RFT is division-selective; preoperatively, patients must be trained to cooperate and localize pain during the procedure; therefore, the patient must remain awake during the procedure. Occasionally, short-acting anesthetic agents, such as propofol, can be used for the induction of sedation. To prevent bradycardia due to the trigeminal depressor response, atropine was administered before the procedure. The patient was positioned supine with slight flexion of the head at 15°. Hartel’s anatomic landmarks were drawn on the ipsilateral face of the patient, inferior to the medial pupil, 3 cm anterior to the external auditory meatus, and 2.5 cm lateral to the lips (Fig. 1). The skin insertion point of the guide cannula was lateral to the lips, with the trajectory guiding towards the other two landmarks. With the help of landmarks and C-arm fluoroscopy, the guide cannula was inserted and advanced along the target trajectory to the foramen ovale (Fig. 2). To prevent violation of the oral cavity and buccal mucosa, the surgeon can direct the cannula with a gloved hand. The electrode penetration should be avoided by placing it 10 mm beyond the clivus because the trochlear and abducens nerves are present there. After confirming the cannula position with C-arm fluoroscopy, the obturator was removed and the electrode was inserted. Test stimulation was performed by asking the patient to localize pain and thereby side effects could be prevented. Alternatively, the patient can be under sedation throughout the surgery, and combination of fluoroscopy and nerve stimulation can be used to verify the appropriate localization. Thermocoagulation was started at the previously tested locations, and lesions were made at a maximum of 0.5 V at 55-80°C for 30 seconds to 2 minutes per cycle. After the patient achieved pain relief, with normal sensory function being conserved, the cannula was removed. The procedure was completed with compression and dressing of the puncture site. The patient remained stable after the procedure and was discharged on the same day.

RFT reduces pain by producing mild hypoalgesia, without causing significant sensory deficits in the affected trigeminal divisions. Pain relief outcomes are favorable, with most studies reporting >95% pain resolution [23,29,59,60]. The reported immediate response rate to RFT was 97.6-99%, and the response rate at 6 months of follow-up was 83.3-89.9% [23]. The recurrence rates varied from 38.2% at 1 year to 10% at 6.5 years of follow-up [23]. A large series reported that after 20 years, 41% of the patients remained in pain control [41]. The degree of sensory reduction correlated with pain control and the duration of pain relief [41,60]. In a prospective study of 154 patients, those with more hypoalgesia had lower rates of pain recurrence but higher rates of dysesthesia, whereas patients with mild hypoalgesia had higher rates of pain recurrence but lower rates of dysesthesia [56]. Type I TN is associated with better outcomes after RFT; in contrast, bilateral pain and psychiatric comorbidities are associated with worse outcomes [20,27]. RFT is effective for multiple sclerosis-related type II TN but requires the induction of anesthesia in the affected trigeminal divisions [41]. The advantage of RFT is that it allows for the most selective targeting of the divisions compared to any other percutaneous procedure [27]. Few studies have examined the predictors of outcomes after RFT [51]. In a study by Kosugi et al. [30], outcomes were superior for isolated third-division pain than for second-or multi-division pain.

Although RFT is durable and has a higher pain control rate than other percutaneous procedures, it is associated with relatively higher rates of side effects [60]. Side effects, such as dysesthesia, corneal anesthesia, and masseter weakness, may occur and seem to be related to the irreversible damage of small, unmyelinated pain fibers in the trigeminal nerve [38,54]. Mild-tomoderate facial numbness is a consequence of RFT, but dysesthesia may also occur [60]. Higher coagulation temperatures for RFT may be related to hypesthesia [15], which lasts for at least 1 month in 3.3% of the patients [27]. There is a 1-20% rate of decreased corneal sensation, with a 1% rate of conjunctivitis [23,29,41,57,60]. Approximately 3-29% of patients experience masseter weakness [60]. Masticatory weakness and diplopia mostly recovered within 6-12 months [60]. Rates of anesthesia dolorosa ranged from 0.6% to 0.8%, which are higher than those for percutaneous balloon compression (PBC) [29,41]. Other complications such as meningitis, hemorrhage, and carotidcavernous fistula formation can occasionally be observed [57,60]. Taha and Tew [57] reported that the use of a curved electrode allowed the creation of more selective lesions and decreased the rates of sensory complications.

The procedure is performed under sedation with a shortacting anesthetic agent such as propofol. Atropine can be preadministered to prevent the trigeminal depressor response. A 20-gauge needle is directed using Hartel’s landmarks to access the foramen ovale and is followed by C-arm fluoroscopy to guide and confirm the needle. After accessing Meckel’s cave through the foramen ovale, cerebrospinal fluid often drains from the needle, confirming its position. Contrast-enhanced cisternography or injection of 1 cm³ of air can be used to outline the trigeminal cistern. The patient’s position is then changed to 60° head-up-tilt, and glycerol is injected. The volume of glycerol injected varies from 0.25 to 0.40 mL. Different volumes can be used to target different divisions. For multidivision pain, injection of the full volume treats all divisions. The needle is withdrawn after the glycerol is injected, and the patient is awakened to remain in a sitting position for 2-3 hours so that the glycerol remains in Meckel’s cave. The patient can be discharged on either the same day or the next day.

The outcomes of GR vary significantly among the studies. The reported immediate and complete pain relief rates are 53-98% [3,5,14,20,23,42]. Outcomes vary in degrees of postoperative hypoalgesia; one study reported mild-to-moderate hypoalgesia in up to 70% of cases but lasting no more than 3-6 months [60]. GR has a mean pain control duration of 11 months [20]. At the 1-year follow-up, there was a 53-63% rate of pain control [5,6,12]. Recurrence rates were 72% after 54 months in one study [24] and 56.5% after 5 years in another study [23]. Another study reported 18.8% in a 10-year follow-up, and most recurrent cases occurred within the initial 3 years [14]. In a study by North et al. [42], female sex, type I TN, prior effectiveness of carbamazepine, shorter pain duration, and intraoperative cerebrospinal fluid outflow were associated with longer pain-free duration. In another study by Pollock [49], non-constant preoperative facial pain, immediate pain during glycerol injection, and new postoperative trigeminal deficits were related to a chance for better outcomes. Other potential prognostic factors for initial pain relief include postoperative hypoesthesia, pain-free intervals, and high preoperative pain scores [7,48]. Intact preoperative facial sensation and no history of prior procedures have been found to be prognostic factors for durable pain relief [60]. Several early reports suggested that cerebrospinal fluid outflow during the procedure indicates improved outcomes and lower recurrence rates; however, in more recent studies, this was not significant [14,45,60]. Three or more previous GR histories were associated with shorter recurrence times [8]. GR may relieve intermittent pain for type I TN but not the constant pain for type II TN; outcomes are worse for type II TN after GR, with a high recurrence rate of 78-100% within a year [12,20]. Another study reported a 59% recurrence rate at a mean of 17 months [48].

Postoperative hypoesthesia occurs in 23.3-72% of cases, and, typically, the facial numbness resolves within hours to days after the operation [8,12,14,23,24,60]. With repeated procedures, higher rates of hypoesthesia occur [24]. Similar to other percutaneous treatments, facial numbness is positively correlated with pain relief. Severe hypoalgesia and analgesia are rare, but patients undergoing repeat procedures are at a higher risk [50]. There is a low risk of anesthesia dolorosa ranging from 0% to 5% and masticatory weakness ranging from 0% to 4.1% [60]. The decreased corneal sensation is more common than in balloon compression, which occurs in up to 16% of cases [12,14,41]. There is a possibility of herpes reactivation in one series, presenting in 77% of postoperative cases [21]. Aseptic or bacterial meningitis can occur in 0.12-3% and 1.5-1.7% of cases, respectively [3,8,12,14]. Carotid puncture occurs at a rate of 0.77% [7,8]. Penetration of the buccal mucosa and buccal hematoma can occur at rates of 1.5% and 7%, respectively [7,8]. Hearing loss was reported at a rate of 1.9% [8].

PBC is performed under general anesthesia. Similar to other percutaneous procedures, a trigeminal depressor response can be evoked. However, atropine is only prepared at the bedside but not administered preoperatively because of intraoperative monitoring of trigeminal compression. Thus, in cases of persistent hypotension and bradycardia, careful monitoring is necessary, and a transcutaneous or transesophageal pacemaker should be attached. With a 15° head elevation, Hartel’s landmarks are used in the same manner as RFT. A 14-gauge needle is inserted and advanced along the target trajectory to the foramen ovale. C-arm fluoroscopy is performed to guide the needle. Although not always observed, entering the trigeminal cistern may result in the outflow of cerebrospinal fluid. Proceeding through the foramen ovale can evoke a trigeminal depressor response. A straight guiding stylet is passed through the cannula under direct visualization using C-arm fluoroscopy. In the anteroposterior view, the stylet is heading toward the porus trigeminus, which is approximately 17 mm beyond the foramen ovale. Pointing the stylet towards the center of the porus is thought to treat second-division or multi-division pain, whereas the lateral position is thought to treat third-division pain and medial for first-division pain. When the stylet is placed in the appropriate position, it is then removed, and a 4-F balloon catheter is inserted and passed towards the target position. The balloon is inflated with 0.7-0.75 mL of contrast agent to a target pressure of 1000-1200 mmHg for 60-90 second. C-arm fluoroscopy visualizes the needle and balloon by showing a typical pear or dumbbell shape in the lateral view, which reflects their position within the porus. During balloon compression, the trigeminal depressor response can be observed again, which is an indicator of appropriate compression. After the procedure, the cannula and catheter are removed and light manual compression and dressing are applied to the skin puncture site. The patient can be discharged the same day or the next day.

The rate of initial pain relief ranged from 82% to 93.8% [3,5,10,23,33,34,52]. Repeat balloon compression for patients with recurrent TN showed a pain relief range of 83-93.8% [13,43]. At a mean follow-up duration of 4 years, the pain control rate was 69.4% [1]. In another study with a median of 64.8 months of follow-up, 54.5% of patients remained pain-free [43]. Rates of pain recurrence range from 20% to 48%, and one report showed that 90% of patients with a history of surgery showed recurrence rates of 64% [32].

Achieving a typical dumbbell shape or pear shape during balloon compression has been considered a prognostic factor in several studies [3,32]. However, the shape eventually represents the correct positioning of the balloon catheter and an appropriate degree of compression pressure [60].

The time to recurrence also varies widely among studies; at 1 year, there is a 25.4% recurrence rate [41], with recurrence rates ranging from 19.2% to 29.5% after 5 years in another study [60], and a 31.9% recurrence rate at a mean of 10.7 years of follow-up [52]. Time to recurrence may be associated with balloon compression time and the duration and degree of hypoesthesia [13,31]. However, some studies have suggested that a longer compression time does not affect pain relief and only increases the complication rate [38]. Recently, trends have changed to shorter balloon compression times, with <60 seconds being considered adequate for achieving pain relief [32,52,60].

Common complications of PBC include numbness, masseter weakness, and dysesthesia. Although higher balloon pressure offers greater pain relief, it is reported that higher balloon pressures are also associated with higher rates of complications [10,11,33,62].

Most patients experience postoperative facial numbness, and 4.6-40% of cases have persistent symptoms that typically resolve within 3 months [23,33,34,52]. Approximately 1-11.4% of patients experience dysesthesia, which tends to decrease with less compression times [33,60]. Most cases of dysesthesia are minor and temporary [15]. Trigeminal nerve dysfunction may be related to pressure rate and compression time [1,23,33]. Anesthesia dolorosa has been reported in 0-3.4% of cases [43,52].

Postoperative masseter weakness due to injury in large myelinated fibers is reported in 10-50% of studies [19,23,33,34,52,60]. However, weakness mostly recovers within a few weeks to months [33]. Postoperative corneal anesthesia is reported to occur in 0-3.1% of cases [1,52,60]. The low rates of corneal anesthesia are associated with nerve fiber selectivity of the procedure, which selectively injures medium to large myelinated fibers while preserving small myelinated or unmyelinated fibers [15]. Repeat PBC procedures are associated with a higher rate of masseter weakness and corneal anesthesia [13,43]. Misplacement of the balloon can result in cranial nerve III, IV, and VI injuries, while the abducens nerve is most commonly injured, resulting in diplopia [2,5]. Herpes reactivation can occur after procedures, which is typically mild and does not require prophylactic antiviral therapy [19,60]. Meningitis is rare, ranging from 0% to 5% of cases [60]. Aseptic meningitis is reported in 0.7% of meningitis cases and bacterial meningitis in 0.7-1% [19,34]. Hearing loss occurs in 2.4-6.3% of cases, cheek hematoma in 3.5-6.7%, and pseudoaneurysm in 1% [1,19,23,33,34,52]. PBC has a higher rate of trigeminal depressor response than other percutaneous procedures, and bradycardia and brief hypotension caused by the trigeminal depressor reflex make it less suitable for patients with underlying cardiac disease [7,15,58].

Anesthesia dolorosa, a serious complication, is relatively rare with rates ranging from 0% to 3.4% of cases [43,52]. Three cases of postoperative corneal keratitis and one case of recoverable blindness have been reported [2,32,43,60].