Efficacy of arterial spin labeling magnetic resonance perfusion in differentiating recurrent or residual high-grade glioma from post-treatment changes: A retrospective hospital-based study

Background: Differentiating true recurrence or residual disease from treatment-related changes in patients treated for high-grade glioma (HGG) remains a major diagnostic challenge, because both produce contrast- enhancing lesions on conventional magnetic resonance imaging (MRI). Arterial spin labeling (ASL) is a non- contrast perfusion technique that quantifies cerebral blood flow using magnetically labeled arterial blood water as an endogenous tracer. We evaluated the efficacy of ASL magnetic resonance (MR) perfusion in distinguishing recurrent or residual HGG from post-treatment changes. Methods: This retrospective hospital-based study analysed 71 patients (>18 years) with histopathologically proven GRADE 3 and 4 GLIOMA who had received post-operative radiotherapy and/or chemotherapy and presented with a focal enhancing lesion. All had undergone 3.0 T MRI, including three-dimensional pseudo- continuous ASL. Relative cerebral blood flow (rCBF) and qualitative perfusion pattern (red/hyperperfusion vs. blue/hypoperfusion) were correlated against a reference standard of histopathology ( n = 22) or a predefined composite clinico-radiological standard based on ≥6 months of serial follow-up imaging ( n = 49); tumors were classified per the World Health Organization 2021 central nervous system classification. Diagnostic performance and a receiver operating characteristic (ROC) cutoff were derived. Results: Recurrence/residual disease was confirmed in 21 patients (29.6%) and post-treatment change in 50 (70.4%). Mean rCBF was significantly higher in recurrence (2.33 ± 0.602) than in post-treatment change (1.058 ± 0.302; Mann–Whitney U, p < 0.0001). The red perfusion pattern was seen in 20 of 21 recurrences and only 1 of 50 post-treatment cases ( p < 0.0001). ASL achieved sensitivity 95.0%, specificity 98.0%, positive predictive value 95.0%, negative predictive value 98.0% and accuracy 97.1%, with almost-perfect agreement with the reference standard (κ = 0.93). ROC analysis yielded an area under the curve of 0.911 with an optimal rCBF cutoff of 2.04 (sensitivity 90.0%, specificity 98.0%). Diagnostic performance was comparable in the histopathology-confirmed subgroup (accuracy 95.5%) and the follow-up-confirmed subgroup (accuracy 98.0%). Conclusion: ASL MR perfusion is a reliable, reproducible and contrast-free technique with excellent accuracy for differentiating recurrent or residual HGG from post-treatment changes, and is well suited to longitudinal surveillance where repeated gadolinium administration is undesirable.