4872

Differentiation of Osteosarcoma and Ewing Sarcoma Using Radiomic AnalysisBased on T2 and CET1 MRI
Yi Dai1,2, Nan Hong2, and Guanxun Cheng1

1Peking University Shenzhen Hospital, Shenzhen, China, 2Peking University People's Hospital, Beijing, China

Synopsis

In this study, we assessed the ability of our newly established radiomic model based on using multiparametric MR data to help differentiate OS from EWS of the pelvis. We evaluated 16 features that were extracted and selected by using the LASSO method. Our radiomics model yielded favorable results and constituted a new technique for the discrimination of OS and EWS. The AUC was high for both T2-FS and CET1. High specificity was achieved when using data both from T2-FS and CET1 (82.9% and 100%, respectively) and the sensitivity was also high from T2-FS (74.2%). In brief, we believe that the methodology developed in this work may serve as a reliable additional tool for differentiation OS from EWS.

Introduction

To determine if osteosarcoma (OS) and Ewing sarcoma (EWS) of the pelvis based on MRI can be differentiated using radiomic analysis.

Materials and Methods

In this study, 3.0 T magnetic resonance (MR) data of 66 patients (40 males and 26 females, mean age 27.6±13.9 years) with pathologically confirmed OS or EWS of the pelvis (35 with OS and 31 with EWS) taken from April 2013 to December 2017 were retrospectively reviewed. T2-weighted fat-saturated (T2-FS) and contrast-enhanced T1-weighted (CET1) images were manually segmented, and imaging features were extracted. Independent-sample t-test, Spearman’s test, and the least absolute shrinkage and selection operator (LASSO) method were used to select the most useful features from the original data set. Logistic regression was applied to build a diagnostic model. The performance of radiomic analysis was investigated by the area under the receiver operating characteristic (ROC) curve (AUC) analysis.

Results

385 initial features were extracted from T2-FS and CET1 MR data. 9 features from T2-FS and 7 features from CET1 were selected by using the LASSO method. The radiomic analysis to differentiate OS and EWS of the pelvis based on T2-FS and CET1 images using the aforementioned selected features achieved AUC values of 0.881 [95% confidence interval (CI): 0.799–0.963] and 0.765 (95% CI: 0.652–0.878), respectively.

Discussion

We assessed the ability of our newly established radiomic model based on using multiparametric MR data to help differentiate OS from EWS of the pelvis. We evaluated 16 features that were extracted and selected by using the LASSO method. Our radiomics model yielded favorable results and constituted a new technique for the discrimination of OS and EWS. The AUC was high for both T2-FS and CET1. High specificity was achieved when using data both from T2-FS and CET1 (82.9% and 100%, respectively) and the sensitivity was also high from T2-FS (74.2%). In brief, we believe that the methodology developed in this work may serve as a reliable additional tool for differentiation OS from EWS.

Conclusion

Radiomic analysis showed a potential in differentiating OS from EWS of the pelvis, in which T2-FS demonstrated better diagnostic value. To differentiate OS from EWS of the pelvis using our multiparametric MRI-based radiomic analysis could preoperatively improve diagnostic accuracy and greatly contribute to therapy planning.

Acknowledgements

No acknowledgement found.

References

1. Moore DD, Luu HH. Osteosarcoma. Cancer treatment and research. 2014;162:65-92.

2. Iannaci G, Luise R, Sapere P, Costanzo RM, Rossiello R. Extraskeletal osteosarcoma: a very rare case report of primary tumor of the colon-rectum and review of the literature. Pathol Res Pract. 2013;209(6):393-6.

3. Sand LG, Szuhai K, Hogendoorn PC. Sequencing Overview of Ewing Sarcoma: A Journey across Genomic, Epigenomic and Transcriptomic Landscapes. Int J Mol Sci. 2015;16(7):16176-215.

4. Neubauer H, Evangelista L, Hassold N, Winkler B, Schlegel PG, Kostler H, et al. Diffusion-weighted MRI for detection and differentiation of musculoskeletal tumorous and tumor-like lesions in pediatric patients. World journal of pediatrics : WJP. 2012;8(4):342-9.

5. Bajpai J, Gamanagatti S, Sharma MC, Kumar R, Vishnubhatla S, Khan SA, et al. Noninvasive imaging surrogate of angiogenesis in osteosarcoma. Pediatric blood & cancer. 2010;54(4):526-31.

6. Gillies RJ, Kinahan PE, Hricak H. Radiomics: Images Are More than Pictures, They Are Data. Radiology. 2016;278(2):563-77.

7. Corino VDA, Montin E, Messina A, Casali PG, Gronchi A, Marchiano A, et al. Radiomic analysis of soft tissues sarcomas can distinguish intermediate from high-grade lesions. J Magn Reson Imaging. 2017.

8. Huang YQ, Liang CH, He L, Tian J, Liang CS, Chen X, et al. Development and Validation of a Radiomics Nomogram for Preoperative Prediction of Lymph Node Metastasis in Colorectal Cancer. J Clin Oncol. 2016;34(18):2157-64.

9. Wu S, Zheng J, Li Y, Yu H, Shi S, Xie W, et al. A Radiomics Nomogram for the Preoperative Prediction of Lymph Node Metastasis in Bladder Cancer. Clin Cancer Res. 2017;23(22):6904-11.

10. Li H, Zhu Y, Burnside ES, Drukker K, Hoadley KA, Fan C, et al. MR Imaging Radiomics Signatures for Predicting the Risk of Breast Cancer Recurrence as Given by Research Versions of MammaPrint, Oncotype DX, and PAM50 Gene Assays. Radiology. 2016;281(2):382-91.

11. Bowen SR, Yuh WTC, Hippe DS, Wu W, Partridge SC, Elias S, et al. Tumor radiomic heterogeneity: Multiparametric functional imaging to characterize variability and predict response following cervical cancer radiation therapy. J Magn Reson Imaging. 2018;47(5):1388-96.

12. Zhao B, Tan Y, Tsai WY, Qi J, Xie C, Lu L, et al. Reproducibility of radiomics for deciphering tumor phenotype with imaging. Scientific reports. 2016;6:23428.

13. Guo W, Li H, Zhu Y, Lan L, Yang S, Drukker K, et al. Prediction of clinical phenotypes in invasive breast carcinomas from the integration of radiomics and genomics data. Journal of medical imaging (Bellingham, Wash). 2015;2(4):041007.

14. Sutton EJ. Histograms and the Zone System.

15. [Available from: https://en.wikipedia.org/wiki/Form_factor_(design).

16. Haralick RM, Shanmugam K, Dinstein I. TEXTURAL FEATURES FOR IMAGE CLASSIFICATION. IEEE Trans Syst Man Cybern. 1973;SMC3(6):610-21.

17. Galloway MM. Texture analysis using Gray level run lengths. Texture analysis using Gray level run lengths. 1974:57 pp- pp.

18. Chu A, Sehgal CM, Greenleaf JF. USE OF GRAY VALUE DISTRIBUTION OF RUN LENGTHS FOR TEXTURE ANALYSIS. Pattern Recognition Letters. 1990;11(6):415-9.

19. Dasarathy BV, Holder EB. IMAGE CHARACTERIZATIONS BASED ON JOINT GRAY LEVEL RUN LENGTH DISTRIBUTIONS. Pattern Recognition Letters. 1991;12(8):497-502.

20. Sauerbrei W, Royston P, Binder H. Selection of important variables and determination of functional form for continuous predictors in multivariable model building. Statistics in medicine. 2007;26(30):5512-28.

21. Hepp T SM, Gefeller O, Waldmann E, Mayr A. Approaches to Regularized Regression - A Comparison between Gradient Boosting and the Lasso. Methods Inf Med. 2016;55:422-30.

22. Eftekhari F. Imaging assessment of osteosarcoma in childhood and adolescence: diagnosis, staging, and evaluating response to chemotherapy. Cancer treatment and research. 2009;152:33-62.

23. Murphey MD, Senchak LT, Mambalam PK, Logie CI, Klassen-Fischer MK, Kransdorf MJ. From the radiologic pathology archives: ewing sarcoma family of tumors: radiologic-pathologic correlation. Radiographics : a review publication of the Radiological Society of North America, Inc. 2013;33(3):803-31.

24. Lim HK, Jee WH, Jung JY, Paek MY, Kim I, Jung CK, et al. Intravoxel incoherent motion diffusion-weighted MR imaging for differentiation of benign and malignant musculoskeletal tumours at 3 T. The British journal of radiology. 2018;91(1082):20170636.

25. Henderson S, Purdie C, Michie C, Evans A, Lerski R, Johnston M, et al. Interim heterogeneity changes measured using entropy texture features on T2-weighted MRI at 3.0 T are associated with pathological response to neoadjuvant chemotherapy in primary breast cancer. Eur Radiol. 2017;27(11):4602-11.

26. Bickelhaupt S, Paech D, Kickingereder P, Steudle F, Lederer W, Daniel H, et al. Prediction of malignancy by a radiomic signature from contrast agent-free diffusion MRI in suspicious breast lesions found on screening mammography. J Magn Reson Imaging. 2017;46(2):604-16.

27. Coroller TP, Agrawal V, Narayan V, Hou Y, Grossmann P, Lee SW, et al. Radiomic phenotype features predict pathological response in non-small cell lung cancer. Radiotherapy and oncology : journal of the European Society for Therapeutic Radiology and Oncology. 2016;119(3):480-6.

28. Kumar V, Gu Y, Basu S, Berglund A, Eschrich SA, Schabath MB, et al. Radiomics: the process and the challenges. Magnetic resonance imaging. 2012;30(9):1234-48.

29. Wu W, Parmar C, Grossmann P, Quackenbush J, Lambin P, Bussink J, et al. Exploratory Study to Identify Radiomics Classifiers for Lung Cancer Histology. Frontiers in oncology. 2016;6:71.

Figures

Fig 1. Workflow of this study. (1) MR images acquired from a qualified study data set. (2) Tumor segmentation was performed on T2-FS and CET1 MR images. Experienced radiologists contoured the tumor areas on MRI slices. (3) 385 features in total were extracted from original MR data. (4) Independent-sample t test, Spearman’s test and the LASSO regression were used to conduct feature reduction. (5) ROC analysis and radiomics nomograms were used to evaluate the established model.

Fig 2. Performance of radiomic analysis using T2-FS and CET1 images. The AUC values are 0.881 (95% CI: 0.799–0.963) and 0.765 (95% CI: 0.652–0.878).

Fig 3. Texture feature derived from T2-FS selected by using the least absolute shrinkage and selection operator (LASSO) binary logistic regression model. (A) Selection of the tuning parameter (λ) in the LASSO model via ten-fold cross-validation based on minimum criteria. (B) LASSO coefficient profiles of the 385 texture features. The nine selected features with nonzero coefficients are indicated in the plot.

Fig 4. Developed radiomics nomogram using T2-FS data.

Proc. Intl. Soc. Mag. Reson. Med. 27 (2019)
4872