ASL Tumor CBF Predicts Early Glioma Progression in Single-Center Analysis

Key Takeaways

• Among diffuse gliomas assessed with preoperative ASL MRI and short-interval follow-up, higher tumor CBF was associated with later increases in enhancing tumor volume.
• Mean tumor CBF was highest in gliomas with increased enhancing tumor volume, intermediate in unchanged cases, and lowest in cases with decreased enhancing tumor volume.
• Tumor CBF remained independently associated with enhancing tumor progression after adjustment, while surrounding and whole-brain perfusion measures were not significantly associated with ET or SNFH change.
• Tumor CBF alone showed good discrimination for early enhancing tumor progression, and adding baseline ET volume improved that discrimination further.

A retrospective single-center analysis of matched University of California San Francisco (UCSF) preoperative and longitudinal post-treatment diffuse glioma datasets included 60 patients with preoperative 3T pseudo-continuous arterial spin labeling (PCASL) MRI and paired follow-up structural MRI. The cohort predominantly included World Health Organization grade 4 glioma cases n=52), and the median interval to follow-up was 56 days (IQR 49-63). Tumor, surrounding, and whole-brain cerebral blood flow (CBF) were quantified from segmented regions of interest, while expert longitudinal segmentations were used to derive changes in enhancing tumor (ET) and surrounding non-enhancing FLAIR-hyperintense (SNFH) volume. ET change was classified as increased, unchanged, or decreased over follow-up, and adjusted models included age, scan interval, and O6-methylguanine-DNA methyltransferase (MGMT) promoter methylation status.

Higher preoperative tumor CBF correlated with greater net ET volume increase, with Spearman r = 0.531 and p < 0.0001. Mean tumor CBF was 147.7 ± 20.2 ml/100 g/min in gliomas with increased ET volume, 124.4 ± 24.1 in unchanged cases, and 115.8 ± 25.4 in decreased cases. Tumor perfusion also exceeded surrounding and whole-brain perfusion at baseline.

In multivariable models, tumor CBF remained associated with ET progression after adjustment for age, scan interval, and MGMT promoter methylation status; the authors report a logistic OR of 2.02 per ml/100 g/min increase in tumor CBF, 95% CI 1.32–3.09, p = 0.001, alongside a significant linear-model coefficient of β = +1.528, 95% CI 0.596 to 2.459, p = 0.002. On receiver operating characteristic analysis, tumor CBF alone yielded AUC 0.805, and adding baseline ET volume improved AUC to 0.858. Surrounding CBF and whole-brain CBF were not significantly associated with ET or SNFH change after adjustment.

Interpretation is bounded by the modest single-center cohort, the retrospective design, and the 52 complete cases available for multivariable regression. Adjuvant treatments varied across patients, and the short follow-up window captured short-interval imaging change rather than longer-term recurrence or survival. The authors also noted that no multiple-testing correction was applied in these exploratory analyses and that fixed physiologic assumptions in ASL quantification can affect absolute tumor CBF estimates.

The authors concluded that preoperative ASL-derived tumor CBF was the tumor-specific perfusion measure carrying prognostic signal for early ET progression in diffuse gliomas, whereas more global perfusion measures did not. They also reported that combining tumor CBF with baseline ET volume produced the highest discrimination observed in this analysis.

Clinician Questions

How strongly did preoperative tumor CBF correlate with later enhancing tumor volume increase in diffuse gliomas?

Higher preoperative ASL-derived tumor CBF correlated with greater net enhancing tumor volume increase in diffuse gliomas on early follow-up imaging after surgery, with Spearman r = 0.531 and p < 0.0001, and the analysis addressed short-interval postoperative change rather than long-term survival.

How did tumor CBF differ across increased, unchanged, and decreased enhancing tumor change groups in glioma follow-up?

In diffuse gliomas, preoperative tumor CBF averaged 147.7 ± 20.2 ml/100 g/min in Increased ET cases, 124.4 ± 24.1 in Unchanged cases, and 115.8 ± 25.4 in Decreased cases.

Did preoperative tumor CBF remain predictive after adjustment for age, scan interval, and MGMT status in gliomas?

Preoperative tumor CBF remained independently associated with enhancing tumor progression in diffuse gliomas after adjustment for age, scan interval, and MGMT promoter methylation status in both multivariable models: linear regression β = +1.528, 95% CI 0.596 to 2.459, p = 0.002 and logistic regression OR = 2.02, 95% CI 1.32–3.09, p = 0.001.

How well did preoperative tumor CBF discriminate early enhancing tumor progression on ROC analysis?

In this diffuse glioma analysis, preoperative tumor CBF alone yielded a 5-fold cross-validated AUC of 0.805 for binary enhancing tumor progression, and adding baseline enhancing tumor volume increased the AUC to 0.858.