Visual field defects (VFDs) are a common side effect of temporal lobe resections for patients with medically intractable mesial temporal lobe epilepsy (MTLE). The frequency and distribution of VFDs after resection ranges from a low of 52% (Tecoma et al., 1993) to a high of nearly 100% (Hughes et al., 1999). More recent studies attempt to correlate VFDs with surgical technique. Nilsson et al. (2004) found that more restricted resections that spared the superior most aspect of the superior temporal gyrus yielded less VFD. Mengesha et al. (2009) found that points along the horizontal meridian of the visual fields tended to be spared with selective amygdalohippocampectomy compared to standard anterior temporal lobectomy.
Gamma knife radiosurgery (RS) may be an alternative to open resection in the treatment of MTLE (Quigg et al., 2012). Seizure remission rates in the U.S. Multicenter Pilot Study are comparable to that historically seen after temporal lobe resection (Barbaro et al., 2009). Additionally, VFDs following radiosurgery for MTLE approximate that seen after standard surgery, with the European prospective trial reporting an incidence of 50% (Régis et al., 2004). RS offers a unique opportunity to study VFD in temporal lobe surgery. Unlike open resection, RS targets deep structures without involvement of surrounding tissue. The radiosurgical target in the U.S. Multicenter Pilot Study was standardized in terms of specific anatomic targets with limits of dose, volume, and safety factors; therefore, neurosurgical variations—a relatively underreported variable in epilepsy surgery—were minimized.
The mechanism of the anticonvulsant effects of RS remains controversial (Quigg et al., 2012). Animal models of limbic epilepsy show that improvement in seizure frequency is not dependent on destruction of the epileptic focus, implying a neuromodulatory effect of RS (Chen et al., 2001). However, spectroscopic data of human subjects suggest that a radiodestructive (rather than neuromodulatory) lesion is important in seizure-free patients (Chang et al., 2010).
In this study, we examined the incidence and severity of VFD in patients who underwent RS for intractable MTLE. Our hypothesis was that the extent of the VFD was correlated to radiosurgical dose or outcome. We also reasoned that examination of the visual fields would provide evidence of destruction to the optic radiations, which would in turn provide evidence of a destructive antiepileptic lesion. Finally, the type of VFD would provide evidence for or against damage to nearby structures (optic nerve or chiasm) related to radiation exposure.
- Top of page
Postoperative visual field testing was available for 24 of the 30 patients enrolled in the trial. Three subjects did not complete the 36-month study: one subject was lost to follow-up; one required urgent temporal lobe surgery 15 months following RS (high dose); and one subject was followed for 24 months, was not seizure-free, and requested temporal lobectomy (low dose). These three patients had no postoperative VF testing. The remaining three patients with no VF testing completed the trial but did not comply with testing.
None of the remaining 24 subjects had abnormal bedside pupillary, visual acuity, or visual field abnormalities, and no patient complained of subjective visual changes. Blinded examination of perimetry results determined that 15 patients (62.5%) developed VFD, all consisting of homonymous superior quadrantanopsias. None of the VFDs was consistent with injury to the optic nerve or chiasm. These interpretations of VFD provided by perimetry correlated significantly with VFDR (mean ± standard deviation: mean VFDR in patients with interpreted VFD = 0.64 ± 0.08 vs. those without VFD = 0.95 ± 0.06, Mann-Whitney U test p-value = 0.0001).
There was no correlation between age and VFDR (Fig. 2A). Neither the dose category (low or high, Fig. 2B) nor the isodose volume into which it was delivered (Fig. 2C) varied with VFDR. Neither RIC (Fig. 2D) nor QOLIE-10 scores (Fig. 2E) varied with VFDR. Finally, patients with seizure remission had smaller (more severe) VFDR compared to patients whose seizures continued (Fig. 2F).
Figure 2. VFDR compared to (A) patient age, (B) dose of radiation, (C) 50% of isodose volume, (D) volume of T2-weighted radiation induced change on MRI, (E) quality of life scale, (F) seizure remission.
Download figure to PowerPoint
- Top of page
The main finding of this prospective study of VFD following RS for MTLE was that the incidence and severity of VFD after RS were similar to those reported for open temporal lobe resection. We conclude that the RS protocol used in the Barbaro et al. (2009) U.S. Multicenter Pilot Study is not only appropriate for its anticonvulsant effect, but the restrictions placed on dose-volume and ophthalmic nerve exposures are effective in preventing excessive morbidity to the visual system. Secondly, the severity of VFD correlated with seizure remission, indicating that the optic radiations lay within the destructive zone of RS. We infer an important anticonvulsant mechanism of RS lies in its focal destruction of tissue within the epileptogenic limbic circuit (Bertram, 2009).
In this study, postoperative VF were measured using automated perimetry methods. The VFDR enables normalization among these methods as it allows comparison of VF abnormalities from side to side in an intrapatient comparison. Use of the VFDR to quantify VFD is also important in that it provides a quantified method of documenting the severity of VFD following surgery involving portions of the visual system. This approach avoids problems with interrater reliability and allows for multiple pools of data to be compared. In addition, the VFDR was validated by comparing it to clinical examination of VF patterns provided by standard VF testing.
The present study is the first to show that quality of life is not adversely affected by the presence of postsurgical VFD when defects are limited to the upper quadrant. This measurement effect agrees with our clinical experience that patients after standard temporal lobe surgery rarely complain of visual symptoms. The lack of clinical findings and the fact that no patient self-reported a visual loss suggests that the morbidity of postoperative VFD when confined to the superior quadrant of vision is minimal, especially compared to the benefits of successful epilepsy surgery. The study emphasizes the value of quantitative visual field measurements, given the absence of abnormalities by standard clinical examination.
The severity of VFD after epilepsy surgery, regardless of surgical technique, is important given that vision may independently limit the ability to drive. Although data on driving status was not collected during the study, the quality of life measure used (the QOLIE-10) includes a scale on improvements in driving. It is unknown whether any of the patients who were seizure-free following the procedure were denied a driving license. Although visual acuity limits are present in all countries and states, requirements for visual fields vary widely and are empirically derived. In the European Union, binocular field requirements generally require ≥120° of intact vision across the 180° hemifield, with no encroachment of vision within 20–30° of the horizontal meridian. Whereas some U.S. states specify no requirements, the majority define limits of 110–140° (Bron et al., 2010). Only one study specifically evaluated driving eligibility related to postsurgical VFD (Manji & Plant, 2000). Of a sample of 24 patients, 11–13 patients (depending on perimetry technique) had “failing” VFDs according to the European standard after temporal lobectomy; 3 (approximately 30%) of those patients could not drive despite seizure freedom (Manji & Plant, 2000). Under the European standard (and converting between VFD definitions), we estimate that 10 (43%) of 23 of our patients would have VFD ≥120 degrees (approximately VFDR < 0.67); all these patients were seizure-free. The ongoing ROSE Trial (Radiosurgery or Open Surgery for Epilepsy), a randomized comparison of temporal lobectomy versus RS, may be able to better provide direct and consistent comparisons across surgical techniques.
Limitations of this study included the small sample size, which was difficult to avoid as RS for MTLE is not widely performed. However, this study includes the largest group of patients treated with RS for MTLE reported thus far.
No association was seen between dose/isodose volume and VFD. We have no comparisons in other studies of RS; the European prospective study did not evaluate dose effects (Régis et al., 2004).
Studies of open surgery show small effects of surgical technique on postoperative VFD. For example, selective amygdalohippocampectomy, a smaller volume technique than anterior temporal lobectomy, tends to spare vision near the horizontal meridian (Mengesha et al., 2009). Subtle differences were found using slightly different surgical techniques (Nilsson et al., 2004). The discrepancies between the present study and open surgery reports attest that “surgical volumes” may not be comparable between RS and resection.
No association was seen between presence of VFD and volume of T2 lesions on MRI at 12 months postoperatively. This emphasizes that the RIC seen in terms of T2 lesion size is transient and has no long-lasting effect on visual function.
One implication of our findings is that it comments on the possible mechanisms of RS in epilepsy surgery. Whereas experimental models of epilepsy and limited human data suggest a neuromodulatory effect (see reviews: Régis et al., 2010; Quigg et al., 2012), in the present study the correlation between radiation dose and VFD, as well as improved seizure freedom with more severe VFD, would support permanent tissue destruction as the mechanism of effective RS. The present findings support the spectroscopic findings (magnetic resonance spectroscopy, MRS) performed by our group that found evidence of ischemia within the radiosurgical target (Chang et al., 2010). An alternative argument would be that the effect on seizure remission is neuromodulatory while the effect on visual fibers is destructive. Although possible, the more parsimonious explanation is that the radiosurgical effect is similar, that is destructive, to all tissues in the treatment volume.
In summary, this study found that radiation dose in RS was not significantly associated with severity of VFD and confirmed that the incidence and severity of VFD were similar to that of open surgery for MTLE. Based on these findings, we speculate that the mechanism of RS involves some degree of tissue damage and is not confined entirely to changes in neuromodulation. In addition, existing tissue tolerance limits, as used in this protocol, are sufficient to protect other important components of the visual system.