PREDICTING KIDNEY TUMOR SUBTYPE FROM CT IMAGES USING RADIOMICS AND CLINICAL FEATURES

Year 2022, , 29 - 37, 30.06.2022

Duygu Şirin , Albert Güveniş

https://doi.org/10.38061/idunas.1084748

Abstract

Purpose: This study aims to evaluate the performance of machine learning methods in predicting the subtype (clear-cell vs. non-clear-cell) of kidney tumors using clinical patient and radiomics data from CT images.
Method: CT images of 192 malignant kidney tumor cases (142 clear-cell, 50 other) from TCIA’s KiTS-19 Challenge were used in the study. There were several different tumor subtypes in the other group, most of them being chromophobe or papillary RCC. Patient clinical data were combined with the radiomic features extracted from CT images. Features were extracted from 3D images and all of the slices were included in the feature extraction process. Initial dataset consisted of 1157 features of which 1130 were radiomics and 27 were clinical. Features were selected using Kruskal Wallis – ANOVA test followed by Lasso Regression. After feature selection, 8 radiomic features remained. None of the clinical features were considered important for our model as a result. Training set classes were balanced using SMOTE. Training data with the selected features were used to train the Coarse Gaussian SVM and Subspace Discriminant classifiers.
Results: Coarse Gaussian SVM was faster compared to Subspace Discriminant with a training time of 0.47 sec and ~11000 obs/sec prediction speed. Training duration of Subspace Discriminant was 4.1 sec with ~960 obs/sec prediction speed. For Coarse Gaussian SVM; validation accuracy was 67,6% while the accuracy of test was 80%, with and AUC of 0.86. Similarly, Subspace Discriminant had 68,8% validation accuracy and 80% test accuracy; AUC was 0.85.
Conclusion: Both models produced promising results on classifying malignant tumors as ccRCC or non-ccRCC. However, Coarse Gaussian SVM might be more preferable because of its training and prediction speed.

Keywords

kidney tumor, clear-cell, machine learning, CT imaging

References

• Clark, Kenneth, Bruce Vendt, Kirk Smith, John Freymann, Justin Kirby, Paul Koppel, Stephen Moore, Stanley Phillips, David Maffitt, Michael Pringle, Lawrence Tarbox, and Fred Prior. 2013. “The Cancer Imaging Archive (TCIA): Maintaining and Operating a Public Information Repository.” Journal of Digital Imaging 26(6):1045–57. doi: 10.1007/s10278-013-9622-7.
• Fedorov, Andriy, Reinhard Beichel, Jayashree Kalpathy-Cramer, Julien Finet, Jean Christophe Fillion-Robin, Sonia Pujol, Christian Bauer, Dominique Jennings, Fiona Fennessy, Milan Sonka, John Buatti, Stephen Aylward, James v. Miller, Steve Pieper, and Ron Kikinis. 2012. “3D Slicer as an Image Computing Platform for the Quantitative Imaging Network.” Magnetic Resonance Imaging 30(9):1323–41. doi: 10.1016/j.mri.2012.05.001.
• Han, Seokmin, Sung il Hwang, and Hak Jong Lee. 2019. “The Classification of Renal Cancer in 3-Phase CT Images Using a Deep Learning Method.” Journal of Digital Imaging 32(4):638–43. doi: 10.1007/s10278-019-00230-2.
• Heller, N., et al. 2019. "Data from C4KC-KiTS" The Cancer Imaging Archive. doi: 10.7937/TCIA.2019.IX49E8NX
• Hoang, Uyen N., S. Mojdeh Mirmomen, Osorio Meirelles, Jianhua Yao, Maria Merino, Adam Metwalli, W. Marston Linehan, and Ashkan A. Malayeri. 2018. “Assessment of Multiphasic Contrast-Enhanced Mr Textures in Differentiating Small Renal Mass Subtypes.” Abdominal Radiology 43(12):3400–3409. doi: 10.1007/s00261-018-1625-x.
• Hsieh, James J., Mark P. Purdue, Sabina Signoretti, Charles Swanton, Laurence Albiges, Manuela Schmidinger, Daniel Y. Heng, James Larkin, and Vincenzo Ficarra. 2017. “Renal Cell Carcinoma.” Nature Reviews Disease Primers 3. doi: 10.1038/nrdp.2017.9.
• Kocak, Burak, Aytul Hande Yardimci, Ceyda Turan Bektas, Mehmet Hamza Turkcanoglu, Cagri Erdim, Ugur Yucetas, Sevim Baykal Koca, and Ozgur Kilickesmez. 2018. “Textural Differences between Renal Cell Carcinoma Subtypes: Machine Learning-Based Quantitative Computed Tomography Texture Analysis with Independent External Validation.” European Journal of Radiology 107:149–57. doi: 10.1016/j.ejrad.2018.08.014.
• Larue, Ruben T. H. M., Janna E. van Timmeren, Evelyn E. C. de Jong, Giacomo Feliciani, Ralph T. H. Leijenaar, Wendy M. J. Schreurs, Meindert N. Sosef, Frank H. P. J. Raat, Frans H. R. van der Zande, Marco Das, Wouter van Elmpt, and Philippe Lambin. 2017. “Influence of Gray Level Discretization on Radiomic Feature Stability for Different CT Scanners, Tube Currents and Slice Thicknesses: A Comprehensive Phantom Study.” Acta Oncologica 56(11):1544–53. doi: 10.1080/0284186X.2017.1351624.
• Le, Valerie H., and James J. Hsieh. 2018. “Genomics and Genetics of Clear Cell Renal Cell Carcinoma: A Mini-Review.” Journal of Translational Genetics and Genomics. doi: 10.20517/jtgg.2018.28.
• Ökmen, Harika Beste, Albert Guvenis, and Hadi Uysal. 2019. “Predicting the Polybromo-1 (Pbrm1) Mutation of a Clear Cell Renal Cell Carcinoma Using Computed Tomography Images and Knn Classification with Random Subspace.” Pp. 30–34 in Vibroengineering Procedia. Vol. 26. JVE International.
• van Oostenbrugge, Tim J., Jurgen J. Fütterer, and Peter F. A. Mulders. 2018. “Diagnostic Imaging for Solid Renal Tumors: A Pictorial Review.” Kidney Cancer 2(2):79–93.
• Tibshiranit, Robert. 1996. Regression Shrinkage and Selection via the Lasso. Vol. 58.
• Tomaszewski, Michal R., Gillies, Robert J. 2021. “The Biological Meaning of Radiomic Features”, Radiology Vol. 298, No. 3. doi: 10.1148/radiol.2021202553
• Wang, Z., Wang, K., An, S. 2011. "Cubic B-spline interpolation and realization." International Conference on Information Computing and Applications pp. 82-89.
• Zhang, G. M. Y., B. Shi, H. D. Xue, B. Ganeshan, H. Sun, and Z. Y. Jin. 2019. “Can Quantitative CT Texture Analysis Be Used to Differentiate Subtypes of Renal Cell Carcinoma?” Clinical Radiology 74(4):287–94. doi: 10.1016/j.crad.2018.11.009.