9 Going from Scan to Plan in a Glioma Surgery
This is the most critical chapter in the book, as it serves as the bridge between the theoretical and applied parts of this book. Most unsuccessful glioma operations fail at the stage of conceptualization of where the tumor and functional systems really are, and of formulating a plan of attack for how to safely remove as much tumor as possible.
In this chapter, I will outline the framework which will serve as the organizing principle for the rest of the book when we describe how to remove gliomas from different parts of the brain. While it’s impossible to create a satisfactory universal system for glioma surgery, this chapter addresses themes common to the planning of most glioma surgery so that planning these cases can proceed in a structured and organized fashion. A plan consisting of “Get in there and take it out” is seldom a recipe for technical excellence.
9.2 The 3 “Ds” of Glioma Surgery
9.3 How Thinking in Terms of the 3 “Ds” Aids with Glioma Surgical Planning
The 3 “Ds” represent the three phases of a glioma operation, and the transition between phases brings changes in surgical technique. More importantly, a transition in phase represents a change in the philosophy of what one is trying to accomplish. For example, dissection is usually slow and cautious during the “divide” phase, and speeds up during the destroy section, when the eloquent regions and subcortical white matter tracts are separate from the planned tumor specimen.
To define our terms (Fig. 9.1):
Define: This is a non-destructive phase of the operation where the principle goals are cognitive, and which begins prior to skin incision and only ends when the brain begins to be cut. The sulcal patterns are studied closely to determine exactly what anatomy is exposed and where the tumor is. Image guidance and DTI are used to tentatively plan a cut and determine the goals of the cortical mapping. Cortical mapping is then done to determine safe entry points, and to determine the cortical termini of white matter.
Divide: This phase is the most challenging, and the most critical portion of the resection. In the division phase, the brain is divided such that the bulk of the tumor specimens is separated from the eloquent brain regions which will be left in place. This is accomplished with ongoing functional mapping of the subcortical white matter and is continued until the cut extends unequivocally past the critical regions.
Destroy: Following the division phase, the remaining tumor volume is essentially a non-eloquent brain tumor and can be removed using standard anatomic techniques, such as temporal lobectomy.
The relative length, complexity, and importance of these three phases are obviously case dependent. For example, the approach to deep tumors expose and transgress a small amount of often tolerant cortex, leaving the “define” phase far shorter than that needed for a tumor of the temporoparietal junction where this phase can be long and complex given the arterial and multimodal functional anatomy. The “division” phase usually is long, but in large insular gliomas we plan for this to last hours.
9.3.1 The “Define” Phase in More Detail
Understand the Cortex
Whether you are doing your third glioma or your five hundredth, it is imperative that you consciously stop and think exactly what gyri are involved with tumor, and where functional areas, white matter tracts and major arteries lie relative to them. Having a false concept of your location is the first step in making incorrect assumptions and, in turn, bad decisions.
Gliomas come in many forms; however, a large number of them greatly expand a portion of the anatomy relative to neighboring brain, often placing the tumor in the region of something that initial raises more concern than warranted. Examples of these distortions include an enlargement of the fornix which appears to involve the thalamus, or enlargement of a parietal gyrus which displaces the motor strip making it appear at first glance that one is dealing with a motor strip glioma when in fact it is not.
Close examination of the T2 weighted images is almost always helpful for sorting these issues out. Sulcal boundaries are more obvious, and a disciplined approach will usually reveal where the tumor really is located. In a system like the one described, which is centered around the idea that anatomic resection is generally the most complete resection, sulci can serve as critical guide posts in a world with few reliable landmarks.
Executing this plan intraoperatively requires you to take an assessment of what gyri you have exposed after you open the dura. While minimalism is fully in line with my philosophy, note that the keyhole principle predicts that you will struggle if superficial targets needed for the surgery (like cortex you plan to map) are not exposed. A small bone flap extension should not be resisted, if warranted.
Also, a mapping case can be derailed quickly into the case if an en passage artery is injured on the way into the subcortex. This is especially true near the Sylvian fissure, where arteries often interdigitate with the opercular clefts on their way to their often critical targets. Even worse are arteries contained within tumor, especially high-grade gliomas, as these are at very high risk for injury by the unaware.
Tumor Spread along White Matter Tracts
It is well known that glioma progression often follows white matter tracts. This is the principle reason these tumors are not surgically curable, but it also provides insight into how to best achieve anatomic resection. The very reason gliomas can be classified and described in a systematic way is that white matter spread so dominates their nature that it forces certain order and predictability on them.
By determining the pattern of spread, it is possible to make sense of the tumor anatomy and to identify a route by which an anatomic resection of these areas of higher tumor density could be removed. In other words, it is helpful to take the tumor out the way it went in, by following the involved tracts as much as possible.
Conversely, tumor spread down certain white matter tracts can often be a trap lying in wait. For example, lateral temporal tumors can often follow the SLF, which generally cannot be removed without deficit. Identifying this high risk situation is the key to avoiding it.
Mapping the Cortex
When I trained, and at many places I am aware of, “cortical mapping” and “awake brain mapping” were basically synonyms, as the patient was put back to sleep at the end of the cortical mapping in most cases. While presently my system reduces cortical mapping from its previous lofty position, it still is a critical step.
Cortical mapping obviously defines certain “no fly zones,” such as speech and motor cortices. More importantly, it demonstrates the point of origin of a functional network which can then be followed inward towards the major tracts. I explain this to trainees as analogous to following a small cortical artery into the Sylvian fissure to locate the MCA trunks. The difference in this case is obviously that one cannot see these fibers, and thus function is your only true map.
As well, cortical mapping is the time to get the patient in the rhythm of doing the tests, when you are not cutting on the brain. Most patients take some time to warm up, and this is the time to get them there.
Task selection is obviously location based: we don’t expect to find naming sites in the medial right frontal lobe and motor firing in the right temporal pole. With that being said; I typically test motor and speech in all cases. First, this is because even if my goal is to avoid injury to motor fibers, I would like to use speech as part of the double task to stress that system during the division phase, and cortical mapping gets the patient up to speed with the naming task.
Also, it’s important to note that most higher human actions engage more than one functional system, and while the site you are operating in may not be the primary site for that function; its removal may interfere with its performance. For example, we know that semantic and lexical memory needed for naming is not primarily located in the posterior right frontal lobe; however, without motor planning ability (it might be there) it is difficult to talk. Thus, while speech mapping in the right frontal lobe is unlikely to yield a true anomia site, it helps to better elucidate the anatomy of that area in different ways. Note that higher activities of the cerebrum involve the use of multiple brain networks.
9.3.2 The “Divide” Phase in More Detail
Division is the most important part of a glioma case as this is where decisions are made regarding the boundaries of what is going to stay and what will go. Here, white matter anatomy becomes key as these tracts often define the boundaries of cuts, and conversely, provide the greatest source of risk to the patient’s function.
I have found that planning the dividing cut often is what defines planning the craniotomy, planning the mapping, and ultimately planning realistic goals of the surgery. One does not need to take the craniotomy up to the temporal tip, but the cut should be front and center in your field.
We keep the patient fully awake and performing functional testing throughout the duration of the division. As described in previous chapters, the test(s) of choice are tailored to the location at risk, and the patient’s functional capacity for the domain at risk. Ideally, the patient should be performing at least two simultaneous tasks using different functional systems at the same time. By splitting attention and cognitive resources between two systems, the actions are more sensitive to any interference, and thus usually providing some warning prior to performing an irreversible injury to critical portions of the white matter pathways. I think the mechanism for this involves mechanical stress on the neighboring tissue which is increasingly transmitted to the important fibers in the adjacent layer causing the test to fail prior to you actually physically coming in contact with the fibers (a diagram of how I think this happens in found in Fig. 9.2). Regardless, in my experience, performing double task type monitoring gives you the warning that we have long wished neurophysiologic monitoring like motor evoked potentials would. In other words, you get one or more warnings prior to doing something you regret.
The default task we utilize is the double task described by Hughes Duffau, which involves moving the arm and naming objects. However, we have found it necessary to tailor the task to the patient and the location at risk. For example, working near the leg motor cortex requires a double task involving leg movement and not arm. When spatiotemporal function is at risk, a neglect centered task, such as target cancelation is better, and can be combined with a different task depending on the patient. One trap I have fallen into once, has taught me to periodically alternate between arm and leg double tasks when the division takes me near the border zone between arm and leg motor cortex. It is easy to get into a sense of security when the arm is working, yet you are unknowingly drifting into leg motor fibers because you aren’t actively monitoring them.
One task type which I have found is ideal for monitoring more complex cases are high level tasks such as knitting, playing an instrument, assembling a mechanical part, etc. When we first began doing it, I thought of it solely as a task of motor function, as it was unclear to me how much someone’s life would really be altered by losing guitar playing abilities in the light of a glioma. However, as I began thinking about the functional anatomy of such a task, I realized that this type of task monitors a great number of functional systems simultaneously, including attention, praxis, fine motor skills, and spatial orientation, making it a superb monitoring task, whenever possible.