Share：
Share this content in WeChat
X
[Chinese][PDF]1883
• Clinical Article •
Predicting lymph-vascular space invasion in cervical cancer based on MR-T2WI with deep learning and radiomic features combined with clinical features
LIN Baojin  LONG Xianfeng  WU Zhaoxia  LIANG Lili  LU Zihong  GAN Wutian  ZHU Chaohua 

Cite this article as: LIN B J, LONG X F, WU Z X, et al. Predicting lymph-vascular space invasion in cervical cancer based on MR-T2WI with deep learning and radiomic features combined with clinical features[J]. Chin J Magn Reson Imaging, 2024, 15(3): 130-136. DOI:10.12015/issn.1674-8034.2024.03.021.


[Abstract] Objective To explore the value of preoperative prediction of cervical cancer lymph-vascular space invasion (LVSI) by combining deep transfer learning features based on MR-T2WI, radiomic features, and clinical characteristics.Materials and Methods Data of 178 patients with cervical cancer by postoperative pathology, including 70 cases with LVSI (+) and 108 cases with LVSI (-) were retrospectively analyzed. The patients were divided into training set [n=142, including 54 LVSI (+) and 88 LVSI (-)] and test set [n=36, including 16 LVSI (+) and 20 LVSI (-)] at a ratio of 8∶2. Univariate logistic regression analysis was conducted on clinical factors to identify independent predictors for LVSI (+) cases. The deep transfer learning (DTL) method and traditional radiomics methods were used to extract the DTL features and radiomics features from the lesions in the sagittal T2WI images. This led to the construction of a deep transfer learning feature dataset, a radiomics feature dataset, and a dataset that merges the DTL features with the radiomics features. Each feature dataset in the training set underwent feature dimension reduction using t-tests, Pearson analysis, and least absolute shrinkage and selection operator (LASSO) regression. The best of these were used to construct radiomics models [radiomics (Rad) model, DTL model, Fusion model of Rad and DTL features (Rad+DTL) model], and the optimal radiomics model was selected. A joint model was constructed based on the best model's radiomics score and independent clinical factors, and a nomogram was drawn. The calibration of the model was evaluated using calibration curves, and the application value of the model was assessed using decision curve analysis.Results Lymph node metastasis and the neutrophil-to-lymphocyte ratio were identified as independent predictors (P<0.05) with LVSI (+). The Rad+DTL model was determined as the optimal radiomics model. The combined model exhibited a higher AUC in the training set compared to the Rad+DTL model (0.984 vs. 0.966, P<0.05), and in the testing set (0.912 vs. 0.759, P=0.05). The combined model showed higher calibration accuracy and greater clinical net benefit.Conclusions The combination of DTL features from MR-T2WI, radiomics features, and clinical characteristics can effectively predict LVSI in cervical cancer.
[Keywords] cervical cancer；lymph-vascular space invasion；radiomics；magnetic resonance imaging；deep transfer learning

LIN Baojin1, 2   LONG Xianfeng1   WU Zhaoxia2   LIANG Lili1   LU Zihong3   GAN Wutian1   ZHU Chaohua1*  

1 Department of Radiation Therapy Physics Laboratory, People's Hospital of Guangxi Zhuang Autonomous Region, Nanning 530021, China

2 Tsinghua University, Beijing 100084, China

3 Department of Radiation Therapy, Peking University Cancer Hospital, Beijing 100142, China

Corresponding author: ZHU C H, E-mail: th0624@163.com

Conflicts of interest   None.

Received  2023-11-30
Accepted  2024-02-23
DOI: 10.12015/issn.1674-8034.2024.03.021
Cite this article as: LIN B J, LONG X F, WU Z X, et al. Predicting lymph-vascular space invasion in cervical cancer based on MR-T2WI with deep learning and radiomic features combined with clinical features[J]. Chin J Magn Reson Imaging, 2024, 15(3): 130-136. DOI:10.12015/issn.1674-8034.2024.03.021.

[1]
BRAY F, FERLAY J, SOERJOMATARAM I, et al. Global cancer statistics 2018: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries[J]. CA Cancer J Clin, 2018, 68(6): 394-424. DOI: 10.3322/caac.21492.
[2]
MA C J, LIU A L. Advance of deep learning and radiomics in the endometrial cancer study[J]. Int J Med Radiol, 2023, 46(3): 303-307. DOI: 10.19300/j.2023.Z20064.
[3]
RAMLI Z, KARIM M K A, EFFENDY N, et al. Stability and reproducibility of radiomic features based on various segmentation techniques on cervical cancer DWI-MRI[J/OL]. Diagnostics, 2022, 12(12): 3125 [2024-02-02]. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9777485/. DOI: 10.3390/diagnostics12123125.
[4]
PAIK E S, LIM M C, KIM M H, et al. Prognostic model for survival and recurrence in patients with early-stage cervical cancer: a Korean gynecologic oncology group study (KGOG 1028)[J]. Cancer Res Treat, 2020, 52(1): 320-333. DOI: 10.4143/crt.2019.124.
[5]
YU Q, LOU X M, HE Y. Prediction of local recurrence in cervical cancer by a Cox model comprised of lymph node status, lymph-vascular space invasion, and intratumoral Th17 cell-infiltration[J/OL]. Med Oncol, 2014, 31(1): 795 [2024-02-02]. https://link.springer.com/article/10.1007/s12032-013-0795-1. DOI: 10.1007/s12032-013-0795-1.
[6]
ZHOU H, LIU Y Y, LUO M, et al. Interpretation of 2023 NCCN clinical practice guidelines for cervical cancer (1st edition)[J]. Chin J Pract Gynecol Obstet, 2023, 39(2): 189-196. DOI: 10.19538/j.fk2023020115.
[7]
ZHANG J, WEN H W. Relationship among lymphovascular space invasion with clinicopathological factors and \rsurvival analysis in early cervical adenocarcinoma[J]. Prog Obstet Gynecol, 2020, 29(11): 819-824, 829. DOI: 10.13283/j.cnki.xdfckjz.2020.11.033.
[8]
ZHANG C Y, JIANG H Y, YUAN L, et al. CircVPRBP inhibits nodal metastasis of cervical cancer by impeding RACK1 O-GlcNAcylation and stability[J]. Oncogene, 2023, 42(11): 793-807. DOI: 10.1038/s41388-023-02595-9.
[9]
DU S, SUN H, GAO S, et al. Relationship between 18F-FDG PET metabolic parameters and MRI intravoxel incoherent motion (IVIM) histogram parameters and their correlations with clinicopathological features ofcervical cancer: evidence from integrated PET/MRI[J]. Clin Radiol, 2019, 74(3): 178-186. DOI: 10.1016/j.crad.2018.11.003.
[10]
MORICE P, PIOVESAN P, REY A, et al. Prognostic value of lymphovascular space invasion determined with hematoxylin-eosin staining in early stage cervical carcinoma: results of a multivariate analysis[J]. Ann Oncol, 2003, 14(10): 1511-1517. DOI: 10.1093/annonc/mdg412.
[11]
NICOLET V, CARIGNAN L, BOURDON F, et al. MR imaging of cervical carcinoma: a practical staging approach[J]. Radiographics, 2000, 20(6): 1539-1549. DOI: 10.1148/radiographics.20.6.g00nv111539.
[12]
DEVINE C, VISWANATHAN C, FARIA S, et al. Imaging and staging of cervical cancer[J]. Semin Ultrasound CT MR, 2019, 40(4): 280-286. DOI: 10.1053/j.sult.2019.03.001.
[13]
ROCKALL A G, QURESHI M, PAPADOPOULOU I, et al. Role of imaging in fertility-sparing treatment of gynecologic malignancies[J]. Radiographics, 2016, 36(7): 2214-2233. DOI: 10.1148/rg.2016150254.
[14]
LI S J, CHENG J L, ZHANG Y, et al. Quantitative analysis of T2 mapping in evaluating pathological features of cervical cancer[J]. Chin J Med Imag Technol, 2019, 35(9): 1365-1369. DOI: 10.13929/j.1003-3289.201903113.
[15]
XU X, ZHANG H L, LIU Q P, et al. Radiomic analysis of contrast-enhanced CT predicts microvascular invasion and outcome in hepatocellular carcinoma[J]. J Hepatol, 2019, 70(6): 1133-1144. DOI: 10.1016/j.jhep.2019.02.023.
[16]
LU Y, PENG W W, SONG J, et al. On the potential use of dynamic contrast-enhanced (DCE) MRI parameters as radiomic features of cervical cancer[J]. Med Phys, 2019, 46(11): 5098-5109. DOI: 10.1002/mp.13821.
[17]
SONG J C, HU Q M, MA Z L, et al. Feasibility of T2WI-MRI-based radiomics nomogram for predicting normal-sized pelvic lymph node metastasis in cervical cancer patients[J]. Eur Radiol, 2021, 31(9): 6938-6948. DOI: 10.1007/s00330-021-07735-x.
[18]
MANGANARO L, NICOLINO G M, DOLCIAMI M, et al. Radiomics in cervical and endometrial cancer[J/OL]. Br J Radiol, 2021, 94(1125): 20201314 [2024-02-02]. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9327743/. DOI: 10.1259/bjr.20201314.
[19]
FANG J, ZHANG B, WANG S, et al. Association of MRI-derived radiomic biomarker with disease-free survival in patients with early-stage cervical cancer[J]. Theranostics, 2020, 10(5): 2284-2292. DOI: 10.7150/thno.37429.
[20]
LI Y, QIU L H. Current status and progresses in MRI radiomics of cervical cancer[J]. Chin J Med Imag Technol, 2023, 39(9): 1407-1410. DOI: 10.13929/j.issn.1003-3289.2023.09.029.
[21]
PRANATA Y D, WANG K C, WANG J C, et al. Deep learning and SURF for automated classification and detection of calcaneus fractures in CT images[J]. Comput Methods Programs Biomed, 2019, 171: 27-37. DOI: 10.1016/j.cmpb.2019.02.006.
[22]
LI J, DONG D, FANG M J, et al. Dual-energy CT-based deep learning radiomics can improve lymph node metastasis risk prediction for gastric cancer[J]. Eur Radiol, 2020, 30(4): 2324-2333. DOI: 10.1007/s00330-019-06621-x.
[23]
DEV K, ASHRAF Z, MUHURI P K, et al. Deep autoencoder based domain adaptation for transfer learning[J]. Multimed Tools Appl, 2022, 81(16): 22379-22405. DOI: 10.1007/s11042-022-12226-2.
[24]
JANG J H, KIM T Y, YOON D. Effectiveness of transfer learning for deep learning-based electrocardiogram analysis[J]. Healthc Inform Res, 2021, 27(1): 19-28. DOI: 10.4258/hir.2021.27.1.19.
[25]
ZHAO X Z, QI S L, ZHANG B H, et al. Deep CNN models for pulmonary nodule classification: model modification, model integration, and transfer learning[J]. J Xray Sci Technol, 2019, 27(4): 615-629. DOI: 10.3233/XST-180490.
[26]
DOVROU A, NIKIFORAKI K, ZARIDIS D, et al. A segmentation-based method improving the performance of N4 bias field correction on T2weighted MR imaging data of the prostate[J]. Magn Reson Imaging, 2023, 101: 1-12. DOI: 10.1016/j.mri.2023.03.012.
[27]
PARK S H, LIM H, BAE B K, et al. Robustness of magnetic resonance radiomic features to pixel size resampling and interpolation in patients with cervical cancer[J/OL]. Cancer Imaging, 2021, 21(1): 19 [2024-02-02]. https://www.ncbi.nlm.nih.gov/pmc/articles/pmid/33531073/. DOI: 10.1186/s40644-021-00388-5.
[28]
ROMERO M, INTERIAN Y, SOLBERG T, et al. Targeted transfer learning to improve performance in small medical physics datasets[J]. Med Phys, 2020, 47(12): 6246-6256. DOI: 10.1002/mp.14507.
[29]
LIANG J, JIANG W P. A ResNet50-DPA model for tomato leaf disease identification[J/OL]. Front Plant Sci, 2023, 14: 1258658 [2024-02-02]. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC10614023/. DOI: 10.3389/fpls.2023.1258658.
[30]
HOSSAIN M B, HASAN SAZZAD IQBAL S M, ISLAM M M, et al. Transfer learning with fine-tuned deep CNN ResNet50 model for classifying COVID-19 from chest X-ray images[J/OL]. Inform Med Unlocked, 2022, 30: 100916 [2024-02-02]. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8933872/. DOI: 10.1016/j.imu.2022.100916.
[31]
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.
[32]
LI Z C, LI H L, WANG S Y, et al. MR-based radiomics nomogram of cervical cancer in prediction of the lymph-vascular space invasion preoperatively[J]. J Magn Reson Imaging, 2019, 49(5): 1420-1426. DOI: 10.1002/jmri.26531.
[33]
WU Q X, SHI D P, DOU S W, et al. Radiomics analysis of multiparametric MRI evaluates the pathological features of cervical squamous cell carcinoma[J]. J Magn Reson Imaging, 2019, 49(4): 1141-1148. DOI: 10.1002/jmri.26301.
[34]
YU H Q, ZHAI J, LIU C L, et al. Radiomics based on MRI for predicting cervical cancer lymph-vascular space invasion[J]. Chin J Med Imag Technol, 2022, 38(3): 421-426. DOI: 10.13929/j.issn.1003-3289.2022.03.023.
[35]
HUA W Q, XIAO T H, JIANG X R, et al. Lymph-vascular space invasion prediction in cervical cancer: exploring radiomics and deep learning multilevel features of tumor and peritumor tissue on multiparametric MRI[J/OL]. Biomed Signal Process Contr, 2020, 58: 101869 [2024-02-02]. https://www.sciencedirect.com/science/article/abs/pii/S1746809420300252. DOI: 10.1016/j.bspc.2020.101869.
[36]
FIELDS E C, WEISS E. A practical review of magnetic resonance imaging for the evaluation and management of cervical cancer[J/OL]. Radiat Oncol, 2016, 11: 15 [2024-02-02]. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4736634/. DOI: 10.1186/s13014-016-0591-0.
[37]
JIANG X R, LI J X, KAN Y Y, et al. MRI based radiomics approach with deep learning for prediction of vessel invasion in early-stage cervical cancer[J]. IEEE/ACM Trans Comput Biol Bioinform, 2021, 18(3): 995-1002. DOI: 10.1109/TCBB.2019.2963867.
[38]
ZHANG X P, ZHANG Y C, ZHANG G J, et al. Deep learning with radiomics for disease diagnosis and treatment: challenges and potential[J]. Front Oncol, 2022, 12: 773840 [2024-02-02]. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8891653/. DOI: 10.3389/fonc.2022.773840.

PREV Multiparametric MRI, MR-HSG and clinical characteristics of uterine factor infertility
NEXT 3.0 T multifrequency magnetic resonance elastography of the kidney: Regional variation and physiological effects on renal stiffness
  


LINK


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