Abstract
Through a systematic review, the research quantitatively evaluates the efficiency of arterial spin labelling (ASL) in identifying glioma grades. EMBASE and MEDLINE were consulted, and 18 studies were selected. In turn, quantitative data were gathered, and a meta-analysis conducted. 8 included studies published CBF values as a mean and SD; 3 published cut-off values and sensitivity/specificity levels; while the remainder addressed both. Tumour blood flow (TBF) references were different, and so were renamed as follows: the TBF was denoted the mean (TBFmean/rTBFmean); the ROI incorporated the whole tumour; or the ROI was chosen with standard images (e.g., T2W). The identified ROI for the perfusion map’s high signal was denoted the maximum (TBFmax/rTBFmax). QUADAS-2 was used for quality appraisal, while statistical analysis was divided as follows: firstly, a random-effects model and a forest plot; and secondly, a system modelling specificity and sensitivity outcomes (owing to the inverse relationship that links them), paired with a hierarchical summary receiver operating characteristic (HROC) curve. The absolute TBF displayed the power to distinguish between HGG and LGG, along with grade-II from grade-IV. Nevertheless, it could not distinguish between grade-II and grade-III or grade-III and grade-IV. Contrastingly, rTBFmax was more effective in glioma grading. An identical outcome was derived from sensitivity and specificity analysis, with rTBFmax showing the highest levels for glioma grading. Estimated effect size for rTBF was compatible to HGG and LGG, as well as grade-II and grade-II ((-1.46, (-2.00, -0.91)), (-1.39, (-1.89, -0.89)), respectively; however, between grade-III and grade-IV, the effect size was reduced (-1.05, (-1.82, -0.27)). This result also derived from the sensitivity and specificity analysis, where ASL demonstrated greater sensitivity when distinguishing between grade-II and grade-III than between grade-III and grade-IV. It follows that ALS is effective for glioma grading when perfusion values show significant differences between glioma grades, especially rTBFmax.