This study is a randomized, single-blind (participant-blinded), placebo-controlled, parallel-group clinical trial conducted at the University of Osaka Hospital in Japan. Because pain is highly susceptible to placebo effects, a sham stimulation group was included as the control group to evaluate the superiority of active stimulation. A parallel-group design was selected instead of a crossover design owing to concerns about potential carryover effects and the risk of unblinding. To further explore potential differences in efficacy, 2 active stimulation sites were designated. The study flowchart is shown in Figure 1.

This study is an investigator-initiated clinical trial. Trial monitoring and auditing will be conducted by an independent academic research organization (Department of Medical Innovation, University of Osaka Hospital). Monitors will perform oversight in accordance with the monitoring plan to ensure data reliability and participant protection, verifying compliance with applicable laws, regulations, and the trial protocol, and will prepare monitoring reports. Data will be captured using an electronic data capture system (REDCap; Research Electronic Data Capture; Vanderbilt University), which minimizes input error, duplication, and missing data while ensuring security and transparency. This paper is based on protocol version 1.3 dated March 25, 2025.

We will enroll patients aged 18 years or older with NP who meet the following inclusion criteria: (1) NP persisting for more than 3 months, (2) pain intensity score of 4 or more on the numerical rating scale (NRS) at screening, and (3) written informed consent to participate in this study. The definition of NP follows the International Association for the Study of Pain terminology [1], adopting a classification of probable or definite NP according to the NP grading system [16]. Exclusion criteria are as follows: (1) dementia with a Mini-Mental State Examination (MMSE) score of 23 or lower; (2) severe aphasia or cognitive dysfunction; (3) serious psychiatric disorder; (4) history of epileptic seizures; (5) use of an implantable stimulator, such as a cardiac pacemaker, except for implantable spinal cord stimulators; (6) metallic implants in the head, except for titanium products; (7) use of an implantable drug delivery system or implantable ventricular assist device; (8) pregnancy; (9) inability to complete the assessment questionnaires; (10) participation in any other clinical trial within 3 months prior to providing consent; and (11) any other condition deemed inappropriate for inclusion by the investigators. Participants will be recruited from outpatients of the Department of Neurosurgery at University of Osaka Hospital.

Participants will be withdrawn from the study if any of the following criteria are met after enrollment: (1) the investigator determines that continuing the study poses an unacceptable risk owing to the occurrence of an adverse event; (2) the participant withdraws consent or requests to discontinue participation; (3) it is discovered that the participant does not meet the eligibility criteria; (4) the participant is unable to complete the required assessments or visits as a result of personal circumstances, such as relocation; or (5) the investigator determines, for any other reason, that discontinuation of the intervention is appropriate.

Participants will be randomly allocated in a 1:1:1 ratio to one of three groups: (1) active stimulation targeting M1, (2) active stimulation targeting PSI, or (3) sham stimulation. Because the primary outcomes rely on participant-reported subjective assessments, this study will adopt a single-blind design in which participants remain blinded to group allocation throughout the trial. Blinding of the investigators delivering the intervention is not feasible with the current device. As this is a preliminary study in which most outcomes are patient-reported, investigators are not blinded. To minimize expectation effects, all groups will undergo identical procedures, including treatment duration, interaction with study personnel, and auditory masking using white noise. The success of participant blinding will be evaluated at the end of the intervention period using a blinding assessment questionnaire.

Randomization will be conducted using a stratified permuted block method, with the underlying cause of NP (central or peripheral origin) as the stratification factor. The allocation table and procedure will be generated by an independent technician who is not involved in patient enrollment or outcome assessment. Assignment will be managed through the REDCap randomization module, which enhances secure handling of allocation information using password protection.

All patients who provide consent are assessed for eligibility by neurosurgeons specializing in pain management. Eligible participants will be randomly allocated to 1 of the 3 groups and undergo baseline evaluations, including demographic and clinical characteristics. The intervention is administered once a week for 8 weeks (Figure 2). This schedule was determined based on previous rTMS studies, which showed that intermittent stimulation—such as weekly sessions—provides significant pain relief for between 4 and 8 weeks or sessions, after which the effect tends to plateau [17-20]. Participants will be asked to record a pain diary from 1 week prior to the intervention until 1 week after completion at week 8. The NRS for pain and paresthesia (shibire in Japanese), along with the Short-Form McGill Pain Questionnaire version 2 (SF-MPQ-2) [21], will be assessed at baseline, immediately after each intervention, and at follow-up visits. Additional measures—including the Brief Pain Inventory (BPI) [22], EQ-5D-5L [23], Hospital Anxiety and Depression Scale (HADS) [24], Pain Catastrophizing Scale (PCS) [25,26], Pain Self-Efficacy Questionnaire (PSEQ) [27], MMSE, quantitative sensory testing (QST), motor cortical excitability, and the shibire questionnaire (validated by our group; A Yamamoto, unpublished data, October 2024) [28]—will be assessed at baseline and immediately after the final intervention at week 8. At the end of the study, Patient Global Impression of Change (PGIC) and blinding assessments will be collected. Adverse events will be monitored continuously from the start of the intervention until follow-up visits at weeks 12 and 16 or until study discontinuation. Potential adverse events associated with TUS will be explained to participants during the informed consent process. In addition, at each treatment session, participants will be asked whether they have experienced any new symptoms or deterioration. The detailed schedule is shown in Table 1.

Given the exploratory nature of this study, the sample size was not intended to provide a definitive or fully powered estimate of treatment efficacy. As no previous clinical trials have examined the effects of repeated sessions of TUS in patients with chronic pain, the sample size calculation was based on a prior clinical trial of rTMS for NP [8]. In that study, the mean reduction in pain score immediately after the final intervention was 34.8 (SD 15.2) in the active stimulation group (n=15). Assuming that the sham group would experience half of this effect (mean reduction of 17.4) [3,8] and that both active TUS groups in our study would yield the same mean reduction (34.8), the SD was set at 15.2 for all groups. On the basis of these assumptions, a one-way ANOVA with an α of .05 and 80% power indicated that 37 participants would provide a reasonable basis for estimating potential between-group differences. Accounting for a 5% dropout rate, the final target sample size was set at 39 (13 per group). The anticipated dropout rate of 5% was based on a previous rTMS trial conducted by our group in patients with NP, in which the dropout rate was below 5% [8].

TUS will be administered using the NeuroFUS system (Sonic Concepts), which consists of a 4-element annular array transducer (CTX-500; Sonic Concepts) with a diameter of 60 mm and a fundamental frequency of 500 kHz. The transducer is powered by a programmable radiofrequency amplifier (TPO-203; Sonic Concepts), which controls the phasing of the 4 elements to adjust the sonication depth. A neuronavigation system (Brainsight; Rogue Research Inc) will be used to identify the stimulation site and precisely monitor and maintain the position and orientation of the transducer throughout all treatment sessions.

The stimulation target will be either M1, corresponding to the most painful region, or the PSI, both located contralateral to the pain. When the M1 hand area is selected, the motor hot spot identified through cortical excitability measurements using transcranial magnetic stimulation (TMS) will be used. The PSI will be localized on individual 3D magnetic resonance imaging (MRI) in the navigation system according to the previously reported “quadrant-within-a-quadrant” method [29]. For all other targets, stimulation sites will be determined on the individual MRI with reference to the Montreal Neurological Institute-Hospital (MNI) space standard atlas integrated into the navigation system [30]. The M1 foot target will be defined using the MNI coordinate (–7 or +7 mm; –27 mm and 59 mm) as the default location and subsequently adjusted based on each participant’s structural MRI. The central sulcus will be identified on axial images and traced medially, whereas the marginal ramus of the cingulate sulcus will be identified on the midsagittal image. The target will be placed in the anterior part of the paracentral lobule, immediately anterior to the central sulcus. If the default MNI coordinate falls within a sulcus, the target will be adjusted to the nearest cortex within the M1 foot area. Because the TUS device achieves optimal efficiency at a focal depth of 48.5 mm, an ultrasound gel pad of appropriate thickness (Echo Gel Pad; Yasojima Proceed Co., Ltd) will be placed between the transducer and the scalp to adjust the distance from the transducer surface to the stimulation target to within 43 to 54 mm. Stimulation parameters will be as follows: pulse duration of 10 ms, pulse repetition interval of 100 ms (10 Hz), pulse train duration of 10 seconds, pulse train repetition interval of 30 seconds, duty cycle of 10%, 20 trains, total stimulation duration of 580 seconds, and spatial peak pulse average intensity in free water of 30 W/cm². The temporal stimulation pattern will be designed with reference to stimulation paradigms used in rTMS for refractory NP, consisting of intermittent 10-Hz pulse trains delivered over approximately 10 minutes [17,18,31]. Selected acoustic parameters, including stimulation frequency and intensity, will be informed in part by the human TUS protocol reported by Osada et al [32] using the same TUS device, although the temporal stimulation parameters will be modified for the present study. The stimulation intensity will be set within a range that has previously been applied without serious adverse events and remained within the safety limits recommended by the recent International Consortium for Transcranial Ultrasonic Stimulation Safety and Standards consensus statement [33]. The mechanical index, estimated using the TUS calculator [34] at a focus depth of 43 mm, was 1.32. On the basis of simulations reported by Osada et al [32], the maximum temperature increase will be estimated to be 0.5 °C in brain tissue and 0.8 °C in the skull. Extrapolating these values according to the effective sonication time used in the present study will yield conservative maximum estimates of 0.83 °C for brain tissue and 1.33 °C for the skull assuming no heat dissipation [32]. As the increase remained below 2 °C and the intervention time was less than 10 minutes, the thermal dose was estimated to be below 0.25 CEM43. The thermal index for the skull was calculated as 1.9. These values are within the safety thresholds recommended by the recent International Consortium for Transcranial Ultrasonic Stimulation Safety and Standards consensus on TUS safety [33].

Treatment will be administered with the participant seated comfortably in a reclining chair. Ultrasound gel will be applied to the scalp, gel pad, and transducer, with all visible air bubbles carefully removed. The transducer will then be affixed to the scalp via a gel pad to ensure complete contact. To maintain blinding, participants will not be shown the TUS device interface, and white noise will be delivered through earphones during stimulation to mask any device sounds. All interventions will be performed under the supervision of investigators who are also neurosurgeons specialized in neuromodulation. In the sham stimulation group, the transducer will be placed over M1, but no ultrasound will be emitted (0 W output), with the same device setup and procedures as in the active stimulation groups. Because TUS is not typically associated with distinct sensory perceptions during stimulation, participants are not expected to distinguish active from sham stimulation based on treatment sensation alone. A single sham condition will be used as a common control group for both active stimulation targets. Because participants will be assigned to only 1 treatment arm throughout the study and will not be exposed to multiple stimulation targets, differences in transducer location will not be expected to substantially affect participant blinding.

To minimize confounding effects, the initiation of new treatments for NP and alterations to existing regimens will, in principle, be prohibited from the time of enrollment until week 8. The use of rescue analgesics will be permitted during the study provided that it does not substantially deviate from the participant’s preintervention medication regimen. Participants will be instructed to maintain their existing analgesic treatment whenever possible throughout the study period. Exceptions may be permitted when required to ensure participant safety.

Efficacy analyses will be performed on the full analysis set, defined as all randomized participants who received at least one stimulation session and had at least one postintervention outcome assessment (intention-to-treat analysis). A per-protocol set excluding participants with major protocol deviations will also be analyzed for sensitivity. Major protocol deviations will include violations of eligibility criteria identified after enrollment, withdrawal of consent, substantial nonadherence to the intervention protocol, and other deviations judged to have a significant impact on efficacy evaluation. The consistency of results between the full analysis set and the per-protocol set will be examined to evaluate the robustness of the primary findings. Any modifications or additions to the analyses after study initiation will follow an evaluation of their validity and potential impact, with corresponding amendments to the statistical analysis plan. Handling of outliers will be prespecified, and depending on the variable, either data transformation or statistical approaches robust to outliers will be applied.

Demographic characteristics and outcomes will be summarized descriptively, with longitudinal outcomes presented as line graphs over time. The primary analysis will use a linear mixed-effects model for repeated measures (MMRM) to evaluate changes in weekly average pain scores from the pain diary. The model will include change from baseline in weekly average pain scores as the response variable, with treatment group, time point, and their interaction as fixed effects and individual participants as a random effect. Baseline pain scores will be included as covariates. The primary treatment effects will be estimated using model-based contrasts comparing M1 stimulation with sham stimulation and PSI stimulation with sham stimulation at week 8. Multiplicity arising from these 2 primary comparisons will be controlled using the Dunnett adjustment. To identify time points showing significant changes from baseline, the Dunnett multiple comparison procedure will be used, with the baseline week serving as the control. The number and proportion of responders will be summarized by treatment group, and between-group comparisons will be performed using logistic regression models adjusted for baseline pain scores. Missing week 8 pain diary data will be imputed using the last observation carried forward method for responder analyses only. Secondary outcomes measured repeatedly, such as pain NRS and SF-MPQ-2, will also be analyzed using MMRM in the same way as the primary outcomes. Outcomes assessed only at baseline and week 8—including BPI, EQ-5D-5L, HADS, PCS, PSEQ, MMSE, QST, and MEP—will be evaluated using analysis of covariance, with change from baseline as the response variable, treatment group as a fixed effect, and baseline value as a covariate. Paresthesia-related outcomes will be analyzed using MMRM, analysis of covariance, or logistic regression depending on data type and timing. PGIC responses at week 8 will be summarized by treatment group and compared using the Fisher exact test. All analyses will be 2 sided, with P values below .05 considered statistically significant.

Patients and members of the public will not be involved in the design, recruitment, or conduct of this study.

This study will be conducted in accordance with the ethical principles of the Declaration of Helsinki and the Clinical Trials Act of Japan. The study protocol was reviewed and approved by the University of Osaka Clinical Research Review Board (approval number S24007). All protocol amendments will be reviewed by the committee. Participants will be provided with detailed information about the study by the investigators, and written informed consent will be obtained before enrollment. Participation is voluntary, and consent may be withdrawn at any time. To protect confidentiality, each participant who provides informed consent will be assigned an identification code, ensuring that individuals cannot be directly identified. The correspondence table linking codes to personal information will be securely stored to prevent any disclosure of personal data outside the trial.

The clinical study report will be made publicly available via the Japan Registry of Clinical Trials website following review by the Clinical Research Review Board and the institutional administrator. Findings will also be disseminated through scientific presentations and peer-reviewed publications, with strict adherence to confidentiality.

During peer review of this protocol manuscript, the Statistical Analysis section was revised to explicitly specify the prespecified primary treatment comparisons (M1 stimulation vs sham stimulation and PSI stimulation vs sham stimulation at week 8) and the planned multiplicity adjustment.