Image Registration in Ultrasound-Assisted Brain Surgery: Its Prospects, Challenges, and Techniques

Image Registration in Ultrasound-Assisted Brain Surgery: Its Prospects, Challenges, and Techniques

Haradhan Chel (Central Institute of Technology Kokrajhar, India) and Prabin Kumar Bora (Indian Institute of Technology Guwahati, India)
Copyright: © 2018 |Pages: 23
DOI: 10.4018/978-1-5225-2829-6.ch007
OnDemand PDF Download:
List Price: $37.50


Image registration is an essential step in the image guided brain surgery. A preoperative magnetic resonance (MR) image guides the neurosurgeon about the size and the location of the tumor inside the brain of the diseased person. Due to several reasons, brain shift occurs during the surgery, results in the shift of the actual position of the tumor. Intra-operative MR imaging is expensive and may not be financially viable for many hospitals. An effective intraoperative US can be used in replacement of MR. For performing registration of US and MR images, the most of the state-of-the-art methods use a suitable similarity or dissimilarity measure, a spline based deformation model, a smoothing technique and an effective fast optimization method. This chapter starts with a discussion on various types of brain tumors and their clinical significance. It also covers on various similarity measures, optimizations and the available database of US and MR brain images.
Chapter Preview


Brain lesion is a broad term which includes Brain Tumor, Cerebral Infarction, Multiple Sclerosis, Abscess etc (Moore,K., Kim, Lyndon, 2009), (Radha, M.L, 2004) . The surgical treatment of a brain lesion is a complex task to a neurosurgeon. Such treatments differ depending on the types, the nature of growth, the criticality of location, shapes, sizes and the malignancy. When a brain tumor patient approaches for a treatment from a neurologists, generally the patient is first advised for a imaging test with either a computer Tomography (CT) or a magneto Resonance (MR) imaging. If a surgery is advised based on those MR or CT images, neurosurgeon plans it accordingly based on those images. In spite of the availability of these high resolution preoperative MR or CT images, it is almost impossible to perform the surgery without the help of an intraoperative image assistance. The main reason is the brain shift during the craniotomy, and the leakage of cerebro spinal fluid (CSF) during surgery. In the field of brain surgery, preoperative and intraoperative imaging have equal importance. Both of these imaging modalities are the necessary assisting elements to a neuro-surgeon. For assisting the doctors with both preoperative and intraoperative imaging for surgery planning and execution, various types of neuro-navigational systems are used. Among all the brain imaging modalities, MR is the most popular modality for its better resolution and visibility of different tissue structures. In the field of clinical investigation of a brain lesion, the shape, the size and the location of the lesion become accurately available from the images. Ultrasound(US) is also suitable modality which can effectively be used for neurosurgical assistance. Experts trained in US imaging can differentiate between normal and abnormal tissues by looking into the echogenicity of different tissue structures. In spite of several advantages of ultrasound, MR imaging is preferred over US imaging as an intraoperative modality in neurosurgical assistance. The main reason is that most of the neuro surgeons are trained in either axial computed tomography slices or magnetic resonance imaging slices only. The limitations of MR images are the higher cost of and the larger time required to scan a volume. Building an infrastructure with intraoperative MR imaging involves a huge investment and the maintenance of such a system requires skilled engineers. This sometime creates financial limitation in many hospitals. For these reasons, such infrastructure is available in a very few locations or big cities in many underdeveloped countries like India. Before the year 2000, the medical community was of the opinion that ultrasound (US) is not a suitable modality for investigating the brain diseases. There are two main reasons behind that. Firstly US imagery has a poor resolution and certain textures of low grade gliomas are not separable from the speckle noises present in the US image. Another reason is that US cannot pass through the cranial bones. US image of brain can only be taken after the craniotomy. By the advancement of the US image acquisition sensors and signal processing techniques, the resolution of the US images have been improved a lot. However, no such technology is reported till now which can generate a better US brain image without craniotomy. In various studies it has shown that the intraoperative ultrasound can replace the intraoperative MR and it directly reduces the infrastructural cost which is easily implementable in all suburban areas. At the present situation, high level professional surgeons who are working in suburban areas face infrastructural issues in using intraoperative MR imaging. This can be minimized by using low cost US modality. For linking the missing space between using the US and MR images, continuous efforts from 1993 (Trobaugh, J.W, 1993) are carried out by this medical community. For using US as an intraoperative modality, this community has developed the state-of-the-art techniques. Beltagy, (2010) proposed a method for registration between pre, during, and post-resection ultrasound images, as well as MR images. With the help of retrospective cohort study of intraoperative US on 25 patients they showed that the intraoperative US was useful in identifying tumor boundaries and minimizing residual tumor. Renovanz (2014), Anokhin,, (2015), Coburger et. al. (2015), and Moiyadi and Shetty (2016) performed a retrospective analysis of the gross total resection (GTR) of patients who underwent ultrasound guided neurosurgery and concluded that ultrasound could be effective in achieving GTR, especially for low-grade glioma. Various literature are available in this field which can categorically be classified from two different views i) clinical point of view: (Moore,K.,Kim, L.,2009), (Radha, M.L, 2004), (Trobaugh,2004), (Kaivukangas 1993), (Maarouf, 1996), (Gronningsaeter, 1996), (Mercier 2011), (Unsgaard, 2002), (Unsgaard, 2005), (Unsgaard., 2002) and, ii) image processing point of view: (Forsberg, D., 2013), (Brown, L. G.,2002),(Maintz, J. B. A. and Viergever M.A., 1998), (Zitova, B., Flusser, J. 2003), (Sotiras, 2013), (Fatma, 2016), (Roche, 1999),(Roche, 2001), (Pennec, 2005),

Complete Chapter List

Search this Book: