Imaging Division, Lawson Health Research Institute, London, Ontario, Canada. Department of Medical Biophysics, The University of Western Ontario, London, Ontario, Canada. Imaging Research Laboratories, Robarts Research Institute, The University of Western Ontario, London, Ontario, Canada.
Phys Med Biol. 2013 Nov 7;58(21):7513-26. doi: 10.1088/0031-9155/58/21/7513. Epub 2013 Oct 8.
Intraventricular hemorrhage (IVH) is a common disorder among preterm neonates that is routinely diagnosed and monitored by 2D cranial ultrasound (US). The cerebral ventricles of patients with IVH often have a period of ventricular dilation (ventriculomegaly). This initial increase in ventricle size can either spontaneously resolve, which often shows clinically as a period of stabilization in ventricle size and eventual decline back towards a more normal size, or progressive ventricular dilation that does not stabilize and which may require interventional therapy to reduce symptoms relating to increased intracranial pressure. To improve the characterization of ventricle dilation, we developed a 3D US imaging system that can be used with a conventional clinical US scanner to image the ventricular system of preterm neonates at risk of ventriculomegaly. A motorized transducer housing was designed specifically for hand-held use inside an incubator using a transducer commonly used for cranial 2D US scans. This system was validated using geometric phantoms, US/MRI compatible ventricle volume phantoms, and patient images to determine 3D reconstruction accuracy and inter- and intra-observer volume estimation variability. 3D US geometric reconstruction was found to be accurate with an error of <0.2%. Measured volumes of a US/MRI compatible ventricle-like phantom were within 5% of gold standard water displacement measurements. Intra-class correlation for the three observers was 0.97, showing very high agreement between observers. The coefficient of variation was between 1.8-6.3% for repeated segmentations of the same patient. The minimum detectable difference was calculated to be 0.63 cm(3) for a single observer. Results from ANOVA for three observers segmenting three patients of IVH grade II did not show any significant differences (p > 0.05) for the measured ventricle volumes between observers. This 3D US system can reliably produce 3D US images of the neonatal ventricular system. There is the potential to use this system to monitor the progression of ventriculomegaly over time in patients with IVH.
脑室内出血(IVH)是早产儿常见的疾病,通常通过二维颅超声(US)进行诊断和监测。患有 IVH 的患者的脑室内通常会有一段时间的脑室扩张(脑积水分流)。这种初始脑室大小的增加要么会自行消退,这通常表现为脑室大小稳定一段时间,最终恢复到更正常的大小,要么是持续的脑室扩张,不会稳定,可能需要介入治疗来减轻与颅内压升高相关的症状。为了更好地描述脑室扩张,我们开发了一种 3D-US 成像系统,该系统可以与常规临床 US 扫描仪一起用于成像有脑积水分流风险的早产儿的脑室系统。专门设计了一个带有电动换能器外壳的系统,用于在孵化器内手持使用,换能器通常用于颅 2D-US 扫描。该系统使用几何体模、US/MRI 兼容的心室体积体模和患者图像进行了验证,以确定 3D 重建准确性和观察者之间及观察者内部的体积估计变异性。3D-US 几何重建的准确性误差小于 0.2%。US/MRI 兼容的心室样体模的测量体积与金标准水置换测量值相差在 5%以内。三位观察者的组内相关系数为 0.97,表明观察者之间非常一致。同一患者的相同部位的重复分割的变异系数在 1.8-6.3%之间。计算出的单观察者最小可检测差异为 0.63 cm3。对三位观察者对三个 IVH 二级患者进行分割的方差分析结果显示,观察者之间的测量心室体积没有任何显著差异(p>0.05)。该 3D-US 系统能够可靠地生成新生儿脑室系统的 3D-US 图像。有可能使用该系统监测 IVH 患者的脑积水分流随时间的进展。