the per-lesion sensitivity for the RCA at 98.4% (179/182) highlights its capability, especially notable given the RCA's susceptibility to motion artifacts due to its perpendicular motion relative to the axial plane of CT scans 14 . Conversely, addressing the false-positive rate remains a crucial step in the clinical application of automated CAC systems. Despite a very low false-negative rate per patient (0.007 [6/918]), which typically does not significantly impact the workflow—since physicians familiar with the patient's medical history can disregard automatic CAC scoring results in patients with coronary stents—the occurrence of false positives when skipping CAC scoring in indicated patients is noteworthy, despite the low overall false-positive rate per patient (0.022 [20/918]). Upon reviewing all cases, we confirmed the following findings. First, all false-positive lesions were indeed coronary calcifications and not extracoronary lesions such as valvular or pericardial calcifications. Second, the false positives were identified as segmental tubular calcifications of the coronary artery, which could mimic the appearance of a coronary stent, especially in cases of in-stent restenosis or stents under coronary calcification. Although we did not conduct an additional reader study to determine whether radiologists could also confuse these lesions with coronary stents when blinded to medical history, our review suggests that these lesions could be challenging for radiologists to distinguish from stents without access to the patient's medical history. Dense calcification can mimic the appearance of a stent, and to mitigate these false positives 8 , we implemented a false-positive reduction process as post-processing. This approach is based on similar principles to those used by Komatsu et al. 8 , who also developed an algorithm, the 3-dimensional plaque map, based on similar principles to ours, to differentiate heavy calcification from stents. We assume that neither radiologists nor radiologic technicians can achieve 100% accuracy in the screening process for CAC scoring candidates, and our algorithm's false-positive rate could be seen as an extension of this intrinsic error. Furthermore, given the lower per-patient false-positive rate observed in patients with coronary stents compared with those without stents, the clinical impact of false-positive results on patient triage in real-world clinical settings could be minimal. However, further technical refinement or the addition of complementary functions, such as incorporating available patient medical records, could assist in more accurately triaging patients for CAC scoring. Considering the current protocol, which requires radiologic technicians to manually confirm the absence of stents in CAC scans before proceeding with CAC scoring, the adoption of Stent_filter could substantially reduce their workload. This reduction is achieved by eliminating the need for this manual confirmation step and subsequent unnecessary manual CAC scoring. Additionally, with the availability of a developed automated CAC system, it becomes possible to implement a fully automated CAC scoring process. This process would not only automate the scoring itself but also include automated triage, streamlining the entire procedure and enhancing efficiency. Our study has several limitations. First, it was fundamentally a retrospective feasibility study validating the automatic identification performance. Its pragmatic effect, in other words, whether and how much it could reduce the workload, was not assessed. Second, its feasibility in non-gated scans was not evaluated. While deep-learning automated CAC has advanced considerably, enabling scoring in non-gated low-dose chest CT scans or enhanced CT coronary angiography scans 15 – 17 , the efficacy of Stent_filter in these non-dedicated scans requires further technology development in line with their protocols. In conclusion, the automated Stent_filter accurately distinguishes coronary stents from preexisting coronary calcifications. This approach holds the potential to serve as a filter within a fully automated CAC scoring workflow, streamlining the process efficiently. Abbreviations AI artificial intelligence CAC coronary artery calcium LAD left anterior descending artery LCX left circumflex coronary artery LM left main coronary artery RCA right coronary artery Declarations Data availability: The datasets used and/or analyzed during the current study available from the corresponding author on reasonable request. Funding: None Author information Contributions: Conceived and designed the analysis: J.L. and D.H.Y.; Collected the data: Y.A., G.J., D.L., C.K., and D.H.Y.; Contributed data or analysis tools: G.J., J.L., and D.H.Y.; Performed the analysis: Y.A. and G.J.; Wrote the paper: Y.A., G.J., J.L., and D.H.Y.; Manuscript editing: Y.A., G.J., D.L., C.K., J.L., and D.H.Y. Ethics declarations Competing interests: The authors declare no competing interests. References Greenland, P., Blaha, M. J., Budoff, M. J., Erbel, R. & Watson, K. E. Coronary Calcium Score and Cardiovascular Risk. J Am Coll Cardiol 72 , 434-447, doi:10.1016/j.jacc.2018.05.027 (2018). Greenland, P. et al. ACCF/AHA 2007 clinical expert consensus document on coronary artery calcium scoring by computed tomography in global cardiovascular risk assessment and in evaluation of patients with chest pain: a report of the American College of Cardiology Foundation Clinical Expert Consensus Task Force (ACCF/AHA Writing Committee to Update the 2000 Expert Consensus Document on Electron Beam Computed Tomography) developed in collaboration with the Society of Atherosclerosis Imaging and Prevention and the Society of Cardiovascular Computed Tomography. J Am Coll Cardiol 49 , 378-402, doi:10.1016/j.jacc.2006.10.001 (2007). Blaha, M. J., Mortensen, M. B., Kianoush, S., Tota-Maharaj, R. & Cainzos-Achirica, M. Coronary Artery Calcium Scoring: Is It Time for a Change in Methodology? JACC Cardiovasc Imaging 10 , 923-937, doi:10.1016/j.jcmg.2017.05.007 (2017). Hecht, H. S. et al. 2016 SCCT/STR guidelines for coronary artery calcium scoring of noncontrast noncardiac chest CT scans: A report of the Society of Cardiovascular Computed Tomography and Society of Thoracic Radiology. J Cardiovasc Comput Tomogr 11 , 74-84, doi:10.1016/j.jcct.2016.11.003 (2017). Lee, J. G. et al. Fully Automatic Coronary Calcium Score Software Empowered by Artificial Intelligence Technology: Validation Study Using Three CT Cohorts. Korean J Radiol 22 , 1764-1776, doi:10.3348/kjr.2021.0148 (2021). van Velzen, S. G. M. et al. Deep Learning for Automatic Calcium Scoring in CT: Validation Using Multiple Cardiac CT and Chest CT Protocols. Radiology 295 , 66-79, doi:10.1148/radiol.2020191621 (2020). Abdelrahman, K. et al. Artificial Intelligence in Coronary Artery Calcium Scoring Detection and Quantification. Diagnostics (Basel) 14 , doi:10.3390/diagnostics14020125 (2024). Komatsu, S. et al. Coronary calcification mimicking the stent: analysis by MDCT Plaque Map. Int J Cardiol 115 , 419-421, doi:10.1016/j.ijcard.2006.01.066 (2007). Yang, D. H. et al. Association between C-reactive Protein and type of coronary arterial plaque in asymptomatic patients: assessment with coronary CT angiography. Radiology 272 , 665-673, doi:10.1148/radiol.14130772 (2014). He, K., Gkioxari, G., Dollár, P. & Girshick, R. in 2017 IEEE International Conference on Computer Vision (ICCV). 2980-2988. McGrath, E. E., Blades, Z. & Anderson, P. B. Chylothorax: aetiology, diagnosis and therapeutic options. Respir Med 104 , 1-8, doi:10.1016/j.rmed.2009.08.010 (2010). Ihdayhid, A. R. et al. Evaluation of an artificial intelligence coronary artery calcium scoring model from computed tomography. Eur Radiol 33 , 321-329, doi:10.1007/s00330-022-09028-3 (2023). Winkel, D. J. et al. Deep learning for vessel-specific coronary artery calcium scoring: validation on a multi-centre dataset. Eur Heart J Cardiovasc Imaging 23 , 846-854, doi:10.1093/ehjci/jeab119 (2022). Ren, P. et al. Motion artefact reduction in coronary CT angiography images with a deep learning method. BMC Med Imaging 22 , 184, doi:10.1186/s12880-022-00914-2 (2022). Kang, H. W. et al. Evaluation of fully automated commercial software for Agatston calcium scoring on non-ECG-gated low-dose chest CT with different slice thickness. Eur Radiol 33 , 1973-1981, doi:10.1007/s00330-022-09143-1 (2023). Lee, J. O., Park, E. A., Park, D. & Lee, W. Deep Learning-Based Automated Quantification of Coronary Artery Calcification for Contrast-Enhanced Coronary Computed Tomographic Angiography. J Cardiovasc Dev Dis 10 , doi:10.3390/jcdd10040143 (2023). Suh, Y. J. et al. Fully automatic coronary calcium scoring in non-ECG-gated low-dose chest CT: comparison with ECG-gated cardiac CT. Eur Radiol 33 , 1254-1265, doi:10.1007/s00330-022-09117-3 (2023). Additional Declarations No competing interests reported. Cite Share Download PDF Status: Published Journal Publication published 28 Oct, 2024 Read the published version in Scientific Reports → Version 1 posted Editorial decision: Revision requested 23 Jul, 2024 Reviews received at journal 17 Jul, 2024 Reviewers agreed at journal 13 Jul, 2024 Reviews received at journal 06 Jul, 2024 Reviewers agreed at journal 02 Jul, 2024 Reviewers agreed at journal 01 Jul, 2024 Reviewers invited by journal 01 Jul, 2024 Editor assigned by journal 24 Jun, 2024 Editor invited by journal 12 Jun, 2024 Submission checks completed at journal 11 Jun, 2024 First submitted to journal 07 Jun, 2024 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-4543450","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":330479794,"identity":"feb4746a-8a96-44a0-a68d-27ac0967a815","order_by":0,"name":"Yura Ahn","email":"","orcid":"","institution":"University of Ulsan College of Medicine, Asan Medical Center","correspondingAuthor":false,"prefix":"","firstName":"Yura","middleName":"","lastName":"Ahn","suffix":""},{"id":330479795,"identity":"3dfa3f7d-e99d-4565-a7ff-32ffb082ebbf","order_by":1,"name":"Gyu-Jun Jeong","email":"","orcid":"","institution":"University of Ulsan College of Medicine, Asan Medical Center","correspondingAuthor":false,"prefix":"","firstName":"Gyu-Jun","middleName":"","lastName":"Jeong","suffix":""},{"id":330479796,"identity":"1c1f23b6-c978-4de1-ae0e-7e3e95bdb65a","order_by":2,"name":"Dabee Lee","email":"","orcid":"","institution":"Dankook University Hospital","correspondingAuthor":false,"prefix":"","firstName":"Dabee","middleName":"","lastName":"Lee","suffix":""},{"id":330479797,"identity":"867c98e4-edaf-470b-a7f0-779769970510","order_by":3,"name":"Cherry Kim","email":"","orcid":"","institution":"Korea University Ansan Hospital","correspondingAuthor":false,"prefix":"","firstName":"Cherry","middleName":"","lastName":"Kim","suffix":""},{"id":330479799,"identity":"1a163f71-2b16-4c93-a588-49d867ef0f5d","order_by":4,"name":"June-Goo Lee","email":"","orcid":"","institution":"University of Ulsan College of Medicine, Asan Medical Center","correspondingAuthor":false,"prefix":"","firstName":"June-Goo","middleName":"","lastName":"Lee","suffix":""},{"id":330479800,"identity":"759985cf-83ee-44bc-aea1-c60940e22552","order_by":5,"name":"Dong Hyun Yang","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA+0lEQVRIiWNgGAWjYDCCAwwMzAheBRAzMzeQouUMSAsjKVoY28Akfi18x3sPvy5sO5zHIJH87OHXebXR/O1ALT8qtuHUInnmXJr1zLbDxQwSaebGstuO5844zNjA2HPmNk4tBjdyzIx52w4nNkgkmElLbjuW2wDUwszYRpSW9G/SknOO5c4nQovxY4iWHDPJjw01uRsIaZE8c8aMmedcemIbz5syaYZjB3I3ArUcxOcXvuM9xp95yqwT+9nTt0n+qKnLnXf+8MEHPypwawECNglGNiApkMDAzMNwGCx0AJ96IGD+wPAHSPEfYGD8wVBHQPEoGAWjYBSMRAAA6WNfUn0NwF8AAAAASUVORK5CYII=","orcid":"","institution":"University of Ulsan College of Medicine, Asan Medical Center","correspondingAuthor":true,"prefix":"","firstName":"Dong","middleName":"Hyun","lastName":"Yang","suffix":""}],"badges":[],"createdAt":"2024-06-07 04:42:56","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-4543450/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-4543450/v1","draftVersion":[],"editorialEvents":[{"content":"https://doi.org/10.1038/s41598-024-76092-8","type":"published","date":"2024-10-28T16:20:40+00:00"}],"editorialNote":"","failedWorkflow":false,"files":[{"id":62654427,"identity":"01f7ac4a-695c-4b0b-9c58-1d1bcd0e287e","added_by":"auto","created_at":"2024-08-17 01:24:30","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":1355862,"visible":true,"origin":"","legend":"\u003cp\u003eThe schematic process of the algorithm for automatic identification of coronary stents (Stent_filter). Step 1. The convolutional neural network for identification of the stent after image pre-processing. Step 2. False-positive reduction algorithm based on the predictive volume (mm\u003csup\u003e3\u003c/sup\u003e) for each pixel of the mask. If the number of pixels (m) is greater than the predicted number of pixels for the suspected stent mask (n), it was considered a true positive. \u0026nbsp;\u003c/p\u003e","description":"","filename":"1.png","url":"https://assets-eu.researchsquare.com/files/rs-4543450/v1/e6f85e8dc53fd44f2fe7b1d8.png"},{"id":62655122,"identity":"b13afbe8-39bc-4594-809f-37201361b0d6","added_by":"auto","created_at":"2024-08-17 01:32:30","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":702161,"visible":true,"origin":"","legend":"\u003cp\u003eIdentifying and distinguishing coronary stents from coronary calcifications. (A) In a coronary artery calcium (CAC) scoring image, both the coronary calcification (black arrow) and coronary stent (white arrow) are aligned in a single file within the left anterior descending artery. (B) A radiologist has annotated the coronary calcification with a red mask and the coronary stent with a green mask for clear identification. (C) Stent_filter accurately identifies the coronary stent (white arrow), differentiating it from the coronary calcification (black arrow).\u003c/p\u003e","description":"","filename":"2.png","url":"https://assets-eu.researchsquare.com/files/rs-4543450/v1/ba1d6d45047755bf81ba7f01.png"},{"id":62654429,"identity":"c50d851f-6065-4394-aaab-0fcf2bd60416","added_by":"auto","created_at":"2024-08-17 01:24:30","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":683880,"visible":true,"origin":"","legend":"\u003cp\u003eFalse-positive identification by Stent_filter: misclassification of coronary calcification as a stent. (A) A coronary artery calcium (CAC) scoring image showcases segment-long calcifications within the proximal segments of the left anterior descending (LAD) and left circumflex arteries (black arrows) (B) A radiologist annotated these calcifications with a red mask. (C) However, the Stent_filter mistakenly recognized the calcification in the LAD as a coronary stent, leading to a false-positive result.\u003c/p\u003e","description":"","filename":"3.png","url":"https://assets-eu.researchsquare.com/files/rs-4543450/v1/12af59b99fd53915263dfa9b.png"},{"id":68207373,"identity":"ab62fdc9-6889-4cb7-a676-d093efa73471","added_by":"auto","created_at":"2024-11-04 16:37:03","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":4562660,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-4543450/v1/96aa2d09-ce20-4b0c-8050-e94b560ea177.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"Automatic Identification of Coronary Stent in Coronary Calcium Scoring CT using Deep Learning","fulltext":[{"header":"Introduction","content":"\u003cp\u003eCoronary artery calcium (CAC) scoring has become a common method for assessing cardiovascular risk in clinical practice \u003csup\u003e\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e,\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e\u003c/sup\u003e. Traditionally, this scoring has been performed using non-enhanced electrocardiography-synchronized cardiac computed tomography (CT), requiring skilled professionals for manual interventions, which can be time-consuming \u003csup\u003e\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e,\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e\u003c/sup\u003e. However, recent advances in artificial intelligence (AI) have facilitated the development of fully automatic calcium scoring system using non-enhanced CT \u003csup\u003e\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e\u003c/sup\u003e, including non-gated low-dose chest CT scans or radiation therapy treatment planning CT \u003csup\u003e\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e,\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eDespite these technological advancements promising to streamline the CAC scoring workflow, one crucial manual step remains: identifying patients with coronary stents. The utility of CAC scoring in patients who have undergone stent or bypass surgery is known to be uncertain. In addition, stents themselves can mimic coronary calcium, leading to an overestimation of calcification burden. As a result, radiologic professionals manually exclude these patients from CAC scoring. Consequently, if it becomes feasible to automatically identify coronary stents on CAC CT scans and triage such patients, it could significantly improve the efficiency of both manual and automated CAC scoring workflows. Despite the potential benefits, no automated algorithm for stent identification from CAC scans exists, to our knowledge. Moreover, the technical challenge arises not only from identifying the stent but also from distinguishing it from the heavy coronary artery calcification that often coexists in patients who have underwent stent insertion \u003csup\u003e\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e\u003c/sup\u003e. Therefore, in this study, we aimed to develop and evaluate a deep learning-based coronary stent filtering algorithm (Stent_filter), along with a false-positive reduction algorithm, for use in CAC scoring CT scans, using a multicenter CAC dataset.\u003c/p\u003e"},{"header":"Methods","content":"\u003cp\u003eThis retrospective study was approved by the institutional review board committee of each participating institution (institutional review board committee of the Asan Medical Center, institutional review board committee of the Dankook University Hospital, and institutional review board committee of the Korea University Ansan Hospital) and informed consent was waived by the institutional review board from all of three institutions due to the retrospective nature of observational study. This study was performed in accordance with the Helsinki Declaration.\u003c/p\u003e \u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003eDevelopment of Stent_filter\u003c/h2\u003e \u003cp\u003eThe development and validation dataset comprised non-enhanced CAC scans obtained from 108 patients (mean age, 60.7\u0026thinsp;\u0026plusmn;\u0026thinsp;11.0 years; range, 37\u0026ndash;84; 84 men [77.8%]) who underwent CT scans for chest pain between January 2016 to December 2020 (Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e). All scans were acquired from the three different scanners from a single vendor, with a slice thickness of 3 mm. Initially, segmentation for the labeled dataset was conducted using a research prototype (AVIEW CAC, Coreline Soft, Co. Ltd.). Subsequently, a cardiothoracic radiologist (Y.A., with 2 years of experience in cardiothoracic radiology) meticulously reviewed the entire set of generated masks and manually differentiated them for stents against those for coronary calcification. In cases when stents and coronary calcification were contiguous, manual adjustments were made to separate the stent from the calcification. In cases where discrimination between stent and calcification was ambiguous, medical records, including invasive coronary angiography records, were reviewed.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eCharacteristics of development and test datasets\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"7\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c7\" colnum=\"7\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCharacteristics\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colspan=\"2\" nameend=\"c3\" namest=\"c2\"\u003e \u003cp\u003eDevelopment Dataset\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colspan=\"4\" nameend=\"c7\" namest=\"c4\"\u003e \u003cp\u003eTest Dataset\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eDevelopment\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eValidation\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eInternal-1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eInternal-2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eExternal-1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eExternal-2\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNo. of patients\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e77\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e31\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e355\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e396\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e105\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e62\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAge (y)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e61.0\u0026thinsp;\u0026plusmn;\u0026thinsp;10.7\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e60.0\u0026thinsp;\u0026plusmn;\u0026thinsp;11.9\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e62.1\u0026thinsp;\u0026plusmn;\u0026thinsp;9.6\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e62.5\u0026thinsp;\u0026plusmn;\u0026thinsp;9.2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e58.3\u0026thinsp;\u0026plusmn;\u0026thinsp;10.7\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e65.3\u0026thinsp;\u0026plusmn;\u0026thinsp;9.3\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNo. of male patients\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e63 (81.8)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e21 (67.7)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e277 (78.0)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e310 (78.3)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e91 (86.7)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e50 (80.6)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTube voltage (kVp)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e100\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0 (0.0)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0 (0.0)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e18 (5.1)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e5 (1.3)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0 (0.0)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0 (0.0)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e120\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e77 (100.0)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e31 (100.0)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e337 (94.9)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e391 (98.7)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e105 (100.0)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e62 (100.0)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTube current (mAs)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e83.0\u0026thinsp;\u0026plusmn;\u0026thinsp;20.4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e81.5\u0026thinsp;\u0026plusmn;\u0026thinsp;19.6\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e72.5\u0026thinsp;\u0026plusmn;\u0026thinsp;28.8\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e32.3\u0026thinsp;\u0026plusmn;\u0026thinsp;42.2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e43.6\u0026thinsp;\u0026plusmn;\u0026thinsp;4.2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e30.2\u0026thinsp;\u0026plusmn;\u0026thinsp;2.6\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSlice thickness\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e2.5 mm\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0 (0.0)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0 (0.0)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1 (0.3)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e252 (63.6)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e102 (97.1)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e62 (100.0)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e3 mm\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e77 (100.0)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e31 (100.0)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e354 (99.7)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e144 (36.4)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0 (0.0)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0 (0.0)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e4 mm\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0 (0.0)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0 (0.0)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0 (0.0)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0 (0.0)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e3 (2.9)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0 (0.0)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eScanners\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eSOMATOM Definition, SOMATOM Definition Flash, SOMATOM Force\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eSOMATOM Definition, SOMATOM Definition Flash, SOMATOM Force\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eSOMATOM Definition, SOMATOM Definition Flash, LightSpeed VCT\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eSOMATOM Definition, SOMATOM Definition Flash, LightSpeed VCT, Discovery CT750 HD\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eIngenuity Core 128, IQon - Spectral CT\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eiCT 256,\u003c/p\u003e \u003cp\u003eIngenuity CT\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMask\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCalcification in LAD\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e56\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e26\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e260\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e391\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e83\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e49\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCalcification in LCX\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e46\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e23\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e266\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e289\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e67\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e42\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCalcification in LM\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e38\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e16\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e193\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e218\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e35\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCalcification in RCA\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e51\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e27\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e295\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e356\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e80\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e57\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eStent\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e62\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e12\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e352\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e103\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e62\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003ctfoot\u003e \u003ctr\u003e\u003ctd colspan=\"7\"\u003eNote.\u0026mdash;Data are presented as numbers with percentage in parenthesis or as means\u0026thinsp;\u0026plusmn;\u0026thinsp;standard deviations. LAD\u0026thinsp;=\u0026thinsp;left anterior descending artery, LCX\u0026thinsp;=\u0026thinsp;left circumflex coronary artery, LM\u0026thinsp;=\u0026thinsp;left main coronary artery, RCA\u0026thinsp;=\u0026thinsp;right coronary artery.\u003c/td\u003e\u003c/tr\u003e \u003c/tfoot\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003eTo develop the false-positive reduction algorithm in the final architecture, 2036 CAC scans with substantial coronary calcium without stents (average Agatston calcium score: 523.8) were utilized. These scans constituted a random subset of asymptomatic individuals who underwent health screening \u003csup\u003e\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec4\" class=\"Section2\"\u003e \u003ch2\u003eTest dataset\u003c/h2\u003e \u003cp\u003eThe portion of the internal test dataset (Internal-2) was derived from a previous study \u003csup\u003e\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e\u003c/sup\u003e, which was utilized for the development and validation of a deep learning system for automatic coronary calcium scoring. This dataset comprised 396 patients and exclusively contained coronary calcifications without stents, serving as a negative control. The remaining three test datasets, obtained from different institutions, included both stents and coronary calcifications. Similar to the development dataset, masks for the labeled dataset were initially generated via software, and all generated masks were reviewed by the same cardiothoracic radiologist (Y.A.). Ultimately, a total of four datasets comprising 918 non-enhanced CAC images, including 517 stents, were employed to evaluate the performance of Stent_filter. Detailed information regarding the test dataset is presented in Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec5\" class=\"Section2\"\u003e \u003ch2\u003eAutomatic coronary stent identification: Stent_filter\u003c/h2\u003e \u003cp\u003eWe developed the Stent_filter methodology to detect coronary stents and calcium separately. This methodology involves two main pipelines: 1) For stent and calcium detection, we used an advanced technique called Mask RCNN, based on the ResNet101 architecture, to accurately identify stents and calcium deposits \u003csup\u003e\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e\u003c/sup\u003e. This model was initialized with pre-trained weights from a large dataset called Microsoft Common Objects in Context (MSCOCO), which helped improve its accuracy. The ResNet101 architecture has special features called residual connections, which effectively capture intricate features while mitigating the issue of vanishing gradients in deep networks. The Mask RCNN model generates potential object candidates by use of a region proposal network. Then, it utilizes the feature pyramid network to detect objects of varying sizes efficiently. To ensure accurate and robust performance, several implementation details were carefully considered. Image preprocessing involved windowing with a window width of 1500 HU and a window level of 300 HU. The input tensor for the model was constructed in 2.5D, incorporating frames from both above and below. Input frames were standardized to RGB channels of size 512 x 512. During training, model parameters were optimized using a stochastic gradient descent optimizer with a mini-batch size of 2 images. Training spanned 100 epochs with an initial learning rate of 1e-3 and included weight decay. Learning rate adjustments were made, reducing by a factor of 0.1 every 25 epochs. The multi-task loss function employed for detection encompassed regression and cross-entropy components. Augmentation techniques, including sharpening, Gaussian blur, additive Gaussian noise, multiplication, and affine transformations (scale, shift, rotate), were also applied. During testing, we set a minimum confidence score of 0.5 for stent detection. 2) For false positive reduction, we applied a rule-based process to further refine the results. We performed 3D connective component analysis to make multiple consecutive detections into one stent candidate lesion. The candidate lesion which is smaller than 100 mm\u003csup\u003e\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e\u003c/sup\u003e in size was discarded.\u003c/p\u003e \u003cp\u003eThe overall procedure is illustrated in Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e. We evaluated the performance of both pipelines using a computer with specifications: an Intel core i7-6700 processor, 32 GB of RAM, and an NVIDIA GeForce GTX 1060 graphics card with 3 GB memory. The time it took for the model to analyze a CT scan with 45 to 60 frames ranged from approximately 20 to 30 seconds.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec6\" class=\"Section2\"\u003e \u003ch2\u003eStatistical analysis\u003c/h2\u003e \u003cp\u003eAnalyses were performed on the per-patient and per-lesion diagnostic performance of Stent_filter. In the per-patient analysis, lesions identified by both Stent_filter and by the radiologist were considered stents, while patients with any overlapping lesion were classified true positives for having stents. Mismatched lesions were classified into false-positive (i.e., coronary calcifications determined as stent by Stent_filter) and false-negative results (i.e., missed stent by Stent_filter). For per-lesion analysis, performance metrics were calculated based on the target coronary artery (i.e., left anterior descending \u003csup\u003e\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e\u003c/sup\u003e and left main coronary artery [LM], left circumflex coronary artery [LCX], and right coronary artery [RCA]). The coronary artery containing the greater portion of the stent mask was designated as the target coronary artery. Per-lesion sensitivity was calculated as (number of true positives /total number of stents in the corresponding vessel) \u0026times; 100. The false-positive ratio was calculated as (number of false-positive lesions/total number of patients). A cardiothoracic radiologist reviewed all mismatched lesions to determine the cause of the mismatch.\u003c/p\u003e \u003c/div\u003e"},{"header":"Results","content":"\u003cdiv id=\"Sec8\" class=\"Section2\"\u003e \u003ch2\u003eStent_filter for coronary stent identification\u003c/h2\u003e \u003cp\u003eIn the per-patient analysis, Stent_filter demonstrated high accuracy in distinguishing stents from coronary calcifications, achieving an overall sensitivity of 98.8% (511/517) and a specificity of 95.0% (381/401) (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e). The overall false-positive ratio was 0.022 (20/918) (Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e). Notably, review of the mismatched lesions revealed that all false-positive results were tubular calcifications mimicking stents without extracoronary lesions (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e)\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab2\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 2\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003ePerformance of Stent_filter: Analysis per patient\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"6\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colspan=\"5\" nameend=\"c6\" namest=\"c2\"\u003e \u003cp\u003eTest Dataset\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePerformance measure\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eInternal-1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eInternal-2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eExternal-1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eExternal-2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eOverall\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNo. of patients\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e355\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e396\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e105\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e62\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e918\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTrue positive\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e349\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eNA\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e101\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e61\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e511\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eFalse negative\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eNA\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e6\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eFalse positive\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e17\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e20\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTrue negative\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e379\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e381\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSensitivity\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.991\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eNA\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.980\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.984\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0.988\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSpecificity\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.000\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.957\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1.000\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.000\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0.950\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eFP ratio\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.008\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.043\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.000\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.000\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0.022\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003ctfoot\u003e \u003ctr\u003e\u003ctd colspan=\"6\"\u003eFP\u0026thinsp;=\u0026thinsp;false positive, NA\u0026thinsp;=\u0026thinsp;not applicable.\u003c/td\u003e\u003c/tr\u003e \u003c/tfoot\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eAcross the three test datasets from different institutions, Stent_filter consistently demonstrated high sensitivity, ranging from 98.0% (External-1; 101/103) to 99.1% (Internal-1; 349/352). It also exhibited a specificity of 95.7% (379/396) in the Internal-2 dataset, which served as a negative control. Notably, the false-positive ratio was significantly lower in the dataset containing stents (Internal-1; 0.008 [3/355]) compared with the dataset without stents (Internal-2; 0.043 [17/396], p\u0026thinsp;=\u0026thinsp;0.008). Furthermore, in the two external test datasets, there were no instances of completely false-positive results for a single patient without stent.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec9\" class=\"Section2\"\u003e \u003ch2\u003ePer-lesion analysis\u003c/h2\u003e \u003cp\u003eOverall, per-lesion sensitivity ranged from 94.5% (449/475) for the LAD and LM to 98.4% (179/182) for the RCA, as detailed in Table\u0026nbsp;\u003cspan refid=\"Tab3\" class=\"InternalRef\"\u003e3\u003c/span\u003e. Among all test datasets, the LAD and LM regions, which had the highest number of stents, also exhibited the highest false-positive ratio, with an overall false-positive ratio for LAD and LM at 0.038 (35/918). Notably, in the external-2 dataset, the false-positive ratio was highest across all coronary arteries: 0.113 for LAD and LM (7/62), 0.065 for the LCX (4/62), and 0.081 for the RCA (5/62), compared with the other three datasets.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab3\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 3\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003ePerformance of Stent_filter: Analysis per lesion\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"16\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c7\" colnum=\"7\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c8\" colnum=\"8\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c9\" colnum=\"9\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c10\" colnum=\"10\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c11\" colnum=\"11\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c12\" colnum=\"12\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c13\" colnum=\"13\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c14\" colnum=\"14\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c15\" colnum=\"15\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c16\" colnum=\"16\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colspan=\"15\" nameend=\"c16\" namest=\"c2\"\u003e \u003cp\u003eTest Dataset\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePerformance measure\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c4\" namest=\"c2\"\u003e \u003cp\u003eInternal-1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c7\" namest=\"c5\"\u003e \u003cp\u003eInternal-2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c10\" namest=\"c8\"\u003e \u003cp\u003eExternal-1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c13\" namest=\"c11\"\u003e \u003cp\u003eExternal-2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c16\" namest=\"c14\"\u003e \u003cp\u003eOverall\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNo. of patients\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c4\" namest=\"c2\"\u003e \u003cp\u003e355\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c7\" namest=\"c5\"\u003e \u003cp\u003e396\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c10\" namest=\"c8\"\u003e \u003cp\u003e105\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c13\" namest=\"c11\"\u003e \u003cp\u003e62\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c16\" namest=\"c14\"\u003e \u003cp\u003e918\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTotal lesions\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c4\" namest=\"c2\"\u003e \u003cp\u003e548\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c7\" namest=\"c5\"\u003e \u003cp\u003eNA\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c10\" namest=\"c8\"\u003e \u003cp\u003e144\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c13\" namest=\"c11\"\u003e \u003cp\u003e87\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c16\" namest=\"c14\"\u003e \u003cp\u003e779\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eVessel\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eLAD, LM\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eLCX\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eRCA\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eLAD, LM\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eLCX\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eRCA\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003eLAD, LM\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003eLCX\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003eRCA\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e \u003cp\u003eLAD, LM\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c12\"\u003e \u003cp\u003eLCX\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c13\"\u003e \u003cp\u003eRCA\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c14\"\u003e \u003cp\u003eLAD, LM\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c15\"\u003e \u003cp\u003eLCX\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c16\"\u003e \u003cp\u003eRCA\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNo. of lesions\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e368\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e81\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e129\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e12\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e84\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e30\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003e46\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e \u003cp\u003e58\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c12\"\u003e \u003cp\u003e20\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c13\"\u003e \u003cp\u003e25\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c14\"\u003e \u003cp\u003e510\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c15\"\u003e \u003cp\u003e131\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c16\"\u003e \u003cp\u003e200\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNo. of stents\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e348\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e79\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e121\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eNA\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eNA\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eNA\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e76\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e27\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003e41\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e \u003cp\u003e51\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c12\"\u003e \u003cp\u003e16\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c13\"\u003e \u003cp\u003e20\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c14\"\u003e \u003cp\u003e475\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c15\"\u003e \u003cp\u003e122\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c16\"\u003e \u003cp\u003e182\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTrue positive\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e327\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e76\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e119\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eNA\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eNA\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eNA\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e72\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e27\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003e40\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e \u003cp\u003e50\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c12\"\u003e \u003cp\u003e15\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c13\"\u003e \u003cp\u003e20\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c14\"\u003e \u003cp\u003e449\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c15\"\u003e \u003cp\u003e118\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c16\"\u003e \u003cp\u003e179\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eFalse negative\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e21\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eNA\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eNA\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eNA\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003e1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e \u003cp\u003e1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c12\"\u003e \u003cp\u003e1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c13\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c14\"\u003e \u003cp\u003e26\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c15\"\u003e \u003cp\u003e4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c16\"\u003e \u003cp\u003e3\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eFalse positive\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e20\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e8\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e12\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e8\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003e5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e \u003cp\u003e7\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c12\"\u003e \u003cp\u003e4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c13\"\u003e \u003cp\u003e5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c14\"\u003e \u003cp\u003e35\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c15\"\u003e \u003cp\u003e9\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c16\"\u003e \u003cp\u003e18\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSensitivity\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.940\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.962\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.983\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eNA\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eNA\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003eNA\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e0.947\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e1.000\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003e0.976\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e \u003cp\u003e0.980\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c12\"\u003e \u003cp\u003e0.938\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c13\"\u003e \u003cp\u003e1.000\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c14\"\u003e \u003cp\u003e0.945\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c15\"\u003e \u003cp\u003e0.967\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c16\"\u003e \u003cp\u003e0.984\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eFP ratio\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.056\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.006\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.023\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.030\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0.003\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0.013\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e0.076\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e0.029\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003e0.048\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e \u003cp\u003e0.113\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c12\"\u003e \u003cp\u003e0.065\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c13\"\u003e \u003cp\u003e0.081\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c14\"\u003e \u003cp\u003e0.038\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c15\"\u003e \u003cp\u003e0.010\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c16\"\u003e \u003cp\u003e0.020\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003ctfoot\u003e \u003ctr\u003e\u003ctd colspan=\"16\"\u003eFP\u0026thinsp;=\u0026thinsp;false positive, NA\u0026thinsp;=\u0026thinsp;not applicable, LAD\u0026thinsp;=\u0026thinsp;left anterior descending artery, LCX\u0026thinsp;=\u0026thinsp;left circumflex coronary artery, LM\u0026thinsp;=\u0026thinsp;left main coronary artery, RCA\u0026thinsp;=\u0026thinsp;right coronary artery.\u003c/td\u003e\u003c/tr\u003e \u003c/tfoot\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec10\" class=\"Section2\"\u003e \u003ch2\u003eStent_filter according to the stent length\u003c/h2\u003e \u003cp\u003eAcross the vessels, sensitivity was highest for medium-length stents (319/334) and lowest for short stents (36/40) (Table\u0026nbsp;\u003cspan refid=\"Tab4\" class=\"InternalRef\"\u003e4\u003c/span\u003e). Notably, Stent_filter exhibited the lowest sensitivity for short (\u0026le;\u0026thinsp;15 mm) stents in the RCA (9/11), while it showed the highest sensitivity for short stents in the LCX (6/6) and medium-length stents (15 mm\u0026thinsp;\u0026lt;\u0026thinsp;stent length\u0026thinsp;\u0026le;\u0026thinsp;30 mm) in the RCA (91/91).\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab4\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 4\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003ePerformance of Stent_filter: according to the stent length\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"13\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c7\" colnum=\"7\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c8\" colnum=\"8\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c9\" colnum=\"9\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c10\" colnum=\"10\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c11\" colnum=\"11\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c12\" colnum=\"12\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c13\" colnum=\"13\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colspan=\"12\" nameend=\"c13\" namest=\"c2\"\u003e \u003cp\u003eTest Dataset\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c4\" namest=\"c2\"\u003e \u003cp\u003eInternal-1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c7\" namest=\"c5\"\u003e \u003cp\u003eExternal-1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c10\" namest=\"c8\"\u003e \u003cp\u003eExternal-2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c13\" namest=\"c11\"\u003e \u003cp\u003eOverall\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eStent length, mm\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003ex\u0026thinsp;\u0026lt;\u0026thinsp;=\u0026thinsp;15\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e15\u0026thinsp;\u0026lt;\u0026thinsp;x\u0026thinsp;\u0026lt;\u0026thinsp;=\u0026thinsp;30\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e30\u0026thinsp;\u0026lt;\u0026thinsp;x\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003ex\u0026thinsp;\u0026lt;\u0026thinsp;=\u0026thinsp;15\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e15\u0026thinsp;\u0026lt;\u0026thinsp;x\u0026thinsp;\u0026lt;\u0026thinsp;=\u0026thinsp;30\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e30\u0026thinsp;\u0026lt;\u0026thinsp;x\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003ex\u0026thinsp;\u0026lt;\u0026thinsp;=\u0026thinsp;15\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e15\u0026thinsp;\u0026lt;\u0026thinsp;x\u0026thinsp;\u0026lt;\u0026thinsp;=\u0026thinsp;30\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003e30\u0026thinsp;\u0026lt;\u0026thinsp;x\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e \u003cp\u003ex\u0026thinsp;\u0026lt;\u0026thinsp;=\u0026thinsp;15\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c12\"\u003e \u003cp\u003e15\u0026thinsp;\u0026lt;\u0026thinsp;x\u0026thinsp;\u0026lt;\u0026thinsp;=\u0026thinsp;30\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c13\"\u003e \u003cp\u003e30\u0026thinsp;\u0026lt;\u0026thinsp;x\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLAD, LM\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e19\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e113\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e170\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e44\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e24\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e20\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003e20\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e \u003cp\u003e23\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c12\"\u003e \u003cp\u003e177\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c13\"\u003e \u003cp\u003e214\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNo. of FN\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e14\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e \u003cp\u003e2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c12\"\u003e \u003cp\u003e10\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c13\"\u003e \u003cp\u003e22\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSensitivity\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.947\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.956\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.918\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.500\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0.955\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0.958\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e1.000\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e0.950\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003e1.000\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e \u003cp\u003e0.913\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c12\"\u003e \u003cp\u003e0.944\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c13\"\u003e \u003cp\u003e0.897\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLCX\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e35\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e27\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e19\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e6\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e12\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003e4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e \u003cp\u003e6\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c12\"\u003e \u003cp\u003e66\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c13\"\u003e \u003cp\u003e37\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNo. of FN\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c12\"\u003e \u003cp\u003e5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c13\"\u003e \u003cp\u003e1\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSensitivity\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1.000\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.886\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1.000\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e1.000\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e1.000\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e1.000\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003eNA\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e0.917\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003e1.000\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e \u003cp\u003e1.000\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c12\"\u003e \u003cp\u003e0.924\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c13\"\u003e \u003cp\u003e0.973\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eRCA\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e7\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e55\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e69\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e27\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e16\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e9\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003e13\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e \u003cp\u003e11\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c12\"\u003e \u003cp\u003e91\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c13\"\u003e \u003cp\u003e98\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNo. of FN\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e \u003cp\u003e2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c12\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c13\"\u003e \u003cp\u003e1\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSensitivity\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.857\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1.000\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.986\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.667\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e1.000\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e1.000\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e1.000\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e1.000\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003e1.000\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e \u003cp\u003e0.818\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c12\"\u003e \u003cp\u003e1.000\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c13\"\u003e \u003cp\u003e0.90\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTotal of no. of lesions\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e31\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e203\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e266\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e6\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e90\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e46\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e41\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003e37\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e \u003cp\u003e40\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c12\"\u003e \u003cp\u003e334\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c13\"\u003e \u003cp\u003e349\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNo. of FN\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e9\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e15\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e \u003cp\u003e4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c12\"\u003e \u003cp\u003e15\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c13\"\u003e \u003cp\u003e24\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSensitivity\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.935\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.956\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.944\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.667\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0.978\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0.978\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e1.000\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e0.951\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c10\"\u003e \u003cp\u003e1.000\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c11\"\u003e \u003cp\u003e0.900\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c12\"\u003e \u003cp\u003e0.955\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c13\"\u003e \u003cp\u003e0.931\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eTotal\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c4\" namest=\"c2\"\u003e \u003cp\u003e500\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c7\" namest=\"c5\"\u003e \u003cp\u003e142\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c10\" namest=\"c8\"\u003e \u003cp\u003e81\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c13\" namest=\"c11\"\u003e \u003cp\u003e723\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNo. of FN\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c4\" namest=\"c2\"\u003e \u003cp\u003e26\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c7\" namest=\"c5\"\u003e \u003cp\u003e5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c10\" namest=\"c8\"\u003e \u003cp\u003e2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c13\" namest=\"c11\"\u003e \u003cp\u003e43\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSensitivity\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c4\" namest=\"c2\"\u003e \u003cp\u003e0..948\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c7\" namest=\"c5\"\u003e \u003cp\u003e0.965\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c10\" namest=\"c8\"\u003e \u003cp\u003e0.975\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"3\" nameend=\"c13\" namest=\"c11\"\u003e \u003cp\u003e0.941\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003ctfoot\u003e \u003ctr\u003e\u003ctd colspan=\"13\"\u003eFN\u0026thinsp;=\u0026thinsp;false negative, NA\u0026thinsp;=\u0026thinsp;not applicable, LAD\u0026thinsp;=\u0026thinsp;left anterior descending artery, LCX\u0026thinsp;=\u0026thinsp;left circumflex coronary artery, LM\u0026thinsp;=\u0026thinsp;left main coronary artery, RCA\u0026thinsp;=\u0026thinsp;right coronary artery\u003c/td\u003e\u003c/tr\u003e \u003c/tfoot\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003c/div\u003e"},{"header":"Discussion","content":"\u003cp\u003eAutomatic pre-screening of pre-existing stents, whose prognostic value remains uncertain, could potentially reduce workload and enhance efficiency. However, such a solution has not yet been developed and validated. In this study, we developed and tested the performance of an automatic identification algorithm for stents. Our results demonstrate highly accurate stent identification (sensitivity of 98.8% [511/517] and specificity of 95.0% [381/401]), along with a false-positive rate of 0.022 across both the internal and external test datasets.\u003c/p\u003e \u003cp\u003eWith recent advancements in AI, automated CAC scoring has reached a state-of-the-art level, demonstrating sensitivities from 93\u0026ndash;98% and specificities from 93\u0026ndash;100%, along with a high interclass correlation coefficient of 1.00 across large multicenter and multivendor cohorts \u003csup\u003e\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e,\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e,\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e\u003c/sup\u003e. As a result, the primary challenge remaining for automated CAC scoring, to minimize additional work for radiologists, lies in the ability of the AI to differentiate non-coronary high-density lesions from vascular calcifications. In particular, the automated identification of coronary stents holds significance beyond simply reducing false positives; coronary stents introduce uncertainty into the prognostic value of CAC scoring, suggesting that their automatic exclusion could substantially decrease workload. However, this area remains under-investigated, and even in studies focused on automated CAC scoring, the presence of a coronary stent has often been a criterion for exclusion \u003csup\u003e\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e,\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e,\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eIn our multicenter and multivendor cohort study involving patients with preexisting coronary calcifications, we evaluated the performance of our Stent_filter. The filter demonstrated a sensitivity exceeding 98% and a specificity ranging from 95.7\u0026ndash;100% across all internal and external datasets, which is considered clinically acceptable compared with the reported performance of automated CAC scoring methods. Specifically, the Internal-2 dataset, characterized by the presence of only coronary artery calcifications, yielded a specificity of 95.7% (379/396), underscoring the effectiveness of the algorithm in discrimination against calcifications. Moreover, the per-lesion sensitivity for the RCA at 98.4% (179/182) highlights its capability, especially notable given the RCA's susceptibility to motion artifacts due to its perpendicular motion relative to the axial plane of CT scans \u003csup\u003e\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eConversely, addressing the false-positive rate remains a crucial step in the clinical application of automated CAC systems. Despite a very low false-negative rate per patient (0.007 [6/918]), which typically does not significantly impact the workflow\u0026mdash;since physicians familiar with the patient's medical history can disregard automatic CAC scoring results in patients with coronary stents\u0026mdash;the occurrence of false positives when skipping CAC scoring in indicated patients is noteworthy, despite the low overall false-positive rate per patient (0.022 [20/918]).\u003c/p\u003e \u003cp\u003eUpon reviewing all cases, we confirmed the following findings. First, all false-positive lesions were indeed coronary calcifications and not extracoronary lesions such as valvular or pericardial calcifications. Second, the false positives were identified as segmental tubular calcifications of the coronary artery, which could mimic the appearance of a coronary stent, especially in cases of in-stent restenosis or stents under coronary calcification. Although we did not conduct an additional reader study to determine whether radiologists could also confuse these lesions with coronary stents when blinded to medical history, our review suggests that these lesions could be challenging for radiologists to distinguish from stents without access to the patient's medical history. Dense calcification can mimic the appearance of a stent, and to mitigate these false positives \u003csup\u003e\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e\u003c/sup\u003e, we implemented a false-positive reduction process as post-processing. This approach is based on similar principles to those used by Komatsu et al. \u003csup\u003e\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e\u003c/sup\u003e, who also developed an algorithm, the 3-dimensional plaque map, based on similar principles to ours, to differentiate heavy calcification from stents.\u003c/p\u003e \u003cp\u003eWe assume that neither radiologists nor radiologic technicians can achieve 100% accuracy in the screening process for CAC scoring candidates, and our algorithm's false-positive rate could be seen as an extension of this intrinsic error. Furthermore, given the lower per-patient false-positive rate observed in patients with coronary stents compared with those without stents, the clinical impact of false-positive results on patient triage in real-world clinical settings could be minimal. However, further technical refinement or the addition of complementary functions, such as incorporating available patient medical records, could assist in more accurately triaging patients for CAC scoring.\u003c/p\u003e \u003cp\u003eConsidering the current protocol, which requires radiologic technicians to manually confirm the absence of stents in CAC scans before proceeding with CAC scoring, the adoption of Stent_filter could substantially reduce their workload. This reduction is achieved by eliminating the need for this manual confirmation step and subsequent unnecessary manual CAC scoring. Additionally, with the availability of a developed automated CAC system, it becomes possible to implement a fully automated CAC scoring process. This process would not only automate the scoring itself but also include automated triage, streamlining the entire procedure and enhancing efficiency.\u003c/p\u003e \u003cp\u003eOur study has several limitations. First, it was fundamentally a retrospective feasibility study validating the automatic identification performance. Its pragmatic effect, in other words, whether and how much it could reduce the workload, was not assessed. Second, its feasibility in non-gated scans was not evaluated. While deep-learning automated CAC has advanced considerably, enabling scoring in non-gated low-dose chest CT scans or enhanced CT coronary angiography scans \u003csup\u003e\u003cspan additionalcitationids=\"CR16\" citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e\u003c/sup\u003e, the efficacy of Stent_filter in these non-dedicated scans requires further technology development in line with their protocols.\u003c/p\u003e \u003cp\u003eIn conclusion, the automated Stent_filter accurately distinguishes coronary stents from preexisting coronary calcifications. This approach holds the potential to serve as a filter within a fully automated CAC scoring workflow, streamlining the process efficiently.\u003c/p\u003e"},{"header":"Abbreviations","content":"\u003cp\u003eAI artificial intelligence\u003c/p\u003e\n\u003cp\u003eCAC coronary artery calcium\u003c/p\u003e\n\u003cp\u003eLAD left anterior descending artery\u003c/p\u003e\n\u003cp\u003eLCX left circumflex coronary artery\u003c/p\u003e\n\u003cp\u003eLM left main coronary artery\u003c/p\u003e\n\u003cp\u003eRCA right coronary artery\u003c/p\u003e\n\n\n\n"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eData availability:\u0026nbsp;\u003c/strong\u003eThe datasets used and/or analyzed during the current study available from the corresponding author on reasonable request.\u003c/p\u003e\n\u003cp\u003e\u0026nbsp;\u003cstrong\u003eFunding:\u003c/strong\u003e None\u003c/p\u003e\n\u003cp\u003e\u0026nbsp;\u003cstrong\u003eAuthor information\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eContributions:\u003c/strong\u003e Conceived and designed the analysis: J.L. and D.H.Y.; Collected the data: Y.A., G.J., D.L., C.K., and D.H.Y.; Contributed data or analysis tools: G.J., J.L., and D.H.Y.; Performed the analysis: Y.A. and G.J.; Wrote the paper: Y.A., G.J., J.L., and D.H.Y.; Manuscript editing: Y.A., G.J., D.L., C.K., J.L., and D.H.Y.\u003c/p\u003e\n\u003cp\u003e\u0026nbsp;\u003cstrong\u003eEthics declarations\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCompeting interests:\u003c/strong\u003e The authors declare no competing interests.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eGreenland, P., Blaha, M. J., Budoff, M. J., Erbel, R. \u0026amp; Watson, K. E. Coronary Calcium Score and Cardiovascular Risk. \u003cem\u003eJ Am Coll Cardiol\u003c/em\u003e \u003cstrong\u003e72\u003c/strong\u003e, 434-447, doi:10.1016/j.jacc.2018.05.027 (2018).\u003c/li\u003e\n\u003cli\u003eGreenland, P.\u003cem\u003e et al.\u003c/em\u003e ACCF/AHA 2007 clinical expert consensus document on coronary artery calcium scoring by computed tomography in global cardiovascular risk assessment and in evaluation of patients with chest pain: a report of the American College of Cardiology Foundation Clinical Expert Consensus Task Force (ACCF/AHA Writing Committee to Update the 2000 Expert Consensus Document on Electron Beam Computed Tomography) developed in collaboration with the Society of Atherosclerosis Imaging and Prevention and the Society of Cardiovascular Computed Tomography. \u003cem\u003eJ Am Coll Cardiol\u003c/em\u003e \u003cstrong\u003e49\u003c/strong\u003e, 378-402, doi:10.1016/j.jacc.2006.10.001 (2007).\u003c/li\u003e\n\u003cli\u003eBlaha, M. J., Mortensen, M. B., Kianoush, S., Tota-Maharaj, R. \u0026amp; Cainzos-Achirica, M. Coronary Artery Calcium Scoring: Is It Time for a Change in Methodology? \u003cem\u003eJACC Cardiovasc Imaging\u003c/em\u003e \u003cstrong\u003e10\u003c/strong\u003e, 923-937, doi:10.1016/j.jcmg.2017.05.007 (2017).\u003c/li\u003e\n\u003cli\u003eHecht, H. S.\u003cem\u003e et al.\u003c/em\u003e 2016 SCCT/STR guidelines for coronary artery calcium scoring of noncontrast noncardiac chest CT scans: A report of the Society of Cardiovascular Computed Tomography and Society of Thoracic Radiology. \u003cem\u003eJ Cardiovasc Comput Tomogr\u003c/em\u003e \u003cstrong\u003e11\u003c/strong\u003e, 74-84, doi:10.1016/j.jcct.2016.11.003 (2017).\u003c/li\u003e\n\u003cli\u003eLee, J. G.\u003cem\u003e et al.\u003c/em\u003e Fully Automatic Coronary Calcium Score Software Empowered by Artificial Intelligence Technology: Validation Study Using Three CT Cohorts. \u003cem\u003eKorean J Radiol\u003c/em\u003e \u003cstrong\u003e22\u003c/strong\u003e, 1764-1776, doi:10.3348/kjr.2021.0148 (2021).\u003c/li\u003e\n\u003cli\u003evan Velzen, S. G. M.\u003cem\u003e et al.\u003c/em\u003e Deep Learning for Automatic Calcium Scoring in CT: Validation Using Multiple Cardiac CT and Chest CT Protocols. \u003cem\u003eRadiology\u003c/em\u003e \u003cstrong\u003e295\u003c/strong\u003e, 66-79, doi:10.1148/radiol.2020191621 (2020).\u003c/li\u003e\n\u003cli\u003eAbdelrahman, K.\u003cem\u003e et al.\u003c/em\u003e Artificial Intelligence in Coronary Artery Calcium Scoring Detection and Quantification. \u003cem\u003eDiagnostics (Basel)\u003c/em\u003e \u003cstrong\u003e14\u003c/strong\u003e, doi:10.3390/diagnostics14020125 (2024).\u003c/li\u003e\n\u003cli\u003eKomatsu, S.\u003cem\u003e et al.\u003c/em\u003e Coronary calcification mimicking the stent: analysis by MDCT Plaque Map. \u003cem\u003eInt J Cardiol\u003c/em\u003e \u003cstrong\u003e115\u003c/strong\u003e, 419-421, doi:10.1016/j.ijcard.2006.01.066 (2007).\u003c/li\u003e\n\u003cli\u003eYang, D. H.\u003cem\u003e et al.\u003c/em\u003e Association between C-reactive Protein and type of coronary arterial plaque in asymptomatic patients: assessment with coronary CT angiography. \u003cem\u003eRadiology\u003c/em\u003e \u003cstrong\u003e272\u003c/strong\u003e, 665-673, doi:10.1148/radiol.14130772 (2014).\u003c/li\u003e\n\u003cli\u003eHe, K., Gkioxari, G., Doll\u0026aacute;r, P. \u0026amp; Girshick, R. in \u003cem\u003e2017 IEEE International Conference on Computer Vision (ICCV).\u003c/em\u003e 2980-2988.\u003c/li\u003e\n\u003cli\u003eMcGrath, E. E., Blades, Z. \u0026amp; Anderson, P. B. Chylothorax: aetiology, diagnosis and therapeutic options. \u003cem\u003eRespir Med\u003c/em\u003e \u003cstrong\u003e104\u003c/strong\u003e, 1-8, doi:10.1016/j.rmed.2009.08.010 (2010).\u003c/li\u003e\n\u003cli\u003eIhdayhid, A. R.\u003cem\u003e et al.\u003c/em\u003e Evaluation of an artificial intelligence coronary artery calcium scoring model from computed tomography. \u003cem\u003eEur Radiol\u003c/em\u003e \u003cstrong\u003e33\u003c/strong\u003e, 321-329, doi:10.1007/s00330-022-09028-3 (2023).\u003c/li\u003e\n\u003cli\u003eWinkel, D. J.\u003cem\u003e et al.\u003c/em\u003e Deep learning for vessel-specific coronary artery calcium scoring: validation on a multi-centre dataset. \u003cem\u003eEur Heart J Cardiovasc Imaging\u003c/em\u003e \u003cstrong\u003e23\u003c/strong\u003e, 846-854, doi:10.1093/ehjci/jeab119 (2022).\u003c/li\u003e\n\u003cli\u003eRen, P.\u003cem\u003e et al.\u003c/em\u003e Motion artefact reduction in coronary CT angiography images with a deep learning method. \u003cem\u003eBMC Med Imaging\u003c/em\u003e \u003cstrong\u003e22\u003c/strong\u003e, 184, doi:10.1186/s12880-022-00914-2 (2022).\u003c/li\u003e\n\u003cli\u003eKang, H. W.\u003cem\u003e et al.\u003c/em\u003e Evaluation of fully automated commercial software for Agatston calcium scoring on non-ECG-gated low-dose chest CT with different slice thickness. \u003cem\u003eEur Radiol\u003c/em\u003e \u003cstrong\u003e33\u003c/strong\u003e, 1973-1981, doi:10.1007/s00330-022-09143-1 (2023).\u003c/li\u003e\n\u003cli\u003eLee, J. O., Park, E. A., Park, D. \u0026amp; Lee, W. Deep Learning-Based Automated Quantification of Coronary Artery Calcification for Contrast-Enhanced Coronary Computed Tomographic Angiography. \u003cem\u003eJ Cardiovasc Dev Dis\u003c/em\u003e \u003cstrong\u003e10\u003c/strong\u003e, doi:10.3390/jcdd10040143 (2023).\u003c/li\u003e\n\u003cli\u003eSuh, Y. J.\u003cem\u003e et al.\u003c/em\u003e Fully automatic coronary calcium scoring in non-ECG-gated low-dose chest CT: comparison with ECG-gated cardiac CT. \u003cem\u003eEur Radiol\u003c/em\u003e \u003cstrong\u003e33\u003c/strong\u003e, 1254-1265, doi:10.1007/s00330-022-09117-3 (2023).\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":true,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"scientific-reports","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"scirep","sideBox":"Learn more about [Scientific Reports](http://www.nature.com/srep/)","snPcode":"","submissionUrl":"","title":"Scientific Reports","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Scientific Reports","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"Coronary artery calcium score, Computed tomography, Coronary stent, Artificial intelligence, Accuracy","lastPublishedDoi":"10.21203/rs.3.rs-4543450/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-4543450/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eAutomatic pre-screening of pre-existing stents, whose prognostic value remains uncertain, could potentially reduce workload and enhance efficiency. However, such a solution has not yet been developed and validated. We aimed to develop and evaluate a deep learning-based coronary stent filtering algorithm (Stent_filter) in CAC scoring CT scans using a multicenter CAC dataset. We developed Stent_filter comprising two main processes: stent identification and false-positive reduction. Development utilized 108 non-enhanced echocardiography-gated CAC scans (including 74 with manually labeled stents), and for false positive reduction, 2063 CAC scans with significant coronary calcium (average Agatston score: 523.8) but no stents were utilized. Stent_filter\u0026rsquo;s performance was evaluated on two independent internal test sets (n\u0026thinsp;=\u0026thinsp;355 and 396; one without coronary stents) and two external test sets from different institutions (n\u0026thinsp;=\u0026thinsp;105 and 62), each with manually labeled stents. We calculated the per-patient sensitivity, specificity, and false-positive rate of Stent_filter. Stent_filter demonstrated a high overall per-patient sensitivity of 98.8% (511/517 cases with stents) and a false-positive rate of 0.022 (20/918). Notably, the false-positive ratio was significantly lower in the dataset containing stents (Internal-1; 0.008 [3/355]) compared with the dataset without stents (Internal-2; 0.043 [17/396], p\u0026thinsp;=\u0026thinsp;0.008). All false-positive identifications were attributed to dense coronary calcifications, with no false positives identified in extracoronary locations. The automated Stent_filter accurately distinguished coronary stents from preexisting coronary calcifications. This approach holds potential as a filter within a fully automated CAC scoring workflow, streamlining the process efficiently.\u003c/p\u003e","manuscriptTitle":"Automatic Identification of Coronary Stent in Coronary Calcium Scoring CT using Deep Learning","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2024-08-17 01:24:25","doi":"10.21203/rs.3.rs-4543450/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Revision requested","date":"2024-07-23T04:03:45+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2024-07-17T19:49:25+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"35276460028213453347213663048235960614","date":"2024-07-13T15:14:19+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2024-07-06T08:32:04+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"291172008094188795082066888501258992625","date":"2024-07-02T07:38:37+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"314656517179838153724908079707733118903","date":"2024-07-01T14:30:43+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2024-07-01T13:25:53+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2024-06-24T17:45:31+00:00","index":"","fulltext":""},{"type":"editorInvited","content":"","date":"2024-06-12T17:16:40+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2024-06-11T05:12:29+00:00","index":"","fulltext":""},{"type":"submitted","content":"Scientific Reports","date":"2024-06-07T04:41:40+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"scientific-reports","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"scirep","sideBox":"Learn more about [Scientific Reports](http://www.nature.com/srep/)","snPcode":"","submissionUrl":"","title":"Scientific Reports","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Scientific Reports","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"76b7aced-6a9c-4947-b1a2-5cace6e35a45","owner":[],"postedDate":"August 17th, 2024","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"published-in-journal","subjectAreas":[],"tags":[],"updatedAt":"2024-11-04T16:29:51+00:00","versionOfRecord":{"articleIdentity":"rs-4543450","link":"https://doi.org/10.1038/s41598-024-76092-8","journal":{"identity":"scientific-reports","isVorOnly":false,"title":"Scientific Reports"},"publishedOn":"2024-10-28 16:20:40","publishedOnDateReadable":"October 28th, 2024"},"versionCreatedAt":"2024-08-17 01:24:25","video":"","vorDoi":"10.1038/s41598-024-76092-8","vorDoiUrl":"https://doi.org/10.1038/s41598-024-76092-8","workflowStages":[]},"version":"v1","identity":"rs-4543450","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-4543450","identity":"rs-4543450","version":["v1"]},"buildId":"qtupq5eGEP_6zYnWcrvyt","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}