Fully automated brain resection cavity delineation for radiation target volume definition in glioblastoma patients using deep learning

Ermiş, Ekin; Jungo, Alain; Poel, Robert; Blatti-Moreno, Marcela; Meier, Raphael; Knecht, Urspeter; Aebersold, Daniel M.; Fix, M.K.; Manser, Peter; Herrmann, Evelyn

doi:10.1186/s13014-020-01553-z

Cited by 50 publications

(55 citation statements)

References 35 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Despite the neuroimaging improvements there is still no definitive consensus for RT treatment volume in GB. Even though several research groups demonstrated high performance of the developed models in terms of Dice metrics (over 0.80) [4], [6], [9]. These results were obtained on the prospective or retrospective data with segmentation masks created under a single guideline.…”

Section: Discussionmentioning

confidence: 95%

“…Automatic segmentation of brain tumours, especially gliomas is of great research interest. Many methods for glioma segmentation were developed under the competitions like Brain Tumor Segmentation Challenge (BraTS) and on the unified prospective datasets [4], [5], some of them even achieving beyond human-level performance [6]. However, a large amount of retrospective data, for instance, stored in radiation treatment planning systems remains unused.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Challenges in Building of Deep Learning Models for Glioblastoma Segmentation: Evidence from Clinical Data

Kurmukov

Dalechina

Saparov

et al. 2021

Studies in Health Technology and Informatics

View full text Add to dashboard Cite

In this article, we compare the performance of a state-of-the-art segmentation network (UNet) on two different glioblastoma (GB) segmentation datasets. Our experiments show that the same training procedure yields almost twice as bad results on the retrospective clinical data compared to the BraTS challenge data (in terms of Dice score). We discuss possible reasons for such an outcome, including inter-rater variability and high variability in magnetic resonance imaging (MRI) scanners and scanner settings. The high performance of segmentation models, demonstrated on preselected imaging data, does not bring the community closer to using these algorithms in clinical settings. We believe that a clinically applicable deep learning architecture requires a shift from unified datasets to heterogeneous data.

show abstract

Section: Discussionmentioning

confidence: 95%

Section: Introductionmentioning

confidence: 99%

Challenges in Building of Deep Learning Models for Glioblastoma Segmentation: Evidence from Clinical Data

Kurmukov

Dalechina

Saparov

et al. 2021

Studies in Health Technology and Informatics

View full text Add to dashboard Cite

show abstract

“…First, we manually delineated all of the slices of the lesion, which was timeconsuming. A CT-based semi-automatic segmentation method was recently used for radiomics analysis of lung tumors [40] and a fully automatic segmentation approach using MRI has been performed for brain cancer [41]. A reliable and stable automatic segmentation method needs to be developed for LHSCC in the future so as to greatly reduce the burden of researchers.…”

Section: Discussionmentioning

confidence: 99%

CT-based radiomics features in the prediction of thyroid cartilage invasion from laryngeal and hypopharyngeal squamous cell carcinoma

Guo

Zhang

et al. 2020

Cancer Imaging

View full text Add to dashboard Cite

Background Laryngeal and hypopharyngeal squamous cell carcinoma (LHSCC) with thyroid cartilage invasion are considered T4 and need total laryngectomy. However, the accuracy of preoperative diagnosis of thyroid cartilage invasion remains lower. Therefore, the purpose of this study was to assess the potential of computed tomography (CT)-based radiomics features in the prediction of thyroid cartilage invasion from LHSCC. Methods A total of 265 patients with pathologically proven LHSCC were enrolled in this retrospective study (86 with thyroid cartilage invasion and 179 without invasion). Two head and neck radiologists evaluated the thyroid cartilage invasion on CT images. Radiomics features were extracted from venous phase contrast-enhanced CT images. The least absolute shrinkage and selection operator (LASSO) and logistic regression (LR) method were used for dimension reduction and model construction. In addition, the support vector machine-based synthetic minority oversampling (SVMSMOTE) algorithm was adopted to balance the dataset and a new LR-SVMSMOTE model was constructed. The performance of the radiologist and the two models were evaluated with receiver operating characteristic (ROC) curves and compared using the DeLong test. Results The areas under the ROC curves (AUCs) in the prediction of thyroid cartilage invasion from LHSCC for the LR-SVMSMOTE model, LR model, and radiologist were 0.905 [95% confidence interval (CI): 0.863 to 0.937)], 0.876 (95%CI: 0.830 to 0.913), and 0.721 (95%CI: 0.663–0.774), respectively. The AUCs of both models were higher than that of the radiologist assessment (all P < 0.001). There was no significant difference in predictive performance between the LR-SVMSMOTE and LR models (P = 0.05). Conclusions Models based on CT radiomic features can improve the accuracy of predicting thyroid cartilage invasion from LHSCC and provide a new potentially noninvasive method for preoperative prediction of thyroid cartilage invasion from LHSCC.

show abstract

“…Furthermore, deep learning algorithms are also frequently used in radiotherapeutic research for automated skull stripping, automated segmentation, or delineation of resection cavities for stereotactic radiosurgery [57][58][59][60]. Despite the ubiquity of highperforming models in clinical research, none has been translated A probability curve of true-positive rates against the false-positive rates at different cutoff points in outcome [1].…”

Section: Discussionmentioning

confidence: 99%

Survival prediction of glioblastoma patients—are we there yet? A systematic review of prognostic modeling for glioblastoma and its clinical potential

et al. 2020

View full text Add to dashboard Cite

Glioblastoma is associated with a poor prognosis. Even though survival statistics are well-described at the population level, it remains challenging to predict the prognosis of an individual patient despite the increasing number of prognostic models. The aim of this study is to systematically review the literature on prognostic modeling in glioblastoma patients. A systematic literature search was performed to identify all relevant studies that developed a prognostic model for predicting overall survival in glioblastoma patients following the PRISMA guidelines. Participants, type of input, algorithm type, validation, and testing procedures were reviewed per prognostic model. Among 595 citations, 27 studies were included for qualitative review. The included studies developed and evaluated a total of 59 models, of which only seven were externally validated in a different patient cohort. The predictive performance among these studies varied widely according to the AUC (0.58–0.98), accuracy (0.69–0.98), and C-index (0.66–0.70). Three studies deployed their model as an online prediction tool, all of which were based on a statistical algorithm. The increasing performance of survival prediction models will aid personalized clinical decision-making in glioblastoma patients. The scientific realm is gravitating towards the use of machine learning models developed on high-dimensional data, often with promising results. However, none of these models has been implemented into clinical care. To facilitate the clinical implementation of high-performing survival prediction models, future efforts should focus on harmonizing data acquisition methods, improving model interpretability, and externally validating these models in multicentered, prospective fashion.

show abstract

Fully automated brain resection cavity delineation for radiation target volume definition in glioblastoma patients using deep learning

Cited by 50 publications

References 35 publications

Challenges in Building of Deep Learning Models for Glioblastoma Segmentation: Evidence from Clinical Data

Challenges in Building of Deep Learning Models for Glioblastoma Segmentation: Evidence from Clinical Data

CT-based radiomics features in the prediction of thyroid cartilage invasion from laryngeal and hypopharyngeal squamous cell carcinoma

Survival prediction of glioblastoma patients—are we there yet? A systematic review of prognostic modeling for glioblastoma and its clinical potential

Contact Info

Product

Resources

About