Abstract
Better Object Recognition and Naming Outcome with MRI-Guided Stereotactic Laser Amygdalohippocampotomy for Temporal Lobe Epilepsy
Drane DL, Loring DW, Voets NL, Price M, Ojemann JG, Willie JT, Saindane AM, Phatak V, Ivanisevic M, Millis S, Helmers SL, Miller JW, Meador KJ, Gross RE. Epilepsia 2015;56(1):101–113.
OBJECTIVES: Patients with temporal lobe epilepsy (TLE) experience significant deficits in category-related object recognition and naming following standard surgical approaches. These deficits may result from a decoupling of core processing modules (e.g., language, visual processing, and semantic memory), due to “collateral damage” to temporal regions outside the hippocampus following open surgical approaches. We predicted that stereotactic laser amygdalohippocampotomy (SLAH) would minimize such deficits because it preserves white matter pathways and neocortical regions that are critical for these cognitive processes. METHODS: Tests of naming and recognition of common nouns (Boston Naming Test) and famous persons were compared with nonparametric analyses using exact tests between a group of 19 patients with medically intractable mesial TLE undergoing SLAH (10 dominant, 9 nondominant), and a comparable series of TLE patients undergoing standard surgical approaches (n = 39) using a prospective, nonrandomized, nonblinded, parallel-group design. RESULTS: Performance declines were significantly greater for the patients with dominant TLE who were undergoing open resection versus SLAH for naming famous faces and common nouns (F = 24.3, p < 0.0001, η2 = 0.57, and F = 11.2, p < 0.001, η2 = 0.39, respectively), and for the patients with nondominant TLE undergoing open resection versus SLAH for recognizing famous faces (F = 3.9, p < 0.02, η2 = 0.19). When examined on an individual subject basis, no SLAH patients experienced any performance declines on these measures. In contrast, 32 of the 39 patients undergoing standard surgical approaches declined on one or more measures for both object types (p < 0.001, Fisher's exact test). Twenty-one of 22 left (dominant) TLE patients declined on one or both naming tasks after open resection, while 11 of 17 right (nondominant) TLE patients declined on face recognition. SIGNIFICANCE: Preliminary results suggest (1) naming and recognition functions can be spared in TLE patients undergoing SLAH, and (2) the hippocampus does not appear to be an essential component of neural networks underlying name retrieval or recognition of common objects or famous faces.
Commentary
Primum non nocere (First do no harm) is one of the bioethical principles medicine is based upon. Nevertheless, in epilepsy surgery we at times accept a neuropsychological or other neurologic deficit and trade it for a good chance for seizure freedom. It has long been known that temporal lobectomy, which to date remains the most effective treatment of temporal drug-resistant epilepsy, can be associated with a memory deficit. In particular, resection of the dominant temporal lobe can lead to significant impairment of verbal memory and object naming (1, 2). Age of seizure onset predicts the extent of postoperative verbal deficit. The extent of the deficit can be quite variable (1). Even more confusing, in some cases improvement and not worsening of cognition can be observed (1). Deficits associated with the nondominant temporal lobe resection are less conspicuous. Lately, problems with facial recognition have been reported with nondominant resections (3).
Drane et al. report the neuropsychological outcome of naming and object/face recognition in a cohort of patients who underwent traditional temporal lobectomy or selective amygdalohippocampectomy (SAH) and a cohort of patients who underwent magnetic resonance imaging guided stereotactic laser amygdalohippocampectomy (SLAH) for temporal lobe epilepsy. By placing a saline-cooled laser applicator in the mesial temporal structures, the target area is destroyed, mimicking the effects of an SAH. This procedure avoids any injury to lateral or subtemporal structures as required for traditional surgical access to the mesial temporal structures. The authors report striking differences in postoperative naming and recognition of famous faces. Patients undergoing SLAH had no deficits, whereas 82% of patients who underwent traditional surgery experienced postoperative deficits.
The authors only report on naming ability as measured with the Boston Naming Test and recognition of famous faces; they did not report on verbal memory, which is frequently measured with the California Verbal Learning test, though this is often the most affected after temporal lobe resection, thereby contributing to significant disability (1). As it is assumed that verbal memory is a hippocampal function, one could assume that there would be no difference in those measures, but confirmation with data would be helpful. Knowledge of long-term memory outcomes is also important as it has been reported that temporal lobe surgery anticipates the memory loss that would eventually occur if seizures continue (4).
The authors hypothesize that the differences in neuropsychological outcomes are related to the fact that this procedure spares the lateral temporal cortex, and they conclude that the hippocampus is not involved in the naming and recognition of famous faces. A previous study found that sparing the superior temporal gyrus with a temporal resection was not advantageous and did not avoid naming deficits (5). In a randomized trial, SAH through the Sylvian fissure resulted in similar verbal memory outcomes but slightly decreased verbal fluency with the transsylvian approach (6). Confrontation naming was not explicitly tested in this study. These results, in addition to those of the Drane et al. study, pose the following questions: What is the function of the anterior temporal neocortex in naming and how does transecting afferent and efferent pathways between the hippocampus and the lateral temporal cortex affect language function. Some information about the preoperative functionality or memory dominance of the resected tissue, including surgical pathology, may have strengthened the results.
The study is not randomized and is observational; in fact, the patients self selected whether they would undergo an SLAH or a temporal lobectomy. The follow-up times for neuropsychological testing are not entirely matched, but if this had an effect it should have affected the results negatively. Patients with temporal lobectomy were given a longer time to recover.
New surgical techniques, especially the ones that require more advanced and sophisticated technology, always seem more attractive and advantageous. The SLAH certainly and without doubt results in shorter hospital stays and less perioperative morbidity, seems much better tolerated, and is certainly less invasive than traditional surgery. However, the major question remains whether the procedure has a comparable seizure outcome as the traditional surgery as any procedure remains a balancing act between the chance of seizure freedom and neuropsychological deficits. The 6-month outcome seems comparable. The authors report Engel Class One outcomes in 57.9% of the patients who underwent SLAH (52.4% with the two previously excluded patients; those patients were excluded due to age and incomplete neuropsychological data) and 61.5% of those who underwent traditional surgery. Those numbers may seem low for an initial 6-month outcome but may reflect the fact that all this included a broad spectrum of patients and not only select populations. On average the surgical outcomes are consistent with the results of some previous studies (7). Nevertheless, it remains to seen whether long-term outcomes remain in the same range or whether the number of seizure-free patients will decrease over time. Other less-invasive surgical procedures, such as radiosurgery of the mesial temporal structures, have been suggested in the past, but over the long-term did not maintain an overwhelmingly positive seizure outcomes (8). A randomized controlled study including long-term outcomes is needed to determine the value of the procedure. However, conducting studies for already available procedures has proven to be difficult in the past (9).
The study also demonstrates that we should consider expanding our routine neuropsychological testing battery to introduce some other measures that are important in assessing temporal lobe epilepsy. This could include more extensive testing of confrontation naming and evaluation of recognition of famous faces.
The discussion about improved neuropsychological outcomes with similar seizure outcomes is reminiscent of past discussions on whether a selective amygdalohippocampectomy is more advantageous than a complete temporal lobectomy (10). It took many studies and patients to demonstrate that a larger resection is likely associated with a slightly better outcome but is also accompanied by greater neuropsychological and psychiatric risk (10). The discussion is ongoing, and reported differences were certainly not as striking as reported in the Drane et al. study (11). Individualized approaches and procedures have been suggested (10). Magnetic resonance imaging–guided SLAH could be one of the individualized approaches. If long-term seizure outcomes remain similar and are confirmed in a randomized controlled trial, we certainly should give patients a chance of better neuropsychological outcome with a far less invasive procedure.