Differential diagnosis of focal changes in the spine using standard and radiomic analysis

Abstract

BACKGROUND: If focal changes in the bones are detected, the radiologist must exclude or confirm the presence of a metastatic lesion. Although the semiotics of metastatic and non-oncological changes according to magnetic resonance imaging (MRI) data is well known, in practice, there may be various combinations of their characteristics that are influenced by other chronic diseases and parallel processes, which significantly complicate interpretation. The use of computer image analysis methods has great prospects and can improve the diagnostic accuracy of standard imaging methods.

OBJECTIVE: To improve the accuracy of diagnosing radiographic findings of focal changes in the spine using additional image evaluation by computer analysis algorithms.

MATERIALS AND METHODS: Thirty patients were examined, and 15 of them had metastatic bone lesions from breast cancer, and 15 had focal changes in the spine of a non-oncological nature. Computer analysis of focal changes in the vertebral bodies was conducted according to T₁WI, T₂WI, and STIR MRI sequences. For the computer analysis, the operator of the complexity of the image Arzela and histogram distribution of brightness were used.

RESULTS: The main differential indicators for hemangioma, conditionally normal areas of the bone marrow, and metastatic foci have been established. The Arzela data image complexity operator was approximately 0.07 for hemangioma, approximately 0.05 for metastases (mts), and approximately 0.04 for vertebrae. The brightness histogram operator was approximately 1.12 for haemangioma and approximately 0.94 for mts. Regarding the difference between indicators, the difference is 20%–25%, between hemangioma and bone marrow and 35% between mts and bone marrow, which make it possible to effectively use these indicators together with other markers.

CONCLUSION: The criteria for differential diagnosis obtained using radiomic analysis showed significant differences between focal changes in the vertebrae of various etiologies. From a mathematical point of view, they are advisory, and the doctor with experience remains at the center of the decision-making system.

Keywords

bone metastases / oncology / MRI / radiomics

References

[1] Padalkar P, Tow B. Predictors of survival in surgically treated patients of spinal metastasis. Indian J Orthop. 2011;45(4):307–313. doi: 10.4103/0019-5413.82333

[2] Padalkar P., Tow B. Predictors of survival in surgically treated patients of spinal metastasis // Indian J Orthop. 2011. Vol. 45, N 4. P. 307–313. doi: 10.4103/0019-5413.82333

[3] Sergeev NI, Fomin DK, Kotlyarov PM, Solodkiy VA. Comparative study of the possibilities of SPECT/CT and whole body MRI in the diagnosis of bone metastases. Vestnik of the Russian Scientific Center of Roentgenoradiology. 2015;15(3):8. (In Russ).

[4] Сергеев Н.И., Фомин Д.К., Котляров П.М., Солодкий В.А. Сравнительное исследование возможностей ОФЭКТ/КТ и магнитно-резонансной томографии всего тела в диагностике костных метастазов // Вестник Российского научного центра рентгенорадиологии. 2015. Т. 15, № 3. С. 8.

[5] Haraldsen A, Bluhme H, Røhl L, et al. Single photon emission computed tomography (SPECT) and SPECT/low-dose computerized tomography did not increase sensitivity or specificity compared to planar bone scintigraphy for detection of bone metastases in advanced breast cancer. Clin Physiol Funct Imaging. 2016;36(1):40–46. doi: 10.1111/cpf.12191

[6] Haraldsen A., Bluhme H., Røhl L., et al. Single photon emission computed tomography (SPECT) and SPECT/low-dose computerized tomography did not increase sensitivity or specificity compared to planar bone scintigraphy for detection of bone metastases in advanced breast cancer // Clin Physiol Funct Imaging. 2016. Vol. 36, N 1. P. 40–46. doi: 10.1111/cpf.12191

[7] Iosefi DYa, Sergeev N, Solodkiy VA, Vinidchenko MA. Possibilities of textural analysis of images (structural heterogeneity coefficient and maps of the apparent diffusion coefficient) for differential diagnosis of metastatic bone lesions according to magnetic resonance imaging. Vestnik of the Russian Scientific Center of Roentgenoradiology. 2022;22(3):1–17. (In Russ).

[8] Иозефи Д.Я., Сергеев Н.И., Солодкий В.А., Винидченко М.А. Возможности текстурного анализа изображений (коэффициента структурной гетерогенности и карт исчисляемого коэффициента диффузии) в дифференциальной диагностике метастатического поражения костей по данным магнитно-резонансной томографии // Вестник Российского научного центра рентгенорадиологии. 2022. Т. 22, № 3. С. 1–17.

[9] Tian Z, Chen C, Fan Y, et al. Glioblastoma and Anaplastic Astrocytoma: Differentiation Using MRI Texture Analysis. Front Oncol. 2019;9:876. doi: 10.3389/fonc.2019.00876

[10] Tian Z., Chen C., Fan Y., et al. Glioblastoma and Anaplastic Astrocytoma: Differentiation Using MRI Texture Analysis // Front Oncol. 2019. N 9. P. 876. doi: 10.3389/fonc.2019.00876

[11] Litvin AA, Burkin DA, Kropinov AA, Paramzin FN. Radiomics and Digital Image Texture Analysis in Oncology (Review). Sovrem Tekhnologii Med. 2021;13(2):97–104. (In Russ). doi: 10.17691/stm2021.13.2.11

[12] Литвин А.А., Буркин Д.А., Кропинов А.А., Парамзин Ф.Н. Радиомика и анализ текстур цифровых изображений в онкологии (обзор) // Современные технологии в медицине. 2021. Т. 13, № 2. С. 97–104. doi: 10.17691/stm2021.13.2.11

[13] Özgül HA, Akin IB, Mutlu U, Balci A. Diagnostic value of machine learning-based computed tomography texture analysis for differentiating multiple myeloma from osteolytic metastatic bone lesions in the peripheral skeleton. Skeletal Radiol. 2023. doi: 10.1007/s00256-023-04333-4. Epub ahead of print.

[14] Özgül H.A., Akin I.B., Mutlu U., Balci A. Diagnostic value of machine learning-based computed tomography texture analysis for differentiating multiple myeloma from osteolytic metastatic bone lesions in the peripheral skeleton // Skeletal Radiol. 2023. doi: 10.1007/s00256-023-04333-4. Epub ahead of print.

[15] Park HJ, Park B, Lee SS. Radiomics and Deep Learning: Hepatic Applications; Korean J Radiol. 2020;21(4):387–401. doi: 10.3348/kjr.2019.0752

[16] Park H.J., Park B., Lee S.S. Radiomics and Deep Learning: Hepatic Applications // Korean J Radiol. 2020. Vol. 21, N 4. P. 387–401. doi: 10.3348/kjr.2019.0752

[17] Shinohara RT, Sweeney EM, Goldsmith J, et al. Statistical normalization techniques for magnetic resonance imaging. Neuroimage Clin. 2014;6:9–19. doi: 10.1016/j.nicl.2014.08.008

[18] Shinohara R.T., Sweeney E.M., Goldsmith J., et al. Statistical normalization techniques for magnetic resonance imaging // Neuroimage Clin. 2014. N 6. P. 9–19. doi: 10.1016/j.nicl.2014.08.008

[19] Pencheva TD. Hochaufgelöste MRT zur Quantifizierung veränderter trabekulärer Knochen-Architektur bei Patientinnen mit und ohne Mammakarzinom und osteoporotischer Wirbelkörper-Fraktur [dissertation]. München: Universitätsbibliothek der TU München, 2013. Available from: https://mediatum.ub.tum.de/doc/1129206/1129206.pdf. Accessed: 14.05.2023. (In German).

[20] Pencheva T.D. Hochaufgelöste MRT zur Quantifizierung veränderter trabekulärer Knochen-Architektur bei Patientinnen mit und ohne Mammakarzinom und osteoporotischer Wirbelkörper-Fraktur: dissertation. München: Universitätsbibliothek der TU München, 2013. Available from: https://mediatum.ub.tum.de/doc/1129206/1129206.pdf. Accessed: 14.05.2023.

[21] Steinhauer V, Steinhauer L. Neuroph and DL4J. Java Magazine. 2021;6:79–82.

[22] Steinhauer V., Steinhauer L. Neuroph and DL4J // Java Magazine. 2021. N 6. P. 79–82.

[23] Morozov SP, Kokina DYu, Pavlov NA, et al. Clinical aspects of using artificial intelligence for the interpretation of chest X-rays. Tuberculosis and Lung Diseases. 2021;99(4):58–64. (In Russ). doi: 10.21292/2075-1230-2021-99-4-58-64

[24] Морозов С.П., Кокина Д.Ю., Павлов Н.А., и др. Клинические аспекты применения искусственного интеллекта для интерпретации рентгенограмм органов грудной клетки // Туберкулез и болезни легких. 2021. Т. 99, № 4. С. 58–64. doi: 10.21292/2075-1230-2021-99-4-58-64

[25] Ye J, Huang H, Jiang W, et al. Tumor Grade and Overall Survival Prediction of Gliomas Using Radiomics. Scientific Programming. 2021;2021:9913466. doi: 10.1155/2021/9913466

[26] Ye J., Huang H., Jiang W., et al. Tumor Grade and Overall Survival Prediction of Gliomas Using Radiomics // Scientific Programming. 2021. Vol. 2021. Article ID 9913466. doi: 10.1155/2021/9913466

[27] Steinhauer V, Sergeev NI. Radiomics in Breast Cancer: In-Depth Machine Analysis of MR Images of Metastatic Spine Lesion. Sovrem Tekhnologii Med. 2022;14(2):16. doi: 10.17691/stm2022.14.2.02.

[28] Штaйнгауэр В., Сергеев Н.И. Радиомика при раке молочной железы: использование глубокого машинного анализа МРТ-изображений метастатического поражения позвоночника // Современные технологии в медицине. 2022. Т. 14, № 2. С. 16. doi: 10.17691/stm2022.14.2.02.

RIGHTS & PERMISSIONS

Eco-Vector