Nomograms for prognostic risk assessment in glioblastoma multiforme: Applications and limitations

Abstract

Glioblastoma multiforme (GBM) is the most common and aggressive form of brain cancer. Prognosis evaluation is of great significance in guiding individualized treatment and monitoring of GBM. By integrating different prognostic variables, nomograms simplify the statistical risk prediction model into numerical estimates for death or recurrence, and are hence widely applied in prognosis prediction. In the past two decades, the application of high-throughput profiling technology and the establishment of TCGA database and other public data deposits have provided opportunities to identify cancer-related molecules and prognostic biomarkers. As a result, both molecular features and clinical characteristics of cancer have been reported to be the key factors in nomogram model construction. This article comprehensively reviewed 35 studies of GBM nomograms, analyzed the present situation of GBM nomograms, and discussed the role and significance of nomograms in personalized risk assessment and clinical treatment decision-making. To facilitate the application of nomograms in the prognostic prediction of GBM patients, a website has been established for the online access of nomograms based on the studies of this review, which is called Consensus Nomogram Spectrum for Glioblastoma (CNSgbm) and is accessible through https://bioinfo.henu.edu.cn/nom/NomList.jsp.

Development and validation of a predictive model in diagnosis and prognosis of primary glioblastoma patients based on Homeobox A family

“…The model's Kaplan–Meier analyses and AUC values were promising in the training cohort and higher in the validation cohort. Compared to previous prediction models [ 47 , 48 ], the performance of our model was acceptable, with the AUC level in the validation group of > 0.75 at 1-year/3-year, which was also similar to the deep-learning and machine-learning models (C-index = 0.68–0.80) [ 48 ]. The prediction performance and net benefits of the model supported its clinical applicability.…”

“…But their net benefits based on these factors were only moderate in CGGA and TCGA groups, as shown in DCA. Previous reviews of nomogram models in GBM also suggested that clinical indicators combined with gene signatures had a better predictive ability [ 47 ]. Building upon these findings, we constructed a nomogram survival prediction model based on novel independent prognostic HOXA genes (HOXA1, HOXA2, HOXA3, HOXA10).…”

Lysyl oxidase-like 1 predicts the prognosis of patients with primary glioblastoma and promotes tumor invasion via EMT pathway

“…The Loxl1-based nomogram exhibited good accuracy, reliability, and clinical benefits in multiple GBM cohorts. Compared to previous GBM prediction models ( Zheng et al, 2022 ) and deep machine learning models ( Poirion et al, 2021 ), the performance of our nomogram and the PRSM was acceptable and practical. In addition, the public databases in our study were built before 2021, and GBMs were defined as IDH1-WT and IDH1-mutant.…”

Development and Validation of CXCR4 Nomogram-Based Immune Infiltration/Tumor Inflammation in Primary Glioblastoma

“…Glioblastoma, which accounts for approximately 15% of all primary brain tumors, is one of the most common and highly lethal central nervous system (CNS) malignancies in adults [ 1 , 2 , 3 ]. The current treatment option for glioblastoma is surgical resection to relieve neurological symptoms, followed by concurrent radiation and adjuvant chemotherapy [ 4 ].…”