{"database":"biostudies-literature","file_versions":[],"scores":null,"additional":{"omics_type":["Unknown"],"volume":["3(2)"],"submitter":["Schock J"],"funding":["START Program of the Faculty of Medicine, RWTH Aachen","Deutsche Forschungsgemeinschaft"],"pubmed_abstract":["<h4>Purpose</h4>To develop and validate a deep learning-based method for automatic quantitative analysis of lower-extremity alignment.<h4>Materials and methods</h4>In this retrospective study, bilateral long-leg radiographs (LLRs) from 255 patients that were obtained between January and September of 2018 were included. For training data (<i>n</i> = 109), a U-Net convolutional neural network was trained to segment the femur and tibia versus manual segmentation. For validation data (<i>n</i> = 40), model parameters were optimized. Following identification of anatomic landmarks, anatomic and mechanical axes were identified and used to quantify alignment through the hip-knee-ankle angle (HKAA) and femoral anatomic-mechanical angle (AMA). For testing data (<i>n</i> = 106), algorithm-based angle measurements were compared with reference measurements by two radiologists. Angles and time for 30 random radiographs were compared by using repeated-measures analysis of variance and one-way analysis of variance, whereas correlations were quantified by using Pearson <i>r</i> and intraclass correlation coefficients.<h4>Results</h4>Bilateral LLRs of 255 patients (mean age, 26 years ± 23 [standard deviation]; range, 0-88 years; 157 male patients) were included. Mean Sørensen-Dice coefficients for segmentation were 0.97 ± 0.09 for the femur and 0.96 ± 0.11 for the tibia. Mean HKAAs and AMAs as measured by the readers and the algorithm ranged from 0.05° to 0.11° (<i>P</i> = .5) and from 4.82° to 5.43° (<i>P</i> < .001). Interreader correlation coefficients ranged from 0.918 to 0.995 (<i>r</i> range, <i>P</i> < .001), and agreement was almost perfect (intraclass correlation coefficient range, 0.87-0.99). Automatic analysis was faster than the two radiologists' manual measurements (3 vs 36 vs 35 seconds, <i>P</i> < .001).<h4>Conclusion</h4>Fully automated analysis of LLRs yielded accurate results across a wide range of clinical and pathologic indications and is fast enough to enhance and accelerate clinical workflows.<i>Supplemental material is available for this article.</i>© RSNA, 2020See also commentary by Andreisek in this issue."],"journal":["Radiology. Artificial intelligence"],"pagination":["e200198"],"full_dataset_link":["https://www.ebi.ac.uk/biostudies/studies/S-EPMC8043357"],"repository":["biostudies-literature"],"pubmed_title":["Automated Analysis of Alignment in Long-Leg Radiographs by Using a Fully Automated Support System Based on Artificial Intelligence."],"pmcid":["PMC8043357"],"pubmed_authors":["Mittelstrass F","Schock J","Truhn D","Nebelung S","Post M","Merhof D","Kuhl C","Conrad S","Abrar DB"],"additional_accession":[]},"is_claimable":false,"name":"Automated Analysis of Alignment in Long-Leg Radiographs by Using a Fully Automated Support System Based on Artificial Intelligence.","description":"<h4>Purpose</h4>To develop and validate a deep learning-based method for automatic quantitative analysis of lower-extremity alignment.<h4>Materials and methods</h4>In this retrospective study, bilateral long-leg radiographs (LLRs) from 255 patients that were obtained between January and September of 2018 were included. For training data (<i>n</i> = 109), a U-Net convolutional neural network was trained to segment the femur and tibia versus manual segmentation. For validation data (<i>n</i> = 40), model parameters were optimized. Following identification of anatomic landmarks, anatomic and mechanical axes were identified and used to quantify alignment through the hip-knee-ankle angle (HKAA) and femoral anatomic-mechanical angle (AMA). For testing data (<i>n</i> = 106), algorithm-based angle measurements were compared with reference measurements by two radiologists. Angles and time for 30 random radiographs were compared by using repeated-measures analysis of variance and one-way analysis of variance, whereas correlations were quantified by using Pearson <i>r</i> and intraclass correlation coefficients.<h4>Results</h4>Bilateral LLRs of 255 patients (mean age, 26 years ± 23 [standard deviation]; range, 0-88 years; 157 male patients) were included. Mean Sørensen-Dice coefficients for segmentation were 0.97 ± 0.09 for the femur and 0.96 ± 0.11 for the tibia. Mean HKAAs and AMAs as measured by the readers and the algorithm ranged from 0.05° to 0.11° (<i>P</i> = .5) and from 4.82° to 5.43° (<i>P</i> < .001). Interreader correlation coefficients ranged from 0.918 to 0.995 (<i>r</i> range, <i>P</i> < .001), and agreement was almost perfect (intraclass correlation coefficient range, 0.87-0.99). Automatic analysis was faster than the two radiologists' manual measurements (3 vs 36 vs 35 seconds, <i>P</i> < .001).<h4>Conclusion</h4>Fully automated analysis of LLRs yielded accurate results across a wide range of clinical and pathologic indications and is fast enough to enhance and accelerate clinical workflows.<i>Supplemental material is available for this article.</i>© RSNA, 2020See also commentary by Andreisek in this issue.","dates":{"release":"2021-01-01T00:00:00Z","publication":"2021 Mar","modification":"2025-04-04T09:55:51.263Z","creation":"2025-04-04T09:55:51.263Z"},"accession":"S-EPMC8043357","cross_references":{"pubmed":["33937861"],"doi":["10.1148/ryai.2020200198"]}}