EA uses ethanol of high purity (95–99%). Ethanol has several hydrogen bonds in a single molecule and dissolves quickly when it comes into contact with water. Its mechanism of action is a combination of the following two effects: 1) coagulative necrosis of tissue due to cell dehydration and protein denaturation due to diffusion, and 2) ischemic necrosis due to small blood vessel thrombus formation and inhibition of enzymatic activity in surrounding tissues (1985), which may result in fibrosis and atrophy of the nodule. The main mechanism is thought to be coagulative necrosis for thyroid cysts, and ischemic necrosis for solid thyroid nodules (86).

It is important to characterize the thyroid nodule and evaluate the surrounding critical anatomical structures through a preprocedural US examination. The size, proportion of solid component, vascularity, and internal contents of the cystic component should be carefully evaluated on US. Three orthogonal diameters (the largest and two diameters perpendicular to it) should be measured by US and nodule volume could be calculated using the equation V = πabc / 6 where V is volume, a is the maximum diameter, and b and c are the other two perpendicular diameters. A laboratory test usually includes a thyroid function test (measurement of thyrotropin [TSH], triiodothyronine, and free thyroxine) (8788), while a blood coagulation battery (bleeding time, prothrombin time, and activated partial thromboplastin time) may be considered based on the patient's medical history or hemorrhagic risk. Nodule-related symptom score on a 10-cm visual analogue scale (0–10) and cosmetic score (1, no palpable mass; 2, a palpable mass but no cosmetic problems; 3, cosmetic problems on swallowing only; 4, readily detected cosmetic problems) should be measured by patients and physician, respectively, as baseline data. An explanation of the procedure and informed consent must be provided before the procedure. Informed consent should include the possibility of adverse events (pain or voice changes are symptoms of nerve injury caused by leakage).

Procedures are generally performed on an outpatient basis. A 16–20 G needle, 10–25 mL syringe, and 95–99% ethanol are necessary for the procedure. There are some differences depending on the operator, but the general method is as follows.

The patient is placed in a supine position on the bed with mild neck extension. After skin sterilization, local anesthetic (1–2% lidocaine) is usually injected at the skin puncture site. Depending on the viscosity of the contents in the cyst, a single 16–20 G needle is inserted into the center of the cystic area under US guidance. A trans-isthmic approach route is recommended to prevent a change in the position of the needle tip and ethanol leakage when the patient swallows or talks (Fig. 1). As much internal fluid is aspirated as possible and residual debris or colloids are removed by saline irrigation. High viscosity fluid of colloid cysts may be difficult to aspirate. Several studies on the evacuation of viscous cystic thyroid nodules have used normal saline irrigation to decrease viscosity, repeated ethanol injection into the nodules with an interval period, a large-bore needle, or even a large pigtail catheter connected to a suction pump (8990919293). Then, the operator slowly injects an appropriate amount of 95–99% ethanol into the cystic space. The amount of ethanol injected depends on the amount of aspirated material in the nodule; approximately 50% of the aspirate volume is usually injected. Total retention amount of ethanol is empirically recommended not to exceed 10 mL, even for large-sized nodules (1928515694). However, if injected ethanol is to be removed after the procedure, a larger amount of ethanol can be used. A sufficient consensus has not yet been established regarding the total amount of ethanol to use for injection. The exact ethanol amount should be tailored according to the size and internal content of the lesion, the opinion of the operator during the procedure, and patient compliance. An ethanol retention time of 2 minutes was found to be sufficient by Kim et al. (95), because ethanol reacts rapidly with cells. However, there is no consensus on whether ethanol should be re-aspirated or not after the procedure. Several studies have suggested that a short retention (at least 2 minutes) and complete removal of ethanol is preferable because longer ethanol retention times may increase patient inconvenience and complications due to possible leakage of ethanol (9596). This also has the advantage that the patient's satisfaction is high, because the size of the nodule is reduced immediately after the procedure. Kim et al. (53) reported that there was no difference in complication or success rated between groups in which ethanol was not aspirated or aspirated after injection. Therefore, ethanol injection with or without aspiration may be used depending on the operator's preference.

If the patient complains of pain during injection, the operator should stop the injection and check for perithyroidal leakage. Perithyroidal leakage is less likely if the needle tip is well located at the center of the lesion and excessive ethanol is not injected. If excessive amounts of ethanol are used, it may be helpful to re-aspirate the injected ethanol for pain relief. When the procedure is finished, quickly pull out the needle and lightly compress the puncture site for 5–10 minutes.

A 1 mL syringe with a 22–25 G needle is usually used for EA of recurrent thyroid cancer (3031323367). Under US guidance, after local anesthesia, 0.1–1 mL ethanol (per session) is injected beginning at the deepest and most peripheral portion of the recurrent cancer and moving to the superficial area. Immediately after the injection of ethanol, the treated area becomes echogenic (Fig. 2). The position of the needle is important to safely and effectively inject ethanol into the entire recurrent thyroid cancer lesion. Doppler US is useful for identifying untreated vascular areas and subsequent repositioning of the needle. Ethanol injections are continued until the entire recurrent cancer turned out to be transient echogenic area (32). If leakage to the outside of the lesion is detected on US or if severe pain occurs, ethanol injection must be terminated immediately.

Current evidence supports the use of EA as a first-line treatment for pure cysts and predominantly cystic thyroid nodules (24285051525354). Clinical outcomes of EA for benign nodules are evaluated by volume reduction and improvement in clinical problems (i.e., improvement in cosmetic problems or pressure symptoms) (505690939497). During the follow-up period, a treated nodule volume of less than 1 mL should be considered treatment success (959899). Previous studies have reported a 38–98.5% volume reduction after EA of benign thyroid nodules (2452545692939499100101102), 85–98.5% volume reduction after EA of cystic nodules (5456939499), 64–73.2% after EA of predominantly cystic nodules (5292100), and 38–47% after EA of solid nodules (2452101102). EA is very effective for cystic thyroid nodules, and has a low recurrence rate of less than 20% (28). In predominantly cystic nodules, recurrence has been reported in large or vascular predominantly cystic nodules (95). Despite the effectiveness of EA in treating pure cysts and predominantly cystic thyroid nodules, about 5–25% of patients have nodules refractory to EA (28535665). A recent study reported that EA was less effective in nodules with more than 20% solid component (55). In these recurrent nodules, RFA has been proposed as the next step treatment (55103) because RFA achieves 91–92% volume reduction and improvement in clinical problems in patients with incompletely resolved clinical problems after EA. Although EA has been used in solid thyroid nodules, the efficacy of EA in these nodules is limited because of the limited volume reduction and requirement for several treatment sessions (525596).

In a 10-year follow-up study of patients with cystic thyroid nodules (97), EA resulted in a 50% volume reduction in 94% (92 of 98) of patients. Moreover, all patients showed a significant improvement in cyst-related symptoms due to the reduction in cyst volume. Only 6.5% (6 of 92) of patients had a recurrent nodule at the 9-year follow-up. In that study, the mean index cyst volume was 35.2 mL (median, 30.5 mL) and the mean number of EA sessions per patient was 1.8.

In another study from Korea (49), EA was used to treat 107 predominantly cystic thyroid nodules and patients were followed-up for 55 months. The recurrence rate was 18.7% (20 of 107) at the 1-month follow-up. Among 87 patients with non-recurrence, 24.1% (21 of 87) showed delayed recurrence. Therefore, recurrence is also far more common for predominantly cystic nodules treated with EA than pure cysts in long term follow up (97).

Although EA is effective and safe for both pure cysts and predominantly cystic nodules, it is more effective for pure cysts than predominantly cystic nodules in terms of volume reduction, therapeutic success, and requirement for additional treatment (525795). In the case of cystic nodules, the volume of ethanol injected is related to the amount of volume reduction; generally 50% of the aspirated volume is recommended (47529095). Vascularity in a solid component is one of the major factors that affects volume reduction of predominantly cystic nodules after EA. According to previous studies (5255576295103), solid components are more resistant to ethanol than cystic components, and the vascularity of solid components may cause the drainage of ethanol, thus limiting the success of EA. In contrast, intranodular echo-staining can be observed during EA when the nodule has poor venous washout of the injected ethanol, and this is considered a sign predictive of the success of EA (48). The vascular solid component can also cause internal bleeding in the nodules during the follow-up period after EA (9596). For both pure cysts and predominantly cystic nodules, the effectiveness of EA can be reduced by a large initial volume (> 10 mL) or insufficient retention time of ethanol (< 2 minutes) (182425284950609295). To successfully treat viscous cystic nodules with EA, vigorous aspiration using a large-bore needle or normal saline irrigation is required as a pre-step for evacuation (4752909295). The success-related factors of EA include the nature of the nodule itself (proportion of solid portion, vascularity, and initial volume), intranodular echo-staining during the procedure, retention time of ethanol and the sufficient evacuation of internal content before injection.

Since Lewis et al. (30) in 2002, a few studies have reported favorable outcomes after EA treatment of recurrent thyroid cancers (293132333435).

In four recent studies with a mean follow-up period ranging from 28.2 months to 5.4 years, three studies reported size reduction in 92.7–100% of cases (293334). In contrast, the study by Guenette et al. (35) reported local progression in 23.8% of cases. Hay et al. (34) and Heilo et al. (33) reported 45.9% and 66.1% complete disappearance rates, respectively. In contrast, Hay et al. (34) and Kim et al. (29) reported recurrence at sites other than the ethanol ablation sites. There were local recurrences at ablation sites in 5.5% and 23.8% of cases enrolled by Heilo et al. (33) and Guenette et al. (35), respectively (2934). In these four studies, the number of treatment sessions varied from two to six. Serum thyroglobulin level fell in 19 of 22 (86%) cases without thyroglobulin autoantibodies in the study by Hay et al. (34).

Concordantly, Kim et al. (77) reported an increase in tumor size in 17.1% of cases while there was no change in size in 24.4% of cases after EA for locally recurrent papillary thyroid carcinoma based on a minimum 60 months of follow-up.

According to the comparison meta-analysis of RFA and EA performed by Suh et al. (76), RFA tended to result in greater volume reduction and a greater complete disappearance rate than EA, while EA tended to show a higher recurrence rate and a higher reduction rate of serum thyroglobulin level. Fontenot et al. (80) compared surgery and EA and reported that the success rate of surgery was higher but local recurrence at the treated site was similar for the two methods.

Studies reporting the use of EA to treat recurrent thyroid cancers are still limited in number making it difficult to generalize outcomes. Further systematic studies with greater numbers of cases are warranted.

EA is a safe treatment option for cystic thyroid nodules and complications are mostly mild and transient, if performed by a person familiar with US-guided procedures (8788). Localized pain in the puncture site is most common, and hematoma, facial flushing, drunken sense, hoarseness, dyspnea, and temporary hyperthyroidism may occur (5795). Most patients complain of mild pain lasting from several minutes to several hours immediately after the procedure. Patients can recover from this pain spontaneously or the pain can be controlled by analgesics. Hematoma is a common complication associated with pain (52). It occurs when vessels in the thyroid capsule are damaged by the needle at the time of puncture (57). However, if the puncture site is tightly compressed for 10 minutes after the procedure, hematoma can be prevented.

Based on a systemic review of the use of EA or RFA to treat recurrent thyroid cancer, the pooled proportion of the complication rate of EA was lower than that of RFA (0.8% vs. 5.8%), although this difference was not significant (p = 0.8479) (76). Several previous investigations have reported the complications of EA of recurrent thyroid cancer (2930313234357578104) (Table 1). Potential complications of EA include local pain or discomfort, hoarseness, radiating pain to the teeth, jaw, head and chest, and tumor implantation through the needle track (78). Delayed complications have not been reported in studies that had performed long-term follow-up (mean, 74 months) (3477).

Local pain or discomfort occurs during injection of ethanol in most cases. Most patients experience mild pain or discomfort at the puncture site, which lasts from several minutes to several hours during or/and immediately after the procedure. This pain can resolve spontaneously or be controlled by analgesics. It can be prevented by sufficient injection of lidocaine to perilesional soft tissue and slow injection of ethanol within the target lesion only.

Hoarseness is an important complication that results from damage to the recurrent laryngeal nerve or vagus nerve. Hoarseness occurs frequently when the target lesion is located in the thyroid bed or around the carotid sheath. Most cases with hoarseness are transient and last from several days to weeks, and recover spontaneously within 3–6 months. Permanent hoarseness has not been reported in cases with recurrent thyroid cancer treated by EA (81). The cause of nerve injury is local spillage of ethanol into the surrounding soft tissue. To prevent this complication, the needle tip should be accurately positioned within the lesion and an extremely small amount of ethanol should initially be injected (32). The injection of ethanol should be discontinued if ethanol spillage is detected on real time US or if the patient complains of considerable pain during ethanol injection. The vagus nerve is also a critical structure related to hoarseness (81). It can be easily identified on US and is located in the carotid space, especially between the carotid artery and internal jugular vein posteriorly. The operator should identify its location and variations during US examination to treat recurrent thyroid cancer located near the carotid space.

Tumor implantation through the needle track is a very rare complication of EA in cases with recurrent thyroid cancer. A case of cutaneous implantation of papillary thyroid cancer originating from a metastatic lymph node in the left paratracheal area that occurred a few months after EA was reported (104). Because this rare complication is clinically relevant, physicians should take into account that this can occur months to years after EA.

Various complications of US-guided EA related to treatment for recurrent thyroid cancers can develop. The operator should have knowledge of these complications and technical tips to avoid procedural-related complications.