Segmentation-based quantitative measurements in renal CT imaging using deep learning

被引:0
作者
Koukoutegos, Konstantinos [1 ,2 ]
Heeren, Richard's [2 ]
De Wever, Liesbeth [2 ]
De Keyzer, Frederik [2 ]
Maes, Frederik [3 ]
Bosmans, Hilde [1 ,2 ]
机构
[1] Katholieke Univ Leuven, Div Med Phys, Dept Imaging & Pathol, Herestr 49, B-3000 Leuven, Belgium
[2] UZ Leuven, Dept Radiol, Herestr 49, B-3000 Leuven, Belgium
[3] Katholieke Univ Leuven, Dept Elect Engn, ESAT, PSI, B-3000 Leuven, Belgium
关键词
Abdomen; Artificial intelligence; Deep learning; Kidney; Tomography (x-ray computed); KIDNEY LENGTH; ULTRASOUND; VOLUME; SIZE; INTEROBSERVER; RELIABILITY; VALUES;
D O I
10.1186/s41747-024-00507-4
中图分类号
R8 [特种医学]; R445 [影像诊断学];
学科分类号
1002 ; 100207 ; 1009 ;
摘要
Background Renal quantitative measurements are important descriptors for assessing kidney function. We developed a deep learning-based method for automated kidney measurements from computed tomography (CT) images. Methods The study datasets comprised potential kidney donors (n = 88), both contrast-enhanced (Dataset 1 CE) and noncontrast (Dataset 1 NC) CT scans, and test sets of contrast-enhanced cases (Test set 2, n = 18), cases from a photon-counting (PC)CT scanner reconstructed at 60 and 190 keV (Test set 3 PCCT, n = 15), and low-dose cases (Test set 4, n = 8), which were retrospectively analyzed to train, validate, and test two networks for kidney segmentation and subsequent measurements. Segmentation performance was evaluated using the Dice similarity coefficient (DSC). The quantitative measurements' effectiveness was compared to manual annotations using the intraclass correlation coefficient (ICC). Results The contrast-enhanced and noncontrast models demonstrated excellent reliability in renal segmentation with DSC of 0.95 (Test set 1 CE), 0.94 (Test set 2), 0.92 (Test set 3 PCCT) and 0.94 (Test set 1 NC), 0.92 (Test set 3 PCCT), and 0.93 (Test set 4). Volume estimation was accurate with mean volume errors of 4%, 3%, 6% mL (contrast test sets) and 4%, 5%, 7% mL (noncontrast test sets). Renal axes measurements (length, width, and thickness) had ICC values greater than 0.90 (p < 0.001) for all test sets, supported by narrow 95% confidence intervals. Conclusion Two deep learning networks were shown to derive quantitative measurements from contrast-enhanced and noncontrast renal CT imaging at the human performance level. Relevance statement Deep learning-based networks can automatically obtain renal clinical descriptors from both noncontrast and contrast-enhanced CT images. When healthy subjects comprise the training cohort, careful consideration is required during model adaptation, especially in scenarios involving unhealthy kidneys. This creates an opportunity for improved clinical decision-making without labor-intensive manual effort.
引用
收藏
页数:15
相关论文
共 42 条
[1]   HOW RELIABLE ARE ULTRASOUND MEASUREMENTS OF RENAL LENGTH IN ADULTS [J].
ABLETT, MJ ;
COULTHARD, A ;
LEE, REJ ;
RICHARDSON, DL ;
BELLAS, T ;
OWEN, JP ;
KEIR, MJ ;
BUTLER, TJ .
BRITISH JOURNAL OF RADIOLOGY, 1995, 68 (814) :1087-1089
[2]   Renal Cortical Thickness Measured at Ultrasound: Is It Better Than Renal Length as an Indicator of Renal Function in Chronic Kidney Disease? [J].
Beland, Michael D. ;
Walle, Nicholas L. ;
Machan, Jason T. ;
Cronan, John J. .
AMERICAN JOURNAL OF ROENTGENOLOGY, 2010, 195 (02) :W146-W149
[3]   Renal length and volume prediction in healthy children [J].
Bianchi, Maria E. ;
Forlino, Daniel ;
Velasco, Gustavo A. ;
Rodriguez, Pablo O. ;
Lopez, German ;
Cusumano, Ana M. .
CLINICAL NEPHROLOGY, 2020, 93 (07) :S31-S35
[4]   How reliable is renal ultrasound to measure renal length and volume in patients with chronic kidney disease compared with magnetic resonance imaging? [J].
Braconnier, Philippe ;
Piskunowicz, Maciej ;
Vakilzadeh, Nima ;
Mueller, Marie-Eve ;
Zuercher, Emilie ;
Burnier, Michel ;
Pruijm, Menno .
ACTA RADIOLOGICA, 2020, 61 (01) :117-127
[5]  
BRENNER BM, 1992, J AM SOC NEPHROL, V3, P162
[6]   An Ensemble of U-Net Models for Kidney Tumor Segmentation With CT Images [J].
Causey, Jason ;
Stubblefield, Jonathan ;
Qualls, Jake ;
Fowler, Jennifer ;
Cai, Lingrui ;
Walker, Karl ;
Guan, Yuanfang ;
Huang, Xiuzhen .
IEEE-ACM TRANSACTIONS ON COMPUTATIONAL BIOLOGY AND BIOINFORMATICS, 2022, 19 (03) :1387-1392
[7]   Normal values for renal length and volume as measured by magnetic resonance imaging [J].
Cheong, Benjamin ;
Muthupillai, Raja ;
Rubin, Mario F. ;
Flamm, Scott D. .
CLINICAL JOURNAL OF THE AMERICAN SOCIETY OF NEPHROLOGY, 2007, 2 (01) :38-45
[8]   Inter-Observer Variation in Ultrasound Measurement of the Volume and Diameter of Thyroid Nodules [J].
Choi, Young Jun ;
Baek, Jung Hwan ;
Hong, Min Ji ;
Lee, Jeong Hyun .
KOREAN JOURNAL OF RADIOLOGY, 2015, 16 (03) :560-565
[9]   Kidney segmentation from computed tomography images using deep neural network [J].
da Cruz, Luana Batista ;
Lima Araujo, Jose Denes ;
Ferreira, Jonnison Lima ;
Bandeira Diniz, Joao Otavio ;
Silva, Aristofanes Correa ;
Sousa de Almeida, Joao Dallyson ;
de Paiva, Anselmo Cardoso ;
Gattass, Marcelo .
COMPUTERS IN BIOLOGY AND MEDICINE, 2020, 123
[10]   3D Slicer as an image computing platform for the Quantitative Imaging Network [J].
Fedorov, Andriy ;
Beichel, Reinhard ;
Kalpathy-Cramer, Jayashree ;
Finet, Julien ;
Fillion-Robin, Jean-Christophe ;
Pujol, Sonia ;
Bauer, Christian ;
Jennings, Dominique ;
Fennessy, Fiona ;
Sonka, Milan ;
Buatti, John ;
Aylward, Stephen ;
Miller, James V. ;
Pieper, Steve ;
Kikinis, Ron .
MAGNETIC RESONANCE IMAGING, 2012, 30 (09) :1323-1341