The most common use of lasers in the nervous system includes the use of argon, ruby, and CO₂ lasers to perform cutting and vaporization, and to achieve tissue hemostasis [26, 27]. The first use of interstitial laser photochemotherapy to nervous tissue, however, traces back to 1987, when Powers et al. [28] described the use of mitochondria-specific rhodamine-123 dye and a 150 mW blue-green argon light for treatment of flank and intracerebral rat gliomas, resulting in progressive central tumor shrinkage and necrosis over time, but eventual tumor recurrence at its periphery, where light penetration was limited by distance [28]. Further studies investigated effects of variations in wavelength and types of thermal energy emitting tips, including the end emitting bare tip versus diffusion emitting sapphire tip optical fiber probes [29] (Fig. 7.2), showing that elevations in tissue temperature are related to surface area of the probe, which directly affects rate of energy delivery, and rate of temperature change near the light source affected distance of heat penetration independent of light wavelength [29]. Laser effect on tissues is also known to be dependent on tissue thermal and optical properties, including thermal conductivity, hemoglobin and fluid content, tissue density, and specific heat, in addition to laser wavelength, mode, and fluence [30].

Despite significant knowledge regarding the factors that influence laser effect on the in vitro tissues, the most significant barrier to the use of laser thermocoagulation remained the inability to predict the size of thermal lesioning in vivo. Its re-invigoration in clinical practice in the past decade is due solely to the development of MR thermometry by Frank Jolesz in 1981 [31–34]. MR thermometry is the ability to interpret MR gradient echo image changes in order to follow thermal tissue changes (Fig. 7.3). Initial feasibility work was performed in cat, dog, rabbit, and pig brains, showing good correlation between changes in gradient echo image information and histological evidence of cell death [21–25]. Further, three histologically distinct zones of post-laser regions of injury, including (a) peri-laser vaporization zone, (b) zone of coagulation necrosis, and (c) zone of peri-lesional edema, can be clearly depicted using gadolinium enhanced T1-weighted spin-echo MR images [35–39] (Fig. 7.4).

Over the past decade, the indications for the use of MR-guided LITT have significantly expanded. Treatment of glial tumors and brain metastases failing radiation and surgery remain the major indications for the use of LITT [40–42] (Fig. 7.5). Within this group also fall the cases of radiation necrosis, with which our center has had significant treatment success [43–45]. Increasingly, however, as the technology advances, LITT has expanded its use into the areas of epilepsy, particularly in patients with mesial temporal sclerosis, or lesion-based seizures [46–48], the area of neurovascular neurosurgery such as LITT use for treatment of cavernous malformations [49, 50], and for the treatment of benign tumors such as meningiomas [51, 52]. The largest advantage of LITT over the more traditional approaches is its ability to minimize the intracranial approach through the normal brain parenchyma in order to access the lesion. In addition, with a stab incision in the skin and a small drill hole in the skull, intra-operative blood loss and postoperative pain associated with the surgery are also significantly minimized. Lastly, the ability to visualize heat delivery based on real-time MRI allows the surgeon to determine adequacy of treatment in real time [35–39] (Fig. 7.3).

LITT procedures can be performed both in the operating room, if an intra-operative MRI is available, and in a diagnostic MRI suite (Fig. 7.1). Prior to starting surgery, at our facility, we spend a significant amount of time planning the laser trajectory (Fig. 7.6). Using our neuro-navigation planning software, we estimate that each laser pass will allow us to cover a cylinder of tissue of up to 25–30 mm in diameter. By overlaying each trajectory onto the lesion, one can plan the number of trajectories required to cover the lesion and the direction from which the laser is best introduced. While there is no current data to support trajectory recommendations, we have found that treatment is usually most easily achieved if the laser is brought along the long axis of the lesion with the structures most at risk at the tip of the fiber.

LITT can be performed under general anesthesia or conscious sedation along with local anesthetics. In our facility, we have access to an intra-operative MRI, and the ability to perform LITT under general anesthesia has allowed us to perform complex multiple trajectory treatments more easily. Prior to starting the procedure, the patient is medicated with 10 mg dexamethasone and antibiotics. Mannitol is not administered in order to minimize brain shift during surgery. Following the administration of general anesthesia, the patient is placed in pins so that when the head is secured to the operating table, the head holder is in neutral position. In addition, the trajectory of the laser needs to be taken into account so that as it protrudes from the head, it is directed as much as possible directly out the bore of the magnet. If this is not possible, in general, it is better for the laser fiber to be pointing off to the side than straight up in the bore. At this time, we obtain a preoperative MRI, since in our experience best results are obtained if the imaging for the procedure is performed the day of surgery. A T2-weighted MRI sequence is usually obtained first, and if the target can be well defined on this sequence, then gadolinium is not administered. If, however, the T2 sequence is insufficient, then a half dose of gadolinium is administered only. Throughout the course of treatment, repeat doses of gadolinium are required and minimizing the amount of gadolinium used initially can decrease potential overall toxicity.

Placement of the laser fiber is then guided by neuro-navigation based on this preoperative MRI scan. For most cases, a stab incision is first made in the scalp and then a twist-drill hole is made in the direction of laser fiber placement. In our experience, not only is it essential to use neuro-navigation to guide the drilling of this hole, but if the hole is not directly perpendicular to the skull, then it is important to ensure that skiving of the drill bit does not occur. In rare cases, if the laser needs to be placed through a burr hole instead of a twist-drill hole, we still recommend drilling the burr hole in the direction of laser placement; otherwise, significant difficulty can occur with having to remove extra bone around the edges of the burr hole. Once the dura is opened, it has been our practice at this time to obtain tissue biopsy, both to confirm the diagnosis and the tumor grade, as well as to obtain tissue in some cases to determine subsequent clinical trial eligibility. Following biopsy, an MR-compatible anchor bolt used for laser probe insertion is placed. The laser is then inserted to a depth determined using the preoperative MRI planning (Fig. 7.6).

At this time, a T1-weighted MPRAGE sequence is obtained with full-dose gadolinium, and in our system, drawing of the target is then performed and the location of the laser within the target is confirmed (Fig. 7.3). Laser heating is then initiated and stopped when adequate heat has been delivered to the margins of the target (Fig. 7.4). Additional ablation can be achieved by retraction, advancement, or rotation of the laser fiber. Once it is felt that lesion ablation is sufficient, T1-weighted MPRAGE images with gadolinium are obtained again to assess thermal injury margin (Fig. 7.4). If thermal injury margin is satisfactory, then the bolt is removed, and the surgical site is closed. All our patients have been admitted to the ICU for overnight observation. In addition to continuing steroids and antibiotics, patients with lesion diameters greater than 4 cm are sometimes administered mannitol overnight. Most patients can be discharged home the following day with a dexamethasone taper over the subsequent 2–3 weeks. Post-ablation MRI imaging is then obtained at 2 weeks postoperatively to help guide postoperative dexamethasone taper and then repeated at 1.5, 3, 6, and 12 months post-LITT (Fig. 7.7).

Two FDA-approved LITT systems are commercially available today in the USA. These are as follows:

The NeuroBlate system uses a 1064 nm neodymium:yttrium aluminum garnet (Nd:YAG) laser that is cooled using CO2 [53]. The system offers two probe types: a diffuse firing tip and a directional firing tip, both of which create a disk-shaped lesion circular around the probe but flat proximally and distally. The directional tip can potentially provide a larger radius of treatment with better conformity to the actual tumor shape but requires a larger cooling catheter. Directional firing seems to be more successful at lower power settings which therefore will make treatment time longer. The laser power settings range between 12 and 16 Watts, and the rapidity of lesional heating is controlled by the ratio of pulsed “time on”: “time off”; less “time off” results in faster heating. Maximum temperatures reached during LITT range from 45 to 70°. It has also been suggested that use of less power may be beneficial for lesions located in the eloquent brain and for post-LITT edema reduction [54, 55]. In addition to options for laser firing direction and power settings, the NeuroBlate system also offers flexibility in laser mounting to the skull. In addition to the anchor bolt, the laser can be held in place using a tripod in the case of salvage surgical situations, where bone may not be present at the site where the laser needs to be introduced.

The Visualase system uses a 15 W 980-nm diode laser and a diffusion-tip probe with a silicone fiberoptic core surrounded by a cooling sheath, and creates an ellipse-shaped LITT lesion centered at the probe tip (Fig. 7.2). Lesion size is time-sensitive due to its ability for rapid tissue heating [48, 56, 57]. Maximum LITT temperatures reach into the 80–90° range. Only the zone of irreversible damage as calculated by the Arrhenius equation is used with this system [39, 58]. Given the hotter temperatures, the laser “on time” tends to be shorter with this system. In our experience, while this system lacks directional control, the heat does tend to conform to the shape of the lesion well in most cases due to the difference in composition of tumor versus surrounding edematous brain.

The innovation that is unique to the NeuroBlate system, however, is an MR-compatible robotic driver that is attached to the laser and allows control of depth and rotational movement of the laser from the control room, thus minimizing the need for the surgeon to manually manipulate the laser fiber and to be able to accurately move the laser to the needed depth and direction without the surgeon having to put their head into the magnet bore and move the fiber an estimated amount, or withdraw the magnet each time a change in laser position is needed.

With the laser turned on, both systems designated planes of imaging. For NeuroBlate, monitoring is performed in 3 parallel planes whereas with the Visualase system, monitoring is performed in 2 orthogonal planes. Both then use the Arrhenius equation [39, 58] to calculate the effect of the heat on the cells surrounding the laser over time. The Visualase system only calculates a single cell kill line compared with the NeuroBlate system that allows calculation of zones of potentially reversible (protein denaturation) versus irreversible (cell kill) injury. In our experience, while the initial heating pattern of both the diffuse tip and the directional firing tip can start off equivalent, we have clearly seen an advantage in using the directional tip in pushing heat out beyond a 15 mm radius in a single direction. Given that many high-grade gliomas lesions are not round but rather quite irregular, this system’s versatility seems advantageous in treating these lesions. What is much needed is a planning prediction model that helps the surgeon decide where best to place the laser fiber based on predicted heat diffusion patterns. With the current imaging versatility and robotic driving system within the magnet bore, development of this technology should hopefully ultimately allow for automation of LITT treatment delivery in the future.

A comprehensive review of the literature describing the results of LITT use for HGGs can be found in the article by Hawasli et al. [59]. The number of studies that describe the risks and benefits of potential LITT use in treatment of patients with HGGs published since the 1990s is less than 20 (Table 7.1 [60–73]). Most of these studies are case series, and there are no large randomized controlled trials available to date.

In summary, a total of 230 patients with 161 high-grade LITT-treated lesions have been described; 55 (34.2%) of these lesions were diagnosed as grade III gliomas and 106 (65.8%) as GBMs, with 40 (24.8%) being newly diagnosed and 107 (66.5%) being recurrent lesions. A wide variation of tumor sizes treated is reported, with diameters of up to 5.5 cm in a few studies, but mostly ranging from 0.4 to 68.9 cm³ [41, 42, 60–73]. Only 21 (13%) of these lesions were in locations where LITT would be considered a first-line approach for treatment, including the insula, corpus callosum, basal ganglia, or the thalamus. What makes it difficult to interpret outcome following LITT is a similarly large range of percentages of lesion heat coverage by LITT. Cell kill coverage with LITT can range from 100% of lesions usually in those less than 3.5 cm in diameter, to frequently quoted ranges of 78–98% (but as low as 28%) in studies that include larger and more complex lesions using a variety (high versus low) of LITT energies [38, 41, 42, 55, 60–69]. This analysis is further complicated by the use of multiple trajectories for lesioning, which in some studies amounted to nearly half of all treated patients [42]. Reported procedural length times can therefore also range significantly from 2 to 16 h.

In studies comparing LITT-treated grade III tumors with LITT-treated GBMs, progression-free survival (PFS) in patients with LITT-treated GBMs was reported as 3–4 months, while PFS in patients with LITT-treated grade III tumors was 10–17 months, with corresponding OS of 7–9 months in patients with LITT-treated GBMs and 30 months in patients with LITT-treated grade III tumors [38, 65, 66] (Table 7.2). Review of other studies that predominantly included patients with GBMs similarly showed average PFS of 4 months and OS of 8.7 months after LITT treatment [41, 60–63]. When looking specifically at LITT use in newly diagnosed HGGs, results are limited with regard to number of patients included across these studies (24.8%, or 40/161), lack of subgroup analysis, inconsistent use of surgical interventions in addition to LITT, and relatively short follow-up times [41, 42, 66]. Studies that predominantly included patients with LITT-treated recurrent GBMs contain significantly larger numbers of patients and reported OS of 6.9–15 months post-LITT [60, 62].