16 Large Multilobar Gliomas

10.1055/b-0039-172176

16 Large Multilobar Gliomas

16.1 Introduction

Conceptually, this chapter introduces no new information. It provides examples of techniques I have used for successfully removing very large gliomas which span more than one brain lobe. However, the previous chapters have provided all that is necessary to think through these cases and to come up with an acceptable plan to tackle these cases, and there is no way a chapter can summarize all possible combinations of spread that gliomas can obtain when they get bad.

This chapter provides some basic guidelines, and follows with some specific examples of what has worked for me in these challenging cases.

16.2 Basic Guidelines for Planning very Large Gliomas

Try to understand how this tumor has spread: Gliomas tend to follow white matter tracts. Smaller tumors may not clearly be following an identifiable tract, but big gliomas usually are.

Common routes of multilobar spread include:

ILF: The common path for tumors extending from temporal to occipital pole.
SLF: A bad way to spread, but unfortunately common, especially when the tumor gains access to the posterolateral temporal lobe, the MFG, IFG or the inferior parietal lobe.
Corpus callosum: Spread across the corpus callosum is one way we can lose control of the tumor. Remember, Bi-anything is bad
Cingulum: This is how tumors get from the PHG to the cingulate gyrus and back. It is a common way splenial butterfly gliomas spread.
Diagonal band of Broca/ventral amygdalofugal tract: This is how tumors can spread from one amygdala to the other, or from the amygdala to the hippocampus.
Descending tracts of the corona radiata / corticospinal tract: This is an uncommon route of spread, thankfully, because it is obviously bad.

By knowing the pathway of spread, you know what parts of the brain are probably out (especially with GBMs) and have an idea of how to tackle this tumor with the minimal amount of damage.

Be realistic: It is always fulfilling when our patients go home with minimal deficits. However, for patients with multilobars tumors, perfection is not a realistic goal in most cases. First, they often come in with deficits, often because the tumor has destroyed a network. Second, they have a grim natural history if you don’t do something meaningful. Third, the idea that you are going to take out two lobes of the brain and nothing noticeable will occur is also misguided. The idea that someone needs to be neurologically perfect or they should be dead is overly rigid. There are meaningful lives led with some neurologic deficits. Where that line is crossed is a complex and personal question.

Remember that these are bad tumors. Be realistic about what you can offer, but also be realistic about how bad the natural history of this problem is.

Think in terms of cuts: This is the foundation of my approach to gliomas. Large multilobar gliomas are no different; however, there are often multiple, complex and/or non-standard cuts involved. The foundation of making these kinds of decisions is grounded in the more fundamental issues described in previous chapters.

Plan around lobectomies/partial lobectomies: This is generally no place for dainty and tailored surgeries. A tumor of this severity needs to be thought of as the extensive cancer it is. Also, a large multilobar cut is going to leave some roads to nowhere, and you need to not leave cancer behind in disconnected nonfunctional brain.

Think about how to combine the cuts: Often, removing a multilobar tumor will need a long cut which is a combination of two subcuts. This does not always need a single large craniotomy along the length of the cut; it can usually be made smaller by working under the bone flap.

One craniotomy or two? If you choose to tackle a glioma this bad, you will be removing a lot of brain, which will make a great deal of room under the skull to work. Thus, I always strive to avoid a large hemicraniectomy type opening, as many of these cases are redo surgery, and I wish to not risk an infection of a flap this large.

There are two ways to exploit this fact. One is to make a medium-sized craniotomy in the center of the cut and to work under the skull to continue the cut in both directions. A second way is to perform two small craniotomies, to make the cut under both and then to extend them under the skull to joint them between the two craniotomies (which I call a pull-through approach). It is rarely necessary to perform a giant craniotomy.

Each case has unique needs, and there are a large number of possible options.

16.2.1 Representative Cases

It is obviously impossible to describe all possible approaches to massive gliomas, as the variability is quite significant. A few case examples are provided to provide some insight and ideas for tackling these complex cases.

Fig. 16.1 demonstrates a recurrent GBM with extension from the occipital lobe, into the anterior temporal lobe. After examining the films, it seemed unlikely that this tumor could be removed while maintaining the visual input to the occipital pole, or the temporal outflow pathways, which are mostly at the back. The patient elected to undergo an occipitotemporal lobectomy.

The tumor is primarily located within the ILF and IFOF systems, as well as the visual system. While it is technically possible to obtain a complete resection without removing the temporal tip, it is unclear what this part of the brain would connect to in a meaningful network. The principle goal of this operation is to maintain the communication between the semantic networks and the SLF system, and ideally the crossed semantic loop. The ipsilateral visual system and IFOF is obviously not working.
We performed this operation via a “pull-through” technique, where occipital and temporal craniotomies were performed to make a cut extending from the posterior parietal midline heading anteriolaterally to the STG roughly paralleling the SLF and continuing into the atrium and temporal horn under the bone between the two bone flaps. After this cut was complete, we put the patient to sleep and performed occipital and temporal lobectomies and removed the intervening disconnected brain under the skull (thus the term “pull through”). This procedure did not incur any additional neurologic morbidity not previously seen preoperatively.
None.
The resection is aggressive but complete and was well tolerated.

Fig. 16.2 demonstrates a case of a patient with a dense anomia but otherwise intact fluent speech from this massive left-sided GBM.

This tumor is a tumor of the left-sided visual system and its extensions including the splenium. Unlike the right-sided case, we elected not to pursue as an aggressive resection involving multilobar resection, but instead elected to base our resection around an occipital lobectomy, and an anterior occipital disconnection, with the medial temporal and splenial tumor being removed through the lobectomy cavity. The primary goal was sparing the lefty SLF.
Given the long extension along the SLF, we performed a biportal type approach, with two small craniotomies centered over the medial parietal cut, and over the cut with the SLF. These cuts were performed awake until the SLF was clear, and then we put the patient to sleep and performed an occipital lobectomy, removed the splenium, and the medial structures.
None.
The resection is good and his speech was preserved; however, he had new problems with judgment which tempered my view of his outcome. I am not certain about why the occipital lobe and splenium play a role in judgment, but I have seen this before in two other cases.

Fig. 16.3 demonstrates a complex fronto-temporo-insular GBM encasing the MCA trunk, which had undergone a minimal resection at another institution.

If you only look at the contrast images, this doesn’t seem like an extremely extensive tumor, but the T2 is more realistically a better measure of the extent of this tumor, though admittedly some of this is edema. It is obviously not possible to completely remove all of these areas; however, temporal lobectomy and marginal resections fit better with the anatomy, but also help isolate the densely encased MCA and allow us to spare the MCA while trimming the fibrous dense enhancing tumor off the multiply encased MCA vessels to leave as little of this as possible.
I began this resection by completing the temporal lobectomy and performing a lateral frontal disconnection and limited frontal lobe resection. This isolated the involved MCA. Using a Doppler probe and techniques borrowed from meningioma surgery, I managed to trim this tumor to an acceptable residual which is not invading the brain (t has been separated from it).
None.
There is some enhancing tumor on the MCA vessels, which simply would not come off the arteries; however, this tumor no longer is attached to the cerebrum, and thus has less opportunity to launch tumor cells into distant white matter tracts.

Fig. 16.4 demonstrates an extreme case of a left fronto-temporo-insula glioma extending into the basal forebrain. While obviously it is unlikely that we will be unable to get all of this out, especially since this patient would not cooperate with brain mapping for very long during the operation, we performed a temporal lobectomy and orbitofrontal cortex resection, and trimmed the anterior insula conservatively based on the IFOF tractography with a reasonable result.

Tumors like this are often termed gliomatosis cerebri because they are in multiple lobes. Obviously, this is potentially applicable in this case, but really this is not consistent with the spirit of what that disease means. This tumor is very extensive glioma of the insula which has maximized all of its routes of spread out of the insula. The temporal lobe and cingulum is densely involved, and the uncinate pathway has filled the orbitofrontal region with tumor. But there is not bilateral disease, or completely unconnected disease. This is a bad glioma in multiple lobes, but it is not really gliomatosis in my opinion.
The craniotomy is more extensive than other similar tumors. It needs to be somewhat anterior as we need significant access to the supraorbital space so that we can comfortably work in the orbital frontal cortex without blind sweeping. This does not need a formal; pterional craniotomy but the spirit of the technique is the same.

The approach follows a lot of similar steps to a temporal-sided insular glioma: Temporal lobectomy, insula resection, following the curve of the uncinated forward, but eventually the pars orbitalis should be entered and a cut of the frontal lobe above the orbitofrontal component should be taken to the midline before subpializing the orbitofrontal brain up the front of M1.

None.
Despite an aggressive resection, there are a lot of things I could have done better on this early case in my experience. First, the orbitofrontal resection is not medial or posterior enough. The M1 is an excellent landmark for staying out of the basal forebrain and lenticulostriates, and since this case, is my stopping point in the coronal plane. The orbitofrontal resection should continue to the midline. There is also some temporal tumor left. Because this patient did not participate for very long during the mapping, I could not make the aggressive posterior temporal disconnection I would normally like to make.

Fig. 16.1 Demonstrates a recurrent glioblastoma multiforme (GBM) with extension from the occipital lobe, into the anterior temporal lobe; (a) preoperative images, (b) images of the approach.

(c) Postoperative images.

Fig. 16.2 A patient with a dense anomia but otherwise intact fluent speech from massive left-sided glioblastoma multiforme (GBM); (a) preoperative images, (b) images of the approach, (c) postoperative images.

Fig. 16.3 A complex fronto-temporo-insular GBM encasing the MCA trunk, which had undergone a minimal resection at another institution; (a) preoperative images, (b) images of the approach.

(c) Postoperative images.

Fig. 16.4 An extreme case of a left fronto-temporo-insula glioma extending into the basal forebrain; (a) preoperative images, (b) images of the approach.

(c) Postoperative images.