Noninvasive Thalamotomy for Refractory Tremor by Frameless, Incisionless Radiosurgery

Abstract

Objective:

To determine whether frameless, incisionless radiosurgical thalamotomy is a safe and effective treatment for essential or parkinsonian tremor refractory to medical therapy.

Methods:

We conducted a single-arm pilot observational prospective trial of adult patients with essential or parkinsonian tremor from 2013 to 2019 with one-year follow-up. Patients were treated with frameless, image-guided unilateral radiosurgical ablation of the thalamic ventral intermediate nucleus. The Fahn-Tolosa-Marin tremor rating scale and the Quality of Life in Essential Tremor or PD Questionnaire were obtained prior to treatment and at 3, 6, 9, and 12 months.

Results:

Thirty-three patients, including 23 with essential tremor and 10 with Parkinson’s disease, were enrolled. Overall treatment response rate was 83% (n=15/18) at 6 months. There was a marked improvement in tremor, with average total FTM reduction of 21% at 3 months (from 46 to 30 points, p=0·003) and 41% at 6 months (from 46 to 24 points, p=0·001); at 6 months, functional decline regressed by 54% (from 15 to 7 points, p=0·001). Quality of life improved by 57% (p=0·001) at 6 months in patients with essential tremor. Mild neurologic adverse events were observed in 6% (n=2/33) of patients. No serious adverse events occurred.

Conclusion:

In a pilot trial, noninvasive, frameless radiosurgical thalamotomy appears safe and efficacious treatment for patients with tremor, with benefits in physical symptoms and quality of life. Randomized controlled trials are merited.

INTRODUCTION

Tremor is the most prevalent movement disorder. For individuals over the age of 65, 5% and 1·8% are affected by essential tremor (ET) and Parkinson’s disease (PD), respectively.^1,2 Tremor can be quite disabling and impair physical and psychosocial wellbeing.³

Efficacy of first-line medical therapy is limited. Medications can reduce symptoms by up to 60% in over 50% of patients with ET, although many patients experience intolerable side effects; likewise, PD tremor is frequently medically refractory.^4–6 Second-line treatment includes targeting of the thalamic ventral intermediate nucleus (VIM) with surgical thalamotomy or deep brain stimulation (DBS).⁷ However, such patients must be fit for anesthesia and surgery, and are at risk of hemorrhage, infection, pneumothorax, hemothorax, and poor wound healing.

In contrast to surgical measures, frame-based radiosurgical thalamotomy has been developed as a similarly efficacious and potentially more cost-effective treatment.^8–11 Frame-based radiosurgical thalamotomy is relatively less invasive compared with surgical thalamotomy and DBS, since there is no surgical tract and no open surgery. However, frame-based radiosurgery is still invasive: a rigid stereotactic headframe must be surgically attached to the cranium by four screws or pins that penetrate the skull at various depths to achieve necessary precision targeting, which may cause discomfort and deter some patients from therapy.¹² Additionally, frame-based radiosurgery (commonly Gamma Knife) is not readily available across radiation therapy centers globally and may not be as cost-efficient compared to linear-accelerator based treatment, which centers routinely utilize for targets across the body (not just the cranium), and does not require the comparably costly maintenance and source replacement of Gamma Knife. MR-guided focused ultrasound (MRgFUS) has been proposed as an experimental alternative, however MRgFUS still requires cranial fixation with screws.¹³ Immobilization with the use of frame-based fixation is not without serious risks, as penetration of the scalp may lead to infections, bleeding, skull fractures, sinus fractures or rarely even penetration of pins into the brain.^14–17 Importantly, there have also been concerning toxicities associated with MRgFUS, likely due to less precise targeting. In a randomized clinical trial of MRgFUS thalamotomy, adverse events in the thalamotomy group included gait disturbance in 36% of patients and paresthesias or numbness in 38%; these adverse events persisted at 12 months in 9% and 14% of patients, respectively.¹³ Such adverse events beg into question the precision and collateral effects of thermal-directed therapies, unlike the precision demonstrated with radiosurgery-based treatment.

In recent years, platforms enabling truly noninvasive radiosurgery with mask-based target systems have emerged. Modern, high quality X-ray image-guidance, with real-time verification of patient anatomy, has been shown to afford frameless radiosurgery accuracy similar to rigid frame-based systems.¹⁸ In a previous work, we described the robust sub-millimeter accuracy of frameless radiosurgical thalamotomy.¹⁹ Aided by these critical advances, frameless radiosurgery has now been safely and effectively utilized for functional neurologic disorders including trigeminal neuralgia.²⁰ Here, we report the short-term safety and efficacy of the first pilot prospective trial of frameless radiosurgical thalamotomy for the treatment of medically refractory ET or PD tremor.

METHODS and MATERIALS

Patient Population

We performed a single-arm, open-label observational prospective trial of frameless radiosurgical thalamotomy for adult patients from 2013 to 2019 at our tertiary referral center to generate foundational evidence necessary before a randomized trial. Institutional review board approval was obtained prior to study enrollment (NCT01734122). All patients gave written informed consent. Patients were required to have ET or PD diagnosed by a movement-disorder neurologist (author P.H. or F.T.P.). Patients were required to have inadequately controlled tremor despite medications, as determined by our movement disorder program including neurologists, neurosurgeons, radiation oncologists, and neuropsychiatrists. Patients were required to have a stated preference to receive radiosurgical thalamotomy rather than DBS or surgical thalamotomy. Patients with contraindications to radiation, prior radiation to the thalamus, contraindications to MRI, prior surgical thalamotomy, estimated life expectancy less than one year, or pregnancy were excluded. Anticoagulation was an exclusion criterion that was removed by protocol amendment after initial safety evaluation in April 2014. Analysis was performed by authors L.W. and C.J.L.

Frameless Radiosurgical Thalamotomy

Our method for frameless radiosurgical thalamotomy has been previously described in a technical manuscript.¹⁹ Briefly, each patient was fitted with a noninvasive, rigid thermoplastic custom-made mask. The patient then underwent simulation head CT obtained in 1·5 mm serial axial slices, which were resampled to provide 0·75 mm targeting accuracy. These post-processed images were then fused to a recent thin-slice MRI. The VIM target (assigned by authors J.N., H.Y., A.J.C., A.N.K) was contralateral to the patient’s most bothersome tremor laterality, typically the dominant hand. A customized plan was designed using iPlan Treatment Planning Software (BrainLab, Munich, Germany), and was reviewed at an independent multidisciplinary quality assurance conference. At the time of treatment delivery, serial orthogonal kilovoltage images were obtained with the patient in the treatment position and confirmed with ExacTrac localization (BrainLab, Munich, Germany). Isocenter alignment was verified by the treating physician, and imaging was performed at each table angle with robotic adjustment whenever beyond a tolerance of 0·5 mm translational and 0·7° rotational. Radiosurgery was performed in a single session with a 4 mm cone collimator using the Novalis TX linear accelerator (Varian Medical Systems, Palo Alto, CA). Patients were evaluated immediately following thalamotomy. The protocol maximum dose (D_max) was 160 Gy, prescribed to the isocenter placed at the VIM.

Outcome Measures

The primary endpoints were efficacy and safety. Efficacy was measured using both the Fahn-Tolosa-Marin (FTM) tremor rating test, performed by a movement-disorder neurologist (author P.H. or F.T.P.), and patient-reported quality of life (QOL) according to the Quality of Life in Essential Tremor (QUEST) or Parkinson’s Disease Questionnaire (PDQ-39). Total FTM scores, total QUEST scores, and total PDQ-39 scores, as well as changes from baseline, were computed. The relative percent change was also calculated, with positive percent changes indicating improvement (a negative numerical change). Safety was measured by adverse events graded according to the US Department of Health and Human Services Common Terminology Criteria for Adverse Events version 4·03 during the trial and reclassified to version 5·0 at the time of manuscript submission. Serious adverse events were defined as ≥ grade 3 events related to radiosurgical thalamotomy. Assessments were obtained at baseline and at 3, 6, 9, and 12 months following treatment, accompanied by unblinded clinical examination by the radiation oncologist and neurologist. Initial post-treatment MRI was obtained at 3 months to document lesion creation and perilesional edema (Figure S1).

Statistical Analysis

Continuous variables were summarized using the 25th, 50th, and 75th percentiles. Wilcoxon signed rank tests were used for paired comparisons between time points. The primary analyses were linear mixed-effects models fit to the FTM score, change in the FTM score, and percentage change in the FTM score. Models were adjusted for baseline. Hypothesis testing was restricted to the primary endpoints without post hoc subset analyses, and thus no adjustments were made for multiplicity of inferences or false discovery. Confidence intervals are reported at 95%, and p-values were calculated using two-sided testing. Significance was defined at an alpha level of 0·05. Analysis was performed using R (R Foundation for Statistical Computing, Vienna, Austria).

RESULTS

Study Patients

39 patients were assessed for eligibility; 6 patients were excluded due to MRI contraindication (n=2), anticoagulation (n=2), and participant wishes (n=2) (Figure S2). Thirty-three patients received the study intervention, including 23 (70%) with ET and 10 (30%) with PD. Median age was 76 years (IQR 71–82 years). Most patients were male (n=22, 67%), Caucasian (n=29, 88%), and without family history of tremor (n=26, 79%) (Table 1). Median disease duration from initial symptoms was 8 years (IQR 4–20 years). Median maximum dose (D_max) was 160 Gy (IQR 156–160 Gy).

Table 1.

Patient demographics.

Demographics	Median/Frequency	IQR/Percent
Age	76	71–82
Gender
Male	22	67%
Female	11	33%
Race
Caucasian	29	88%
African American	4	12%
Diagnosis
ET	23	70%
PD	10	30%
Duration of tremor from initial symptoms (yrs)	8	4–20
Family History	7	21%
Treated Extremity Contralateral to Brain Target
Left	5	15%
Right	28	85%
Dmax (Gy)	160	156–160

Abbreviations: IQR (interquartile range); ET (essential tremor); PD (Parkinson’s disease); Gy (Gray); D_max (maximum dose).

Objective Tremor Severity

Patients experienced significant reductions in tremor severity at all follow-ups compared to baseline (Table 2). At 3 months, 76% of patients (n=16/21) had a reduction in total FTM score, and median decrease in total FTM score was 9 points (IQR 1 to 20, p=0·003) representing a 21% improvement (IQR 2% to 43%) (Figure 1). 83% of patients (n=15/18) had improvements in FTM score by 6 months (Figure 1). Compared to baseline, the median decrease at 6 months was 20 points (IQR 10 to 28, p=0·001) for an improvement of 41% (IQR 23% to 58%) (Table S1). Regardless of diagnosis, improvements occurred by 6 months (Figure S3). Improvements in total FTM score remained durable, including 44% (IQR 11% to 49%, p=0·004) at 9 months and 50% (IQR 22% to 65%, p=0·003) at 12 months (Figure 2A). Linear mixed-effects modeling suggested maximum benefit occurred at 6 months by a median of 18 points (95% CI 13 to 23 points) or 37% (95% CI 25% to 49%) (Table S2). The model further suggested that improvement was maintained at 9 months (30%, 95% CI 18% to 24%) and 12 months (35%, 95% CI 21% to 48%).

Table 2.

Fahn-Tolosa-Marin tremor rating scale scores over time following frameless radiosurgical thalamotomy.

Domain	Baseline	3 months	p-value	6 months	p-value	9 months	p-value	12 months	p-value
	N=29	N=26; N=25*		N=20; N=19*		N=17; N=16*		N=12; N=11*
Part A: Tremor Severity during rest, posture, action positions	12 (9–14)	9 (6–13)	0·121	7 (5–8)	0·001	10 (7–11)	0·017	7 (5–11)*	0·083
Part B: Handwriting and Pouring	18 (13–25)	13 (7–17)*	0·003	12 (7–15)	0·001	14 (10–18)	0·01	10 (5–18)	0·005
Part C: Functional Disability	15 (13–17)	10 (6–16)	0·004	7 (6–10)*	0·001	7 (6–11)*	0·005	7 (6–11)*	0·006
Total	46 (38–55)	30 (21–45)*	0·003	24 (19–31)	0·001	32 (26–40)	0·004	26 (14–39)	0·003

^*Scores are summarized by median (interquartile range). Statistically significant p-values, obtained by Wilcoxon Signed Rank test comparing each post-treatment timepoint to the pre-treatment baseline, are bolded.

Figure 1.

Patient-level numerical changes in total Fahn-Tolosa-Marin (FTM) score following radiosurgical thalamotomy at (A) 3 months, (B) 6 months, (C) 9 months, and (D) 12 months.

Figure 2.

Tremor is reduced and overall disease-specific quality of life is improved following radiosurgical thalamotomy at 3, 6, 9, and 12 months as depicted by smoothed conditional mean and 95% confidence band. (A) Fahn-Tolosa-Marin (FTM) total score. (B) FTM Part A score representing tremor severity during rest, posture and action positions. (C) FTM Part B score representing the ability to write and pour liquids. (D) FTM Part C representing functional disability. (E) Quality of Life in Essential Tremor (QUEST) total score. (F) Parkinson’s Disease Questionnaire 39 (PDQ) total score.

Patients experienced significant benefits in each domain (Table 2). Tremor severity during rest, posture, and action positions decreased at 6, 9, and 12 months, with a median improvement of 39% at 6 months (IQR 7% to 63%, p=0·001) (Figure 2B). The ability to write and pour liquids similarly improved over time, including by 38% at 6 months (IQR 27% to 58%, p=0·001) (Figure 2C; see Figure 3 for a writing example). Functional disability also regressed with median benefit of 54% at 6 months (IQR 36% to 61%, p=0·001) (Figure 2D).

Figure 3.

Improvement in Fahn-Tolosa-Marin (FTM) measured writing. Example of patient drawing exercise before frameless radiosurgical thalamotomy (A) and 3 months after frameless radiosurgical thalamotomy (B).

Quality of Life

For patients with ET, QOL improved by an average of 29% at 3 months (p=0·003), 57% at 6 months (p=0·001), 44% at 9 months (p=0·019), and 37% at 12 months (p=0·027) (Figure 2E). The greatest QOL benefit appeared to be physical, although there was also improvement in the hobbies/leisure and psychosocial domains (Table 3). For patients with PD, QOL showed neither improvements nor decrements (Figure 2F; Table S3).

Table 3.

Quality of Life in Essential Tremor scores over time following frameless radiosurgical thalamotomy. Scores are summarized by median (interquartile range).

Domain	Baseline	3 months	6 months	9 months	12 months
	N=22	N=18	N=13	N=13	N=9
Communication	0 (0–25)	0 (0–8)	0 (0–8)	0 (0–0)	0 (0–17)
Work and Finances	0 (0–20)	0 (0–6)	0 (0–0)	0 (0–4)	0 (0–0)
Hobbies and Leisure	50 (12–92)	54 (17–67)	0 (0–33)	0 (0–67)	33 (0–83)
Physical Score	81 (72–89)	42 (15–73)	22 (6–47)	17 (0–67)	58 (25–72)
Psychosocial	33 (19–54)	14 (6–34)	6 (0–17)	3 (0–14)	6 (0–17)
Total*	39 (25–55)	25 (18–31)	10 (6–16)	19 (1–28)	25 (14–30)

^*Compared to baseline, total score was significantly reduced at 3 months (p=0.003), at 6 months (p=0.001), at 9 months (p=0.019), and at 12 months (p=0.027) according to Wilcoxon Signed Rank test.

Safety

During the 12-month follow-up period, the incidence of toxicity was 6% (n=2/33); zero patients had serious toxicity. One patient experienced a grade 2 headache relieved by non-steroidal anti-inflammatories. The other patient experienced a grade 2 headache controlled by dexamethasone.

DISCUSSION

In the first prospective trial of frameless, noninvasive, incisionless radiosurgical thalamotomy, patients with medically refractory tremor experienced significant reduction of tremor severity, improved QOL, and minimal sequelae. The benefits of frameless radiosurgical thalamotomy, assessed by a movement-disorder neurologist and by disease-specific patient-reported QOL measures, appeared at 3 months, increased maximally at 6 months, and remained durable at 9 and 12 months. Given the robust benefit demonstrated and the absence of severe adverse events, this pilot trial suggests that frameless radiosurgical thalamotomy appears safe and effective for the treatment of patients with essential and parkinsonian tremor.

Our findings are highly impactful when compared to more invasive treatments. Our approach achieves an excellent treatment response rate, on par with surgical thalamotomy and DBS.⁸ Complications related to the invasiveness of surgical thalamotomy and DBS, such as hemorrhage, infection, seizures, pulmonary embolism, pneumothorax or hemothorax, pneumocephalus, stroke, and death, are obviated by our incisionless approach.^21–23 Less invasive treatments, including frame-based radiosurgery and MRgFUS, do not pose the magnitude of risks of open neurosurgery, but are associated with other notable sequelae, as they are still invasive. For example, 30–35% of patients undergoing stereotactic frame fixation for MRgFUS experienced pin-site pain, edema, or bruising, none of which occur with our frameless technique.¹³ Cranial fixation may also cause bleeding with subcutaneous or subgaleal hematoma, infection, sinus fractures, or even penetration of the pins through the inner cortex into the brain parenchyma, and these serious adverse events secondary to piercing screws appear mitigated by our frameless approach.^24,25 In addition, there are significant concerns with the accuracy of MRgFUS given the high toxicities reported. Most recently in a randomized control trial, 36% of patients experienced post-MRgFUS gait unsteadiness and 38% of patients experienced paresthesias or numbness.¹³ These are higher rates of toxicity than seen in the radiosurgery literature, accustomed to sub-mm accuracy.

Patients are able to take breaks during frameless thalamotomy, since there is no incision required, in contrast to frame-based approaches in which the entire treatment must be completed without interruption. Frameless thalamotomy by linear accelerators may also be more cost-effective. Linear accelerators are widely available and used for general radiation treatment of all body sites. In contrast, dedicated MRgFUS units or highly specialized platforms for frame-based radiosurgery that require frequent cobalt source exchanges can only be used for cranial targets and are less cost effective. This cost benefit may be even greater when considering the potential for reduced complications, such as bleeding or infection, and no requirement for operating room time or hospital admission. Finally, use of linear accelerators, which are the most prevalent system for radiation delivery, may increase access to treatment for patients.

Paradoxically, frame-based fixation has also been associated with MRI distortion due to metal artifact and mechanical distortion from torque during fixation, both of which may reduce accuracy.^25,26 Skull slippage despite pin placement may also introduce error unaccounted for during treatment as frame-based radiosurgery does not have real-time on-board imaging.^27,28 Our method of frameless radiosurgery does not suffer from these concerns and includes real-time orthogonal x-rays during radiation delivery, which confirm that the target is aligned at every table angle during treatment delivery.¹⁹

In our study, patients with ET had significant improvements in QOL, most prominently in the physical and psychosocial domains, which are plausibly driven directly by radiosurgical thalamotomy. Such improvements are impressive considering unilateral treatment, and we will assess in future studies whether QOL can be further enhanced with bilateral thalamotomy. Interestingly, however, patients with PD had neither improvement nor worsening in PD-specific QOL. Importantly, patients with PD suffer from a more global neurodegenerative disease, where QOL is affected by more than tremor alone. Consequently, our measurement of PD-specific QOL may not have been sensitive enough to detect improvements in tremor-related QOL. Furthermore, while the PDQ-39 is validated in PD, the scale was designed to measure the effects of pharmacologic treatments rather than unilateral interventions, further reducing instrument resolution. Patients with PD may also experience disease progression in areas unaddressed by thalamotomy, such as cognition, pain, communication, rigidity, and/or bradykinesia, confounding potential QOL benefit from thalamotomy. Finally, combined with these limitations, our study evaluated a small number of patients with PD, whose follow-up declined by 60% at 1 year, limiting the power of PD-related QOL hypothesis testing. Thus, while there is moderate reassurance that QOL does not seem to be worsened and that adverse events are minimal, further studies are needed to more fully characterize PD-specific QOL following frameless radiosurgical thalamotomy.

Compared to MRgFUS, DBS, and surgical thalamotomy, radiosurgical thalamotomy has a delayed time to benefit. In our study, patients were found to experience initial tremor reduction at first follow-up of 3 months, progressing at 6 months, before plateauing at 9 and 12 months. While our study’s first formal follow-up measurements occurred at 3 months, in institutional experience, patients report improvement even within 4 weeks after treatment, earlier than our study was designed to detect. Nonetheless, the apparent delay is due to inherent differences between the mechanisms of radiosurgical thalamotomy and MRgFUS, DBS, and surgical thalamotomy, and patients should be counseled accordingly.⁹ Moreover, unlike the real-time biological feedback of DBS or MRgFUS, radiosurgical thalamotomy only allows for intraprocedural validation of the target through image-guidance. While we have shown previously that modern systems allow frameless radiosurgery to achieve accuracy and precision similar to frame-based radiosurgery, and while we demonstrate here that frameless radiosurgical thalamotomy is safe and effective, the lack of real-time biological feedback poses a comparative limitation.¹⁹ While not addressed here, future studies could combine the advantages of MRgFUS and frameless radiosurgery. For instance, reversible, low-power, non-ablative sonications (150 to 250 W), which may not require frame-based cranial fixation (NIH Project Number 1U18EB029351-01), could enhance target elucidation prior to definitive frameless radiosurgical thalamotomy.

Limitations include an observational, unblinded, non-randomized, and non-comparative design. Because our study design was observational, rather than interventional, there was no control group, and placebo effect was not evaluated separately from radiosurgery effect since comparison was made pre and post-treatment. Generalizability is limited by our study design of one-year of follow-up, as well as by missing survey data which may have underpowered our evaluation of adverse effects, although late effects, such as radionecrosis or secondary malignancy, appear rarely following radiosurgery for benign disease, particularly in this age group.^29,30 The unilateral nature of our study treatment may bias potential modality comparisons using global scales in favor of commonly bilateral treatments, such as DBS, and thus current benefits may be underreported. Certainly, frameless radiosurgery can be performed bilaterally with caution, although as the first prospective study, our trial prudently sought to first examine the efficacy and safety of unilateral thalamotomy, and future bilateral frameless radiosurgical thalamotomy may have even more robust effects on global tremor measures. Moreover, 17% of patients did not respond to thalamotomy, and our study was not powered to evaluate for predictive factors, such as disease, age, or radiation dose. We are actively investigating the hypothesis that lack of response is due to choice of isocenter target, which may be improved using MRI tractography with diffusion tensor imaging or atlas-based libraries. Notwithstanding these limitations, this study represents a key initial step in generating foundational evidence needed before proceeding to larger, more definitive randomized trials comparing treatment approaches.

In summary, frameless radiosurgical thalamotomy appears to be a safe and effective treatment of tremor worthy of further exploration in randomized trials.

Data sharing:

Research data are stored in an institutional repository and will be shared upon reasonable request to the corresponding author.