17 Cases that Taught Me Lessons
Our field for years has had a problem with blunt full disclosure. This is not to say no one talks about complications at meetings or reports them in papers. To the contrary, “disaster sessions” are a common event. It is a common practice for people to show cases that went bad and talk about them. The issue is that in most cases, many people show themselves tackling massive or highly dangerous tumors and getting into trouble (“Oh, no kidding, you did a 9cm intraventricular tumor and had a complication”) or show cases when they did something impressive to salvage a bad mistake. Alternately, they demonstrate a complication which most people would acknowledge is hard to always avoid (i.e., a CSF leak), or something so rare that its interesting, but does not help people learn how to avoid flawed thinking and bad plans, which are at the root of a plurality of failures in neurosurgery.
It is all good and well to know that bad tumors have a lot of risk, and that master surgeons can try some impressive things to try to bail themselves out. It’s just not that educational. What’s often more useful are a blunt disclosure of true mistakes (i.e., real screw-ups). Cases where we often cringed looking at a post-op scan with a poor resection, or never forgot something we did that hurt a patient that we would never do again knowing what we know now. In other words, these are the real learning curve of experienced surgeons.
While the fact that tackling a 9 cm tumor or bypassing the carotid in an emergency allow people to admit that they have complications without losing face, the true screw-ups are the pearls we hang onto that keep us out of trouble. It’s hard to publically show ourselves screwing up, especially if you wish to be seen as a master. Perhaps there are people who have reached the point where they do not make mistakes like that, but having worked with several outstanding surgeons, and talked to many more, and their trainees, I am highly skeptical that those people exist, or they may have a practice limited to one of two specific cases. At minimum, they probably have many mistakes that others could learn from if they were studied objectively.
At its core, I have tried to keep this book as honest as possible. While the premise of the book is that with better techniques, and a better understanding of the organization of the cerebrum we can do better glioma surgery and address tumors which we have not considered trying before, I have tried to always point out limits in my own knowledge, limits to our current techniques, and the importance of defining reasonable goals. We cannot defeat everyone’s glioma, and that is not the claim of the book, just that we should try harder than we have been in the past.
I also have littered this book with my own opinions and ideas, often without a case series or citation. Evidence-based medicine is an important goal; however, in many cases it is beginning to lobotomize doctors, as not all questions can reasonably be answered by a randomized control trial, and not all randomized trials provide meaningful answers. For example, one shouldn’t wait until they have a statically meaningful case series before concluding that deviating into the midbrain reticular activating system during a deep glioma resection is a bad idea. Opinions do not have the mandate.
This chapter is a frank and honest look into my learning curve told through very memorable cases. Given that many of the tumors in this book are not ones that very many surgeons that I know take on regularly, some of this learning curve involved figuring out things that were not in books, or taught at courses, or in some cases not even something you could find someone and ask for advice about. Other parts of the learning curve just involved making stupid mistakes in simple cases (that usually didn’t seem stupid at the time), or finding out the hard way that the brain is more complex than I previously gave it credit for. Other parts are fascinating cases which went well, but demonstrate an idea I have never thought of prior to that case.
17.2 Lesson One
17.2.1 My First Serious Insular Glioma and the Genesis of the Third Law
After years of thinking about how to do an insular glioma, studying the anatomy, watching cases and assisting, and preparing for my opportunity to tackle one of these, I felt prepared for the large left insular glioma which began my experience.
During the case, I felt I had done an aggressive resection. At the time I decided to stop, I was convinced I had taken out an extremely big part of the deep brain core, and was impressed how clear the cavity seemed. I also began to get concerned that I was too deep in the brain and was doing something crazy. I was happy that the patient woke up well; however, as many young tumor surgeons feel with their first serious attempt at a challenging brain tumor, my impressions of my own ability with insular tumors was smashed by the reality of the postop scan, demonstrating that I had done a glorified biopsy, and left a large part of the tumor. I always cringe when I see how little I got out.
When I see a patient who has undergone surgery elsewhere for a glioma which has left a substantial portion of the tumor, I imagine that the surgeon probably had the same impression I had in that case, that they were deep in the brain, that they had removed a large amount of brain/tumor, and that they should stop before they go too far.
When I considered the reasons I failed to do an adequate resection, the conclusion I came to, which I still base my surgeries on, were 1) we cannot expect good resections based on eyeball impressions of when it is time to stop; only when we reach anatomic landmarks, or have to stop to save function, can we be sure it is time to stop, 2) gliomas are substantially bigger tumors than we believe them to be until we have been aggressively resecting them for quite a while, and that you need to continue resecting well beyond where your instinct tells you enough is, 3) a real glioma resection should be shocking to the untrained eye.
Presently, I continue to resect the insula until I reach the circular sulcus on all sides, and have removed it down until it is flush with the temporal horn in the sagittal plane. Only the IFOF and occasionally the SLF limit this resection. My failure to do this in this case, led me to believe I had cleared out the insula when in reality I had cleared out part of a volumetrically large tumor.
17.3 Lesson Two
17.3.1 Classic Teachings about the Anatomy of the Language System are Incomplete
For well over a century, most of us have believed that language arose from two main places in the brain, Broca’s and Wernicke’s areas, and as surgeons, avoidance of damaging these areas was the principle way to avoid aphasia during surgery. A careful student of neurology might note that other areas might cause speech problems of various strains, but basically, I graduated my residency with the idea that if I identified the “speech cortices” and stayed out of them, then the patient would talk afterwards. I (and I assume others in our field) didn’t pay much heed to how exactly the information left these areas to cause a patient to talk, or how relevant inputs got there. I just focused on probability maps of rates of where the language sites were likely to be in glioma patients, and based my decision about awake vs asleep on those maps.
The case in Fig. 17.2 is a recurrent low grade glioma which I compared against the maps, determined that it was safe from a speech standpoint, based on the maps, and elected to resect asleep. I was stunned to find that this patient developed a dense expressive aphasia and anomia which only partially improved in long term follow-up. Perhaps others might have considered this a higher risk speech case than I did at the time; however, it did not make sense based on my understanding of how language is organized in the human brain at that time.
I unfortunately do not have preoperative tractographic data for this patient, so I can only speculate on the cause based on a large number of cases I have done since with similar features. First, the SLF can take abnormal and unpredictable courses and end fairly far posteriorly in the temporo-occipital lobe. Second, you cannot look at a brain and its gyral pattern and know what the language network looks like and where it is exactly. It is not random (i.e., the right occipital pole is not the site of language), but you cannot with confidence look at a glioma patient and know how they are wired up. Third, the brain is more complex than models made in the 1800s described it as being, and using outdated models to do brain surgery can lead to bad outcomes that we don’t understand.
This case was one of the key cases which prompted my interest in network based models of brain function.