Share：
Share this content in WeChat
X
[Chinese][PDF]960142
• Clinical Article •
Application of different radiomics models based on MRI conventional T2WI in preoperative tri-classification of ovarian epithelial tumors
HU Yan  LIU Yang  ZHENG Yineng  XIAO Zhibo  CHEN Liping  ZHANG Jian  DAI Mengying  LI Guanghui  ZHONG Yuqing  MA Si  LÜ Fajin 

Cite this article as: Hu Y, Liu Y, Zheng YE, et al. Application of different radiomics models based on MRI conventional T2WI in preoperative tri-classification of ovarian epithelial tumors[J]. Chin J Magn Reson Imaging, 2021, 12(12): 34-38, 54. DOI:10.12015/issn.1674-8034.2021.12.007.


[Abstract] Objective Conventional T2WI sequences based on MRI were used to compare the diagnostic efficacy of the radiomics models established by different machine learning algorithms in preoperative tri-classification of epithelial ovarian tumors. Materials and Methods: Preoperative MR images of 300 patients (100 benign, 100 borderline and 100 malignant) with pathologically confirmed ovarian epithelial tumors were retrospectively analyzed, and all the data were randomly divided into training sets and testing sets according to the ratio of 8∶2. Image features are extracted from the volume of interest (VOI) manually drawn on the axial T2WI, and screening them. Four feature selection methods and seven machine learning classifiers were pairwise combined to construct 28 classification models. AUC and accuracy were used to evaluate the prediction performance of all models.Results The best performance among 28 classification models is the "RFE-KNN" model that combines recursive feature elimination (RFE) and K nearest neighbor (KNN) classifiers. AUC of benign, borderline and malignant group was 0.94, 0.93 and 0.96.Conclusions Quantitative radiomics features extracted from T2WI have a good performance in differentiating benign, borderline, and malignant epithelial ovarian tumors.
[Keywords] ovarian epithelial tumor；magnetic resonance imaging；radiomics；machine learning；T2-weighted imaging

HU Yan1   LIU Yang1   ZHENG Yineng1, 2   XIAO Zhibo2   CHEN Liping2   ZHANG Jian1   DAI Mengying1   LI Guanghui1   ZHONG Yuqing1   MA Si1   LÜ Fajin1*  

1 State Key Laboratory of Ultrasound in Medicine and Engineering, College of Biomedical Engineering, Chongqing Medical University, Chongqing, 400016, China

2 Department of Radiology, the First Affiliated Hospital of Chongqing Medical University, Chongqing 400016, China

Lü FJ, E-mail: fajinlv@163.com

Conflicts of interest   None.

Received  2021-07-01
Accepted  2021-09-24
DOI: 10.12015/issn.1674-8034.2021.12.007
Cite this article as: Hu Y, Liu Y, Zheng YE, et al. Application of different radiomics models based on MRI conventional T2WI in preoperative tri-classification of ovarian epithelial tumors[J]. Chin J Magn Reson Imaging, 2021, 12(12): 34-38, 54. DOI:10.12015/issn.1674-8034.2021.12.007.

[1]
Mutch DG, Prat J. 2014 FIGO staging for ovarian, fallopian tube and peritoneal cancer[J]. Gynecol Oncol, 2014, 133(3): 401-404. DOI: 10.1016/j.ygyno.2014.04.013.
[2]
Hauptmann S, Friedrich K, Redline R, et al. Ovarian borderline tumors in the 2014 WHO classification: evolving concepts and diagnostic criteria[J]. Virchows Arch, 2017, 470(2): 125-142. DOI: 10.1007/s00428-016-2040-8.
[3]
Guillaume A, Pirrello O. Preservation of fertility in surgery of benign and borderline malignant ovarian tumors[J]. J Visc Surg, 2018, 155 (Supplement 1): S17-S21. DOI: 10.1016/j.jviscsurg.2018.04.001.
[4]
El Din AA, Badawi MA, Aal SE, et al. DNA Cytometry and Nuclear Morphometry in Ovarian Benign, Borderline and Malignant Tumors[J]. Open Access Maced J Med Sci, 2015, 3(4): 537-544. DOI: 10.3889/oamjms.2015.104.
[5]
Sun Y, Xu J, Jia XM. The Diagnosis, Treatment, Prognosis and Molecular Pathology of Borderline Ovarian Tumors: Current Status and Perspectives[J]. Cancer management and research, 2020, 12: 3651-3659. DOI: 10.2147/CMAR.S250394.
[6]
Du Bois A, Trillsch F, Mahner S, et al. Management of borderline ovarian tumors[J]. Annals of oncology: official journal of the European Society for Medical Oncology, 2016, 27(supplement 1): i20-i22. DOI: 10.1093/annonc/mdw090.
[7]
Foti PV, Attina G, Spadola S, et al. MR imaging of ovarian masses: classification and differential diagnosis[J]. Insights Imaging, 2016, 7(1): 21-41. DOI: 10.1007/s13244-015-0455-4.
[8]
Gity M, Parviz S, Saligheh Rad H, et al. Differentiation of Benign from Malignant Adnexal Masses by Dynamic Contrast-Enhanced MRI (DCE-MRI): Quantitative and Semi-quantitative analysis at 3-Tesla MRI[J]. Asian Pac J Cancer Prev, 2019, 20(4): 1073-1079. DOI: 10.31557/APJCP.2019.20.4.1073.
[9]
Denewar FA, Takeuchi M, Urano M, et al. Multiparametric MRI for differentiation of borderline ovarian tumors from stage I malignant epithelial ovarian tumors using multivariate logistic regression analysis[J]. Eur J Radiol, 2017, 91: 116-123. DOI: 10.1016/j.ejrad.2017.04.001.
[10]
Li HM, Feng F, Qiang JW, et al. Quantitative dynamic contrast-enhanced MR imaging for differentiating benign, borderline, and malignant ovarian tumors[J]. Abdom Radiol (NY), 2018, 43(11): 3132-3141. DOI: 10.1007/s00261-018-1569-1.
[11]
Zhang H, Zhang GF, He ZY, et al. Prospective evaluation of 3T MRI findings for primary adnexal lesions and comparison with the final histological diagnosis[J]. Arch Gynecol Obstet, 2014, 289(2): 357-364. DOI: 10.1007/s00404-013-2990-x.
[12]
Ma FH, Zhao SH, Qiang JW, et al. MRI appearances of mucinous borderline ovarian tumors: pathological correlation[J]. Journal of magnetic resonance imaging: JMRI, 2014, 40(3): 745-751. DOI: 10.1002/jmri.24408.
[13]
Flicek K, Vanburen W, Dudiak K, et al. Borderline epithelial ovarian tumors: what the radiologist should know[J]. Abdom Radiol (New York), 2021, 46(6): 2350-2366. DOI: 10.1007/s00261-020-02688-z.
[14]
Lambin P, Rios-Velazquez E, Leijenaar R, et al. Radiomics: extracting more information from medical images using advanced feature analysis[J]. Eur J Cancer, 2012, 48(4): 441-446. DOI: 10.1016/j.ejca.2011.11.036.
[15]
Gillies RJ, Kinahan PE, Hricak H. Radiomics: Images Are More than Pictures, They Are Data[J]. Radiology, 2016, 278(2): 563-577. DOI: 10.1148/radiol.2015151169.
[16]
Rogers W, Thulasi Seetha S, Refaee TAG, et al. Radiomics: from qualitative to quantitative imaging[J]. Br J Radiol, 2020, 93(1108): 20190948. DOI: 10.1259/bjr.20190948.
[17]
Mayerhoefer ME, Materka A, Langs G, et al. Introduction to Radiomics[J]. J Nucl Med, 2020, 61(4): 488-495. DOI: 10.2967/jnumed.118.222893.
[18]
Xie HH, Hu J, Zhang XD, et al. Preliminary utilization of radiomics in differentiating uterine sarcoma from atypical leiomyoma: Comparison on diagnostic efficacy of MRI features and radiomic features[J]. Eur J Radiol, 2019, 115: 39-45. DOI: 10.1016/j.ejrad.2019.04.004.
[19]
Xiao G, Rong WC, Hu YC, et al. MRI Radiomics Analysis for Predicting the Pathologic Classification and TNM Staging of Thymic Epithelial Tumors: A Pilot Study[J]. AJR Am J Roentgenol, 2020, 214(2): 328-340. DOI: 10.2214/AJR.19.21696.
[20]
Peerlings J, Woodruff HC, Winfield JM, et al. Stability of radiomics features in apparent diffusion coefficient maps from a multi-centre test-retest trial[J]. Sci Rep, 2019, 9(1): 4800. DOI: 10.1038/s41598-019-41344-5.
[21]
Zhang H, Mao YF, Chen XJ, et al. Magnetic resonance imaging radiomics in categorizing ovarian masses and predicting clinical outcome: a preliminary study[J]. Eur Radiol, 2019, 29(7): 3358-3371. DOI: 10.1007/s00330-019-06124-9.
[22]
Kazerooni AF, Malek M, Haghighatkhah H, et al. Semiquantitative dynamic contrast-enhanced MRI for accurate classification of complex adnexal masses[J]. J Magn Reson Imaging, 2017, 45(2): 418-427. DOI: 10.1002/jmri.25359.
[23]
Nougaret S, Tardieu M, Vargas HA, et al. Ovarian cancer: An update on imaging in the era of radiomics[J]. Diagn Interv Imaging, 2019, 100(10): 647-655. DOI: 10.1016/j.diii.2018.11.007.
[24]
Li YA, Jian JM, Pickhardt PJ, et al. MRI-Based Machine Learning for Differentiating Borderline From Malignant Epithelial Ovarian Tumors: A Multicenter Study[J]. J Magn Reson Imaging, 2020, 52(3): 897-904. DOI: 10.1002/jmri.27084.
[25]
Jian JM, Li YA, Pickhardt PJ, et al. MR image-based radiomics to differentiate type Iota and type IotaIota epithelial ovarian cancers[J]. Eur Radiol, 2021, 31(1): 403-410. DOI: 10.1007/s00330-020-07091-2.
[26]
Song XL, Ren JL, Zhao D, et al. Radiomics derived from dynamic contrast-enhanced MRI pharmacokinetic protocol features: the value of precision diagnosis ovarian neoplasms[J]. Eur Radiol, 2021, 31(1): 368-378. DOI: 10.1007/s00330-020-07112-0.
[27]
Yang ZW, Guo T, Xie HB, et al. Application of radiomics in the grading of brain tumor[J]. Chin J Magn Reson Imaging, 2018, 9(6): 439-445. DOI: 10.12015/issn.1674-8034.2018.06.008.
[28]
Mu JH, Zhang YW, Wu ZG. Application of different radiomics dignostic models based on conventional MR images in the preprotive grading of brain glioma[J]. Chin J Magn Reson Imaging, 2020, 11(1): 55-59. DOI: 10.12015/issn.1674-8034.2020.01.012.
[29]
Zhang YP, Zhang BR, Liang F, et al. Radiomics features on non-contrast-enhanced CT scan can precisely classify AVM-related hematomas from other spontaneous intraparenchymal hematoma types[J]. Eur Radiol, 2019, 29(4): 2157-2165. DOI: 10.1007/s00330-018-5747-x.
[30]
Almustafa KM. Prediction of heart disease and classifiers' sensitivity analysis[J]. BMC Bioinformatics, 2020, 21(1): 242-252. DOI: 10.1186/s12859-020-03626-y.
[31]
Lopez-Rincon A, Mendoza-Maldonado L, Martinez-Archundia M, et al. Machine Learning-Based Ensemble Recursive Feature Selection of Circulating miRNAs for Cancer Tumor Classification[J]. Cancers, 2020, 12(7): 1785. DOI: 10.3390/cancers12071785.
[32]
Zhang F, Petersen M, Johnson L, et al. Recursive Support Vector Machine Biomarker Selection for Alzheimer's Diseas[J]. Journal of Alzheimer's Disease: JAD, 2021, 79(4): 1691-1700. DOI: 10.3233/jad-201254.
[33]
Tahmassebi A, Wengert GJ, Helbich TH, et al. Impact of Machine Learning With Multiparametric Magnetic Resonance Imaging of the Breast for Early Prediction of Response to Neoadjuvant Chemotherapy and Survival Outcomes in Breast Cancer Patients[J]. Invest Radiol, 2019, 54(2): 110-117. DOI: 10.1097/RLI.0000000000000518.
[34]
Wang XH, Wan Q, Chen HJ, et al. Classification of pulmonary lesion based on multiparametric MRI: utility of radiomics and comparison of machine learning methods[J]. Eur Radiol, 2020, 30(8): 4595-4605. DOI: 10.1007/s00330-020-06768-y.
[35]
Kamal S, Ripon SH, Dey N, et al. A MapReduce approach to diminish imbalance parameters for big deoxyribonucleic acid dataset[J]. Computer Methods and Programs in Biomedicine, 2016, 131(7): 191-206. DOI: 10.1016/j.cmpb.2016.04.005.
[36]
Maillo J, Ramírez S, Triguero I, et al. kNN-IS: An Iterative Spark-based design of the k-Nearest Neighbors classifier for big data[J]. Knowledge-Based Systems, 2017, 117: 3-15. DOI: 10.1016/j.knosys.2016.06.012.
[37]
Deng ZY, Zhu XS, Cheng DB, et al. Efficient kNN classification algorithm for big data[J]. Neurocomputing, 2016, 195: 143-148. DOI: 10.1016/j.neucom.2015.08.112.
[38]
Gallego AJ, Calvo-Zaragoza J, Valero-Mas J, et al. Clustering-based k-nearest neighbor classification for large-scale data with neural codes representation[J]. Pattern Recognition, 2018, 74: 531-543. DOI: 10.1016/j.patcog.2017.09.038.
[39]
Fischerova D, Zikan M, Dundr P, et al. Diagnosis, treatment, and follow-up of borderline ovarian tumors[J]. Oncologist, 2012, 17(12): 1515-1533. DOI: 10.1634/theoncologist.2012-0139.
[40]
Taylor E, Irshaid L, Mathur M. Multimodality Imaging Approach to Ovarian Neoplasms with Pathologic Correlation[J]. Radiographics:a review publication of the Radiological Society of North America, Inc, 2021, 41(1): 289-315. DOI: 10.1148/rg.2021200086.

PREV Multiparametric magnetic resonance imaging to characterize pathological grading and stage of cervical squamous cell carcinoma
NEXT Radiomics features of sub-attention region on the diagnosis of the extracapsular extension of the prostate cancer on magnetic resonance imaging
  


LINK


Tel & Fax: +8610-67113815    E-mail: editor@cjmri.cn