Alessandro Moiraghi, Gregory Zegarek, and Karl Schaller
Geneva University Hospitals Department of Neurosurgery
University of Geneva Faculty of Medicine
Importance of Extent of Resection in Glioma Surgery
The goal of surgical treatment of cerebral gliomas is maximizing the extent of resection (EOR) while preserving the function of relevant structures in order to ensure a good neurological outcome for the patient. The impact of the extent of resection on survival of patients with high grade (HGG) and low grade gliomas (LGG) remains a controversial subject, but in the last two decades, many studies have shown the benefit of a greater EOR on survival. It is therefore common practice today that neurosurgeons attempt to achieve the highest level of resection feasible with minimal residual tumor volume (preferably zero), while preserving neurological function and quality of life of the patient.
In order to improve the EOR and rate of complete resection, many different methods and technologies have been proposed. Well known examples include 5-ALA fluorescence for direct tumor visualization, intra-operative magnetic resonance imaging (ioMRI), and intra-operative ultrasound (ioUS). In this letter, we explore these three technologies, and compare the advantages and disadvantages of ioUS in comparison to the other techniques.
5-ALA fluorescence in Glioma Surgery
Despite many advances in the field of intra-operative imaging, 5-ALA fluorescence is the only technique so far that has been proven to be effective by a randomized multicenter clinical trial and supported by evidence of class IIb. In a phase III clinical trial of 5-ALA-guided surgery, Gross Total Resection (GTR) was reached in 65% of cases operated with 5-ALA, as compared with 36% in the control white-light group (p < 0.0001).1
In light of these results, the pursuit of an ever more radical resection seems feasible thanks to fluorescence-guided surgery. However, it’s necessary to debate the consequences of obtaining more complete and aggressive resections when this typically includes areas of lower fluorescence whose specificity for tumor remains somewhat indeterminate. Authors have made attempts to standardize the subjective description of fluorescence intensity with terminology from “strong” to “weak or vague” and with 4 point grading scales.2 Nevertheless, we know from experience that tissue without clear fluorescence may have tumor on histopathology. So, while a helpful tool, 5-ALA fluorescence is certainly not a holy grail for delimiting tumor margins intra-operatively. The main disadvantages of fluorescence techniques are the need for the administration of a drug, and the fact that 5-ALA is suitable only for HGG, limiting its applicability.
Advances in intra-operative MRI
Intra-operative MRI is considered today to be the most accurate intra-operative imaging technique available, allowing increased EOR in patients with glioma without an increase of postoperative complications. ioMRI has high accuracy rate (GTR of 96% vs. 68% for the control group, p=0.023).3 Of course, there are important limitations to keep in mind regarding ioMRI. It is extremely expensive, time consuming, requires entirely dedicated operating rooms and instruments, and it is not a real-time imaging technique. Globally, its use is very limited as a very small percentage of operative centers have access to ioMRI.
Comparisons and Considerations on intra-operative Ultrasound
Intra-operative ultrasound has many important advantages over other techniques, such as low cost, availability, and ease of use. In the existing literature, this technique has been reported to allow for obtaining a gross total resection in the majority of patients. There have recently been a net increase in the number of studies reporting the usefulness of ultrasound-guided resections of gliomas; however, no one study regarding ioUS has managed to bring a high level of evidence to the technique.
Applications to High Grade Gliomas
Studies on the EOR of gliomas operated with intra-operative ultrasound have been encouraging with improving results over time. Solheim et al have reported a 37% rate of GTR in a series of 156 HGG consecutively operated in 2010, with a worsening of neurological functions to 4-6 weeks in 13% of cases. In a subgroup of patients analyzed with unifocal lesions, contrast-enhancing, non-invasive basal ganglia, thalamus or corpus callosum, and with KPS> 70, the percentage of cases with GTR grows to 63%.4
In a study of Moiyadi et al from 2013, the use of intraoperative 3D ultrasound in a series of 51 HGG has shown a GTR in 47% of cases, with a frequency of complications similar to those obtained with other intra-operative imaging techniques. Dividing the series into resectable and unresectable tumors, the fraction of GTR increased up to 88% in tumors deemed radiologically to be resectable.5
Another interesting application related to ioUS is the use of contrast-enhanced ultrasound (CEUS). In 2017, Prada et al reported their experience comparing CEUS and preoperative gadolinium-enhanced T1-weighted MR imaging in a series of glioblastomas showing superimposable intraoperative images regarding location, margins, morphologic features, and dimensions, with a similar enhancement pattern in most cases.6
Based on recent literature, the potential use of CEUS in brain tumor surgery and in other neurosurgical applications has been recognized in the recent update of the guidelines released by the European Federation for Ultrasound in Medicine and Biology (EFSUMB).7
Applications to Low Grade Gliomas
Impressively, Coburger et al in a recent prospective study from 2015 report that modern linear high-frequency ioUS reaches a degree of accuracy close to iMRI in detecting residual tumor masses during intraoperative resection control in LGG. They used a linear array transducer with an extended frequency range of 7-15 MHz, characterized by a hockey-stick design with a small footprint conceived for intra-operative applications and observed a sensitivity slightly higher in iMRI (83%) than in ioUS (79%) and the same specificity (67%) for both.8
Personal Experience combining Intra-Operative Ultrasound and Neuronavigation
Our experience with ioUS and real-time neuro-navigation (NN) confirms that the technique allows surgeons to plan and execute precise targeted craniotomies. The fusion of two different imaging techniques makes it possible to combine the strengths and overcome the limitations of both: the lack of panoramic views of ultrasonography has been compensated from the large visual field offered by the NN, and the interpretation of ultrasound images has been simplified by the possibility to compare them with corresponding co-planar MRI slices. Real-time ultrasound images allow the operator to select any desirable plane free in space and follow tissue shifting and deformation.9 Furthermore, Rasmussen et al reported that functional NN with preoperative fMRI and DTT combined with updated anatomic information with 3D ultrasound gives the neurosurgeon an advantage when resecting brain lesions and preserving eloquent cortex.10
Re-calibration and fusion of ultrasound imaging and MRI has demonstrated an excellent ability to neutralize the loss of accuracy due to the well-known "brain shift". On the other hand, distortion and deformation of the parenchyma as a result of surgical maneuvers, called "brain deformation", often becomes so extensive that a precise overlap between images becomes impossible. In this case the surgeon refers exclusively to the ultrasound image. All the above means that it would be not possible to base a tumor resection solely on the images derived from the initial registration, a fact emphasizing in a paramount way the drawbacks of classical NN.9,11
In our clinical experience and by becoming gradually more familiar with the technology, the registration procedures and fine-tuning could be performed progressively faster and can be completed within a few minutes. The technique of fusion imaging also allows to correctly interpret the ultrasound images thanks to the comparison of ultrasound to MRI NN, more familiar to the neurosurgeon. Thus, the learning curve was accelerated significantly to a point that eventually allows operating with only ultrasound guidance. Obviously learning curves vary between different individuals, but our experience so far indicates that after approximately 2 dozen ultrasound-guided resections, the technique of the surgeon is advanced enough to allow him to operate a majority of cases solely guided by ioUS and employ Fusion Imaging mainly on complex cases for better guidance.
Today, due to the current state that ioMRI is effectively very limited globally, we believe that the use of ioUS could be complementary to 5-ALA fluorescence, allowing a well-trained neurosurgeon to discriminate between the limitation of both techniques and making them symbiotic to improve extent of resection and ultimately, outcomes in glioma surgery.
Fig. 1 Example of real-time fusion imaging in a case of right temporal GBM: 1a 2D B-mode ultrasound scan showing in high definition the explored lesion and the surrounding structures. 1b co-planar T1 weighted contrast-enhanced volumetric pre-operative MRI. With the aid of the co-planar MRI slice, it is easier to appreciate all the anatomical landmarks and their alignment (t= tumour, e= tentorial edge, c=cerebellum, v= vermis and p= pons). 1c the four projection of virtual navigation from left to right on sagittal, coronal and axial planes and the real oblique axis determined by the real-time reconstruction of the co-planar plane determined by the US probe direction represented on a 3D volume rendering of the pre-operative MRI.