Radiofrequency ablation (RFA) is an effective procedure with a low complication rate (1–3), indicating for those patients who with benign thyroid nodule (4, 5). However, risks exist with every procedure, thus planned follow-ups are essential. RFA-associated complications include voice changes, skin burns, hematoma, brachial plexus injury, nodule rupture, and transient hyper-/hypo-thyroidism (5, 6). The complication rate for benign nodules was 3.3%-5.6% (1, 6–8). More extensive physician experience is directly related to fewer complications (9).

Among complications, nodule rupture sometimes requires invasive interventions, including aspiration, drainage or even debridement or lobectomy (6, 10, 11). In one recently-published study involving 12 patients with thyroid nodule rupture after RFA, the nodule rupture could be further classified by the location of rupture (12). Although the majority of ruptures can be managed conservatively, some indeed require further intervention, creating a burden for both patients and surgeons. Currently, studies of the clinical factors affecting treatment post rupture are limited. The study aimed to evaluate the relationship among management, the clinical manifestations, and imaging features. The data of 9 patients with post-RFA goiter rupture were collected from two of the largest thyroid RFA centers in our country, while the 17 other ones were collected from literature review. The total 26 patients were divided into two groups, based on having received invasive or conservative treatment. The parameters affecting initial management were determined, and the odds ratio of each parameter was calculated by logistic regression analysis.

The baseline characteristics of the 9 cases in the study (7 females and 2 males, with mean age 38.9 ± 7.3 years) are summarized in Table 1 . The complication rate of nodule rupture was 1.1% (9/818). The ruptured nodules were predominantly solid (mean volume 73.9 ml; maximum 198ml). The mean ablation time and mean total energy per RFA session were 29 minutes (range, 9.4 to 50.7) and 16.6 kcal (range, 1.5 to 33.3), respectively. The median energy per RFA session was 50 W. The volume reduction rate (VRR) was 83.6 ± 5.8%. All patients were in euthyroidism status before RFA. The clinical and imaging features are demonstrated in Table 2 . The most common symptoms were sudden neck swelling, erythema, and pain, and one patient reported fever. The mean time from RFA to nodule rupture was 68.6 ± 55.7 days (range, 16 to 195). 3 patients were initially diagnosed with computed tomography (CT), these patients initially went to the emergency department to seek treatment. The other 6 patients were evaluated and diagnosed by US following the development of symptoms. The anterior type was the most common type of rupture (8/9, 89%), which was characterized by anterior nodule capsule disruption; CT revealed heterogeneous density of soft tissue bulging to the anterior side of the ablated nodule and tracheal deviation, which was shown in Figure 1 .

Only 2 patients were initially treated by analgesics, followed by close observation. Meanwhile, 7 patients (78%) received antibiotics treatment initially, and 6 of them required subsequent invasive procedures, including aspiration (1), incision and drainage (3), and surgical debridement (2). 4 patients required hospitalization, with the duration between 7-20 days. None of the patients experienced hypothyroidism after RFA. The review of literatures was presented in Table 3 . After combining the data of 9 patients in the 2 of our institutions with data of other 17 patients extracted from the previously published studies (10, 12), the data of 26 patients (mean age 41.5 ± 14.3 years, 17 females, 9 males) were evaluated. There were 13 patients receiving conservative therapies, and the other 13 ones required invasive management. The average duration to rupture was 54.8 ± 43 days. Most patients with rupture were diagnosed via ultrasound (23/26, 88.5%), which reflects its convenience and utility as a follow-up tool. Symptoms were primarily neck swelling and pain, while 4 patients suffered neck erythema and 2 presented with fever (7.7%). The anterior rupture was the most common type (22/26, 84.6%).

Nonparametric statistics identified significant differences in nodule diameter and the ablation time of the course prior to rupture (RUP time) between the two groups (p = 0.045 and 0.008, respectively) ( Table 4 ). The p-value of the nodule volume was marginally significant (p = 0.054), which may be limited by the relatively small sample size and maintained after further regression analysis. Two-tailed Pearson’s correlation revealed the initial nodule volume is moderately correlated with RUP time (r = 0.643), with high statistical significance (p <0.001). Logistic regression analysis indicated the nodule diameter and RUP time are associated with the requirement of invasive management, with OR 1.99 (1.07-3.67, p = 0.029) and 1.11 (1.01-1.22, p = 0.025) respectively (additional data are given in online resource, Supplementary Table 1 ). Based on the ROC curve analysis, when a nodule with the initial maximum diameter of >4.5cm ruptured, invasive management was suggested, with sensitivity of 69% and specificity of 79% (the area under curve, AUC 0.75). If the ruptured nodule occurred with an ablation time of >20mins, invasive management was suggested, with sensitivity of 62% and specificity of 69% (AUC 0.81). However, an ablation time of >20 mins appears not to be uncommon for a single RFA session. This has led to the diminished clinical application significance of this parameter. Therefore, we will focus on the impact of initial nodule size in the following discussion. All patients recovered without long-term sequelae.

Despite the frequently favorable prognosis after nodule rupture, some patients still require hospitalization for administration of intravenous antibiotics, or possible surgical intervention. If these patients can be identified earlier, it will result in reduced medical expenses and increased patient satisfaction. The present retrospective study enrolled patients from the 2 major thyroid RFA medical centers in the country. Furthermore, the literature review that identified patients having suffered nodule rupture post RFA and factors related to the determination of treatment methods were investigated. Notably, this study revealed that a larger initial nodule size and longer procedural time were associated with increased need for invasive management.

Previous studies have reported the complication rate of ablated nodule rupture was 0.19-2.5% ( Table 3 ). In one study which enrolled 1459 patients treated with RFA, the complication rate of nodule rupture was 0.19% (3/1543 nodules); indeed, the size of the ablated nodule had gradually decreased prior to rupture (6). The diagnosis of nodule rupture is based on clinical symptoms and images; while the most common symptom is sudden painful bulging neck mass (10, 12). In our study, the common symptoms were sudden neck swelling (89%), erythema (50%), and pain (33%). The US features included breakdown of the thyroid capsule with formation of fluid accumulation or new mass, indicating the interaction between intra- and extra-thyroidal spaces. Hematoma or intra-nodular hemorrhage during RFA can be controlled by direct compression and/or direct ablation at the bleeding point with the RF electrode (18). In 2019, various types of rupture were classified according to localization of rupture by Chung et al. (12). Although the anterior type was the most common, it was typically exhibited in the strap muscle layer. By comparison, with the contents bulging medially toward to the tracheoesophageal (TE) groove, the medial type may cause a mass effect on the trachea, leading to a compromised airway.

As compared to a multicenter retrospective study in Korea which included 1491 patients with 2616 RFA sessions wherein six cases (0.2%) with rupture were reported (10), an acceptable rupture rate was noted in our study. By comparing our patients with the 17 patients identified from previous studies (10, 12), a younger patient age (38.9 ± 7.8 vs. 39.4 ± 15.3, p 0.033) and larger ruptured nodule size (diameter 5.7 ± 2.3 vs. 4.1 ± 1.4 cm, p 0.018 and volume 73.9 ± 69.4 vs. 20 ± 20.8 ml, p <0.001) were noted in patients of the present study, which may be related to patient selection ( Table 5 ). Additionally, the RUP time was longer (29.3 ± 13.5 vs. 16.2 ± 11.8 mins, p= 0.896), and the median energy per RFA session was lower (50 (35-60) vs. 60 (50-85) W, p =0.16 with nonparametric statistics). The above findings reflect the current status of nodule rupture management, providing a valuable and practical reference for clinicians. Since larger nodule size and longer procedure time are associated with an increased need for invasive management upon rupture event, when a patient’s condition meets the abovementioned clinical criteria, initial intervention is recommended.

We compared the characteristics of our 9 patients, basing on need and not need of invasive management (6 and 3 patients, respectively), and no significant difference was noted on nodule diameter (p= 0.905), ablation time (p= 0.905), RF power (p= 0.548) nor time to rupture (p= 0.548). A longer duration between the RFA and rupture event was noted in the patients from our institutions (68.6 ± 55.7 vs. 47.5 ± 34.6 days, p 0.068), indicating that for larger nodules, well-timed and longer duration of follow-ups are necessary. Furthermore, detailed postoperative explanations, including self-care and vigilance of risk signs, including fever, sudden pain, and erythema should be communicated to the patient.

RFA leads to the formation of a fibrous capsule around the area of the coagulative necrosis, while central scattered hyaline sclerosis and scarring can be observed microscopically (21, 22). In addition, spontaneous intra-nodular hemorrhage may occur under pressure (such as coughing, crying, neck pressure or twisting) (23). The possible mechanisms include abnormal vessel anatomy with weakening of the veins and intra-nodular arterio-venous shunting, resulting in extravasation into the nodule under pressure, especially for patients with hypervascularity prior to RFA (24, 25). However, the cause of nodule rupture after RFA remains unclear, and is not necessarily related to trauma or delayed bleeding. Furthermore, the use of higher energy during the procedure, larger amounts of necrotic tissue after RFA, and more severe tissue edema may all contribute to higher risk of rupture for large nodules. The more obvious compression symptoms caused by a large nodule may also impel the physician to actively manage the event. One multicenter study published by Baek et al. suggests that nodule rupture may be caused by acute volume expansion due to hemorrhage, and conservative management was suggested (6). In our study, one female with a large thyroid mas (8.1 x 7. 9 x 5.9 cm) was treated with RFA. Sudden swelling and erythema was noted 19 days later, with CT demonstrating anterior nodule rupture ( Figure 1 ). During debridement, large volumes of mucous and turbid content were noted, without delayed hemorrhage. Therefore, imaging can be an effective diagnostic tool, however the evaluation of delayed bleeding primarily depends on the surgical or aspiration findings. Of the 9 patients in our study, 3 were prescribed oral analgesics, and 1 with oral steroid. 7 patients initially received treatment with antibiotics, but the improvements were limited; 4 patients required hospitalization for intravenous antibiotics administration, and 6 patients required subsequent invasive procedures.

This is currently the most comprehensive study to analyze the factors determining invasive treatment for nodule rupture after RFA. Still our findings should be interpreted in light of some limitations. Although the nodular size will affect the treatment, whether the use of steroids or other medications to reduce tissue swelling can reduce the hospitalization days or even prevent rupture requires further verification. The low probability of nodule rupture after RFA and diverse management modalities may mask unknown factors that can affect further treatment. Certainly, these two factors identified (initial nodule volume and ablation time) would probably not be significant, if applied in our own cohort of 818 patients or future larger cohort. Larger scale, multicenter studies will be necessary for more detailed evaluation. However, even a limited number of patients can yield meaningful results; we believe the results herein can be generally applied, at least for the Asia patients. In addition, the clarification of correlations between rupture, preclinical history, and laboratory data changes (white blood cell count, high sensitivity C-reactive protein, or erythrocyte sedimentation rate) should be a focus of future research.

The determination of treatment of post-RFA ruptured nodule depends on the nodule size and procedure time. Though some ruptured nodules can be managed conservatively, those with a maximum initial diameter of >4.5cm may necessitate more invasive management. Understanding these features with appropriate risk assessment will help physicians to accurately diagnose and to improve patient’s satisfaction.

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

The studies involving human participants were reviewed and approved by the institutional review boards of Kaohsiung Chang Gung Memorial Hospital and the institutional review boards of Chung Shan Medical University Hospital. Written informed consent for participation was not required for this study in accordance with the national legislation and the institutional requirements.

For the contribution of each author, Wen-CC and W-CL conceived and designed the analysis. Wen-CC and K-LC collected the data. Y-HC and C-KC contributed data or analysis tools. Wei-CC and S-DL performed the analysis. Wen-CC and K-LC wrote the paper. All authors contributed to the article and approved the submitted version.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.