A radiomic signature as a non-invasive predictor of progression-free survival in patients with lower-grade gliomas

Abstract

ObjectiveThe aim of this study was to develop a radiomics signature for prediction of progression-free survival (PFS) in lower-grade gliomas and to investigate the genetic background behind the radiomics signature.MethodsIn this retrospective study, training (n = 216) and validation (n = 84) cohorts were collected from the Chinese Glioma Genome Atlas and the Cancer Genome Atlas, respectively. For each patient, a total of 431 radiomics features were extracted from preoperative T2-weighted magnetic resonance images. A radiomics signature was generated in the training cohort, and its prognostic value was evaluated in both the training and validation cohorts. The genetic characteristics of the group with high-risk scores were identified by radiogenomic analysis, and a nomogram was established for prediction of PFS.ResultsThere was a significant association between the radiomics signature (including 9 screened radiomics features) and PFS, which was independent of other clinicopathologic factors in both the training (P < 0.001, multivariable Cox regression) and validation (P = 0.045, multivariable Cox regression) cohorts. Radiogenomic analysis revealed that the radiomics signature was associated with the immune response, programmed cell death, cell proliferation, and vasculature development. A nomogram established using the radiomics signature and clinicopathologic risk factors demonstrated high accuracy and good calibration for prediction of PFS in both the training (C-index, 0.684) and validation (C-index, 0.823) cohorts.ConclusionsPFS can be predicted non-invasively in patients with LGGs by a group of radiomics features that could reflect the biological processes of these tumors.

A systematic review reporting quality of radiomics research in neuro-oncology: toward clinical utility and quality improvement using high-dimensional imaging features

“…Many of studies performed validation using datasets from the same or a different institute (35 out of 51, 68.6%). Six studies earned the full 5 points for validation [21,24,31,36,54,55], using data from three datasets from distinct institutes or public dataset.…”

Integrated regulatory models for inference of subtype‐specific susceptibilities in glioblastoma

“…In addition, log-rank tests of survival of MYBL2-high versus MYBL2low patients in the TCGA dataset (as demarcated by median expression within each subtype) show that MYBL2-low patients show significantly higher survival rates only in the Proneural subtype (log-rank test P = 0.035, Figs 5G and EV4D). We observed a similarly significant trend in an independent set of 172 GBM samples from the Chinese Glioma Genome Atlas (CGGA) project (Wang et al, 2015;Liu et al, 2018), where low MYBL2 levels correspond to higher survival only in the Proneural subtype (log-rank test P = 0.028, Figs 5G and EV4D). Interestingly, our F value correlation analysis showed that MYBL2 displays high correlation with the Proneural signature TFs TFCP2 and MXI1, suggesting that subtypespecific TF interactions may be key to maintaining cell state and viability in the corresponding subtype.…”

“…We therefore performed all of our downstream analysis in this study using the TCGA bulk expression dataset, based on its high gene coverage and sample abundance compared with sparse singlecell data. Additionally, we compared models built from an independent RNA-seq set of 172 GBM samples from the Chinese Glioma Genome Atlas (CGGA) project (Wang et al, 2015;Liu et al, 2018) that were classified using the same method as we did for TCGA samples and found that there is a significant overlap between the regulatory edges inferred by the model on the TCGA and CGGA data. This overlap occurs despite vastly different sample sizes and expression profiling platforms ( Figs 1E and EV1F), suggesting that the regression models are also robust with respect to the size and type of expression data used.…”

MRI-based radiomics model for preoperative prediction of 5-year survival in patients with hepatocellular carcinoma

“…Accumulating evidence also suggests the predictive value of an MRI-based radiomics model in nasopharyngeal carcinoma, breast cancer, glioma and cervical cancer. [29][30][31][32][33][34] Given the widespread application and predictive value of MRI, it is necessary to build an MRI-based prognosis model for treatment guidance in resectable HCC.…”