BACKGROUND: Manually performed diameter measurements on ECG-gated CT-angiography (CTA) represent the gold standard for diagnosis of thoracic aortic dilatation. However, they are time-consuming and show high inter-reader variability. Therefore, we aimed to evaluate the accuracy of measurements of a deep learning-(DL)-algorithm in comparison to those of radiologists and evaluated measurement times (MT). METHODS: We retrospectively analyzed 405 ECG-gated CTA exams of 371 consecutive patients with suspected aortic dilatation between May 2010 and June 2019. The DL-algorithm prototype detected aortic landmarks (deep reinforcement learning) and segmented the lumen of the thoracic aorta (multi-layer convolutional neural network). It performed measurements according to AHA-guidelines and created visual outputs. Manual measurements were performed by radiologists using centerline technique. Human performance variability (HPV), MT and DL-performance were analyzed in a research setting using a linear mixed model based on 21 randomly selected, repeatedly measured cases. DL-algorithm results were then evaluated in a clinical setting using matched differences. If the differences were within 5 mm for all locations, the cases was regarded as coherent; if there was a discrepancy >5 mm at least at one location (incl. missing values), the case was completely reviewed. RESULTS: HPV ranged up to ±3.4 mm in repeated measurements under research conditions. In the clinical setting, 2,778/3,192 (87.0%) of DL-algorithm's measurements were coherent. Mean differences of paired measurements between DL-algorithm and radiologists at aortic sinus and ascending aorta were -0.45±5.52 and -0.02±3.36 mm. Detailed analysis revealed that measurements at the aortic root were over-/underestimated due to a tilted measurement plane. In total, calculated time saved by DL-algorithm was 3:10 minutes/case. CONCLUSIONS: The DL-algorithm provided coherent results to radiologists at almost 90% of measurement locations, while the majority of discrepent cases were located at the aortic root. In summary, the DL-algorithm assisted radiologists in performing AHA-compliant measurements by saving 50% of time per case. 2021 Quantitative Imaging in Medicine and Surgery. All rights reserved.
BACKGROUND: Manually performed diameter measurements on ECG-gated CT-angiography (CTA) represent the gold standard for diagnosis of thoracic aortic dilatation. However, they are time-consuming and show high inter-reader variability. Therefore, we aimed to evaluate the accuracy of measurements of a deep learning-(DL)-algorithm in comparison to those of radiologists and evaluated measurement times (MT). METHODS: We retrospectively analyzed 405 ECG-gated CTA exams of 371 consecutive patients with suspected aortic dilatation between May 2010 and June 2019. The DL-algorithm prototype detected aortic landmarks (deep reinforcement learning) and segmented the lumen of the thoracic aorta (multi-layer convolutional neural network). It performed measurements according to AHA-guidelines and created visual outputs. Manual measurements were performed by radiologists using centerline technique. Human performance variability (HPV), MT and DL-performance were analyzed in a research setting using a linear mixed model based on 21 randomly selected, repeatedly measured cases. DL-algorithm results were then evaluated in a clinical setting using matched differences. If the differences were within 5 mm for all locations, the cases was regarded as coherent; if there was a discrepancy >5 mm at least at one location (incl. missing values), the case was completely reviewed. RESULTS: HPV ranged up to ±3.4 mm in repeated measurements under research conditions. In the clinical setting, 2,778/3,192 (87.0%) of DL-algorithm's measurements were coherent. Mean differences of paired measurements between DL-algorithm and radiologists at aortic sinus and ascending aorta were -0.45±5.52 and -0.02±3.36 mm. Detailed analysis revealed that measurements at the aortic root were over-/underestimated due to a tilted measurement plane. In total, calculated time saved by DL-algorithm was 3:10 minutes/case. CONCLUSIONS: The DL-algorithm provided coherent results to radiologists at almost 90% of measurement locations, while the majority of discrepent cases were located at the aortic root. In summary, the DL-algorithm assisted radiologists in performing AHA-compliant measurements by saving 50% of time per case. 2021 Quantitative Imaging in Medicine and Surgery. All rights reserved.
Entities:
Keywords:
Deep learning; aortic aneurysm; computed tomography angiography; dimensional measurement accuracy; observer variation; time management
Authors: Loren F Hiratzka; George L Bakris; Joshua A Beckman; Robert M Bersin; Vincent F Carr; Donald E Casey; Kim A Eagle; Luke K Hermann; Eric M Isselbacher; Ella A Kazerooni; Nicholas T Kouchoukos; Bruce W Lytle; Dianna M Milewicz; David L Reich; Souvik Sen; Julie A Shinn; Lars G Svensson; David M Williams Journal: Circulation Date: 2010-03-16 Impact factor: 29.690
Authors: Andreas Biesdorf; Karl Rohr; Duan Feng; Hendrik von Tengg-Kobligk; Fabian Rengier; Dittmar Böckler; Hans-Ulrich Kauczor; Stefan Wörz Journal: Med Image Anal Date: 2012-06-21 Impact factor: 8.545
Authors: Ryan R Davies; Lee J Goldstein; Michael A Coady; Shawn L Tittle; John A Rizzo; Gary S Kopf; John A Elefteriades Journal: Ann Thorac Surg Date: 2002-01 Impact factor: 4.330
Authors: Judith Z Goldfinger; Jonathan L Halperin; Michael L Marin; Allan S Stewart; Kim A Eagle; Valentin Fuster Journal: J Am Coll Cardiol Date: 2014-10-21 Impact factor: 24.094
Authors: Leslie E Quint; Peter S Liu; Anna M Booher; Kuanwong Watcharotone; James D Myles Journal: Int J Cardiovasc Imaging Date: 2012-08-03 Impact factor: 2.357
Authors: J Rueckel; P Reidler; N Fink; J Sperl; T Geyer; M P Fabritius; J Ricke; M Ingrisch; B O Sabel Journal: Eur J Radiol Date: 2020-11-21 Impact factor: 3.528
Authors: Somdatta Goswami; David S Li; Bruno V Rego; Marcos Latorre; Jay D Humphrey; George Em Karniadakis Journal: J R Soc Interface Date: 2022-08-31 Impact factor: 4.293
Authors: Maurice Pradella; Rita Achermann; Jonathan I Sperl; Rainer Kärgel; Saikiran Rapaka; Joshy Cyriac; Shan Yang; Gregor Sommer; Bram Stieltjes; Jens Bremerich; Philipp Brantner; Alexander W Sauter Journal: Front Cardiovasc Med Date: 2022-08-22