Applying radiomics in computed tomography data analysis to predict sarcopenia

I. A. Schmidt ; E. D. Kotina

Vestnik Sankt-Peterburgskogo universiteta. Prikladnaâ matematika, informatika, processy upravleniâ, Tome 20 (2024) no. 3, pp. 376-390

Résumé

This article presents an algorithm implementing a radiomics approach to processing computed tomography (CT) data for diagnosing sarcopenia. The proposed method includes region of interest extraction, automatic muscle segmentation using deep learning models, extraction of radiomic features from CT-images, construction of correlation matrices, and selection of criteria for classification. The results show that the obtained radiomic parameters have a significant correlation with the presence of sarcopenia, allowing the construction of accurate classification models based on machine learning. This approach can significantly improve the diagnosis of sarcopenia, providing reliable non-invasive analysis methods.

Détail

Keywords: radiomics, texture analysis, machine learning, sarcopenia.

Bibliographie

[1] Chen Y. C., Hsieh J. W., Yang Y. H., Lee C. H., Yu P. Y., Chen P. Y., San A. S., “Towards deep learning-based sarcopenia screening with body joint composition analysis”, 2021 IEEE International Conference on Image Processing (ICIP) (Anchorage, AK, USA), 2021, 3807–3811 | DOI

[2] Chung H., Jo Y., Ryu D., Jeong C., Choe S. K., Lee J., “Artificial-intelligence-driven discovery of prognostic biomarker for sarcopenia”, Journal of Cachexia, Sarcopenia and Muscle, 12:6 (2021), 2220–2230 | DOI

[3] Castillo-Olea C., Garcia-Zapirain S. B., Carballo L. C., Zuniga C., “Automatic classification of sarcopenia level in older adults: A case study at Tijuana General Hospital”, International Journal of Environmental Research and Public Health, 16:18 (2019), 3275 | DOI

[4] Ackermans L. L. G. C., Rabou J., Basrai M., Schweinlin A., Bischoff S. C., Cussenot O., Cancel-Tassin G., Renken R. J., Gomez E., Sanchez-Gonzalez P., Rainoldi A., Boccia G., Reisinger K. W., Bosch J. A. T., Blokhuis T. J., Screening, diagnosis and monitoring of sarcopenia: When to use which tool?, Clin. Nutr. ESPEN, 48 (2022), 36–44 | DOI

[5] Xie H., Gong Y., Kuang J., Yan L., Ruan G., Tang S., Gao F., Gan J., “Computed tomography-determined sarcopenia is a useful imaging biomarker for predicting postoperative outcomes in elderly colorectal cancer patients”, Cancer Research and Treatment, 52:3 (2020), 957–972 | DOI

[6] Jalal M., Campbell J. A., Wadsley J., Hopper A. D., “Computed tomographic sarcopenia in pancreatic cancer: Further utilization to plan patient management”, Journal of Gastrointest Cancer, 52:3 (2021), 1183–1187 | DOI

[7] Smorchkova A. K., Petraikin A. V., Semenov D. S., Sharova D. E., “Sarcopenia: modern approaches to solving diagnosis problems”, Digital Diagnostics, 3:3 (2022), 196–211 | DOI

[8] Ueki H., Hara T., Okamura Y., Bando Y., Terakawa T., Furukawa J., Harada K., Nakano Y., Fujisawa M., “Association between sarcopenia based on psoas muscle index and the response to nivolumab in metastatic renal cell carcinoma: A retrospective study”, Investig. Clin. Urol., 63:4 (2022), 415–424 | DOI

[9] Kim S., Kim T.-H., Jeong C.-W., Lee C., Noh S., Kim J. E., Yoon K.-H., “Development of quantification software for evaluating body composition contents and its clinical application in sarcopenic obesity”, Scientific Reports, 10 (2020), 10452 | DOI

[10] Chicklore S., Goh V., Siddique M., Roy A., Marsden P. K., Cook G. J. R., “Quantifying tumour heterogeneity in 18F-FDG PET/CT imaging by texture analysis”, European Journal of Nuclear Medicine and Molecular Imaging, 40:1 (2012), 133–140 | DOI

[11] Cook G., Siddique M., Taylor B., Yip C., Chicklore S., Goh V., “Radiomics in PET: principles and applications”, Clinical and Translational Imaging, 2:3 (2014), 269–276 | DOI

[12] Schmidt I., Kotina E., Buev P., “Deep learning muscle segmentation model for CT images in DICOM format”, Cybernetics and Physics, 12:3 (2023), 201–206 | DOI

[13] Shmidt I. A., Kotina E. D., Kamyshanskaya I. G., Makarenko B. G., “Radiomics in the study of sarcopenia using CT images”, Diagnostic and Interventional Radiology, 18:S2.1 (2024), 94–99 (In Russian)

[14] Shmidt Y. A., Kotina E. D., Kamyshanskaya I. G., Makarenko B. G., “Application of radiomics criteria in the study of sarcopenia based on abdominal computed tomography data”, Diagnostic Radiology and Radiotherapy, S(15) (2024), 195–196

[15] Islam S., Kanavati F., Arain Z., Costa O. F. D. Crum W., Aboagye E. O., Rockall A. G., “Fully-automated deep learning slice-based muscle estimation from CT images for sarcopenia assessment”, Clinical Radiology, 77:5 (2022), e363–e371 | DOI

[16] Ha J., Park T., Kim H. K., Shin Y., Ko Y., Kim D. W., Sung Y. S., Lee J., Ham S. J., Khang S., Jeong H., Koo K., Lee J., Kim K. W., “Development of a fully automatic deep learning system for L3 selection and body composition assessment on computed tomography”, Scientific Reports, 11:1 (2021), 21656 | DOI

[17] Zwanenburg A., Leger S., Vallieres M., Löck S., Image biomarker standardisation initiative, 2016, arXiv: 1612.07003 | Zbl

[18] Löfstedt T., Brynolfsson P., Asklund T., Nyholm T., Garpebring A., “Gray-level invariant Haralick texture features”, PLoS One, 14:2 (2019), e0212110 | DOI