Risk of Tract Seeding Following Laser Interstitial Thermal Therapy for Brain Tumors

Abstract

BACKGROUND:

The management of intracranial oncological disease remains a significant challenge despite advances in systemic cancer therapy. Laser interstitial thermal therapy (LITT) represents a novel treatment for local control of brain tumors through photocoagulation with a stereotactically implanted laser fiber. Because the use of laser interstitial thermal therapy continues to increase within neurosurgery, characterization of LITT is necessary to improve outcomes.

OBJECTIVE:

To quantify the risk of tumor seeding along the laser fiber tract in patients receiving LITT for primary or metastatic brain tumors at a high-volume treatment center.

METHODS:

We retrospectively reviewed all patients receiving LITT from 2015 to 2021 at our medical center. Patients with biopsy-confirmed tumors were included in this study. Tract seeding was identified as discontinuous, newly enhancing tumor along the LITT tract.

RESULTS:

Fifty-six patients received LITT for biopsy-confirmed tumors from 2015 to 2021, with tract seeding identified in 3 (5.4%). Twenty-nine (51.8%) patients had gliomas, while the remainder had metastases, of which lung was the most common histology (20 patients, 74%). Tract seeding was associated with ablation proceeding inward from superficial tumor margin closest to the cranial entry point (P = .03). Patients with tract seeding had a shorter median time to progression of 1.1 (0.1-1.3) months vs 4.2 (2.2-8.6) months (P = .03).

CONCLUSION:

Although the risk of tract seeding after LITT is reassuringly low, it is associated with decreased progression-free survival. This risk may be related to surgical technique or experience. Follow-up radiosurgery to the LITT tract has the potential to prevent this complication.

ABBREVIATIONS:

CNS

central nervous system

FU

follow-up

KPS

Karnofsky Performance Status

LITT

laser interstitial thermal therapy

LOI

lesion of interest

RT

radiotherapy

SRS

stereotactic radiosurgery

WBRT

whole-brain radiation therapy.

Despite recent advances in systemic cancer therapy, the management of intracranial disease remains a significant challenge in oncology. Novel, successful therapies for extracranial malignancies have failed to produce the same benefit in primary brain tumors or brain metastases.^1-4 The incidence of intracranial disease in the United States is expected to increase as patients with treated, quiescent systemic disease age.⁵ Typically, multiple modalities, including stereotactic radiosurgery (SRS), open resection, and systemic therapy, are used in the treatment of these lesions.⁵ However, not all patients and tumor histologies are optimal candidates for these approaches. Minimally invasive ablation through laser interstitial thermal therapy (LITT) has been increasingly applied for deep-seated, unresectable tumors or in the management of recurrent disease or radiation necrosis after SRS. Continued characterization of the benefits and risks of LITT is essential as the technology matures within neurosurgical oncology.

For both available commercial LITT systems, a laser fiber is stereotactically introduced into the lesion to deliver infrared light and produce coagulative necrosis. Heating and spread of the resultant ablation are monitored with intraoperative MRI. In addition to its cytoreductive properties, LITT also facilitates biopsy of the target lesion, making it particularly useful in situations where tissue diagnosis guides subsequent therapy. In gliomas, molecular characterization by biopsy is necessary to guide therapy and aid in prognostication even in the setting of otherwise unresectable disease.^6,7 Biopsy is less common in the case of newly diagnosed metastatic lesions but is indicated when metastasis occurs in the setting of otherwise undetectable systemic disease.⁵ In the recurrent setting, tissue-based distinction between radiation necrosis and tumor progression is a key inflection point for further therapy.⁸ With favorable overall costs and outcomes relative to open resection, rapid postoperative recovery, and the ability to target unresectable lesions, the utilization of LITT has continued to increase among neurosurgeons.^9-11 However, a nuanced understanding of potential complications is important for adequate patient selection.

Historically, there was concern regarding the theoretical risk of seeding the biopsy tract with viable tumor cells during traditional biopsy procedures.¹² In gliomas, this risk is believed to be mitigated by the fact that patients are treated with larger radiation fields which routinely include the biopsy tract.^12,13 In early studies, the rate of biopsy-associated tumor seeding in gliomas was shown to be exceedingly low, on the order of 0.3%.^14,15 However, in a recent study of biopsies of intracranial metastases by Carnevale et al,¹³ the authors reported an unexpected 50% rate of tract seeding in a small cohort of patients with greater than 3 months of radiographic follow-up. It is unknown whether a similarly high rate of tract seeding occurs after LITT procedures.

Indeed, as with traditional biopsy, there is a risk that the LITT procedure results in tumor seeding along the tract and subsequent growth outside the field of ablation. Quantifying the risk of tumor seeding is critical to neurosurgical oncology and may have broad clinical implications for adjuvant therapy after LITT. We therefore sought to characterize the risk of tract seeding in patients receiving LITT for biopsy-confirmed primary or metastatic brain tumors at one of the highest volume LITT centers in the United States.

METHODS

A retrospective review of patients treated at a single National Cancer Institute–designated Comprehensive Cancer Center was conducted in accordance with a protocol approved by the health system Institutional Review Board (IRB # Pro00107226); consent was not required for participation. Patients receiving LITT for biopsy-confirmed, contrast-enhancing primary or metastatic brain tumors from 2015 to 2021 were included in this study. Patient demographics, systemic disease status, tumor characteristics, preoperative and postoperative interventions, hospital course, and survival outcomes were collected. Post-treatment progression-free and overall survival was defined as the time from LITT to confirmed radiographic progression of the index lesion or death, respectively. Detailed procedural data were also collected, including number of trajectories, rounds of lasing, direction and extent of ablation, and tract length. Tract length was defined as the length from the most superficial lesion edge to the inner calvarial table in the center of the burr hole as visualized on intraoperative magnetic resonance (MR) imaging and measured using neuronavigation software (iPlan 3.0, Brainlab AG). Tract seeding was defined as discontinuous, newly enhancing areas of tumor along the biopsy tract and was confirmed by 2 independent reviewers. Statistical analyses were performed using GraphPad Prism 9 (GraphPad Software). Significance was set at P = .05. Categorical variables were compared using Fisher exact or χ² tests, with Mann-Whitney U test used to compare continuous data. Survival data were estimated using Kaplan-Meier analysis and compared using log-rank tests.

RESULTS

A total of 127 patients received LITT for a suspected brain tumor from 2015 to 2021. Of these, 56 patients with biopsy-proven primary or metastatic brain tumors met inclusion criteria. Sixty-six excluded cases were radiation necrosis, 3 had no contrast-enhancing tumor sufficient for tract analysis, and 2 were nondiagnostic. The median age of patients in the cohort was 58.5 years (range 13-87 years). The median pre-LITT Karnofsky Performance Status was 90, and 21 (37.5%) patients were female. Most (29 patients, 51.8% of the cohort) received LITT for primary brain tumors, of which 24 were high grade (82.7% of primary tumors) and 5 were low grade (17.2%). 27 (48.2%) patients received LITT for the treatment of brain metastases. The most common histologies were lung (20 patients, 74% of metastases), followed by breast (3 patients, 11.2%) and skin (2 patients, 7.4%) cancers. The most common location for tumors was the frontal lobe (25 patients, 44.6%), followed by deep structures (eg, basal ganglia, thalamus: 10, 17.9%) and the temporal lobe (12, 21.4%). Most lesions were recurrent at the time of LITT (48 tumors, 85.7%). Seventeen (30.3%) lesions were periventricular, defined as less than 0.5 cm from any of the ventricular structures or foramina. 47 (83.9%) patients had previously received SRS to the lesion of interest (LOI), while a further 18 (32.1%) and 4 (7.1%) patients had previous surgery and whole-brain radiotherapy, respectively. Complete clinical characteristics are presented in Table 1.

TABLE 1.

Clinical Characteristics of LITT Cohort

Characteristic	Total (n = 56)	No tract seeding (n = 53)	Tract seeding (n = 3)	P
Median age (range)	58.5 (13-87)	59.0 (13-87)	58.0 (50-60)	.83
Female sex	21 (37.5%)	19 (35.8%)	2 (66.7%)	.55
Median Pre-LITT KPS (range)	90 (60-100)	90 (60-100)	90 (70-90)	.53
Tumor type				a
Brain	29 (51.8%)	28 (52.8%)	1 (33.3%)
Low grade	−5 (17.2%)	−5 (17.9%)	−0 (0.0%)
High grade	−24 (82.7%)	−23 (82.1%)	−1 (100.0%)
Metastasis	27 (48.2%)	25 (47.2%)	2 (66.6%)
Lung	−20 (74.0%)	−20 (80.0%)	−0 (0.0%)
Skin	−2 (7.4%)	−1 (4%)	−1 (50.0%)
Breast	−3 (11.2%)	−2 (8%)	−1 (50.0%)
Other	−2 (7.4%)	−2 (8%)	−0
Status of brain lesion				.99
Newly diagnosed	8 (14.3%)	8 (15.1%)	0 (0.0%)
Recurrent	48 (85.7%)	45 (84.9%)	3 (100.0%)
Location				a
Frontal	25 (44.6%)	25 (47.1%)	0 (0.0%)
Temporal	12 (21.4%)	11 (20.8%)	1 (33.3%)
Parietal	5 (8.9%)	3 (5.7%)	2 (66.6%)
Occipital	3 (5.4%)	3 (5.7%)	0 (0.0%)
Deep	10 (17.9%)	10 (19.0%)	0 (0.0%)
Skull base	1 (1.8%)	1 (1.9%)	0
Periventricular				.16
Yes	17 (30.3%)	15 (28.3%)	2 (66.6%)
No	39 (69.6%)	38 (71.7%)	1 (33.3%)
Presence of multiple intracranial lesions	32 (57.1%)	30 (56.6%)	2 (66.6%)	.99
Prior SRS or fractionated RT to LOI	47 (83.9%)	44 (83.0%)	3 (100.0%)	.99
Prior surgery to LOI	18 (32.1%)	17 (32.1%)	1 (33.3%)	.99
Prior WBRT	4 (7.1%)	4 (7.6%)	0 (0.0%)	.99
Median preoperative max lesion diameter (IQR)	2.4 (1.7-3.0)	2.3 (1.7-3.0)	2.9 (2.5-3.1)	.25

KPS, Karnofsky Performance Status; LOI, lesion of interest; LITT, laser interstitial thermal therapy; RT, radiotherapy; SRS, stereotactic radiosurgery; WBRT, whole-brain radiotherapy.

^aData in these columns were not compared due to invalidity of χ² analyses for groups with frequency <1.

There was radiographic evidence of new, discontinuous tumor growth along the LITT tract in 3 (5.4%) patients, 2 of which were metastases (1.1 months, NRAS-positive, BRAF V600E/KIT-negative melanoma; 1.3 months, triple-negative breast cancer), with the remaining patient harboring an IDH wildtype, 06-methylguanine-DNA methyltransferase unmethylated recurrent glioblastoma (0.4 months). A representative schema for analysis of patient imaging is shown in Figure 1. T1 contrast-enhanced sequences of the lesion were obtained before LITT, at the time of radiographic confirmation of tract seeding, and depicting interval growth for each affected patient (Figure 2A and 2C). Between patients with and without tract seeding, there were no significant differences in baseline clinical characteristics, including age, sex, pre-LITT Karnofsky Performance Status, tumor histology or status, periventricular location, intracranial tumor burden, previous treatments, or lesion diameter (Table 1). We then analyzed whether any technical aspects of the LITT procedure were associated with risk of tract recurrence (Table 2). 54 of 56 patients had biopsy performed concomitantly with the LITT procedure; of the remaining 2 patients, 1 patient (without tract seeding) underwent biopsy showing anaplastic astrocytoma and deferred further intervention for approximately 6 months before LITT. One patient (with tract seeding) had biopsy confirmation of glioblastoma multiforme 5 weeks before LITT. Most patients with and without tract seeding had similar likelihood of being treated with a single LITT trajectory (3/3 vs 48/53, P = .99). Similarly, patients with and without tract seeding were equally likely to have received multiple rounds of lasing (3/3 vs 45/53, P = .99). Interestingly, there was a trend toward longer tract length in patients with tract seeding (4.0 cm vs 2.2 cm, P = .16). Ablation technique was also associated with tract seeding, as all 3 (100%) patients with tract seeding received ablations proceeding from superficial to deep within the lesion, whereas only 12 of 53 (22.6%) patients without tract seeding were ablated in this direction (P = .03).

FIGURE 1.

Imaging analysis for tract seeding. Patients receiving LITT for biopsy-confirmed brain tumors were surveilled for tract seeding on MR imaging by 2 independent reviewers. Intraoperative scans were used to analyze the length and directionality of the LITT tract and compare with follow-up post-LITT imaging. When available, radiation dosimetry was also reviewed to determine whether the LITT tract was treated before identification of tract seeding. From left to right, representative A, pre-LITT, B, intraoperative, C, post-LITT, D, radiation dosimetry, and E, tract seeding images are shown for the patient with recurrent melanoma. LITT, laser interstitial thermal therapy.

FIGURE 2.

Imaging characteristics of patients with tract seeding. Shown are representative contrast-enhanced T1-weighted magnetic resonance images of patients with A, glioblastoma, B, recurrent melanoma, and C, recurrent breast cancer prelaser interstitial thermal therapy (left) at the time of tract seeding (middle) and with interval progression at the seeded site (right).

TABLE 2.

Treatment Variables

Characteristic	Total (n = 56)	No tract seeding (n = 53)	Tract seeding (n = 3)	P a
Biopsy result
Recurrent metastasis	27 (48.2%)	25 (47.2%)	2 (66.6%)
New primary	8 (14.3%)	8 (15.1%)	0 (0.0%)
Recurrent primary	21 (37.5%)	20 (37.7%)	1 (33.3%)
Trajectories per lesion
1	51 (91.1%)	48 (90.6%)	3 (100.0%)	.99
2	4 (7.1%)	4 (7.5%)	0 (0.0%)
3	1 (1.8%)	1 (1.9%)	0 (0.0%)
Multiple rounds of lasing	47 (83.9%)	45 (83.0%)	3 (100.0%)	.99
Ablation technique
Deep-superior	32 (57.1%)	32 (60.3%)	0 (0%)	.03
Superior-deep	15 (26.8%)	12 (22.6%)	3 (100%)
Unknown	9 (16.1%)	9 (17.0%)	0 (0%)
Median length of tract (cm, IQR)	2.3 (1.4-3.6)	2.2 (1.3-3.5)	4.0 (2.5-4.2)	.16

^aData in these columns were not compared due to invalidity of χ² analyses for groups with frequency <1.

P<.05 is considered significant and bolded.

The immediate postoperative course of patients with and without tract seeding was uncomplicated and is presented in Table 3. Patients received surveillance MRI scans every 3 months, with no significant differences in median follow-up between groups. Patients were equally likely to be transferred from the neurosurgical intensive care unit on postoperative day (POD) 0 regardless of future tract recurrence, although median POD of discharge trended later in the group with tract seeding (POD 1 vs 2, P = .18). There were no significant differences in 30-day readmission rate. There were no differences in the frequency of reoperation on the LOI (1/3 vs 3/53, P = .20) or receipt of postoperative radiation (2/3 vs 20/53, P = .55) in patients with tract seeding. For patients with tract seeding who received radiation, SRS did not include the LITT tract in one patient (Figure 1) and was performed after resection of the LOI for tract seeding in the other patient. On Kaplan-Meier analysis (Figure 3), progression at the site of tract seeding was an early phenomenon, with a median time to progression of 1.1 months (range 0.4-1.3 months) vs 4.2 months (2.2-8.6 months) in patients without seeding of the tract (P = .03, logrank test). There was also a trend toward decreased median overall survival in patients with tract seeding (5.4 vs 12.7 months, P = .18).

TABLE 3.

Postoperative Course

Characteristic	Total (n = 56)	No tract seeding (n = 53)	Tract seeding (n = 3)	P
Median radiographic FU (IQR)	6.1 (2.9-15.0)	6.3 (2.8-15.6)	5.1 (4.7-12.7)	.92
Median POD of discharge (IQR)	1 (1-2)	1 (1-2)	2 (1-8)	.18
Median POD of ICU transfer (range)	0 (0-1)	0 (0-1)	0 (0-1)	.99
Readmission within 30 d	7 (12.5%)	6 (11.3%)	1 (33.3%)	.34
Postoperative SRS or fractionated RT to LOI	22 (39.3%)	20 (37.7%)	2 (33.3%)	.55
Reoperation on LOI	4 (7.1%)	3 (5.7%)	1 (33.3%)	.20
Tract seeding Median time to tract seeding (mo, range)	3 (5.4%) X	0 (0.0%) X	3 (100.0%) 1.1 (0.4-1.3)	X
Progression Median time (mo, range)	23 (41.1%) 3.2 (2.0-8.0)	20 (37.7%) 4.2 (2.2-8.6)	3 (100.0%) 1.1 (0.4-1.3)	.03
Median overall survival by KM analysis (mo, range)	12.7 (0.6-43.9)	12.7 (0.6-43.9)	5.4 (5.1-14.1)	.18

FU, follow-up; ICU, intensive care unit; KM, Kaplan-Meier; LOI, lesion of interest; POD, postoperative day.

P<.05 is considered significant and bolded.

FIGURE 3.

Overall Survival of patients with and without tract seeding after LITT. Kaplan-Meier curves of median overall survival after LITT based on logrank testing of A, the entire study cohort (n = 57, 5.4 vs 12.7 months, P = .18) and B, patients with recurrent metastases only (n = 27, 5.3 vs 8.7 months, P = .31). Patients without tract seeding are shown in red, and tract seeding is shown in blue. LITT, laser interstitial thermal therapy.

Given the differing natural history of brain metastases and gliomas, we conducted a subgroup analysis including the 27 patients with metastases (Table 4). No significant differences were identified, although there were trends toward increased preoperative lesion diameter (3.0 vs 2.2 cm, P = .08) and tract length (3.3 vs 2.0 cm, P = .17) as well as shorter median time to progression (1.2 vs 8.8 months, P = .09) for tract seeding patients.

TABLE 4.

Recurrent Metastasis Subgroup Analysis

Characteristic	Total (n = 27)	No tract seeding (n = 25)	Tract seeding (n = 2)	P
Preoperative max lesion diameter	2.3 (1.7-2.9)	2.2 (1.7-2.7)	3.0 (2.9-3.1)	.08
Tract length	2.1 (1.4-3.1)	2.0 (1.3–3.0)	3.3 (2.5-4.2)	.17
Postoperative SRS	17 (63.0%)	15 (60.0%)	2 (100.0%)	.52
Tract seeding Median time (mo)	2 (7.4%) X	0 (0.0%) X	2 (100.0%) 1.2 (1.1-1.3)	X
Local progression Median time (mo)	7 (25.9%) 8.0 (1.5-21.8)	5 (20.0%) 8.8 (3.1-21.8)	2 (100.0%) 1.2 (1.1-1.3)	X .09
Overall survival by KM analysis (mo)	5.9 (1.2–43.9)	8.7 (1.2–43.9)	5.3 (5.1-5.4)	.31

KM, Kaplan-Meier; SRS, stereotactic radiosurgery.

DISCUSSION

Previously, tract seeding was believed to be a potential complication of traditional biopsy, especially in the setting of aggressive glioblastomas.¹² However, historical series have shown tract seeding from biopsy alone to be infrequent, although understudied in metastases.^14,15 It is possible that the large radiation treatment fields used in the treatment of primary central nervous system (CNS) neoplasms better mitigate spreading along the biopsy tract.¹³ However, more conformal radiation therapies such as SRS using intensity-modulated or proton-based techniques may simultaneously spare tissue outside of the target lesion while increasing the risk of seeding along the biopsy tract.^8,16 As a result, understanding and managing tract seeding as a complication of LITT is essential to maximizing its utility.

To the best of our knowledge, this is the first study to report rates of tumor seeding after LITT procedures for primary and metastatic brain tumors. In our cohort, tract seeding was a rare event, only affecting 3 of 56 (5.4%) patients with biopsy-confirmed tumors from 2015 to 2021. The studied LITT patient cohort was closely split between primary and metastatic tumors (51.8% vs 48.2%, respectively), reflective of the overall epidemiology of malignant CNS tumors in adults.^5,17 In addition, the proportion of recurrent or difficult to access tumors (85.7% and 19.7%, respectively) was consistent with current use cases for LITT. Many patients had received prior stereotactic radiation therapy (83.9%) or craniotomy (32.1%), with LITT offered as a later line of therapy. Despite the heavily pretreated nature of this patient population, we redemonstrated that patients tolerated the procedure well, with a median POD of transfer from intensive care unit and discharge at 0 and 1, respectively. Only 1 patient without tract seeding had a complication attributed to LITT, with the development of diabetes insipidus in the initial postoperative period.

On analysis of risk factors, our data suggested that ablation technique—whether the ablation proceeded from the superficial portion of the lesion to the deepest or vice versa—was associated with seeding along the fiber tract. All 3 patients with tract seeding had ablations consistent with the former. The mechanism of this phenomenon is unknown, as standard practice has favored a superficial to deep progression. It is possible that initial superficial ablation allows for reflux of viable tumor cells as the probe is withdrawn. A logical solution to this problem is to ablate in the opposite pattern, from deep to superficial. Furthermore, 54/56 patients in this study received biopsy during the LITT procedure, rendering it difficult to parse out whether seeding risk is specifically associated with the biopsy or LITT portion of the combined procedure. Of the 2 patients who received biopsy separately from LITT, 1 ultimately developed tract seeding along the intersecting tracts <1 month post-LITT, obscuring which procedure ultimately led to tract seeding. Interestingly, we noted that all instances of tract seeding occurred from 2016 to 2018, with 2 cases being within the first 15 LITT procedures ever performed at our institution, suggesting a role for surgeon experience. In addition, there was a trend toward longer tract length in patients who developed tract seeding. Longer tract lengths disrupt larger volumes of brain, increasing the probability for implantation of viable tumor.

Our analyses showed that tract seeding represents a unique phenomenon, separate from contiguous recurrence at the LOI. LITT tract seeding vs recurrence at the LOI occurred at a median of 1.1 vs 4.2 months (P = .03). We also identified a trend toward shorter median overall survival, with tract seeding patients surviving a median of 5.4 vs 12.7 months post-LITT. These differences are multifactorial and include the above discrepancies in ablation technique as well as unclear standards for post-LITT adjuvant therapy, including radiation. Although the overall low rate of tract seeding in our cohort is reassuring, strategies to reduce its occurrence are essential for improved local control. Our group previously reported improved outcomes with the combination of LITT plus SRS in contrast to SRS or LITT alone for patients with recurrent brain metastases.¹⁸ Our institutional practice is to offer adjuvant SRS following LITT procedures for recurrent metastatic disease, with recent initiation of an associated randomized trial (NCT05124912). Prophylactic inclusion of the LITT tract in radiation treatment fields could mitigate risk of tumor seeding but requires evidence. In addition, adjustments in ablation technique could contribute to further reduction. Although our analyses focused on risks because of the LITT procedure itself, it is likely that patient debility plays a role in the speed of disease progression and risk of tract recurrence; future studies are necessary to validate these findings.

Limitations

This study is limited by its single-institution, retrospective nature and small sample size. The low incidence of tract seeding in our study precludes drawing definitive conclusions regarding risk factors. In addition, because this study focused on the LITT procedure itself, we included both metastatic and primary brain tumors, which are distinct and diverse populations. However, on analysis of metastases alone, tract seeding risk remained low.

CONCLUSION

In recent years, the utilization of LITT has increased, with indications in neurosurgical oncology, focal epilepsy, and movement disorders.^9,10 In addition, there is compelling evidence of LITT-related effects on the perilesional microenvironment, including immune infiltration and alteration of blood-brain barrier permeability that could license native immune responses and delivery of systemic therapies.^19-24 Maximizing the safety and clinical efficacy of LITT is critical to fully realizing its potential benefits. Here, we demonstrate that LITT tract seeding is a rare event potentially associated with operative technique and predictive of shortened time to local recurrence. Standardization of LITT technique and tract inclusion in subsequent radiation treatment fields could mitigate this risk. These findings represent an important contribution to the understanding of LITT-related complications with broad implications for neuro-oncology.

Funding

Research reported in this publication was supported by the National Cancer Institute of the National Institutes of Health under Award No. 1R38-CA245204-01 (to Eugene J. Vaios).

Disclosures

Peter E. Fecci is a consultant for Monteris Medical and Synaptive Medical. The other authors have no personal, financial, or institutional interest in any of the drugs, materials, or devices described in this article.