Share this content in WeChat
Clinical Article
The value of machine learning model for predicting prostate cancer bone metastases based on MRI radiomics
LI Kejian  ZHANG Juntao  REN Kaixuan  FANG Caiyun  SHANG Hui  JIAO Tianyu  ZENG Qingshi 

Cite this article as: LI K J, ZHANG J T, REN K X, et al. The value of machine learning model for predicting prostate cancer bone metastases based on MRI radiomics[J]. Chin J Magn Reson Imaging, 2023, 14(1): 100-104, 115. DOI:10.12015/issn.1674-8034.2023.01.018.

[Abstract] Objective To develop and validate MRI-based machine learning models for predicting bone metastases in patients with prostate cancer (PCa).Materials and Methods The clinical and MRI data of 150 patients with pathologically confirmed PCa in the First Affiliated Hospital and Affiliated Provincial Hospital of Shandong First Medical University were retrospectively obtained from January 2018 to January 2022. According to the ratio of 7∶3, the samples were randomly divided into training set (n=105) and testing set (n=45). The apparent diffusion coefficient (ADC) and fat saturated T2 weighted imaging (FS-T2WI) of each patient were manually outlined for the tumor's region of interest and extracted for imaging histological features, respectively. Dimension reduction and feature selection were performed using intra-class correlation coefficients (ICC), feature importance and minimal-redundancy-maximal-relevance (mRMR). The filtered features were used to establish radiomics models using generalized linear model (GLM), random forest (RF), support vector machine (SVM) and extreme gradient boosting (XGB) methods. The models were evaluated using the following metrics: area under the curve (AUC) of receiver operating characteristic (ROC), calibration curve, decision curve analysis and Delong test.Results Seventeen features were selected and models were constructed using GLM, XGB, SVM and RF. In the training set, the mean AUC were 0.714, 0.845, 0.768 and 0.858, respectively. In the testing set, the AUC were 0.796, 0.729, 0.755 and 0.765, respectively. Calibration curve and Delong test indicated no significant differences between the four models. Decision curve analysis showed that the four models had similar clinical applications.Conclusions The MRI-based radiomics features allowed GLM, SVM, XGB and RF classifiers to be used as a promising tool for predicting bone metastases in PCa patients, providing potentially valid information for clinical management.
[Keywords] prostate cancer;bone metastasis;magnetic resonance imaging;radiomics;prediction

LI Kejian1, 2   ZHANG Juntao3   REN Kaixuan4   FANG Caiyun1, 2   SHANG Hui1, 2   JIAO Tianyu1, 2   ZENG Qingshi1*  

1 Department of Radiology, Shandong Provincial Qianfoshan Hospital, the First Hospital Affiliated Hospital of Shandong First Medical University, Jinan 250014, China

2 Shandong First Medical University and Shandong Academy of Medical Sciences, Jinan 271016, China

3 GE Healthcare Shanghai Co., Ltd., Shanghai 210000, China

4 Department of Medical Imaging, Shandong Provincial Hospital Affiliated to Shandong First Medical University, Jinan 250022, China

Corresponding author: Zeng QS, E-mail:

Conflicts of interest   None.

Received  2022-08-02
Accepted  2022-12-05
DOI: 10.12015/issn.1674-8034.2023.01.018
Cite this article as: LI K J, ZHANG J T, REN K X, et al. The value of machine learning model for predicting prostate cancer bone metastases based on MRI radiomics[J]. Chin J Magn Reson Imaging, 2023, 14(1): 100-104, 115. DOI:10.12015/issn.1674-8034.2023.01.018.

SUNG H, FERLAY J, SIEGEL R L, et al. Global cancer statistics 2020: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries[J]. CA Cancer J Clin, 2021, 71(3): 209-249. DOI: 10.3322/caac.21660.
HINZPETER R, BAUMANN L, GUGGENBERGER R, et al. Radiomics for detecting prostate cancer bone metastases invisible in CT: a proof-of-concept study[J]. Eur Radiol, 2022, 32(3): 1823-1832. DOI: 10.1007/s00330-021-08245-6.
KFOURY Y, BARYAWNO N, SEVERE N, et al. Human prostate cancer bone metastases have an actionable immunosuppressive microenvironment[J]. Cancer Cell, 2021, 39(11): 1464-1478. DOI: 10.1016/j.ccell.2021.09.005.
LIU C Y, MA C, YIN X R, et al. Risk factors of bone metastasis in patients with newly diagnosed prostate cancer[J]. Eur Rev Med Pharmacol Sci, 2022, 26(2): 391-398. DOI: 10.26355/eurrev_202201_27863.
MCGREGOR B, ZHANG L, GRAY K P, et al. Bone targeted therapy and skeletal related events in the era of enzalutamide and abiraterone acetate for castration resistant prostate cancer with bone metastases[J]. Prostate Cancer Prostatic Dis, 2021, 24(2): 341-348. DOI: 10.1038/s41391-020-00280-6.
TOMBAL B, LECOUVET F. Modern detection of prostate cancer's bone metastasis: is the bone scan era over?[J/OL]. Adv Urol, 2012, 2012: 893193 [2022-06-15]. DOI: 10.1155/2012/893193.
MCARTHUR C, MCLAUGHLIN G, MEDDINGS R N. Changing the referral criteria for bone scan in newly diagnosed prostate cancer patients[J]. Br J Radiol, 2012, 85(1012): 390-394. DOI: 10.1259/bjr/79184355.
GILLIES R J, KINAHAN P E, HRICAK H. Radiomics: images are more than pictures, they are data[J]. Radiology, 2016, 278(2): 563-577. DOI: 10.1148/radiol.2015151169.
LAMBIN P, LEIJENAAR R T H, DEIST T M, et al. Radiomics: the bridge between medical imaging and personalized medicine[J]. Nat Rev Clin Oncol, 2017, 14(12): 749-762. DOI: 10.1038/nrclinonc.2017.141.
LAFATA K J, WANG Y, KONKEL B, et al. Radiomics: a primer on high-throughput image phenotyping[J]. Abdom Radiol (NY), 2022, 47(9): 2986-3002. DOI: 10.1007/s00261-021-03254-x.
MIN X, LI M, DONG D, et al. Multi-parametric MRI-based radiomics signature for discriminating between clinically significant and insignificant prostate cancer: Cross-validation of a machine learning method[J/OL]. Eur J Radiol, 2019, 115: 16-21 [2022-06-15]. DOI: 10.1016/j.ejrad.2019.03.010.
BLEKER J, KWEE T C, DIERCKX R A J O, et al. Multiparametric MRI and auto-fixed volume of interest-based radiomics signature for clinically significant peripheral zone prostate cancer[J]. Eur Radiol, 2020, 30(3): 1313-1324. DOI: 10.1007/s00330-019-06488-y.
QI Y F, ZHANG S T, WEI J W, et al. Multiparametric MRI-based radiomics for prostate cancer screening with PSA in 4-10 ng/mL to reduce unnecessary biopsies[J]. J Magn Reson Imaging, 2020, 51(6): 1890-1899. DOI: 10.1002/jmri.27008.
SUSHENTSEV N, RUNDO L, BLYUSS O, et al. Comparative performance of MRI-derived PRECISE scores and delta-radiomics models for the prediction of prostate cancer progression in patients on active surveillance[J]. Eur Radiol, 2022, 32(1): 680-689. DOI: 10.1007/s00330-021-08151-x.
ZHENG H X, MIAO Q, LIU Y K, et al. Multiparametric MRI-based radiomics model to predict pelvic lymph node invasion for patients with prostate cancer[J]. Eur Radiol, 2022, 32(8): 5688-5699. DOI: 10.1007/s00330-022-08625-6.
ZHAO S Y, LIU X B, Guo Y, et al. Application progress of imaging in prostate cancer[J]. Chin J Med Imaging, 2021, 29(4): 411-414. DOI: 10.3969/j.issn.1005-5185.2021.04.030.
LIU J F, ZENG P E, GUO W, et al. Prediction of high-risk cytogenetic status in multiple myeloma based on magnetic resonance imaging: utility of radiomics and comparison of machine learning methods[J]. J Magn Reson Imaging, 2021, 54(4): 1303-1311. DOI: 10.1002/jmri.27637.
WANG X H, WAN Q, CHEN H J, 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.
WANG Y R, YU B, ZHONG F, et al. MRI-based texture analysis of the primary tumor for pre-treatment prediction of bone metastases in prostate cancer[J]. Magn Reson Imaging, 2019, 60: 76-84. DOI: 10.1016/j.mri.2019.03.007.
DONG H, LU G M. The predictive value of bp-MRI radiomics model in the first diagnosis of bone metastasis for prostate cancer[J]. J Pract Radiol, 2021, 37(6): 968-971. DOI: 10.3969/j.issn.1002-1671.2021.06.023.
ZHANG W, MAO N, WANG Y, et al. A Radiomics nomogram for predicting bone metastasis in newly diagnosed prostate cancer patients[J/OL]. Eur J Radiol, 2020, 128 [2022-06-15]. DOI: 10.1016/j.ejrad.2020.109020.
ARNOLD K F, DAVIES V, de KAMPS M, et al. Reflection on modern methods: generalized linear models for prognosis and intervention-theory, practice and implications for machine learning[J]. Int J Epidemiol, 2021, 49(6): 2074-2082. DOI: 10.1093/ije/dyaa049.
SONG R, CUI Y, REN J, et al. CT-based radiomics analysis in the prediction of response to neoadjuvant chemotherapy in locally advanced gastric cancer: A dual-center study[J]. Radiother Oncol, 2022, 171: 155-163. DOI: 10.1016/j.radonc.2022.04.023.
ERICKSON B J, KORFIATIS P, AKKUS Z, et al. Machine learning for medical imaging[J]. Radiographics, 2017, 37(2): 505-515. DOI: 10.1148/rg.2017160130.
KOO T K, LI M Y. A. A guideline of selecting and reporting intraclass correlation coefficients for reliability research[J]. J Chiropr Med, 2016, 15(2): 155-163. DOI: 10.1016/j.jcm.2016.02.012.
PENG H C, LONG F H, DING C. Feature selection based on mutual information criteria of max-dependency, max-relevance, and Min-redundancy[J]. IEEE Trans Pattern Anal Mach Intell, 2005, 27(8): 1226-1238. DOI: 10.1109/TPAMI.2005.159.
EROGLU Y, YILDIRIM M, CINAR A. Convolutional Neural Networks based classification of breast ultrasonography images by hybrid method with respect to benign, malignant, and normal using mRMR[J/OL]. Comput Biol Med, 2021, 133: 104407 [2022-06-17]. DOI: 10.1016/j.compbiomed.2021.104407.
CHEN X, WANG H R, HUANG K P, et al. CT-based radiomics signature with machine learning predicts MYCN amplification in pediatric abdominal neuroblastoma[J/OL]. Front Oncol, 2021, 11: 687884 [2022-06-17]. DOI: 10.3389/fonc.2021.687884.
JING R, WANG J T, LI J B, et al. A wavelet features derived radiomics nomogram for prediction of malignant and benign early-stage lung nodules[J/OL]. Sci Rep, 2021, 11(1): 22330 [2022-06-19]. DOI: 10.1038/s41598-021-01470-5.
ZHENG Y L, ZHOU D, LIU H, et al. CT-based radiomics analysis of different machine learning models for differentiating benign and malignant parotid tumors[J]. Eur Radiol, 2022, 32(10): 6953-6964. DOI: 10.1007/s00330-022-08830-3.
MITCHELL-HAY R N, AHEARN T S, MURRAY A D, et al. Investigation of the inter- and intrascanner reproducibility and repeatability of radiomics features in T1-weighted brain MRI[J]. J Magn Reson Imaging, 2022, 56(5): 1559-1568. DOI: 10.1002/jmri.28191.
ZHU Y D, YANG L, SHEN H L. Value of the application of CE-MRI radiomics and machine learning in preoperative prediction of sentinel lymph node metastasis in breast cancer[J/OL]. Front Oncol, 2021, 11: 757111 [2022-06-20]. DOI: 10.3389/fonc.2021.757111.
DUAN C F, LIU F, GAO S, et al. Comparison of radiomic models based on different machine learning methods for predicting intracerebral hemorrhage expansion[J]. Clin Neuroradiol, 2022, 32(1): 215-223. DOI: 10.1007/s00062-021-01040-2.
KAYADIBI Y, KOCAK B, UCAR N, et al. Radioproteomics in Breast Cancer: Prediction of Ki-67 Expression With MRI-based Radiomic Models[J/OL]. Acad Radiol, 2022, 1: S116-S125 [2022-06-20]. DOI: 10.1016/j.acra.2021.02.001.

PREV Preoperative prediction of vascular invasion in rectal cancer patients without lymph node metastasis based on multimodal MRI imaging features combined with clinical risk factors
NEXT Multimodal MRI manifestations and pathological basis of peritumoral infiltration of glioma in rat

Tel & Fax: +8610-67113815    E-mail: