-
Quantitative Imaging in Medicine and... Aug 2023Positron emission tomography (PET) imaging is a promising molecular neuroimaging technique and has been proposed as one of the criteria for glioma management. However,...
BACKGROUND
Positron emission tomography (PET) imaging is a promising molecular neuroimaging technique and has been proposed as one of the criteria for glioma management. However, there is some controversy concerning the diagnostic accuracy of PET using different radiotracers to differentiate between glioma pseudoprogression (PsP) and true progression (TPR). The purpose of this meta-analysis was to systematically evaluate the methodological quality and clinical value of original studies for distinguishing PsP from TPR in glioma.
METHODS
The Medline, Web of Science, Embase, Cochrane Library, and ClinicalTrials.gov were searched from inception until September 1, 2022. Retrieved clinical studies only investigated the PsP cases but did not include the cases of radiation necrosis or other treatment-related changes. Eligible studies were screened for data extraction and evaluated by 2 independent reviewers using the Quality Assessment of Diagnostic Accuracy Studies 2 (QUADAS-2) tool. A random effects model was used to describe summary receiver operating characteristics. Meta-regression and subgroup analyses were applied to identify any sources of heterogeneity.
RESULTS
The meta-analysis included 20 studies, comprising 317 (30.9%) patients with PsP and 708 (69.1%) with TPR. The summary sensitivity and specificity of general PET for identifying PsP were 0.86 [95% confidence interval (CI): 0.77-0.91] and 0.84 (95% CI: 0.79-0.88), respectively. The statistical heterogeneity was explained by sample size, study design, World Health Organization (WHO) grade, gold standard, and radiotracer type. The summary sensitivity and specificity of O-(2-F-fluoroethyl)-L-tyrosine (F-FET PET) were 0.80 (95% CI: 0.68-0.88) and 0.81 (95% CI: 0.75-0.85), respectively. The maximum tumor-to-brain ratio (TBRmax) and the mean tumor-to-brain ratio (TBRmean) both showed excellent diagnostic performance in F-FET studies, the summary sensitivity was 0.83 (95% CI: 0.72-0.91) and 0.79 (95% CI: 0.65-0.98), respectively, and the specificity was 0.76 (95% CI: 0.68-0.84) and 0.78 (95% CI: 0.64-0.88), respectively.
CONCLUSIONS
PET imaging is generally accurate in identifying glioma PsP. Considering the credibility of meta-evidence and the practicability of using radiotracer, F-FET PET holds the highest clinical value, while TBRmax and TBRmean should be regarded as reliable parameters. PET used with the radiotracers and multiple-parameter combinations of PET with magnetic resonance imaging (MRI) and radiomics analysis have broad research and application prospects, whose diagnostic values for identifying glioma PsP warrant further investigation.
PubMed: 37581048
DOI: 10.21037/qims-22-1340 -
Critical Reviews in Oncogenesis 2024Deep learning (DL) is poised to redefine the way medical images are processed and analyzed. Convolutional neural networks (CNNs), a specific type of DL architecture, are...
Deep learning (DL) is poised to redefine the way medical images are processed and analyzed. Convolutional neural networks (CNNs), a specific type of DL architecture, are exceptional for high-throughput processing, allowing for the effective extraction of relevant diagnostic patterns from large volumes of complex visual data. This technology has garnered substantial interest in the field of neuro-oncology as a promising tool to enhance medical imaging throughput and analysis. A multitude of methods harnessing MRI-based CNNs have been proposed for brain tumor segmentation, classification, and prognosis prediction. They are often applied to gliomas, the most common primary brain cancer, to classify subtypes with the goal of guiding therapy decisions. Additionally, the difficulty of repeating brain biopsies to evaluate treatment response in the setting of often confusing imaging findings provides a unique niche for CNNs to help distinguish the treatment response to gliomas. For example, glioblastoma, the most aggressive type of brain cancer, can grow due to poor treatment response, can appear to grow acutely due to treatment-related inflammation as the tumor dies (pseudo-progression), or falsely appear to be regrowing after treatment as a result of brain damage from radiation (radiation necrosis). CNNs are being applied to separate this diagnostic dilemma. This review provides a detailed synthesis of recent DL methods and applications for intratumor segmentation, glioma classification, and prognosis prediction. Furthermore, this review discusses the future direction of MRI-based CNN in the field of neuro-oncology and challenges in model interpretability, data availability, and computation efficiency.
Topics: Humans; Glioma; Prognosis; Brain Neoplasms; Neural Networks, Computer; Deep Learning; Magnetic Resonance Imaging; Image Processing, Computer-Assisted
PubMed: 38683153
DOI: 10.1615/CritRevOncog.2023050852