Safety and Efficacy of Radiofrequency Ablation of Thyroid Nodules

Expanding Treatment Options in the United States

Iram Hussain; Fizza Zulfiqar; Xilong Li; Shahzad Ahmad; Jules Aljammal


J Endo Soc. 2021;5(8) 

In This Article


Study Population and Design

A retrospective chart review was conducted to identify all patients who underwent RFA of their thyroid nodules between November 2018 and January 2021, at The Thyroid Clinic, in Salt Lake City, Utah. A total of 53 patients with 58 nodules (30 NFTNs and 28 AFTNs) were identified.

The criteria for treatment with RFA was as follows:

  1. NFTNs were treated if they were causing compressive symptoms or cosmetic concerns; and were predominantly solid or solid-cystic (>25% solid) on ultrasonographic imaging.

  2. AFTNs were treated in cases of overt hyperthyroidism or subclinical hyperthyroidism with symptoms and/or risk factors for adverse outcomes, namely, advanced age, decreased bone density, and heart disease (heart failure, coronary artery disease, or atrial flutter/fibrillation); or in case of symptoms regardless of thyroid function.

In patients with more than one thyroid nodule, only the nodules that met the above criteria were treated with RFA. Nodules not treated with RFA were followed with ultrasonographic imaging.

Data were extracted from the electronic medical record to determine demographics of the patients, ultrasound characteristics of the nodules, compressive symptoms, cosmetic scores, preablation volume, postablation volume, complications, and effect on thyroid function studies.

Treatment success was defined as a volume reduction of more than 50% within 6 months after RFA with improvement in symptoms and/or cosmetic concerns, and in case of AFTNs, normalization of thyroid function tests and resolution of symptoms within 12 months after RFA.

The protocol of this study was reviewed and approved by the University of Texas Southwestern Institutional Review Board.

Preablation Assessment

All patients were evaluated clinically and had neck ultrasonography and laboratory blood tests prior to the RFA procedure. Patients with presence of metallic hardware/pacemaker, bleeding diathesis, pregnancy, or nodules that appeared suspicious on ultrasonographic imaging but had benign results on ultrasound-guided fine needle aspiration (UG-FNA) were not considered to be candidates for RFA. Written and verbal informed consent was obtained from all patients prior to performing RFA.

Thyroid function tests including serum thyrotropin (TSH; thyroid-stimulating hormone) and free thyroxine (T4) levels were performed, along with blood platelet count and coagulation tests (including prothrombin time and activated partial thromboplastin time) prior to RFA in all patients, with the exception of 1 patient who had a normal TSH level but did not have a free T4 level. Eighteen patients also underwent testing for thyroid peroxidase antibodies (TPO Ab) prior to RFA, as elevated levels of TPO Ab can increase risk of developing hypothyroidism.[30]

All patients with suppressed TSH underwent a pretreatment radioactive iodine uptake and scan, and autonomously functioning thyroid nodules were identified. One patient had symptoms of hyperthyroidism in the presence of a thyroid nodule on ultrasonographic imaging but had normal levels of TSH and free T4. She also underwent a radioactive iodine uptake and scan that identified the AFTN.

Patients with NFTNs underwent 2 separate ultrasound-guided fine needle aspiration (UG-FNA) biopsies showing benign cytopathology (Bethesda II), with the exception of 1 patient who had one UG-FNA biopsy with benign cytopathology (Bethesda II) and a second UG-FNA biopsy with indeterminate cytopathology (Bethesda III) followed by negative molecular marker studies (Thyroseq v3 genomic classifier showed low probability of malignancy). Patients with AFTNs each underwent 1 UG-FNA biopsy that showed benign cytopathology (Bethesda II), except for 1 patient whose cytopathology showed atypia of undetermined significance (Bethesda III) followed by Thyroseq v3 genomic classifier showing low probability of malignancy.

Ultrasonography was performed using the MyLab Gamma ultrasound system (Esaote North America) with a linear matrix array transducer (SL1543) operating at 4 to 14 MHz and included color Doppler imaging. Each nodule was measured in 3 dimensions and the nodule volume was calculated using the ellipsoid volume formula:

An objective cosmetic score was obtained using a 1 to 4 scale as reported in the 2017 thyroid RFA guidelines by the Korean Society of Thyroid Radiology, where 1 is no palpable mass; 2 is no cosmetic problem but presence of palpable mass; 3 is a cosmetic problem on swallowing only; and 4 is a readily visible cosmetic problem.[31] The presence or absence of compressive symptoms was also recorded in all patients.

Radiofrequency Ablation Procedure

RFA involves insertion of an internally cooled electrode into the target nodule. The electrode is connected to a generator that produces a high-frequency alternating current. This causes vibration of ions in the tissue in contact with the exposed tip, resulting in thermal injury and coagulative necrosis in the target nodule. The ablation is followed by shrinkage of the lesion over time.

The procedure was performed in an outpatient setting using standard aseptic techniques and local anesthesia for pain control. Patients were placed in supine position with hyperextended neck, and target nodule and vital cervical structures were visualized with ultrasonography in real time. To ensure safety and achieve maximum efficacy, the RFA was performed using limited hydrodissection, a trans-isthmic approach, and the 'moving shot' technique.[31]

Anesthesia. Under ultrasound guidance, using a 25-gauge needle, 3 to 8 cc of 2% lidocaine without epinephrine was injected under the skin and into the thyroid capsule and perithyroidal area to provide adequate analgesia. Patients also had additional mild conscious sedation with midazolam and/or fentanyl, and 1 patient required ketamine, administered by an anesthesiologist who monitored the patients' hemodynamics throughout the procedure. There were no problems reported with anesthesia. All patients remained alert, coherent, and verbal throughout the procedure, which is important for early detection of vocal cord paresis/paralysis during the procedure.

Hydrodissection. Hydrodissection is a technique used to separate the target lesion from surrounding structures in the neck, such as the carotid artery, recurrent laryngeal nerve, and anterior cervical muscles. It involves injecting either lidocaine (a total volume of 20 mL should not be exceeded in any patient to avoid lidocaine toxicity) or dextrose 5% in water in between the nodule and adjacent structures to create a safety margin that prevents thermal damage to these critical structures.[31] In our study, limited hydrodissection was performed for all patients. The 2% lidocaine without epinephrine used for anesthesia was injected to separate the anterior thyroid capsule (and anterior cervical muscles) from the thyroid parenchyma. Additional hydrodissection (eg, to separate the nodule from neurovascular bundle) was not required due to the location of the nodules.

Trans-isthmic Approach. The trans-isthmic approach refers to inserting the electrode from the midline of the neck and advancing it laterally into the target nodule, thus limiting heat exposure to the recurrent laryngeal nerve located in the tracheoesophageal groove and/or esophagus (Figure 1). In addition, this makes it harder for the electrode to change positions when the patient talks or swallows, compared with the lateral to medial approach, and also prevents leakage of hot fluid into the perithyroidal area.[32]

Figure 1.

Ultrasonographic image (transverse view) of radiofrequency ablation electrode in thyroid nodule. White arrow pointing towards electrode inserted in the middle part of the thyroid nodule (TN) using the 'trans-isthmic' approach (from medial to lateral) with tip surrounded by hyperechoic area indicating ablated zone. In this approach, the electrode tip points away from the trachea (T) limiting heat exposure to the recurrent laryngeal nerve located between the trachea (T) and the inferior pole of the thyroid lobe. Inserting the electrode from medial to lateral, rather than lateral to medial also avoids the neurovascular bundle. Abbreviations: ACSM, anterior cervical strap muscles; CCA, common carotid artery.

Moving Shot Technique. In this technique, the nodule is ablated bit by bit, starting at the inferior-most posterolateral part and gradually moving medially and anteriorly, with the electrode pulled back along its longitudinal axis following the same track as the initial advancement as the ablation is continued.[33] As each area is ablated, it becomes hyperechoic on ultrasonography, indicating that the electrode must then be repositioned within the nodule to continue the ablation. The process is then repeated, with overlapping, to completely ablate the inferior part of the nodule, followed by the middle part and then the superior part, until the entire nodule is ablated. This technique was used for all patients.

Electrodes and Generator. In 55 nodules, RFA was performed with an 18-gauge internally cooled electrode (STARMed, Seoul, South Korea), that was 7 cm in length with a 0.7-cm active tip and was powered by the VIVA RF generator (STARMed). In 3 nodules, similar 18-gauge internally cooled electrodes, 7 cm in length with 0.7-cm active tips from RF Medical powered by RF Ablation System V-1000 (RGS Healthcare), were used. For all patients, an initial power of 30 watts (W) was used for the ablation and this was increased in increments of 5 to 10 W every 10 seconds, up to a maximum of 55 W. The median power used was 35 W (range, 30–55 W), with median active ablation time of 3 minutes (range, 32 seconds to 11 minutes 13 seconds).

A total of 53 patients underwent a single RFA session for 58 nodules (5 patients had RFA of 2 separate nodules during the same treatment session on the same day). One patient with an AFTN required 2 RFA sessions 1 year apart.

Postablation Assessment

All patients were contacted by telephone 1 to 2 days after the RFA procedure to evaluate for immediate complications including pain, fever, hematoma or swelling, voice change (both immediate and after 24 hours), and onset of dysphagia.

Patients were evaluated with repeat ultrasonography and thyroid function tests 6 to 10 weeks after the RFA procedure; and were asked to return for follow-up at 6- and 12-month intervals. The volume reduction percentage (VRP) for each nodule was calculated using the following equation:

An assessment was made with regards to presence or absence of compressive symptoms following the RFA procedure, and postablative cosmetic score was also recorded. Patients were also evaluated for late-onset complications, including hoarseness of voice, transient thyroiditis, and nodule rupture. Complications were classified as either minor (no or nominal therapy with no long-term consequence) or major (requiring therapy, leading to hospitalization, or causing permanent adverse sequelae) based on previously reported criteria.[34]

Statistical Analysis

The categorical variables are presented by frequency (percentage) and continuous variables are presented by mean ± standard deviation for Gaussian distribution, and median (minimum and maximum values) for non-Gaussian distributed data. Comparisons between the AFTNs and NFTNs were made using the Fisher exact test for categorical variables, one-way analysis of variance (ANOVA) for Gaussian, and Kruskal-Wallis for non-Gaussian distributed data from continuous variables. The paired t test was applied to compare the pre- and post-RFA thyroid nodule volume changes, cosmetic score changes, and changes in laboratory values (TSH and free T4). A linear regression model was used to detect the association of baseline volume and volume changes after the procedure, adjusted by energy used, and Pearson correlation was used to see the relationship between volume changes and energy used.

Statistical analyses were performed using SAS 9.4 (SAS Institute, Cary, NC). A P value of < 0.05 was considered statistically significant.