Introduction

Papillary thyroid cancer (PTC) is the most common subtype of thyroid cancer, representing approximately 80% of all thyroid cancers [1]. Although the prognosis of patients with PTC is generally good [1], up to 13% of localized PTCs completely excised during initial surgery were found to recur over 40 postoperative years [2]. The standard treatment for recurrent tumors consists of repeat operation, followed by treatment with radioactive iodine and thyroid hormone. However, repeated neck operations are usually challenging due to the distortions of normal tissue planes and fibrosis caused by scar tissue formation in the surgical bed, thus increasing the risk of complications [3]. Additionally, small recurrent lesions may be difficult to detect without ultrasound (US) examination [3].

Nonsurgical localized therapies, including thermal ablation [4,5,6,7,8] and ethanol ablation (EA) [9,10,11], may be an effective alternative to repeat operation in patients with recurrent PTC. Recent guidelines recommend these localized treatments for patients with a single or a few metastases or with metastases at high risk of local complications [12, 13]. A meta-analysis of the efficacy and safety of radiofrequency ablation (RFA) for treating locally recurrent thyroid cancer shows that these treatments have therapeutic success rates (volume reduction rate [VRR] > 50%) of 89.5–100% [14], with 68.8% of these lesions completely disappearing after RFA. However, this meta-analysis was limited by the relatively short-term follow-up periods of the included studies. To our knowledge, no study to date has investigated the longer-term efficacy of RFA in patients with locally recurrent PTC. These findings suggested the need to evaluate the longer-term efficacy of RFA to verify the completeness of treatment. Previously, our institution reported the outcome of RFA for locally recurrent PTC after a mean follow-up period of 26.4 months [15]. The purpose of this study was to evaluate the longer-term efficacy of US-guided RFA for treating locally recurrent PTC.

Materials and methods

The retrospective study protocol was approved by our Institutional Review Board for human investigation, and written informed consent was obtained from all patients prior to US-guided fine-needle aspiration biopsy (FNAB) and RFA.

This data were retrospectively collected from the cohort of our previous short-term follow-up study (mean, 26.4 months) for treating recurrent PTC in the neck using RFA between September 2008 and April 2012 [15]. Previous study included 61 recurrent PTCs of 39 patients. This retrospective study finally included 29 patients who underwent RFA for 46 recurrent PTCs and were followed up for at least 5 years. Two of 10 patients who lost to follow-up died from lung cancer and pneumonia, respectively, and another eight patients were followed up at another hospital nearby their home. All patients undergone total thyroidectomy and central compartment node dissection, and 20 had also undergone lateral neck dissection. All patients received postoperative radioiodine therapy and supplemental levothyroxine thyrotropin-suppressing therapy. All RFA procedures were performed according to the request of physician for three reasons: (1) radioiodine refractory PTC (n = 17), (2) history of repeat operation more than two times (n = 11), and (3) high risk for general anesthesia (n = 1).

All patients were evaluated by US examination, US-guided FNAB of the recurrent tumors, contrast-enhanced neck CT, and laboratory tests before RFA. The presence of a recurrent tumor was confirmed by cytological smears and/or by a finding of increased thyroglobulin (Tg) levels in FNAB samples. Tumor diameters prior to RFA were measured by US examination, and the volume (V) of each tumor was calculated as V = πabc / 6, where a is the largest diameter and b and c are the two perpendicular diameters [16].

US-guided RFA was performed by one of two radiologists with 11 and 16 years of experience, respectively. We used a radiofrequency generator (Cool-Tip RF system, Radionics; SSP-2000, Taewoong Medical; VIVA RF system, STARmed) and an 18- or 19-gauge internally cooled electrode with 0.5-, 0.7-, and 1-cm active tips (Cool-Tip, Radionics; Well-point RF electrode, STARmed; VIVA, STARmed), depending on the size of the targeted tumor. Moving-shot and hydrodissection techniques were used, and the surrounding normal tissue was also ablated to prevent marginal recurrence [15, 17]. Ablation was started with 10 W of power in a 0.5-cm active tip, 15 W in a 0.7-cm active tip, and 30 W in a 1-cm active tip. If a transient hyperechoic zone did not form at the electrode tip within 5–10 s, the radiofrequency power was increased in 5- to 10-W increments up to 50 W. The ablation was terminated when all portions of the treated tumor had changed to transient hyperechoic zones.

The patients were followed up for at least 5 years, by US and serum Tg level at 1, 3, 6, and 12 months and every 6–12 months thereafter. US examinations were performed by the two radiologists who had performed RFA. The percentage reduction in volume was calculated as VRR = ([initial volume − final volume] × 100) / initial volume [18].

All patients underwent contrast-enhanced computed tomography (CT) scans after ablation. Additional RFA was performed if the 1-month follow-up US showed the presence of power Doppler signals, suggesting that a remnant of the treated nodule was viable, or if immediate follow-up CT detected an enhanced portion of the treated tumor.

All data were analyzed using SPSS 23 for Microsoft Windows version 23.0 statistical software (SPSS). The Wilcoxon signed-rank test was used to compare the largest tumor diameter, tumor volume, and serum Tg levels before RFA and at final follow-up. Differences were considered statistically significant when the p value was < 0.05.

Results

Table 1 shows the baseline characteristics of the 29 patients, which included 12 men and 17 women, mean age 51.8 years (range, 21–84 years). Of the 46 recurrent tumors, 23 (50%) were occurred in the thyroid operation bed or neck level 6 and 23 (50%) were occurred in the lateral neck lymph node. The mean number of treatment sessions for each tumor was 1.1 ± 0.4 (range, 1–3), and the mean follow-up duration after RFA was 80 ± 17.3 months (range, 60–114 months).

Table 1 Baseline patient characteristics
Full size table

Tumor volume, largest tumor diameter, and serum Tg concentrations observed at each follow-up evaluation are shown in Figs. 1 and 2. The volume of the recurrent tumors decreased significantly, from 0.25 ± 0.42 mL before ablation to 0.01 ± 0.08 mL (p < 0.001) at final evaluation, with a mean tumor VRR of 99.5% ± 2.9% (range, 81%–100%). The mean serum Tg level also decreased significantly, from 2.55 ± 4.7 ng/dL before ablation to 0.75 ± 1.83 ng/dL (p < 0.001) at final follow-up, and was below the level of detection (0.08 ng/dL) in 15 (51.7%) of the 29 patients at the last follow-up.

Fig. 1
figure 1

Changes in metastatic tumor volume and the largest diameter before RFA and at each follow-up

Full size image
Fig. 2
figure 2

Changes of serum thyroglobulin level before RFA and at each follow-up

Full size image

The 35 tumors that had completely disappeared in our previous study [15] did not recur after longer-term follow-up. Previous study reported 11 incompletely disappeared lesions on US images in short-term follow-up [15]. We summarized the longer-term results of these 11 lesions in Table 2. Of the 11 such lesions described in our previous study, 10 were followed up for at least 5 years. In addition, one lesion, which had been missed previously, was added (no. 11). Of these 11 recurrent tumors, seven completely disappeared after RFA (Fig. 3); however, four were still visible on US and/or CT. Overall, of the 46 tumors, 42 (91.3%) completely disappeared. Of the remaining four tumors, one presented as a macrocalcification on US examination and CT, with no change in size for 110 months and absence of malignant cells on FNAB; one gradually increased in size on follow-up CT despite three sessions of RFA and was confirmed as an anaplastic carcinoma after incisional biopsy; one was located too deep in the paratracheal area for complete RFA and was resected 4 months after RFA; and one showed a residual enhanced lesion adherent to trachea on follow-up CT. Of the 29 patients, 8 had new locoregional recurrent lesions, two had distant metastatic lesions, and the other 19 showed no evidence of diseases during follow-up. None of these patients experienced any delayed complications related to RFA during a mean follow-up period of 80 months.

Table 2 Long-term follow-up results of 11 patients with residual lesions on our short-term follow-up data
Full size table
Fig. 3
figure 3

A 48-year-old female with recurrent papillary thyroid cancer at left level 6 after thyroidectomy and RI therapy. Transverse ultrasound image shows a 1.8-cm-sized hypoechoic mass at left level 6 (arrows) (a). An internally cooled electrode with a 1-cm-sized active tip is inserted into the recurrent tumor (b). After completion of RFA, the ablated zone is larger than the initial tumor size (c). 18 months after RFA, the transverse and longitudinal US scans show the treated tumor is decreased in size (d, e). 84 months after RFA, the treated tumor is not observed on US (f). RI radioactive iodine, RFA radiofrequency ablation, US ultrasound

Full size image

Discussion

This study shows the longer-term effectiveness of US-guided RFA for controlling locally recurrent PTC. The efficacy of RFA was indicated by the 99.5% VRR, the complete disappearance of 91.3% of the treated tumors, and the significantly decreased serum Tg concentrations. Moreover, RFA shows no delayed complications during a mean follow-up period of 80 months. These results suggest that RFA is an acceptable nonsurgical treatment option for local control of recurrent PTCs.

Previous studies report the short-term effectiveness of RFA in the treatment of recurrent thyroid cancer [4,5,6,7,8, 15, 19]. To our knowledge, however, our study is the first to present longer-term (> 5 years) follow-up data. Previous shorter-term RFA studies reported acceptable local tumor control effect with a mean tumor VRR of 50.9–96.4% and the complete disappearance of 25–94% of treated tumors [5,6,7,8,9, 15]. Our longer-term study revealed similar efficacy and safety. RFA induces coagulation necrosis, hyaline sclerosis, and scarring within the tumor [20]. This process is slow, with a long period of time required for the tumor to decrease in size. Of the 29 patients in this study, four had residual lesions after RFA. Residual tumors are induced by incomplete treatment because of anaplastic transformation or deep location of tumor. In addition, mean serum Tg level decreased abruptly at 1 month after RFA, and the level remained low until the last follow-up. Recent meta-analysis and current RFA guideline revealed that the serum Tg decreased after RFA [14, 18]. The meta-analysis revealed that the pooled proportion of reduction in serum level of Tg after RFA was 71.6% ([CI 63.5–79.7%]; I2 = 0.0%) [14]. Furthermore, previous studies revealed various early complications of RFA [15] for recurrent cancers. In our study population, the early complication rate was reported in three patients. All three complications were voice change which had recovered in all patients within 2 months after RFA. Four patient complained self-limiting local pain during ablation. During longer-term follow-up period, no patient experienced delayed complications. A recent meta-analysis compared the efficacy of EA and RFA for recurrent thyroid cancers [14]. That study reported that RFA was more effective, with a higher tumor VRR ≥ 50%, a higher rate of complete lesion disappearance after treatment, and a lower rate of local tumor recurrence. Moreover, RFA required fewer treatment sessions than EA. However, the complication rate was lower for EA than for RFA. These differences may be due to differences in treatment areas, with RFA treating a larger area, including the tumor and surrounding normal soft tissue [14, 15]. When we compared our results with the long-term effects of EA for recurrent PTC [21], we found that the rate of tumor size increase (17.1% vs. 2.2%) and the mean number of ablations (2.7 vs. 1.1) were higher after EA than after RFA, suggesting that RFA is more effective than EA for local control of recurrent PTC.

Repeat operation is undoubtedly a definitive curative treatment for recurrent tumors. Current guidelines recommend evaluation of lesions > 0.8 cm because of the indolent nature of thyroid cancer, the technical difficulties of repeat operation, and the risks of general anesthesia [10, 12, 22, 23]. Although repeat operation may be delayed [5, 24], untreated tumors may suddenly transformed to aggressive subtype [25]. By contrast, RFA can be performed in outpatient clinics without general anesthesia and can be applied relatively easily, repeatedly, and safely. The reported rates of permanent recurrent laryngeal nerve paralysis and hypocalcemia after repeat surgery range from 0 to 12% and from 0 to 7%, respectively [8, 26, 27]. By comparison, the incidence of voice change after RFA for recurrent thyroid cancers was reported to be 7.95%, but most of these were transient, and there have been no reports of hypoparathyroidism [28]. The lower incidence of complications with RFA might be associated with the ability of high-resolution US imaging to target the tumor exactly, knowledge of US-based nerve anatomy [29], thyroid-dedicated small active tips (0.38 and 0.5 cm) [21], and the use of hydrodissection techniques [21, 30].

Concerns associated with RFA include recurrent tumors that may not be found by imaging modalities and are only detected in pathologic specimens after surgical dissection; however, the clinical significance of these undetected tumors remains unknown [8]. In our study, eight patients had new locoregional recurrent lesions and two had distant metastatic lesions during follow-up among 29 patients. Furthermore, less than 50% of patients showed decreased but still elevated serum Tg level after treatment. Thus, RFA is effective to the local control of ablated metastatic lymph node, but has a limited role in controlling the disease and its recurrences at other regions.

This study had several limitations. First, it is a retrospective design and included only patients followed for > 5 years, which may have resulted in a selection bias. However, because this study included 11 of the 12 patients with residual lesions in our previous study [15], the efficacy of treatment was likely not overestimated due to selection bias. Second, this study was based on data from a single, specialist center; different results may occur in studies of patients treated at community hospitals with less expertise.

In conclusion, this study demonstrated the longer-term efficacy of RFA for locally recurrent PTC, with a mean tumor VRR of 99.5% and the complete disappearance of 91.3% of treated tumors. These results suggest that RFA seems to be an effective minimally invasive therapy for the treatment of locally recurrent PTC even in the longer-term period.