Cost‐effectiveness of laser interstitial thermal therapy vs. anterior temporal lobectomy for drug‐resistant mesial temporal lobe epilepsy: A decision analytic model

Abstract

Objective

To evaluate the comparative cost‐effectiveness of anterior temporal lobectomy (ATL) and magnetic resonance–guided laser interstitial thermal therapy (LITT) in patients with drug‐resistant mesial temporal lobe epilepsy (mTLE).

Methods

A decision‐analytic model was constructed to simulate the risks, benefits, and costs of ATL versus LITT over a 1‐year time horizon from a healthcare system perspective. Model inputs, including seizure freedom rates, complication probabilities, and cost estimates, were derived from recent meta‐analyses and published cost comparison studies. The main outcome measure was the probability of a positive outcome, which was defined as seizure freedom without major complication. Incremental cost‐effectiveness ratios (ICERs) were calculated, and both one‐way and two‐way sensitivity analyses were conducted to assess parameter uncertainty.

Results

In the base‐case analysis, ATL achieved a 61% probability of a positive outcome compared with 55% for LITT. The mean per‐patient cost was $145 847 for ATL and $108 730 for LITT, resulting in an incremental cost of $37 117. Although ATL provided a modest 6% absolute increase in favorable outcomes, the ICER was $574 521 per additional positive outcome gained relative to LITT. Sensitivity analyses demonstrated that LITT remained the more cost‐effective option under most plausible parameter values, with ATL becoming favorable only under conditions of unusually high seizure freedom and low complication rates.

Significance

Although ATL yields slightly superior seizure outcomes, its substantially higher costs render LITT the more cost‐effective strategy for drug‐resistant mTLE at a 1‐year horizon. These findings support the growing role of minimally invasive techniques in epilepsy surgery, though further real‐world and long‐term data are needed to confirm the durability of LITT's clinical and economic benefits.

Plain Language Summary

This study compared two surgical treatments for drug‐resistant mesial temporal lobe epilepsy: anterior temporal lobectomy (ATL) and laser interstitial thermal therapy (LITT). Using a decision‐analytic model over 1 year, ATL showed slightly better seizure outcomes but at a substantially higher cost. Overall, LITT was more cost‐effective in most scenarios due to lower costs and fewer complications. These findings suggest that LITT may offer a practical and economically favorable option for many patients, although treatment decisions should still consider individual clinical factors.

Key points.

1. A decision‐analytic model compared ATL vs. LITT for drug‐resistant mTLE over 1 year from a healthcare system perspective.

2. ATL had higher effectiveness (61% vs. 55%) but higher cost ($145 847 vs. $108730), yielding an ICER of $574 521 per additional positive outcome.

3. Sensitivity analyses showed LITT was more cost‐effective in most scenarios, with ATL favored only under high efficacy and low complication rates.

1. INTRODUCTION

Temporal lobe epilepsy represents the most common form of medically refractory epilepsy, with mesial temporal lobe epilepsy (mTLE) due to hippocampal sclerosis comprising nearly 70% of cases. ¹ For decades, anterior temporal lobectomy (ATL) has been regarded as the gold‐standard surgical treatment, conferring substantial seizure control and notable improvements in quality of life for patients with drug‐resistant mTLE. ² Compared with pharmacotherapy, surgical intervention delivers markedly superior outcomes, with seizure freedom achieved in up to 80% of individuals. ³ , ⁴ Nevertheless, despite its proven efficacy, ATL remains a highly invasive procedure, accompanied by the risk of treatment‐related complications such as cognitive impairment, visual field deficits, intracranial hemorrhage, and, on rare occasions, mortality. ⁵

In response to these concerns, less invasive surgical strategies have been introduced, most notably radiofrequency thermocoagulation (RF‐TC) and magnetic resonance–guided laser interstitial thermal therapy (LITT). ⁶ , ⁷ LITT utilizes real‐time MRI guidance to perform focal laser ablation of the epileptogenic zone, thereby enabling precise lesioning while limiting collateral injury. As a minimally invasive technique, LITT has yielded favorable outcomes in drug‐resistant mTLE, particularly in patients contraindicated for, or declining, open craniotomy. ⁸ Although seizure freedom rates following LITT are generally lower than those achieved with ATL, the procedure is associated with a superior safety profile with fewer postoperative complications. ⁹

Given the differences in efficacy, safety, and procedural invasiveness, the associated healthcare costs are also expected to vary. Accordingly, an evaluation of the relative cost‐effectiveness of ATL and LITT—accounting for both clinical outcomes and complication risks—may facilitate more informed, individualized surgical decision‐making. ¹⁰ , ¹¹ To this end, we developed a decision‐analytic model to compare the cost‐effectiveness of ATL and LITT over a 1‐year time horizon in the management of drug‐resistant mTLE.

A decision‐analytic model provides a mathematical framework for simulating complex clinical scenarios in which multiple competing outcomes occur simultaneously. By structuring these possibilities into decision trees, the model assigns corresponding probabilities and costs to each potential clinical pathway. The primary advantage of this approach lies in its ability to integrate diverse data into a single, actionable metric. However, a significant limitation is that the model inherently represents a simplification of clinical reality; its accuracy is entirely contingent upon the quality of the input parameters, and it may not fully capture the specific biological variations of individual patients. Despite these trade‐offs, it remains a reliable tool for guiding healthcare policy and personalized surgical planning.

2. METHODS

This study was designed and reported in accordance with the Consolidated Health Economic Evaluation Reporting Standards (CHEERS) statement, which provides best practice guidelines for conducting and presenting healthcare economic evaluations. ¹²

2.1. Decision model

A decision‐analytic model was developed using TreeAge Pro (Version 2022 R1.0; Williamstown, MA, USA) to simulate the risks, benefits, and costs of ATL and LITT in patients with drug‐resistant mTLE (Figure 1). The model evaluated a binary treatment strategy (ATL vs. LITT) and projected clinical and economic outcomes over a 1‐year time horizon. This interval was chosen to capture short‐term postoperative outcomes while minimizing uncertainty related to long‐term recurrence and chronic care costs. The model incorporated key clinical events, including seizure freedom, major surgical complications, and procedure‐related mortality. A “positive outcome” was defined as seizure freedom (Engel Class I) without major complication; all other outcomes were categorized as unfavorable.

FIGURE 1.

Decision analytic model. The diagram illustrates the structure of the decision analytic model used for cost‐effectiveness analysis of drug‐resistant mTLE patients who are considered surgical candidates.

2.2. Clinical inputs and cost parameters

Key model parameters and data sources are summarized in Table 1. Estimates for postoperative seizure freedom and complication rates were derived from a recent meta‐analysis comparing ATL and LITT. ¹³ For the mortality analysis of patients with mTLE, the mortality risk of ATL was obtained from a published decision analysis model, whereas there were no relevant reports on the mortality of LITT. ¹⁴ Only major complications—defined as events resulting in significant disability, prolonged hospitalization, or the need for reoperation—were incorporated in the model, whereas minor complications unlikely to influence long‐term costs or outcomes were excluded. ¹³

TABLE 1.

Clinical parameters and data sources.

Variable	Value	Range	References
Clinical parameters
Mortality risk for ATL	0.3%	0–0.75%	Choi et al., 2008
Mortality risk for LITT	N/A	N/A
Complication risk for LITT	3.8%	1.9–5.7%	Kohlhase et al., 2021
Complication risk for ATL	10.9%	5.45–16.35%	Kohlhase et al., 2021
Seizure‐free after LITT	57%	53–61%	Kohlhase et al., 2021
Seizure‐free after ATL	69%	62–75%	Kohlhase et al., 2021
Cost estimates
LITT	104 930	85 622–124 238	Hines et al., 2022
ATL	134 980	126 166–143 794	Hines et al., 2022
Complication cost for LITT	100 000	25 000–200 000	Conservative assumption
Complication cost for ATL	100 000	25 000–200 000	Conservative assumption

Abbreviations: ATL, anterior temporal lobectomy; LITT, laser interstitial thermal therapy.

Cost inputs were obtained from a recent published single‐center study that compares the real‐world cost of LITT to ATL. ¹⁵ The costs of ATL and LITT encompassed hospital charges, surgical supplies, imaging, anesthesia, and physician services. The analysis was conducted from a healthcare system perspective and incorporated only direct medical costs. Due to the limited availability of standardized complication cost estimates, a base‐case value of $100 000 was assigned to major complications. Wide parameter ranges were examined in sensitivity analyses to account for uncertainty. This approach ensures that the model's conclusions are primarily driven by clinical event rates rather than speculative cost differences between types of adverse events. This estimate was informed by reports of craniotomy‐associated adverse events and was considered a conservative assumption. ¹⁶ , ¹⁷

2.3. Cost‐effectiveness analysis

A cost‐effectiveness analysis was conducted to estimate the cost per positive outcome, specifically defined as seizure free without major complications. Surgical interventions were assessed using the incremental cost‐effectiveness ratio (ICER). The ICER was defined as the difference in mean expected costs between ATL and LITT divided by the difference in the cumulative probability of achieving a positive outcome (seizure freedom without major complications). Specifically, the probability of a positive outcome for each arm was calculated by aggregating the probabilities of all clinical pathways leading to that state in the decision tree. This approach ensures that all risks and benefits simulated in the model are integrated into the final cost‐effectiveness estimate. Given the 1‐year time horizon adopted for the model, no discounting of costs or outcomes was applied in the analysis.

2.4. Sensitivity analyses

To evaluate the robustness of the model and the impact of parameter uncertainty, both one‐way and two‐way sensitivity analyses were performed. In one‐way analyses, individual parameters—including seizure freedom rates, complication probabilities, and procedure costs—were varied independently across ranges informed by published literature. A two‐way sensitivity analysis was then conducted on the two most influential parameters to examine their joint effect on the ICER.

3. RESULTS

3.1. Base case cost‐effectiveness analysis

Base‐case cost‐effectiveness results are presented in Table 2. ATL achieved a 61% probability of a positive outcome (seizure freedom without major complications), compared with 55% for LITT. The mean per‐patient cost was $145 847 for ATL and $108 730 for LITT, yielding an incremental cost of $37 117 in favor of LITT. Although ATL provided a 6% absolute increase in the probability of achieving a favorable outcome, the corresponding ICER was $574 521 per additional positive outcome gained relative to LITT.

TABLE 2.

Base case cost‐effectiveness analysis.

Strategy	Cost	Incremental cost	Effectiveness a	Incremental effectiveness	ICER
ATL	145 847	37 117	0.61	0.06	574 521
LITT	108 730		0.55

Abbreviation: ICER, incremental cost‐effectiveness ratio.

^a Effectiveness is the probability of a positive outcome.

3.2. Sensitivity analyses

One‐way sensitivity analyses were conducted across all model parameters, including procedure costs, complication‐related costs, seizure freedom rates, and morbidity probabilities. The results are illustrated in the tornado diagram (Figure 2), which indicated that the ICER was most sensitive to variation in the probability of seizure freedom following ATL. Other influential drivers included the probability of complications with ATL and the probability of seizure freedom with LITT. By contrast, changes in direct procedural costs, complication costs, and the probabilities of death or complications associated with LITT had comparatively minor effects on the ICER. These findings highlight the substantial influence of efficacy and safety parameters on cost‐effectiveness outcomes.

FIGURE 2.

Tornado diagram depicting results of one‐way sensitivity analyses of model parameters. The blue bar represents the ICER when the parameter increases from the base‐case value. The orange bar represents the ICER when the parameter decreases from the base‐case value.

To further assess the interaction between key clinical variables, a two‐way sensitivity analysis was performed on the probability of seizure freedom and complication rate associated with ATL (Figure 3). The Willingness‐to‐Pay (WTP) threshold represents the maximum cost the healthcare system is willing to incur for an additional unit of effectiveness. Assuming a WTP threshold of $250 000 per positive outcome, LITT remained the preferred strategy across most parameter combinations. ATL became cost‐effective only under conditions where seizure freedom exceeded 71% and complication rates were below 9%. Taken together, both one‐way and two‐way sensitivity analyses demonstrated that ATL would be considered cost‐effective relative to LITT only under limited combinations of high efficacy and low complication rates.

FIGURE 3.

Results of two‐way sensitivity analysis. Two‐way sensitivity analysis varied probability of seizure freedom by ATL and probability of complication by ATL. The analysis adopted a willingness‐to‐pay threshold of $250 000 per positive outcome gained. The red area represents parameter combinations for which LITT is the preferred strategy, whereas the blue area represents those for which ATL is favored.

4. DISCUSSION

In this cost‐effectiveness analysis of ATL versus LITT for the management of drug‐resistant mTLE, ATL was associated with a modestly higher probability of seizure freedom, whereas LITT emerged as the more cost‐effective strategy at the 1‐year time horizon. The resulting ICER of $574 521 per additional positive outcome with ATL indicates that, under commonly accepted WTP thresholds, LITT represents the economically favorable option in the majority of clinical scenarios.

In the present study, the comparison between ATL and LITT highlights the trade‐off between clinical effectiveness and economic impact. While ATL is associated with a higher probability of seizure freedom, LITT offers a more favorable safety profile and lower overall cost, which together contribute to its economic advantage within the modeled time horizon. This balance between efficacy, procedural risk, and resource utilization represents a central consideration in contemporary epilepsy surgery. ¹⁸

To date, no randomized clinical trial has directly compared LITT with traditional surgical approaches such as ATL. Evidence from recent meta‐analyses indicates that ATL is associated with a higher likelihood of seizure freedom relative to LITT, albeit at the cost of greater invasiveness and an increased risk of major complications. ¹³ , ¹⁹ , ²⁰ Conversely, LITT provides a minimally invasive alternative characterized by reduced perioperative morbidity, shorter hospitalization, and a more favorable safety profile, though with modestly lower rates of seizure freedom. ²¹ This trade‐off between efficacy and safety remains central to surgical decision‐making, particularly when the incremental benefits of seizure control must be balanced against heightened operative risk and healthcare resource utilization.

The economic advantage of LITT demonstrated in our model was primarily attributable to lower procedural and hospitalization costs, together with a reduced incidence of complications requiring additional care. Although ATL yielded a higher proportion of favorable outcomes, the incremental clinical benefit did not outweigh the substantial increase in cost, particularly from a healthcare system perspective. These findings are consistent with a growing body of literature highlighting the economic and clinical value of minimally invasive surgical strategies in neurosurgery, especially in the management of chronic conditions such as epilepsy. ²² , ²³ , ²⁴

Sensitivity analyses further confirmed the robustness of the model. The ICER was most sensitive to variations in seizure freedom and complication rates associated with ATL, underscoring the critical role of patient selection and surgical expertise in determining optimal outcomes. Importantly, even when these parameters were varied across plausible ranges, LITT remained the more cost‐effective strategy in most scenarios. These findings suggest that LITT may represent a cost‐effective alternative to ATL in the majority of clinical contexts, particularly when elevated surgical risk or patient preference limits the feasibility of open resection. ²⁵

Although this analysis suggests that LITT provides greater cost‐effectiveness over a 1‐year time horizon, surgical decision‐making in drug‐resistant mTLE must remain individualized. The choice between ATL and LITT should not rely solely on aggregate economic outcomes but must also account for patient‐specific clinical, anatomical, and cognitive factors. For elderly patients or those with significant medical comorbidities who face elevated perioperative risk with craniotomy, LITT may be particularly advantageous. ²⁶ Patients with dominant hemispheric epilepsy foci are relatively more likely to develop neuropsychological deficits after surgery. ²⁷ LITT can reduce this risk. In a comparative study, Drane et al. reported significantly lower declines in famous face and common noun naming among patients undergoing LITT relative to those treated with ATL. ²⁸ By contrast, for younger adults in whom long‐term seizure freedom is prioritized, ATL may confer greater benefit. Barot et al. observed that seizure freedom rates after LITT declined from 60% at short‐term follow‐up (6–12 months) to 53% at a mean follow‐up exceeding 24 months. ²⁰ Over extended periods, recurrence rates following LITT may surpass those of ATL, likely reflecting incomplete ablation of epileptogenic tissue. ²⁹ Such incomplete resection may result in insufficient disruption of epileptogenic networks and potential reorganization, thereby contributing to seizure recurrence. ³⁰ Cost‐effectiveness findings should be interpreted as a guide for healthcare resource allocation rather than a substitute for individualized clinical and anatomical considerations in surgical decision‐making.

Several limitations merit consideration. First, the model employed a 1‐year time horizon, which may underestimate the long‐term benefits of ATL, particularly for patients who achieve durable seizure freedom. Although our clinical composite endpoint provides an intuitive measure for short‐term surgical success, it does not account for long‐term health‐related quality of life. Improvements in quality of life and baseline healthcare utilization costs were not incorporated. Future analyses should incorporate quality‐adjusted life years (QALYs) and life cycle cost trajectories, which can provide a more comprehensive assessment of the comparative value of these treatment strategies. To further explore the impact of the time horizon on our findings, it is important to consider the potential for time‐horizon bias. While our model emphasizes the immediate economic benefits of LITT driven by lower procedural costs and a more favorable safety profile, clinical evidence suggests that ATL may offer more durable seizure control over several years. In a longer‐term or lifetime model, the sustained seizure‐free status associated with ATL could lead to significant cost offsets through reduced long‐term anti‐seizure medication use and fewer seizure‐related emergency visits. Consequently, the ICER of ATL might decrease over time, potentially rendering it the more favorable option for patients with a longer life expectancy who prioritize definitive seizure remission. Future Markov models with multi‐year states are required to identify the exact cross‐over point where ATL's clinical superiority outweighs its initial economic burden. Second, cost inputs were derived from US‐based healthcare data from a single institution. This reliance on a specific institutional framework may limit the generalizability of our findings to other health systems with differing cost structures and reimbursement mechanisms. There is a critical necessity for future multicenter and international cost‐effectiveness analyses. Third, the analysis was based on aggregate estimates from published meta‐analyses, which are subject to heterogeneity in patient populations, surgical techniques, and outcome definitions. Finally, the model adopts a relatively simple decision‐tree structure and does not incorporate more complex disease trajectories, such as seizure recurrence, repeat interventions, or long‐term follow‐up outcomes. Although the model provides a structured comparison of key short‐term outcomes and costs, its results should be interpreted as a simplified representation of clinical reality rather than a comprehensive evaluation of all relevant treatment effects.

5. CONCLUSIONS

In this decision‐analytic model comparing ATL and LITT for the management of drug‐resistant mTLE, ATL was associated with a modestly higher probability of seizure freedom without major complication but incurred substantially greater costs. LITT, as a minimally invasive alternative, demonstrated favorable cost‐effectiveness across a broad range of clinical scenarios and remained the economically preferred strategy under most conditions. Future research incorporating real‐world data and long‐term follow‐up is warranted to validate these findings and to clarify the durability of LITT's clinical and economic benefits.

AUTHOR CONTRIBUTIONS

Zhihao Guo: Writing – original draft, writing – review and editing, methodology, software, validation, formal analysis, visualization, funding acquisition. Yihe Wang: Writing – review and editing, methodology, software. Kailiang Wang: Writing – review and editing, data curation. Yongzhi Shan: Writing – review and editing, supervision, investigation. Guoguang Zhao: Conceptualization, resources, writing – original draft, writing – review and editing, supervision, data curation, investigation.

FUNDING INFORMATION

This work was sponsored by the Beijing Natural Science Foundation (L256053), Capital's Funds for Health Improvement and Research‐Brain (2024‐3‐2061), and Xuanwu Hospital Elite Cultivation Program (YC20250106).

CONFLICT OF INTEREST STATEMENT

The authors declare no conflict of interest. We confirm that we have read the Journal’s position on issues involved in ethical publication and affirm that this report is consistent with those guidelines.

DATA AVAILABILITY STATEMENT

The datasets generated and analyzed during the current study are available from the corresponding author upon reasonable request.