Neurocrit Care DOI 10.1007/s12028-014-0071-z
TRANSLATIONAL RESEARCH
Modeling the Pattern of Contrast Extravasation in Acute Intracerebral Hemorrhage Using Dynamic Contrast-Enhanced MR R. Liu • T. J. Huynh • Y. Huang • D. Ramsay K. Hynynen • R. I. Aviv
•
Ó Springer Science+Business Media New York 2014
Abstract Background Contrast extravasation (CE) in spontaneous intracerebral hemorrhage (ICH), coined the spot sign, predicts hematoma expansion (HE) and poor clinical outcome. The dynamic relationship between CE and the mode of ICH growth are poorly understood. We characterized the in vivo pattern and rate of HE using a novel animal model of acute ICH. Methods Basal ganglia ICH was created in 14 Yorkshire swine utilizing a novel MRI integrated model, permitting real-time CE observation using dynamic contrast-enhanced (DCE) MRI. Computerized planimetry measured CE volume at each time point. Spatial vector analysis along three
Electronic supplementary material The online version of this article (doi:10.1007/s12028-014-0071-z) contains supplementary material, which is available to authorized users. R. Liu T. J. Huynh R. I. Aviv (&) Department of Medical Imaging, Sunnybrook Health Sciences Centre, 2075 Bayview Ave., Room AG-31E, Toronto, ON M4N 3M5, Canada e-mail: [email protected] T. J. Huynh K. Hynynen R. I. Aviv University of Toronto, 27 King’s College Circle, Toronto, ON M5S 1A1, Canada Y. Huang K. Hynynen Imaging Research, Sunnybrook Research Institute, Toronto, ON, Canada D. Ramsay Department of Pathology, London Health Sciences Centre, London, ON, Canada K. Hynynen Medical Biophysics, University of Toronto, Toronto, ON, Canada
orthogonal axes determined distance vectors. Maximizing and minimizing the coefficient of determination defined the temporal phases of growth and stability, respectively. CE rate was calculated using a Patlak model. Results Asymmetric growth and variable rates of expansion characterized HE defining three distinct growth phases and patterns. A primary growth phase (duration 160 s; IQR 50–130) demonstrated rapid linear growth (0.04 mm/s IQR 0.01–0.10) accounting for 85 ± 15 % of total HE. The stationary phase demonstrated stability (duration 145 s; IQR 0–655). A secondary growth phase (duration 300; 130–600 s) accounted for 23 ± 8 % of total HE. In the primary and secondary growth phase, asymmetric growth occurred in the anterior–posterior (AP) planes (0.056 mm/ s; p = 0.026 and 0.0112 mm/s; p = 0.03). Monophasic 2 (14 %), biphasic 4 (35 %) (primary followed by secondary growth), and triphasic 8 (56 %) patterns (primary, stationary, and secondary growth phase) were observed. Conclusions A novel model of ICH provides real-time study of the dynamics and rate of CE. This data facilitates the understanding of pattern and rate of ICH formation. Keywords Animal model Spot sign Focused-ultrasound MRI
Introduction Intracerebral hemorrhage (ICH) accounts for 10–30 % of stroke associated with significant morbidity and increased risk of death [1]. Hematoma expansion is a strong predictor of poor outcome and is associated with mortality rates over 50 % [2]. Developing effective therapies to reduce clinical outcome requires an understanding of the pathophysiological basis of ICH formation and growth. The mechanism
123
Neurocrit Care
of ICH growth is unknown but is contributed to by continued [3, 4] or delayed bleeding [5] of the initially ruptured vessel. Expansion occurring via secondary shearing of blood vessels surrounding a hematoma may be complicit [6]. Secondary vessel injury may trigger a ‘‘domino’’ pattern of hematoma expansion, recently reproduced on a computer simulation study [7]. Hematoma location, volume, and expansion are strong predictors of poor post-ICH outcome [2]. Since location and initial volume are determined at presentation, prevention, or attention of hematoma expansion remains a promising therapeutic target [8]. Clinical studies have demonstrated biological activity of prothrombotic agents such as recombinant Factor VIIa (rFVIIa), but have yet to yield improvement on the functional outcome. Increased effort to identify patients most likely to benefit from intervention is the focus of ongoing clinical studies dichotomizing patients by spot sign or contrast extravasation presence; an established marker of hematoma growth [9, 10]. The spot sign is a morphological marker of contrast extravasation but is associated with significant differences in rate of leakage [11]. A better understanding of hematoma expansion dynamics including both the rate and pattern of expansion is essential. Such knowledge will potentially inform future studies evaluating therapeutic options for prevention or attenuation of hematoma expansion. Only 20–30 % of acute ICH patients demonstrate a spot sign on first pass CTA studies. Few centers routinely measure contrast extravasation rate with CT permeability [11] or dynamic techniques [12, 13]. We have previously created and reported a novel real-time MRI porcine ICH model of acute hematoma growth and contrast extravasation to bridge this gap [14]. The method of ICH induction closely resembles vessel disruption seen in human ICH and produces reliable sized hematomas comparable to that seen in clinical practice. In this paper, we investigate the spatial and temporal characteristics of ICH-associated contrast extravasation and provide insights into the pattern and pace of ICH expansion.
Methods Experimental Design This study was conducted with the approval of the Sunnybrook Hospital Research Institute Animal Care Committee (Animal Use Protocol #12-435) in compliance with the guidelines established by the Canadian Council on Animal Care and the Animals for Research Act of Ontario, Canada. The ICH model methodology is previously described. Briefly Yorkshire male swine were chemically immobilized and maintained under general anesthesia with
123
2–3 % isoflurane in oxygen (>2 L/min). Physiological monitoring at 5 min intervals included clinical observation, tail pulse oximetry, rectal temperature and blood pressure measurement through the femoral arterial line. The animal was placed in a supine position with the head immersed in degassed water serving as a conduction medium for ultrasound emitted from a upwardly-positioned modified clinical MR-Guided Focused Ultrasound (MRgFUS) brain system (ExAblate 4000, Insightec, Tirat Carmel, Israel) comprising a 1024-element hemispherical phased-array with a diameter of 30 cm. Accurate anatomical localization of vessels within the basal ganglia was achieved using MRI localization sequences. Following 0.08 mL/kg (1.5 mL) Perflutren lipid microsphere infusion (DefinityÒ, Lantheus Medical Imaging Inc, N. Billerica, MA, USA), focused low frequency ultrasound (230 kHz) was delivered. A Brain MRI was obtained before and after focused ultrasound vessel disruption using a 3.0T MRI (Discovery MR750, GE Health care, Milwaukee, WI, USA) and a 6 channel phased-array flexible coil (General Electric, Milwaukee, WI, USA) as follows: FSE T2 (TR/TE 3000/71, ETL 4, FOV 16 9 16 cm, ST 2 mm, 256 9 192); 3D T1 (TR/TE 6/2, FOV 12 9 12 cm, ST 2 mm); Ax T2 FLAIR (TR/TE 8000/127, 12 9 12 cm, ST 2 mm, 128 9 128), Fast GRE (TR/TE 100/13, 12 9 12 cm, ST 2 mm). A dynamic contrast-enhanced sequence (3d T1 DCE (TR/TE 6/2, 12 9 12 cm, ST 3 mm, 128 9 128; temporal sampling 10 s) commenced 30 s after insonation. Gadobutrol 1 mmol/mL (Gadovist Bayer, Toronto, Canada) was injected via a pump injector (Medrad Spectris Solaris, Bayer HealthCare, Pennsylvania, USA) at 0.5 mL/kg and 5 mL/s followed by a 10 cc normal saline bolus. Image Processing and Analysis Contrast margins were manually traced at each time point on DCE-MRI using computerized planimetry (MIPAV, version 7.0; National Institutes of Health) at each time point. Utilizing time-series and Fast fourier transform (FFT) analyses we characterized the hemorrhages by the rate of leakage, duration, and volume expansion in three dimensional axes using Python (Python Software Foundation 2013). Spatial vectors were determined in the medial–lateral (M–L), anterior–posterior (A–P), and superior–inferior (S–I) planes via a spherical model. The diameter was calculated using the maximum and minimum points in each respective plane. We determined the temporal phases of growth and stability through visual inspection, maximizing the coefficient of determination through a least-squares method in the linear growth phase. A stationary phase was defined by a period of zero to minimal growth exhibiting
TRANSLATIONAL RESEARCH
Modeling the Pattern of Contrast Extravasation in Acute Intracerebral Hemorrhage Using Dynamic Contrast-Enhanced MR R. Liu • T. J. Huynh • Y. Huang • D. Ramsay K. Hynynen • R. I. Aviv
•
Ó Springer Science+Business Media New York 2014
Abstract Background Contrast extravasation (CE) in spontaneous intracerebral hemorrhage (ICH), coined the spot sign, predicts hematoma expansion (HE) and poor clinical outcome. The dynamic relationship between CE and the mode of ICH growth are poorly understood. We characterized the in vivo pattern and rate of HE using a novel animal model of acute ICH. Methods Basal ganglia ICH was created in 14 Yorkshire swine utilizing a novel MRI integrated model, permitting real-time CE observation using dynamic contrast-enhanced (DCE) MRI. Computerized planimetry measured CE volume at each time point. Spatial vector analysis along three
Electronic supplementary material The online version of this article (doi:10.1007/s12028-014-0071-z) contains supplementary material, which is available to authorized users. R. Liu T. J. Huynh R. I. Aviv (&) Department of Medical Imaging, Sunnybrook Health Sciences Centre, 2075 Bayview Ave., Room AG-31E, Toronto, ON M4N 3M5, Canada e-mail: [email protected] T. J. Huynh K. Hynynen R. I. Aviv University of Toronto, 27 King’s College Circle, Toronto, ON M5S 1A1, Canada Y. Huang K. Hynynen Imaging Research, Sunnybrook Research Institute, Toronto, ON, Canada D. Ramsay Department of Pathology, London Health Sciences Centre, London, ON, Canada K. Hynynen Medical Biophysics, University of Toronto, Toronto, ON, Canada
orthogonal axes determined distance vectors. Maximizing and minimizing the coefficient of determination defined the temporal phases of growth and stability, respectively. CE rate was calculated using a Patlak model. Results Asymmetric growth and variable rates of expansion characterized HE defining three distinct growth phases and patterns. A primary growth phase (duration 160 s; IQR 50–130) demonstrated rapid linear growth (0.04 mm/s IQR 0.01–0.10) accounting for 85 ± 15 % of total HE. The stationary phase demonstrated stability (duration 145 s; IQR 0–655). A secondary growth phase (duration 300; 130–600 s) accounted for 23 ± 8 % of total HE. In the primary and secondary growth phase, asymmetric growth occurred in the anterior–posterior (AP) planes (0.056 mm/ s; p = 0.026 and 0.0112 mm/s; p = 0.03). Monophasic 2 (14 %), biphasic 4 (35 %) (primary followed by secondary growth), and triphasic 8 (56 %) patterns (primary, stationary, and secondary growth phase) were observed. Conclusions A novel model of ICH provides real-time study of the dynamics and rate of CE. This data facilitates the understanding of pattern and rate of ICH formation. Keywords Animal model Spot sign Focused-ultrasound MRI
Introduction Intracerebral hemorrhage (ICH) accounts for 10–30 % of stroke associated with significant morbidity and increased risk of death [1]. Hematoma expansion is a strong predictor of poor outcome and is associated with mortality rates over 50 % [2]. Developing effective therapies to reduce clinical outcome requires an understanding of the pathophysiological basis of ICH formation and growth. The mechanism
123
Neurocrit Care
of ICH growth is unknown but is contributed to by continued [3, 4] or delayed bleeding [5] of the initially ruptured vessel. Expansion occurring via secondary shearing of blood vessels surrounding a hematoma may be complicit [6]. Secondary vessel injury may trigger a ‘‘domino’’ pattern of hematoma expansion, recently reproduced on a computer simulation study [7]. Hematoma location, volume, and expansion are strong predictors of poor post-ICH outcome [2]. Since location and initial volume are determined at presentation, prevention, or attention of hematoma expansion remains a promising therapeutic target [8]. Clinical studies have demonstrated biological activity of prothrombotic agents such as recombinant Factor VIIa (rFVIIa), but have yet to yield improvement on the functional outcome. Increased effort to identify patients most likely to benefit from intervention is the focus of ongoing clinical studies dichotomizing patients by spot sign or contrast extravasation presence; an established marker of hematoma growth [9, 10]. The spot sign is a morphological marker of contrast extravasation but is associated with significant differences in rate of leakage [11]. A better understanding of hematoma expansion dynamics including both the rate and pattern of expansion is essential. Such knowledge will potentially inform future studies evaluating therapeutic options for prevention or attenuation of hematoma expansion. Only 20–30 % of acute ICH patients demonstrate a spot sign on first pass CTA studies. Few centers routinely measure contrast extravasation rate with CT permeability [11] or dynamic techniques [12, 13]. We have previously created and reported a novel real-time MRI porcine ICH model of acute hematoma growth and contrast extravasation to bridge this gap [14]. The method of ICH induction closely resembles vessel disruption seen in human ICH and produces reliable sized hematomas comparable to that seen in clinical practice. In this paper, we investigate the spatial and temporal characteristics of ICH-associated contrast extravasation and provide insights into the pattern and pace of ICH expansion.
Methods Experimental Design This study was conducted with the approval of the Sunnybrook Hospital Research Institute Animal Care Committee (Animal Use Protocol #12-435) in compliance with the guidelines established by the Canadian Council on Animal Care and the Animals for Research Act of Ontario, Canada. The ICH model methodology is previously described. Briefly Yorkshire male swine were chemically immobilized and maintained under general anesthesia with
123
2–3 % isoflurane in oxygen (>2 L/min). Physiological monitoring at 5 min intervals included clinical observation, tail pulse oximetry, rectal temperature and blood pressure measurement through the femoral arterial line. The animal was placed in a supine position with the head immersed in degassed water serving as a conduction medium for ultrasound emitted from a upwardly-positioned modified clinical MR-Guided Focused Ultrasound (MRgFUS) brain system (ExAblate 4000, Insightec, Tirat Carmel, Israel) comprising a 1024-element hemispherical phased-array with a diameter of 30 cm. Accurate anatomical localization of vessels within the basal ganglia was achieved using MRI localization sequences. Following 0.08 mL/kg (1.5 mL) Perflutren lipid microsphere infusion (DefinityÒ, Lantheus Medical Imaging Inc, N. Billerica, MA, USA), focused low frequency ultrasound (230 kHz) was delivered. A Brain MRI was obtained before and after focused ultrasound vessel disruption using a 3.0T MRI (Discovery MR750, GE Health care, Milwaukee, WI, USA) and a 6 channel phased-array flexible coil (General Electric, Milwaukee, WI, USA) as follows: FSE T2 (TR/TE 3000/71, ETL 4, FOV 16 9 16 cm, ST 2 mm, 256 9 192); 3D T1 (TR/TE 6/2, FOV 12 9 12 cm, ST 2 mm); Ax T2 FLAIR (TR/TE 8000/127, 12 9 12 cm, ST 2 mm, 128 9 128), Fast GRE (TR/TE 100/13, 12 9 12 cm, ST 2 mm). A dynamic contrast-enhanced sequence (3d T1 DCE (TR/TE 6/2, 12 9 12 cm, ST 3 mm, 128 9 128; temporal sampling 10 s) commenced 30 s after insonation. Gadobutrol 1 mmol/mL (Gadovist Bayer, Toronto, Canada) was injected via a pump injector (Medrad Spectris Solaris, Bayer HealthCare, Pennsylvania, USA) at 0.5 mL/kg and 5 mL/s followed by a 10 cc normal saline bolus. Image Processing and Analysis Contrast margins were manually traced at each time point on DCE-MRI using computerized planimetry (MIPAV, version 7.0; National Institutes of Health) at each time point. Utilizing time-series and Fast fourier transform (FFT) analyses we characterized the hemorrhages by the rate of leakage, duration, and volume expansion in three dimensional axes using Python (Python Software Foundation 2013). Spatial vectors were determined in the medial–lateral (M–L), anterior–posterior (A–P), and superior–inferior (S–I) planes via a spherical model. The diameter was calculated using the maximum and minimum points in each respective plane. We determined the temporal phases of growth and stability through visual inspection, maximizing the coefficient of determination through a least-squares method in the linear growth phase. A stationary phase was defined by a period of zero to minimal growth exhibiting
