Model incorporating multiple diffusion MRI features: development and validation of a radiomics-based model to predict adult-type diffuse gliomas grade

European Radiology
Peng Wang1, Shaolei Xie2, Qiong Wu2, Lixin Weng1, Zhiyue Hao2, Pengxuan Yuan2, Chi Zhang2, Wei Gao2, Shaoyu Wang3, Huapeng Zhang3, Yang Song3, Jinlong He2, Yang Gao2
1Inner Mongolia Medical University, Hohhot, China
2Department of Radiology, Affiliated Hospital of Inner Mongolia Medical University, Hohhot, China
3MR Scientific Marketing, Siemens Healthineers, Shanghai, China

Tóm tắt

Abstract Objectives To develop and validate a radiomics-based model (ADGGIP) for predicting adult-type diffuse gliomas (ADG) grade by combining multiple diffusion modalities and clinical and imaging morphologic features. Methods In this prospective study, we recruited 103 participants diagnosed with ADG and collected their preoperative conventional MRI and multiple diffusion imaging (diffusion tensor imaging, diffusion kurtosis imaging, neurite orientation dispersion and density imaging, and mean apparent propagator diffusion-MRI) data in our hospital, as well as clinical information. Radiomic features of the diffusion images and clinical information and morphological data from the radiological reports were extracted, and multiple pipelines were used to construct the optimal model. Model validation was performed through a time-independent validation cohort. ROC curves were used to evaluate model performance. The clinical benefit was determined by decision curve analysis. Results From June 2018 to May 2021, 72 participants were recruited for the training cohort. Between June 2021 and February 2022, 31 participants were enrolled in the prospective validation cohort. In the training cohort (AUC 0.958), internal validation cohort (0.942), and prospective validation cohort (0.880), ADGGIP had good accuracy in predicting ADG grade. ADGGIP was also significantly better than the single-modality prediction model (AUC 0.860) and clinical imaging morphology model (0.841) (all p < .01) in the prospective validation cohort. When the threshold probability was greater than 5%, ADGGIP provided the greatest net benefit. Conclusion ADGGIP, which is based on advanced diffusion modalities, can predict the grade of ADG with high accuracy and robustness and can help improve clinical decision-making. Clinical relevance statement Integrated multi-modal predictive modeling is beneficial for early detection and treatment planning of adult-type diffuse gliomas, as well as for investigating the genuine clinical significance of biomarkers. Key Points • Integrated model exhibits the highest performance and stability. • When the threshold is greater than 5%, the integrated model has the greatest net benefit. • The advanced diffusion models do not demonstrate better performance than the simple technology.

Từ khóa


Tài liệu tham khảo

GBD 2016 Brain and Other CNS Cancer Collaborators (2019) Global, regional, and national burden of brain and other CNS cancer, 1990-2016: a systematic analysis for the Global Burden of Disease Study 2016. Lancet Neurol 18(4):376–393. https://doi.org/10.1016/S1474-4422(18)30468-X

Jiang T, Nam DH, Ram Z et al (2021) Clinical practice guidelines for the management of adult diffuse gliomas. Cancer Lett 499:60–72. https://doi.org/10.1016/j.canlet.2020.10.050

Louis DN, Perry A, Wesseling P et al (2021) The 2021 WHO classification of tumors of the central nervous system: a summary. Neuro Oncol 23(8):1231–1251. https://doi.org/10.1093/neuonc/noab106

Chuntova P, Chow F, Watchmaker PB et al (2021) Unique challenges for glioblastoma immunotherapy-discussions across neuro-oncology and non-neuro-oncology experts in cancer immunology. Meeting report from the 2019 SNO immuno-oncology think tank. Neuro Oncol 23(3):356–375. https://doi.org/10.1093/neuonc/noaa277

Smits M (2021) MRI biomarkers in neuro-oncology. Nat Rev Neurol 17(8):486–500. https://doi.org/10.1038/s41582-021-00510-y

Wang P, He J, Ma X et al (2022) Applying MAP-MRI to identify the WHO grade and main genetic features of adult-type diffuse gliomas: a comparison of three diffusion-weighted MRI models. Acad Radiol 30(7):1238–1246. https://doi.org/10.1016/j.acra.2022.10.009

Gao A, Zhang H, Yan X et al (2022) Whole-tumor histogram analysis of multiple diffusion metrics for glioma genotyping. Radiology 302(3):652–661. https://doi.org/10.1148/radiol.210820

Jiang R, Jiang S, Song S et al (2021) Laplacian-regularized mean apparent propagator-MRI in evaluating corticospinal tract injury in patients with brain glioma. Korean J Radiol 22(5):759–769. https://doi.org/10.3348/kjr.2020.0949

Saleem KS, Avram AV, Glen D et al (2021) High-resolution mapping and digital atlas of subcortical regions in the macaque monkey based on matched MAP-MRI and histology. Neuroimage 245:118759. https://doi.org/10.1016/j.neuroimage.2021.118759

Lambin P, Leijenaar RTH, Deist TM et al (2017) Radiomics: the bridge between medical imaging and personalized medicine. Nat Rev Clin Oncol 14(12):749–762. https://doi.org/10.1038/nrclinonc.2017.141

Limkin EJ, Sun R, Dercle L et al (2017) Promises and challenges for the implementation of computational medical imaging (radiomics) in oncology. Ann Oncol 28(6):1191–1206. https://doi.org/10.1093/annonc/mdx034

Jian A, Jang K, Manuguerra M, Liu S, Magnussen J, Di Ieva A (2021) Machine learning for the prediction of molecular markers in glioma on magnetic resonance imaging: a systematic review and meta-analysis. Neurosurgery 89(1):31–44. https://doi.org/10.1093/neuros/nyab103

He J, Ren J, Niu G et al (2022) Multiparametric MR radiomics in brain glioma: models comparation to predict biomarker status. BMC Med Imaging 22(1):137. https://doi.org/10.1186/s12880-022-00865-8

Halligan S, Menu Y, Mallett S (2021) Why did European Radiology reject my radiomic biomarker paper? How to correctly evaluate imaging biomarkers in a clinical setting. Eur Radiol 31(12):9361–9368. https://doi.org/10.1007/s00330-021-07971-1

Song Y, Zhang J, Zhang YD et al (2020) FeAture Explorer (FAE): A tool for developing and comparing radiomics models. PLoS One 15(8):e0237587. https://doi.org/10.1371/journal.pone.0237587

Calabrese E, Rudie JD, Rauschecker AM et al (2022) Combining radiomics and deep convolutional neural network features from preoperative MRI for predicting clinically relevant genetic biomarkers in glioblastoma. Neurooncol Adv 4(1):vdac060. https://doi.org/10.1093/noajnl/vdac060

Weller M, van den Bent M, Preusser M et al (2021) EANO guidelines on the diagnosis and treatment of diffuse gliomas of adulthood. Nat Rev Clin Oncol 18(3):170–186. https://doi.org/10.1038/s41571-020-00447-z

Bobholz SA, Lowman AK, Barrington A et al (2020) Radiomic features of multiparametric MRI present stable associations with analogous histological features in patients with brain cancer. Tomography 6(2):160–169. https://doi.org/10.18383/j.tom.2019.00029

Ho Y, Pepyne D (2002) Simple explanation of the no-free-lunch theorem and its implications. J Optim Theory Appl 115:549–570. https://doi.org/10.1023/A:1021251113462

Ning L, Laun F, Gur Y et al (2015) Sparse reconstruction challenge for diffusion MRI: validation on a physical phantom to determine which acquisition scheme and analysis method to use? Med Image Anal 26(1):316–331. https://doi.org/10.1016/j.media.2015.10.012

Feng L, Liu Z, Li C et al (2022) Development and validation of a radiopathomics model to predict pathological complete response to neoadjuvant chemoradiotherapy in locally advanced rectal cancer: a multicentre observational study. Lancet Digit Health 4(1):e8–e17. https://doi.org/10.1016/S2589-7500(21)00215-6

Seow P, Hernowo AT, Narayanan V et al (2021) Neural fiber integrity in high- versus low-grade glioma using probabilistic fiber tracking. Acad Radiol 28(12):1721–1732. https://doi.org/10.1016/j.acra.2020.09.007

Barajas RF Jr, Hess CP, Phillips JJ et al (2013) Super-resolution track density imaging of glioblastoma: histopathologic correlation. AJNR Am J Neuroradiol 34(7):1319–1325. https://doi.org/10.3174/ajnr.A3400

Cruz Rivera S, Liu X, Chan AW et al (2020) Guidelines for clinical trial protocols for interventions involving artificial intelligence: the SPIRIT-AI extension. Nat Med 26(9):1351–1363. https://doi.org/10.1038/s41591-020-1037-7

Muftuler LT, Nencka AS, Koch KM (2022) Diffusion propagator metrics are biased when simultaneous multi-slice acceleration is used. Magn Reson Imaging 86:46–54. https://doi.org/10.1016/j.mri.2021.11.003

Chung AW, Seunarine KK, Clark CA (2016) NODDI reproducibility and variability with magnetic field strength: a comparison between 1.5 T and 3 T. Hum Brain Mapp 37(12):4550–4565. https://doi.org/10.1002/hbm.23328

Thông tin
Thông tin xuất bản

European Radiology

Thông tin tác giả