P-035 Effect of pre-operative embolization on AVM blood flow

Introduction Both surgical resection and endovascular embolization are important tools in the treatment of arteriovenous malformations (AVMs). Prior to surgical resection of AVMs, pre-operative embolization is often selectively employed for large or high-flow AVMs, to reduce risk of intra-operative hemorrhage and mitigate the potential for bleeding related to “normal perfusion breakthrough” in the peri-operative period. A primary goal of the embolization is to sequentially and gradually reduce flow in the AVM; the success in achieving this goal is primarily determined through subjective and qualitative assessment of the angiographic runs pre and post embolization. We sought to directly measure the effects of embolization on AVM flow using Quantitative Magnetic Resonance Angiography (QMRA). Methods Medical charts, cerebral angiograms, and flow measurements were reviewed in patients with QMRA measurements of AVM flow pre and post embolization. Total AVM flow was calculated based on the aggregate flow within the primary arterial feeders, or flow measured from single draining veins. AVM flows were compared pre and post embolization. All sessions of embolization were performed with N-BCA glue except three sessions where Onyx was used. Results 51 AVM patients were evaluated with QMRA. Among those, 17 patients had flow measurements pre and post embolization. A total of 42 sessions of embolization were performed (range 1–6 per patient). Average AVM flow prior to treatment was 439 cc/min, ranging from 129 to 952 cc/min in AVMs from 2.3–32 cc in size (median 6.4 cc). The median number of pedicles embolized per session was 3 (range 1–5). Flow dropped by a mean of 133 cc/min (range 14–410 cc/min) per each session of embolization, representing an average 40% flow decline per session relative to flows immediately prior to the embolization. In larger AVMs, overall flow changes were sometimes not evident until the second or third sessions of embolization, suggesting re-channeling of flow through remaining compartments of the nidus. As compared to the baseline, there was an overall 67% (range 27%–99%) reduction in AVM flow at the end of multiple sessions of embolization, with mean AVM flow dropping to 103 cc/min (range 5–300 cc/min) before the surgery. Interestingly, the number of pedicles embolized per session did not correlate with the flow drop, which also correlated poorly with subjective determinations of the extent of embolization at time of angiography. Conclusion Flow alterations induced by embolization can be quantified, and may provide a more robust strategy to determine the number and efficacy of preoperative embolization sessions. Competing interests A Alaraj: ev3, Cordis-Codman. V Aletich: Micrus. ev3, Cordis-Codman. A Ivanov: None. A Carlson: None. G Oh: None. F Charbel: VasSol Inc. S Amin-Hanjani: None.