Doppler Ultrasound Flow Reversal in the Superior Sagittal Sinus to Detect Cerebral Venous Congestion in Vein of Galen Malformation

VGAM consists of high-flow arteriovenous connections between the persistent median prosencephalic vein of Markowski and cerebral arterial feeders, causing intracerebral shunting, loss of arterial resistance, and risk of high-output cardiac failure. Despite advances, neonatal cases carry high mortality and poor neurodevelopmental outcomes. Beyond impaired arterial perfusion, elevated cerebral venous pressure is central to progressive parenchymal brain injury via venous arterialization, outflow obstruction, congestion, venous ischemia, and hydrocephalus malresorptivus. Venous hypertension may be worsened by high shunt volumes, right heart failure, and jugular bulb stenosis from flow-related vessel wall maturation disorders. Current noninvasive tools (cMRI) only show indirect signs or established parenchymal damage and are not well suited for serial monitoring. The study aimed to evaluate whether serial venous Doppler ultrasound of the SSS and cortical veins can detect and track increased cerebral venous pressure in VGAM.

Prior literature emphasizes the role of elevated venous pressure in VGAM-related brain injury, with invasive venous pressures in the Galen/straight sinus complex often markedly above normal and associated with congestion-related sequelae. cMRI can demonstrate indirect and direct signs of venous congestion (congested veins, edema, microhemorrhages, calcifications, atrophy) but typically reflects late changes. Jugular bulb narrowing and reduced SSS diameter have been linked to poor outcomes, and case reports describe endovascular management of venous stenoses. Doppler characterization of neonatal cerebral venous flow has been described in normal and pathologic states, including retrograde flow associated with elevated right atrial pressure and intraventricular hemorrhage in preterms, and Ikeda et al. proposed venous flow waveform classifications. However, a practical, noninvasive bedside approach for early detection and serial monitoring of venous congestion in VGAM has been lacking.

Design: Retrospective, single-center study including all neonates with VGAM <28 days old admitted to the tertiary NICU at Sana Hospital Duisburg from January to June 2022. Ethical approval was obtained (University Hospital Essen, 22-10801-BO). Clinical severity was assessed using the Bicêtre Neonatal Evaluation Score (BNES) before the first cMRI. Ultrasound protocol: Serial Doppler assessments were analyzed at predefined timepoints: T1 (before first cMRI/embolization), T2 (1 day after first cMRI/embolization), T3 (5–7 days after embolization), and T4 (before discharge). Additional post-discharge examinations were analyzed when available. Equipment included GE LOGIQ S8/E10s R3 with high-resolution linear probes (ML5-15-D/ML4-20-D). SSS Doppler was performed in a median sagittal plane, with angle-corrected pulsed Doppler sampling a posterior SSS segment before any VGAM drainage junctions, ensuring the sample volume encompassed the full vessel diameter and optimizing Doppler parameters. Cortical vein Doppler used similar technique. Flow profiles (FPs) were evaluated only when signal quality was sufficient for ≥5 seconds with correct parameters. Flow profile classification: Following Ikeda et al., venous FPs were categorized: FP1 (low undulation; Vmin ≥ 0.5 Vmax), FP2 (increased undulation; Vmin < 0.5 Vmax), FP3 (undulation with Vmin to 0 cm/s or brief retrograde), FP4 (pulsatile retrograde component), FP5 (high retrograde; ≥ one-third of total flow retrograde), FP6 (complete retrograde). MR imaging: First cMRI performed under the same anesthesia as the initial endovascular procedure; a blinded neuroradiologist used a structured report and an established scoring system to classify parenchymal damage (none, mild, moderate, severe), focusing on atrophy, medullary lesions, ischemic/hemorrhagic infarcts, hemorrhages, and venous congestion-related injury. Follow-up cMRI occurred at 1–10 months. Endovascular therapy: Routine combined transarterial and transvenous approaches were used to precisely occlude high-flow AV shunts at the shunting point using coils and/or ethylene-vinyl alcohol. Techniques included superselective arterial probing and retrograde transvenous access with looping or “kissing microcatheter” techniques to target the entry point at the dilated persistent MPV. Statistics: Categorical variables as counts/percentages; continuous as median (range). Spearman correlation assessed association between SSS FP category and BNES. Analyses performed with SAS Enterprise Guide 8.3.

- Cohort: 7 neonates with VGAM; median gestational age 38 1/7 weeks (range 34 2/7–41 6/7); median birth weight 3380 g (2220–4250 g). Median age at admission 2 days (1–13); median BNES 18 (7–21). VGAM types: choroidal 3/7 (42.9%), mural 3/7 (42.9%), mixed 1/7 (14.3%). One palliative patient died in week 2; 6 underwent endovascular therapy at median 4.5 days (1–14) and were discharged; one had intraventricular hemorrhage with posthemorrhagic hydrocephalus. - Ultrasound dataset: 44 SSS Doppler examinations and 36 cortical vein examinations from 7 patients (overall). During initial therapy series, 26 SSS and 25 cortical vein exams were analyzed. - Pre-intervention correlation: SSS Doppler FP severity correlated strongly with disease severity (BNES) before therapy: Spearman r = −0.97 (95% CI −0.995 to −0.787), P < .0001. Lower BNES (worse condition) associated with more pathologic FPs. - Flow reversal prevalence: At T1, 4/7 (57.1%) patients had a retrograde flow component in the SSS. After embolization (T2), no patient had retrograde SSS flow. At T3–T4, one patient (patient 5) developed increasing retrograde flow progressing to complete retrograde flow in SSS and cortical veins despite prior embolization. - cMRI correlation: Only patients with high retrograde SSS flow (FP5–FP6; ≥ one-third retrograde or complete retrograde; n = 2) showed severe venous congestion damage on cMRI. Others without high-grade retrograde SSS flow lacked severe congestion-related damage. - Longitudinal case (patient 5): Developed arterialized retrograde flow in SSS and cortical veins; bilateral sinu-jugular stenosis diagnosed. Endovascular balloon dilations and left-sided stent implantation progressively normalized Doppler flow direction and reduced arterialization in real time. At 9 months, recurrent arterialized retrograde flow and worsening cMRI congestion damage occurred without jugular restenosis; after >90% occlusion of VGAM, SSS FP normalized.

Serial Doppler assessment of superficial cerebral venous flow in VGAM provides a bedside, noninvasive window into venous hemodynamics. Retrograde components in the SSS reflect elevated venous pressures and/or mechanical outflow obstruction and correlate with clinical severity (BNES) and with cMRI evidence of venous congestion–related brain injury, particularly when retrograde flow constitutes a large proportion of total flow (FP5–FP6). After effective shunt reduction via embolization, normalization of venous FPs supports pressure relief in the cerebral venous system. Conversely, persistent or progressive retrograde, arterialized waveforms prompted identification of downstream venous outflow obstruction (bilateral sinu-jugular stenosis) and guided targeted endovascular correction, with immediate Doppler improvement during intervention. Potential mechanisms for retrograde venous Doppler signals include transmission of elevated right atrial pressure pulsations (A and V waves) to intracranial veins in the context of high central venous pressure and right heart failure, and arterialized pressure within the confluence and straight sinus complex in VGAM. Obstructive venous lesions can initially increase prestenotic pulsatile reflux and, when severe, reverse flow within the SSS and cortical veins, contributing to venous congestion and parenchymal injury if collateral outflow is insufficient. These findings underscore Doppler venous waveforms as an adjunct to guide timing and evaluation of endovascular and supportive therapies, complementing cMRI, which captures later structural consequences and is less feasible for frequent monitoring.

Flow reversal in the superior sagittal sinus measured with spectral Doppler ultrasound is a promising, noninvasive parameter to assess cerebral venous congestion in VGAM. High-grade retrograde SSS flow correlates with disease severity and with venous congestion–related parenchymal injury on cMRI, while normalization follows effective shunt reduction or relief of venous outflow obstruction. Integrating cerebral venous Doppler waveform assessment into routine VGAM care may improve monitoring and therapeutic decision-making. Future prospective studies should validate these findings, refine waveform interpretation, and explore applications during intensive care, interventional management, and prenatal assessment.

- Small sample size inherent to the rarity of VGAM and retrospective, single-center design limit generalizability and advanced statistical analyses. - cMRI, used as the reference for venous congestion–related injury, detects indirect signs and parenchymal damage with a time lag from hemodynamic changes. - Cerebral venous Doppler waveforms are influenced by central venous pressure and cardiocirculatory status (volume, inotropes/vasodilators/sedation, ductus arteriosus status), potentially confounding serial comparisons. - Ultrasound does not provide direct venous pressure measurements and reflects mainly superficial venous outflow; deep venous system hemodynamics cannot be assessed with this method. - Collateral venous pathways can cause segmental retrograde flow and may persist even after shunt closure, complicating interpretation.