Skip to main content
SpringerLink
Log in

Combining multiparametric MRI features-based transfer learning and clinical parameters: application of machine learning for the differentiation of uterine sarcomas from atypical leiomyomas

  • Imaging Informatics and Artificial Intelligence
  • Published:
European Radiology Aims and scope Submit manuscript

    We’re sorry, something doesn't seem to be working properly.

    Please try refreshing the page. If that doesn't work, please contact support so we can address the problem.

Abstract

Objectives

To explore the feasibility and effectiveness of machine learning (ML) based on multiparametric magnetic resonance imaging (mp-MRI) features extracted from transfer learning combined with clinical parameters to differentiate uterine sarcomas from atypical leiomyomas (ALMs).

Methods

The data of 86 uterine sarcomas between July 2011 and December 2019 and 86 ALMs between June 2013 and June 2017 were retrospectively reviewed. We extracted deep-learning features and radiomics features from T2-weighted imaging (T2WI) and diffusion weighted imaging (DWI). The two feature extraction methods, transfer learning and radiomics, were compared. Random forest was adopted as the classifier. T2WI features, DWI features, combined T2WI and DWI (mp-MRI) features, and combined clinical parameters and mp-MRI features were applied to establish T2, DWI, T2-DWI, and complex multiparameter (mp) models, respectively. Predictive performance was assessed with the area under the receiver operating characteristic curve (AUC).

Results

In the test set, the T2, DWI, T2-DWI and complex mp models based on transfer learning (AUCs range from 0.76 to 0.81, 0.80 to 0.88, 0.85 to 0.92, and 0.94 to 0.96, respectively) outperformed the models based on radiomics (AUCs of 0.73, 0.76, 0.79, and 0.92, respectively). Moreover, the complex mp model showed the best prediction performance, with the Resnet50-complex mp model achieving the highest AUC (0.96) and accuracy (0.87).

Conclusions

Transfer learning is feasible and superior to radiomics in the differential diagnosis of uterine sarcomas and ALMs in our dataset. ML models based on deep learning features of nonenhanced mp-MRI and clinical parameters can achieve good diagnostic efficacy.

Key Points

• The ML model combining nonenhanced mp-MRI features and clinical parameters can distinguish uterine sarcomas from ALMs.

• Transfer learning can be applied to differentiate uterine sarcomas from ALMs and outperform radiomics.

• The most accurate prediction model was Resnet50-based transfer learning, built with the deep-learning features of mp-MRI and clinical parameters.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Log in via an institution

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Institutional subscriptions

Fig. 1
Fig. 2
Fig. 3
Fig. 4

Similar content being viewed by others

Abbreviations

ADC:

Apparent diffusion coefficient

ALMs:

Atypical leiomyomas

AUC:

Area under the curve

CNN:

Convolutional neural network

DWI:

Diffusion weighted imaging

DCE-MRI:

Dynamic contrast-enhanced magnetic resonance imaging

Grad-CAM:

Gradient-weighted class activation mapping

LASSO:

Least absolute shrinkage selector operator

ML:

Machine learning

Mp-MRI:

Multiparametric magnetic resonance imaging

RF:

Random forest

ROC:

Receiver operating characteristic curve

ROI:

Region of interest

T2WI:

T2-weighted imaging

3.References

  1. Bi Q, Xiao Z, Lv F, Liu Y, Zou C, Shen Y (2018) Utility of clinical parameters and multiparametric MRI as predictive factors for differentiating uterine sarcoma from atypical leiomyoma. Acad Radiol 25:993–1002

    Article  Google Scholar 

  2. Tamai K, Koyama T, Saga T et al (2008) The utility of diffusion-weighted MR imaging for differentiating uterine sarcomas from benign leiomyomas. Eur Radiol 18:723–730

    Article  Google Scholar 

  3. Mendez RJ (2020) MRI to differentiate atypical leiomyoma from uterine sarcoma. Radiology 297:372–373

    Article  Google Scholar 

  4. Aleksandrovych V, Bereza T, Sajewicz M, Walocha J, Gil K (2015) Uterine fibroid: common features of widespread tumor. Folia Med Cracov 55

  5. Yang S, Kong F, Hou R et al (2017) Ultrasound guided high-intensity focused ultrasound combined with gonadotropin releasing hormone analogue (GnRHa) ablating uterine leiomyoma with homogeneous hyperintensity on T 2 weighted MR imaging. Br J Radiol 90:20160760

  6. Brito Pires NM, Godoi ET, Oliveira DC et al (2017) Impact of pelvic magnetic resonance imaging findings in the indication of uterine artery embolization in the treatment of myoma. Ginekol Pol 88:129–133

    Article  Google Scholar 

  7. Owen C, Armstrong AY (2015) Clinical management of leiomyoma. Obstet Gynecol Clin North Am 42:67–85

    Article  Google Scholar 

  8. Mbatani N, Olawaiye AB, Prat J (2018) Uterine sarcomas. Int J Gynaecol Obstet 143(Suppl 2):51–58

    Article  Google Scholar 

  9. Namimoto T, Yamashita Y, Awai K et al (2009) Combined use of T2-weighted and diffusion-weighted 3-T MR imaging for differentiating uterine sarcomas from benign leiomyomas. Eur Radiol 19:2756–2764

    Article  Google Scholar 

  10. Thomassin-Naggara I, Dechoux S, Bonneau C et al (2013) How to differentiate benign from malignant myometrial tumours using MR imaging. Eur Radiol 23:2306–2314

  11. Abdel Wahab C, Jannot AS, Bonaffini PA et al (2020) Diagnostic algorithm to differentiate benign atypical leiomyomas from malignant uterine sarcomas with diffusion-weighted MRI. Radiology 297:361–371

    Article  Google Scholar 

  12. Causa Andrieu P, Woo S, Kim TH, Kertowidjojo E, Hodgson A, Sun S (2021) New imaging modalities to distinguish rare uterine mesenchymal cancers from benign uterine lesions. Curr Opin Oncol 33:464–475

    Article  CAS  Google Scholar 

  13. Malek M, Tabibian E, Dehgolan MR et al (2020) A diagnostic algorithm using multi-parametric MRI to differentiate benign from malignant myometrial tumors: machine-learning method. Sci Rep 10:1–12

  14. Nakagawa M, Nakaura T, Namimoto T et al (2019) A multiparametric MRI-based machine learning to distinguish between uterine sarcoma and benign leiomyoma: comparison with (18)F-FDG PET/CT. Clin Radiol 74:167 e161-167 e167

  15. Xie H, Hu J, Zhang X, Ma S, Liu Y, Wang X (2019) Preliminary utilization of radiomics in differentiating uterine sarcoma from atypical leiomyoma: comparison on diagnostic efficacy of MRI features and radiomic features. Eur J Radiol 115:39–45

    Article  Google Scholar 

  16. Xie H, Zhang X, Ma S, Liu Y, Wang X (2019) Preoperative differentiation of uterine sarcoma from leiomyoma: comparison of three models based on different segmentation volumes using radiomics. Mol Imaging Biol 21:1157–1164

    Article  Google Scholar 

  17. Nakagawa M, Nakaura T, Namimoto T et al (2019) Machine learning to differentiate T2-weighted hyperintense uterine leiomyomas from uterine sarcomas by utilizing multiparametric magnetic resonance quantitative imaging features. Acad Radiol 26:1390–1399

    Article  Google Scholar 

  18. Liao D, Xiao Z, Lv F, Chen J, Qiu L (2020) Non-contrast enhanced MRI for assessment of uterine fibroids' early response to ultrasound-guided high-intensity focused ultrasound thermal ablation. Eur J Radiol 122:108670

    Article  Google Scholar 

  19. Chan HP, Samala RK, Hadjiiski LM, Zhou C (2020) Deep learning in medical image analysis. Adv Exp Med Biol 1213:3–21

    Article  Google Scholar 

  20. LeCun Y, Bengio Y, Hinton G (2015) Deep learning. Nature 521:436–444

    Article  CAS  Google Scholar 

  21. Pan SJ, Yang Q (2010) A survey on transfer learning. IEEE Transactions on Knowledge and Data Engineering 22:1345–1359

    Article  Google Scholar 

  22. Sharif Razavian A, Azizpour H, Sullivan J, Carlsson S (2014) CNN features off-the-shelf: an astounding baseline for recognition Proceedings of the IEEE conference on computer vision and pattern recognition workshops, pp 806-813

  23. Liu W, Cheng Y, Liu Z et al (2021) Preoperative prediction of Ki-67 status in breast cancer with multiparametric MRI using transfer learning. Acad Radiol 28:e44–e53

    Article  Google Scholar 

  24. Deepak S, Ameer PM (2019) Brain tumor classification using deep CNN features via transfer learning. Comput Biol Med 111:103345

    Article  CAS  Google Scholar 

  25. Tian S, Niu M, Xie L, Song Q, Liu A (2021) Diffusion-tensor imaging for differentiating uterine sarcoma from degenerative uterine fibroids. Clin Radiol 76:313 e327-313 e332

  26. van Griethuysen JJM, Fedorov A, Parmar C et al (2017) Computational radiomics system to decode the radiographic phenotype. Cancer Res 77:e104–e107

    Article  Google Scholar 

  27. Balagurunathan Y, Kumar V, Gu Y et al (2014) Test-retest reproducibility analysis of lung CT image features. J Digit Imaging 27:805–823

    Article  Google Scholar 

  28. Hu Y, Xie C, Yang H et al (2021) Computed tomography-based deep-learning prediction of neoadjuvant chemoradiotherapy treatment response in esophageal squamous cell carcinoma. Radiother Oncol 154:6–13

    Article  CAS  Google Scholar 

  29. Muthukrishnan R, Rohini R (2016) LASSO: A feature selection technique in predictive modeling for machine learning2016 IEEE international conference on advances in computer applications (ICACA). IEEE, pp 18-20

  30. Gerges L, Popiolek D, Rosenkrantz AB (2018) Explorative investigation of whole-lesion histogram MRI metrics for differentiating uterine leiomyomas and leiomyosarcomas. AJR Am J Roentgenol 210:1172–1177

    Article  Google Scholar 

  31. Truhn D, Schrading S, Haarburger C, Schneider H, Merhof D, Kuhl C (2019) Radiomic versus convolutional neural networks analysis for classification of contrast-enhancing lesions at multiparametric breast MRI. Radiology 290:290–297

    Article  Google Scholar 

  32. Liu S, Zheng H, Feng Y, Li W (2017) Prostate cancer diagnosis using deep learning with 3D multiparametric MRIMedical imaging 2017: computer-aided diagnosis. International Society for Optics and Photonics, p 1013428

    Google Scholar 

  33. Lundervold AS, Lundervold A (2019) An overview of deep learning in medical imaging focusing on MRI. Z Med Phys 29:102–127

    Article  Google Scholar 

  34. Takahashi M, Kozawa E, Tanisaka M, Hasegawa K, Yasuda M, Sakai F (2016) Utility of histogram analysis of apparent diffusion coefficient maps obtained using 3.0T MRI for distinguishing uterine carcinosarcoma from endometrial carcinoma. J Magn Reson Imaging 43:1301–1307

    Article  Google Scholar 

  35. Testa AC, Di Legge A, Bonatti M, Manfredi R, Scambia G (2016) Imaging techniques for evaluation of uterine myomas. Best Pract Res Clin Obstet Gynaecol 34:37–53

    Article  Google Scholar 

  36. Ghassemi M, Oakden-Rayner L, Beam AL (2021) The false hope of current approaches to explainable artificial intelligence in health care. Lancet Digit Health 3:e745–e750

    Article  Google Scholar 

  37. Barral M, Place V, Dautry R et al (2017) Magnetic resonance imaging features of uterine sarcoma and mimickers. Abdom Radiol (NY) 42:1762–1772

    Article  Google Scholar 

  38. Wang T, Gong J, Li Q et al (2021) A combined radiomics and clinical variables model for prediction of malignancy in T2 hyperintense uterine mesenchymal tumors on MRI. Eur Radiol 31:6125–6135

    Article  Google Scholar 

  39. Xie H, Hu J, Zhang X, Ma S, Liu Y, Wang X (2019) Preliminary utilization of radiomics in differentiating uterine sarcoma from atypical leiomyoma: comparison on diagnostic efficacy of MRI features and radiomic features. Eur J Radiol 115:39–45

    Article  Google Scholar 

  40. Chen I, Firth B, Hopkins L, Bougie O, Xie RH, Singh S (2018) Clinical characteristics differentiating uterine sarcoma and fibroids. JSLS 22

Download references

Acknowledgements

The author would like to thank Yang Liu for his contribution to the collection of patients. Thanks to Yan Hu, Guanghui Li, and Jian Zhang who participated in the discussion of this research together. Thanks to the public code provided by Yihuai Hu et al.

Funding

This study has received funding from Chongqing Postgraduate Research and Innovation Project (NO. CYB20165).

Author information

Author notes

  1. Mengying Dai and Yang Liu contributed equally to this work.

Authors and Affiliations

  1. State Key Laboratory of Ultrasound in Medicine and Engineering, College of Biomedical Engineering, Chongqing Medical University, Yixueyuan Road, Yuzhong District, Chongqing, 400016, China

    Mengying Dai, Yang Liu, Yan Hu, Guanghui Li, Jian Zhang, Zhibo Xiao & Fajin Lv

  2. Chongqing Key Laboratory of Biomedical Engineering, Chongqing Medical University, Chongqing, 400016, China

    Mengying Dai, Yang Liu, Yan Hu, Guanghui Li, Jian Zhang & Fajin Lv

  3. Department of Radiology, The First Affiliated Hospital of Chongqing Medical University, Youyi Road, Yuzhong District, Chongqing, 400016, China

    Yang Liu, Zhibo Xiao & Fajin Lv

  4. Institute of Medical Data, Chongqing Medical University, Chongqing, 400016, China

    Fajin Lv

Authors
  1. Mengying Dai
    View author publications

    You can also search for this author in PubMed Google Scholar

  2. Yang Liu
    View author publications

    You can also search for this author in PubMed Google Scholar

  3. Yan Hu
    View author publications

    You can also search for this author in PubMed Google Scholar

  4. Guanghui Li
    View author publications

    You can also search for this author in PubMed Google Scholar

  5. Jian Zhang
    View author publications

    You can also search for this author in PubMed Google Scholar

  6. Zhibo Xiao
    View author publications

    You can also search for this author in PubMed Google Scholar

  7. Fajin Lv
    View author publications

    You can also search for this author in PubMed Google Scholar

Corresponding authors

Correspondence to Zhibo Xiao or Fajin Lv.

Ethics declarations

Guarantor

The scientific guarantor of this publication is Fajin Lv.

Conflict of interest

The authors of this manuscript declare no relationships with any companies, whose products or services may be related to the subject matter of the article.

Statistics and biometry

No complex statistical methods were necessary for this paper.

Informed consent

Written informed consent was waived by the Institutional Review Board.

Ethical approval

Institutional Review Board approval was obtained.

Methodology

• Retrospective

• diagnostic or prognostic study

• performed at one institution

Additional information

Publisher’s note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary information

ESM 1

(DOCX 250 kb)

Rights and permissions

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Dai, M., Liu, Y., Hu, Y. et al. Combining multiparametric MRI features-based transfer learning and clinical parameters: application of machine learning for the differentiation of uterine sarcomas from atypical leiomyomas. Eur Radiol 32, 7988–7997 (2022). https://doi.org/10.1007/s00330-022-08783-7

Download citation

  • Received:

  • Revised:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1007/s00330-022-08783-7

Keywords

Search

Navigation