A Comparison of Clinical Outcomes for Subependymal Giant Cell Astrocytomas Treated with Laser Interstitial Thermal Therapy, Open Surgical Resection, and mTOR Inhibitors

Tuberous sclerosis complex (TSC) is a multi-organ genetic condition involving the development of benign tumors (hamartomas) in brain, heart, kidney, liver, and lung tissue [1]. TSC displays autosomal dominant inheritance and is characterized by dysregulation of the mammalian target of rapamycin (mTOR) signaling pathway [2]. This is typically due to mutations in TSC1 or TSC2, leading to constitutive activation of mTOR with subsequent upregulation of cell growth and proliferation [2]. One common tumor is the subependymal giant cell astrocytoma (SEGA), which can affect between 5 and 24% of patients with TSC [1, 2]. These tumors commonly occur within the first 2 decades of life, arise from benign subependymal nodules near the foramen of Monroe, and may be either unilateral or bilateral [1]. SEGAs are WHO grade 1 lesions composed of mixed glioneuronal tissue [2]. Despite their typically slow growth pattern, their proximity to the ventricular system may lead to life-threatening obstructive hydrocephalus [2].

Historically, management of SEGAs has been limited to surgical resection, often via transcortical/transcallosal routes [3, 4]. While potentially curative and often associated with satisfactory neurological outcome, these substantial surgeries in children have been shown to have significant risk of long-term morbidity or mortality. Complication rates as high as 49% have been reported [5]. Additionally, many lesions are not amenable to gross total resection due to size and location adjacent to critical venous or deep gray matter structures, and residual tumor or recurrence following resection has been reported in up to 34% of patients. Complication rates increase in tandem with tumor size [5, 6], although more recent reviews of surgical outcomes with updated techniques have shown improved morbidity and mortality [7]. In patients presenting with acute clinical deterioration due to acute hydrocephalus or intratumoral hemorrhage, surgery remains the only expedient option [3].

More recently, there has been substantial progress in the medical management of SEGAs due to the identification of mTOR as a key protein kinase driving development of these tumors, with subsequent deployment of mTOR inhibitors such as rapamycin (sirolimus) and everolimus. The first reported use of sirolimus in TSC patients in 2006 demonstrated significant reduction in SEGA volume (46–63%), with further studies supporting the utility of these inhibitors [8, 9]. In symptomatic patients, medical treatment with mTOR inhibitors is indicated in cases of episodic headache with mild ventriculomegaly, in the absence of papilledema or cranial nerve dysfunction [3]. Though surgery is traditionally reserved for patients with acute hydrocephalus, mTOR inhibitor therapy alone has also been used in these cases with reported rapid tumor shrinkage and improvement in symptoms [10]. Medical treatment is favored when tumors are recurrent, bilateral, or when gross total resection is not feasible as residual tumors frequently continue to grow [3].

One promising new modality in the management of SEGAs is laser ablation. Laser interstitial thermal therapy (LITT) is a minimally invasive procedure that involves the use of a laser system in conjunction with magnetic resonance imaging (MRI) to form nonionizing radiation as a coherent beam of light energy to thermally ablate specific targets. First introduced as a therapeutic modality in the 1990s, LITT has been used to treat a wide variety of intracranial pathologies, including gliomas, metastases, and epileptogenic foci, such as those occurring in mesial temporal sclerosis [11‒13]. More recently, LITT has been shown to have great potential in the treatment of SEGAs, with reports of 70% or greater reduction in tumor burden [14, 15]. LITT is currently being used in various centers, including our own, for management of SEGAs [14‒16].

The purpose of our study was to identify patients with SEGAs treated at our institution by open surgical resection, LITT, mTOR medical management, or a combination of these modalities, determine the clinical outcomes associated with their treatment modality, and compare clinical outcomes across modalities. The primary study outcome was tumor volume at most recent follow up compared with tumor volume at initiation of treatment. Secondary outcomes were clinical complications resulting from the different treatment modalities. To our knowledge, this is the first study to compare clinical outcomes in LITT for SEGA management with both surgical resection and mTOR treatment.

This study protocol was reviewed and approved by the Seattle Children’s Hospital Institutional Review Board, approval number STUDY00003148. Our Institutional Review Board did not require written informed consent for this retrospective research, which is compliant with institutional requirements.

After Institutional Review Board approval, retrospective chart review was performed to identify patients with SEGAs treated at our institution from January 2010 to March 2021 via open surgical resection, LITT, medical management via mTOR inhibitors, or a combination of these modalities. Pediatric patients were included if they had a diagnosis of SEGA, received treatment for their SEGA at Seattle Children’s Hospital, and had both baseline MRI prior to treatment initiation and at least 1 follow-up MRI. Medical treatment decisions were made by a multidisciplinary team including subspecialists in pediatric neurosurgery, oncology, and neurology. Decisions on which surgical modality used were at the discretion of the 2 treating neurosurgeons, while the decision to initiate or continue medical therapy was at the discretion of each patient’s neurologist/oncologist in accord with patient and family preferences. Demographic and clinical data were recorded, including age at treatment initiation, gender, TSC diagnosis confirmed with genetic testing, duration of follow-up, and presenting symptoms.

Patients undergoing surgical resection were treated by 1 of 2 neurosurgeons (JGO and JSH). Open resection consisted of a standard craniotomy and microsurgical interhemispheric transcallosal approach for resection with frameless stereotactic navigation (Medtronic Stealth). All patients underwent external ventricular drain (EVD) placement at the time of surgery and were monitored post-operatively in the pediatric intensive care unit. EVD was weaned prior to discharge. One patient in the open group required conversion to a VP shunt after failing multiple EVD weaning attempts.

Patients who underwent LITT treatment were treated using the Visualase system (Medtronic, Minneapolis, MN, USA). The laser fiber was placed using stereotactic guidance in the operating room. Target and entry coordinates were set using the preoperative MRI such that the catheter traversed the long axis of the lesion without entry into the ventricle or passage through a vascular structure. Patients were then fixed to a rigid skull clamp and registered using Framelink software (Medtronic) or using the ROSA ONE robotic platform (Zimmer Biomet Warsaw IN, USA). Following fiber placement, the patient was taken to the MRI suite. Laser ablation was then performed using the Visualase workstation under the direction of the attending surgeon. MR thermometry on a Siemens 3T MRI scanner enabled real-time visualization of the ablation.

The laser parameters for each ablation are listed in Table 1. Ablation power and duration depend on the volume of tumor and its proximity to critical structures. The Visualase workstation enables designation of critical temperature limits (detected by thermometry) to prevent unwanted thermal injury to surrounding structures. These limits are set by the surgeon prior to ablation, and if the target temperature is detected in that region, the laser is automatically shut off. As such, ablations are typically conducted in serial intervals shorter than 3 min. The total number of “burns” is determined by the surgeon and is guided by the ablation maps provided by the Visualase software (Fig. 1). Specifically, the general objective is to achieve diffuse cell death and necrosis within the entire volume of the lesion (corresponding to temperatures above 50°C). This is visually represented on the Visualase interface by thermal damage estimate mapping, which colors orange the pixels overlying any area that has exceeded the user-set ablation temperature (Fig. 1). Once the entire lesion is covered by the thermal damage estimate map, or once the lesion is maximally ablated given limitations of surrounding structures, the procedure is stopped. We use a temperature of 42°C for critical structures, such as the optic apparatus, brainstem, and hypothalamus.

Following ablation, post-procedural MRI sequences are obtained to demonstrate the final volume of injury (using T1 with contrast, DWI, and T2 FLAIR sequences), as well as identify any areas of hemorrhage (SWI or GRE sequence). The fiber is then removed and the patient observed overnight.

Medical mTOR inhibitor therapy with everolimus was prescribed according to FDA labeling and a standard institutional protocol for therapy and side effect monitoring. All patients were treated with everolimus at our institution. Dosing was based on patient body surface area at 5 mg/m² initial dose. Dose adjustments were made for toxicity and drug levels were monitored (goal 5–15 ng/mL concentration) and assessed with periodic drug level testing throughout follow-up.

All follow-up imaging was obtained on a Siemens 3T MRI scanner with pre- and post-gadolinium contrast sequences. T1 pre- and post-contrast, T2, FLAIR, and DWI imaging sequences were obtained for all patients.

The primary study outcome was tumor volume reduction following treatment. Tumor volumes were calculated from post-gadolinium contrast MRI sequences obtained at initiation of treatment and at most recent cranial imaging using the volumetric formula of (4/3)*π*r¹*r²*r³ where r¹⁻³ are the radii of the lesion in the axial, coronal, and sagittal planes, respectively. Secondary outcomes included need for shunt placement for CSF diversion as part of disease treatment and occurrence of complications following surgical or medical therapy. Follow-up duration was defined as the time in months between treatment initiation and most recent cranial imaging.

Demographic variables were collected. Mean follow-up duration differences between groups were analyzed using Kruskal-Wallis nonparametric testing due to small sample size of each group and non-normal distributions of all groups as determined by pre-analysis Shapiro-Wilk testing. Differences in tumor volume comparing cranial imaging prior to treatment and at most recent follow-up imaging were analyzed using Kruskal Wallis. The 1 patient who underwent all three modalities was removed from follow-up analysis due to multiple treatment start dates. Additionally, this patient was removed from the surgical treatment group analysis due to incomplete preoperative imaging data and from the mTOR inhibitor treatment group analysis due to potential confounding from prior LITT treatment. Significance was assessed at p < 0.05. Statistical analysis was performed using R statistical software version 4.0.5 (R Foundation for Statistical Computing, Vienna, Austria).

A total of 12 patients identified as having SEGAs were followed at our institution during the duration of the study period (Table 2). One patient received the majority of SEGA treatment at other institutions and was excluded from the analyses. Of the remaining 11 patients, 3 underwent LITT (27%), 3 underwent surgical resection (27%), 4 were treated with mTOR inhibitors only (36%), and 1 patient was treated with all three modalities (9%). Within the surgical groups, 2 patients had previously received mTOR inhibitor therapy, but this was discontinued due to side effects. Mean age at initiation of treatment was 5.6 years, and 6 of 11 were male (54%). The majority of patients (73%) presented with unilateral lesions, with only 3 showing bilateral SEGAs at presentation. All patients’ lesions were situated adjacent to the foramen of Monroe within the frontal lobe(s). All patients in the cohort presented with seizures as the primary symptom of their underlying TSC. All 3 patients who underwent surgical resection had evidence of hydrocephalus on their cranial imaging prior to resection; no patients in the other groups had evidence of hydrocephalus before or after therapy initiation. Of those with hydrocephalus, 1 reported headache as a significant symptom, but no patients were reported to have vision changes or other significant neurologic changes. All patients met 2021 diagnostic criteria for TSC [17] and 54% of patients had a diagnosis of TSC confirmed with genetic testing.

Mean follow-up duration for the entire cohort was 23 months (range 3–78 months). The patient that received all 3 modalities of therapy was excluded from follow-up analysis due to differing start dates for each treatment. By modality, mean follow-up duration was 9 months for LITT (range 4–19 months), 52 months for surgical resection (range 11–78 months), and 11.5 months following initiation of mTOR inhibitors (range 3–27 months). Within the limitations imposed by our small sample sizes, there was no significant difference in follow-up duration between groups (p = 0.223).

For patients who underwent open surgical resection (including the patient who underwent all 3 procedures), 2 had gross total resections and 2 had subtotal resections. Volumetric analysis showed no statistically significant difference in absolute tumor volume change between treatment groups (p = 0.156). For relative volume change (Table 3), there was no significant difference between the three treatment groups with Kruskal-Wallis analysis (p = 0.0513). Two patients developed recurrence or demonstrated progression following treatment, including 1 patient in the mTOR group who discontinued everolimus due to side effects, and the patient who underwent all 3 modalities to treat bilateral lesions. There were no serious complications or post-operative neurologic deficits observed in the surgical groups. No patients in the LITT or mTOR groups required CSF diversion. During follow-up, 1 patient in the open resection cohort ultimately required permanent CSF diversion via shunt placement. Two patients, (1 resection, 1 LITT) discontinued everolimus due to complications (stomatitis in one case, pneumonitis in the other), while 1 required dose reduction due to everolimus-associated stomatitis. One patient in the medical treatment group had issues with everolimus compliance due to cost and access to care.

Management of SEGAs in patients with TSC has evolved considerably over the last decade as minimally invasive approaches and medical therapies have proven effective. While the literature on institutional experiences using individual treatment modalities for TSC SEGAs offers evidence for efficacy, few reports directly compare them. We sought to better understand the comparative benefits of three major approaches to the management of SEGAs in patients with TSC: microsurgical resection, LITT, and mTOR inhibitor therapy.

Laser ablation treatment for intracranial pathologies has become increasingly utilized within neurosurgery, from tumor treatment of deep lesions that would have unacceptably high morbidity with open resection [18, 19], to management of epilepsy, such as in mesial temporal sclerosis [20]. Use of LITT has already been reported in TSC patients for treatment of SEGAs [15] and for management of cortical tubers causing refractory epilepsy [21]. To the best of our knowledge, our series is the first to compare LITT treatment with other current approaches to SEGA management. Our data indicate that LITT may be an effective option to control tumor growth in select cases compared with the current standard of care and was well tolerated in terms of morbidity or complications; however, more data with longer follow-up duration is needed to support this.

Our results are consistent with prior reported SEGA volume reductions of approximately 68% with LITT therapy alone [15]. Though there was no significant difference in relative reduction between all three groups (p = 0.0513), our results did show larger average volume reduction in the surgical group (mean 90%), compared with LITT (48%). However, given the limited sample size of our cohort, this result is lacking statistical power. Nevertheless, this trend can be attributed to the greater number of patients who had gross total resection (2) within the resection cohort as this skewed the relative tumor reduction higher for this group. This represents an obvious advantage of open resection over LITT and medical management, though as discussed, this is balanced by the invasiveness of this option.

While other authors have identified complications related to LITT treatment for SEGAs [16], none of the patients in our LITT group had reported complications. Similarly, none of the patients in our mTOR therapy-only group reported related complications. However, it should be noted that 3 of our other patients (1 in the surgical group, 1 in the LITT group, and 1 treated with all 3 modalities) required discontinuation or dose reduction of everolimus due to side effects including severe stomatitis and pneumonitis. In addition, 1 patient in the medical cohort had substantial issues with everolimus compliance due to care access and medication cost.

In addition to LITT, treatment options for SEGAs include surgical resection, mTOR inhibitors, and radiotherapy. Recently, use of endoscopic techniques in SEGA resection has shown promise as an additional more minimally invasive surgical approach, though this is frequently limited by tumor size or approach trajectory to certain select cases [22]. Stereotactic radiosurgery has been proposed for management of small or residual SEGAs following resection, but limited data exist at this time, and it has not been described as first-line therapy [23]. Additionally, with stereotactic radiosurgery, reduction of tumor size is not immediate [4] and development of malignant transformation after adjuvant radiation therapy has been reported in some cases [24]. Presently, surgical resection is used to target lesions that demonstrate serial growth over time or in patients that present with symptomatic obstructive hydrocephalus from their tumors [4, 25]. Surgical resection represents an opportunity to fully eliminate a SEGA lesion, but due to location and the required surgical approach, gross total resection is not always feasible. Surgical resection of SEGAs potentially carries high risk of morbidity [1, 3, 5, 6], with some series showing resection of tumors >3 cm associated with 67% rate of surgical complications [5], although more recent studies suggest improved outcomes as surgical techniques have evolved [7].

Medical therapy with mTOR inhibitors fills an important gap as treatment for patients with growing residual tumors, tumors that are unsafe to resect, or in patients who are not surgical candidates [25]. However, as our results also indicate, these drugs are not always well tolerated [26]. Additionally, lesions frequently continue to grow following therapy discontinuation [4], leading to the need for prolonged or even life-long therapy which can be a substantial burden to patients. Reports of successful mTOR inhibitor use in patients presenting with large lesions with acute obstructive hydrocephalus highlight the need for clear and safe management guidelines for medical therapy for SEGAs [10].

In 2012, a European consensus meeting convened in Rome discussed treatment approaches for SEGA. The recommendations included surgical indications, including for patients with symptomatic signs of increased intracranial pressure or growth of tumor tissue on serial imaging, with or without ventricular enlargement. Recommendations of first-line treatment with mTOR inhibitors were for patients over 3 years of age with contraindications for surgery or anesthesia, or if there was risk of subtotal resection [1]. Given emerging data on the safety and effectiveness of LITT in the treatment of SEGAs, it will be desirable to update these recommendations to include this treatment modality.

At our institution, multiple factors are considered in decisions about the best modality for treatment of SEGAs in a particular patient (Fig. 2). Size, specific location of tumor, and patient presentation are the most relevant issues when deciding whether LITT may be a successful option. Large tumors (>3 cm) may not be able to be completely ablated with a single catheter [17], making surgery longer and potentially increasing anesthetic and other surgical risk. Tumors situated squarely in the foramen of Monroe may not always be appropriate targets for LITT as post-ablation edema may result in obstructive hydrocephalus [14]. In these cases, concurrent septostomy or EVD placement can be planned ahead of surgery to control for this possibility [15], though none of the patients in this cohort required these procedures at the time of LITT therapy.

SEGAs are slow growing lesions and for this reason may grow quite large and cause significant CSF obstruction before patients present [1]. Patient condition at presentation is taken into consideration above all else. For patients presenting with symptoms of acute obstruction (headache, nausea, vomiting, papilledema on ophthalmologic exam, depressed mental status), open surgical resection is preferred over LITT. For those few patients who present with asymptomatic progressive ventricular enlargement, our treating neurosurgeons would in general advocate for open resection with concurrent EVD placement and conversion to VP shunt if deemed necessary, though none of the patients in our cohort fit this clinical picture. In general, age is only considered as a factor if the treating surgeon or anesthesiologist felt that a patient was too young to safely tolerate a procedure, which was not the case in any of our cohort. In our experience, patients are usually old enough to tolerate either medical or surgical interventions at the time of their initial presentation.

Many of the patients undergoing surgical resection at our institution have already failed medical therapy or found it intolerable, as was the case for multiple patients in our cohort. While mTOR inhibitors have shown great efficacy in the literature, they require long-term administration, with regrowth on discontinuation [15]. This increases cost, risk of adverse reactions, and ultimately patient burden. Frequently, patients and parents request surgical options given these drawbacks to mTOR therapy, and this preference is always considered during our treatment decisions.

If patients undergoing resection have residual tumor, further clinical and imaging surveillance is warranted as these frequently continue to grow. In this case, mTOR therapy can be considered if patients have not previously failed it, or patients can be considered for LITT. In patients presenting with smaller or bilateral lesions, upfront LITT in combination with mTOR therapy is typically preferred. All patients who cannot continue mTOR therapy due to side effects or cost should be considered for LITT treatment.

This study has several limitations, the most critical being the small sample size of treatment groups. As such, our study is not sufficiently powered to detect statistical differences at our stated p value, and the generalizability of our data is limited. Nevertheless, statistical analysis with Kruskal-Wallis testing was included as a way to quickly describe overall data trends within our sample. The small patient sample steered us toward the use of nonparametric tests for our analysis. Though robust, the Kruskal-Wallis test (the nonparametric equivalent of ANOVA) requires transformation of data into ranked results, reducing the granularity of our analysis. This test is traditionally performed with randomized samples; however, this was not possible with our retrospective analysis. Regardless, this test is preferred in the setting of small sample size and non-normal data distributions, as in the case of our subgroups.

The number of treating surgeons (2) introduces some bias into clinical decision-making between patients. The retrospective study design is limited by selection bias and incomplete data in some cases. However, all patients included in analysis had pre- and post-treatment imaging data available for volumetric analysis of tumors, and the volumetric calculation was derived from that used in other studies [15]. Although not statistically significant, length of follow-up time varied by modality, suggesting lengthier follow-up with additional patients is warranted. Additionally, management preferences and supportive care likely shifted over the study duration as newer techniques evolved; however, one of the patients in our LITT group was treated in 2013, early in the timeframe encompassed by our series.

Our results indicate that LITT should be strongly considered as a treatment option for neurosurgeons treating patients with SEGAs. Decisions about the use of this modality must be based on patient-specific factors, such as tumor size, location, presence of obstructive hydrocephalus, and response to prior therapies. Additional prospective, multi-institution studies with larger cohorts and longer follow-up duration are desirable to further evaluate the safety and efficacy of LITT in the treatment of SEGAs, compared with other treatment modalities.

This study protocol was reviewed and approved by the Seattle Children’s Hospital Institutional Review Board, approval number STUDY00003148. The Institutional Review Board did not require written informed consent for this retrospective research, which is appropriately compliant with institutional requirements.

Dr. Hauptman is a consultant for Medtronic and BK Medical.

There was no funding for this research.

Scott Boop and Jason Hauptman designed the study. Scott Boop and David Bonda collected and analyzed the data. Jason Hauptman provided study supervision. Scott Boop, David Bonda, Stephanie Randle, Sarah Leary, Nicholas Vitanza, Erin Crotty, Edward Novotny, Seth Friedman, Richard Ellenbogen, Sharon Durfy, Hannah Goldstein, Jeffrey Ojemann, and Jason Hauptman drafted and/or critically revised the manuscript and reviewed and approved the final manuscript.

All data generated or analyzed during this study are included in this article. Further inquiries can be directed to the corresponding author.