Influence factors and nomogram for volume reduction rate in benign thyroid nodule after thermal ablation

Abstract Background Thermal ablation is a minimally invasive and safe treatment for benign thyroid nodules, and the volume reduction rate (VRR) of nodule is a primary clinical efficacy indicator. Purpose To screen factors influencing VRR in benign thyroid nodules after thermal ablation and establish a predictable nomogram. Materials and methods This retrospective study enrolled 238 patients with benign thyroid nodules who underwent thermal ablation between January 2016 and September 2021. Clinical information and imaging characteristics in routine ultrasound (US) and contrast-enhanced ultrasound (CEUS) were evaluated. Factors influencing the VRR ≥75% were screened using multivariate logistic regression, and a predictable nomogram was established. Results At the 12-month follow-up, the VRR of nodule was 77.0 ± 20.6% (18.4–100%). Seven factors influencing the VRR ≥75.0% were identified: echogenicity, component, calcification, enhancement degree, enhancement defect, ring enhancement, and energy of ablation. A nomogram was established based on the above factors, and the predictive ability of the model was confirmed by internal validation with 1000 bootstrap repetitions. The area under the receiver-operating characteristic curve (AUC) of the model was 0.926, and the calibration curve and decision curve analysis (DCA) revealed that this model demonstrated predictive ability. Conclusion Seven factors influencing VRR in benign thyroid nodules after thermal ablation were screened out in the present study and used to establish a nomogram to predict the probability of VRR ≥75% at the 12-month follow-up. It would be beneficial to make personalized medical decisions to trigger thermal ablation in patients with benign thyroid nodules.


Introduction
Thyroid nodule is a common disease with a high detection rate of approximately 50-60% in healthy populations with ultrasound (US) examination [1], and approximately 95% of thyroid nodule is benign [2], including adenoma, benign follicular nodules, and nodular goiter.Most benign thyroid nodules do not require treatment.However, some large benign thyroid nodules require treatment because of local compression symptoms, anxiety, or esthetic demands.
Surgical resection is the standard treatment for benign thyroid nodules [3].However, the disadvantages of surgery include scars, wide peripheral soft tissue injury, and hypothyroidism [4].Therefore, surgery for benign thyroid nodules could affect the patients' quality of life.
Thermal ablation is a minimally invasive and safe treatment for various solid tumors recommended for benign thyroid nodules by the 2020 European Thyroid Association Clinical Practice Guidelines [5].Several studies have revealed that thermal ablation for benign thyroid nodules does not influence thyroid function [6] or quality of life, and its safety can be compared with that of conventional thyroidectomy [6,7].
Compared with thyroidectomy, the greatest challenge in thermal ablation of benign thyroid nodules is that the residual ablation zone requires a long time for absorption to become smaller or disappear; otherwise, the residual ablation zone could still have compression of the surrounding thyroid tissue, inducing local compression symptoms.Therefore, the volume reduction rate (VRR) of nodule is the primary clinical efficacy indicator of thermal ablation.According to previous reports, the routine US characteristics of thyroid nodules, such as volume [8,9], macrocystic/solid component [9][10][11], and peripheral blood flow on color Doppler US [11] influence the VRR of thyroid nodules after thermal ablation.However, few studies have attempted to combine routine US and contrast-enhanced ultrasound (CEUS) characteristics to screen for more comprehensive factors influencing VRR.
In the present study, the VRR of benign thyroid nodules after thermal ablation was summarized, the factors influencing VRR were explored in routine US and CEUS characteristics, and a nomogram was established based on the relevant factors.

General information and patients enrollment
This retrospective study was approved by our hospital's ethics committee (No.S2019-283-02).The requirement for informed consent was waived due to the retrospective nature of the study, and the clinical and imaging data were allowed to be published anonymously.Written informed consent for operation was signed by each patient before the ablation procedure.The data of patients who underwent thermal ablation for benign thyroid nodules in our department between January 2016 and September 2021 were summarized.
The inclusion criteria were as follows: (a) benign thyroid nodules confirmed by fine needle aspiration (Bethesda II [12] A total of 238 patients were enrolled in the present study according to the inclusion and exclusion criteria (Figure 1).A total of 187 of 238 patients have been previously reported [13], which dealt with the complication of thermal ablation in thyroid nodules, whereas in the present study, we focused on the VRR.

Preoperative evaluation
Before ablation, routine US and CEUS (Perflubutane, Sonazoid, Daiichi-Sankyo Co. Ltd.Tokyo, Japan) examinations were performed on all enrolled patients using a LOGIQ E9 scanner (GE Healthcare).Laboratory tests, including routine blood, thyroid function, thyroid antibody, calcitonin, and coagulation function tests, were performed before ablation.
The parameters in routine US include the nodule boundary, morphology, echogenicity, component, nodule calcification, and maximum diameter.The nodule volumes were calculated based on the formula of the ellipsoid model volume: The different characteristics of CEUS, including the degree of enhancement, uniformity of enhancement, mode of enhancement, defect of enhancement, and surrounding ring enhancement of the nodules in the arterial phase were further evaluated (Table 1).(determined based on the experience and habits of doctors) was 30-40 W for MWA and 25-60 W for RFA.The hydrodissection technique was performed intraoperatively to protect vital structures such as the trachea, recurrent laryngeal nerve, and large cervical vessels surrounding the thyroid [14].

Technique of thermal ablation
When a hyperechoic zone completely covered the thyroid nodules, ablation was terminated, and CEUS examination was performed to evaluate the therapeutic effect to ensure complete ablation.Complete ablation was defined as a nonenhanced ablation zone completely covering the entire target nodule, or else further ablation was performed immediately.

Follow-up and outcome assessment
Generally, patients undergo follow-up every 3 months in the first year, every 6 months in the next 4 years, and annually thereafter.Routine US and thyroid function tests were performed at each follow-up examination.In addition, the volume of nodules was calculated 12 months after the ablation.The VRR of nodule (after thermal ablation) was calculated using the following equation:

Statistical analysis
Statistical analysis was performed using the R software (version 4.2.1;R Statistical Software, R Foundation for Statistical Computing, Vienna, Austria).Continuous variables are presented as means and standard deviations (mean ± SD), and discrete variables are presented as frequency counts and percentages.The chi-squared test was used to compare discrete variables, and the Wilson rank-sum test was used to compare continuous variables.Statistical significance was set at p < .05.
The VRR of nodule at 12 months after thermal ablation was converted to dichotomous variables (VRR !75.0%, VRR <75.0%) [9], and logistic regression was used to screen out the influencing factors of VRR !75.0%.There were 15 candidate variables in the logistic regression model, including sex, age, ablation technique, energy of ablation, six features of benign thyroid nodules in routine US examination, and five in CEUS examination.Categorical variables were converted to dummy variables.Table 1 presents the definitions of the variables in this study.
The results of logistic regression were expressed as odds ratios (OR), 95% confidence intervals (CI), and p values.A nomogram was established to visualize the logistic regression model based on these factors.The predictive ability of the model was evaluated by internal validation using the bootstrapping technique (1000 resamples).The area under the receiver-operating characteristic curve (AUC), calibration curve, and decision curve analysis (DCA) were used to demonstrate the accuracy of the model.

Demographic and clinical information
Demographic, clinical, and imaging information of the patients enrolled in this study are presented in Table 2.All the enrolled patients underwent complete ablation in one session.According to the statistical results, the VRR of nodule was 77.0 ± 20.6% (18.4-100%) in the 12 months after ablation.Furthermore, VRR !75% occurred in 145 (60.9%) cases

Establishment and evaluation of nomogram
Based on the results of multivariable logistic regression, seven variables in routine US and CEUS (echogenicity, calcification, nodule components, enhancement degree, enhancement defect, and ring enhancement), and energy of ablation were identified and used to establish the nomogram.The nomogram is shown in Figure 3.The influencing factors included in the nomogram had vertically corresponding points on a top-point scale.The sum of these points had a corresponding percentage at the last line of the nomogram, which provided a probability of VRR !75% for patients with benign thyroid nodules after thermal ablation.
Prediction model performance was evaluated using a bootstrap internal validation procedure with 1000 bootstrap repetitions.AUC, calibration curve, and DCA analyses demonstrated the predictive ability of the model.The AUC of the model was 0.926 (Figure 4(a)).The calibration curve of the model, which reveals promising predictive accuracy between the predicted and actual probabilities, is presented in Figure 4(b).DCA results, indicating that patients with benign thyroid nodules who underwent thermal ablation based on this model could lead to a higher net benefit, are presented in Figure 4(c).

Discussion
Thermal ablation has become an alternative or even primary treatment for benign thyroid nodules in clinical practice [1,[5][6][7][15][16][17].The main treatment indications for benign thyroid nodules are local compression symptoms and esthetic demands, which are generally associated with large nodules.The severity of these symptoms primarily depends on the nodule volume.A higher VRR corresponds to increased absorption and reduced volume of the residual ablation zone, resulting in milder compression symptoms and less impact on appearance.Therefore, VRR is a major concern for both doctors and patients during follow-up.A prediction of VRR before ablation will benefit evidence-based medical decisions, and the maximum diameter of the nodule that triggers thermal ablation can be speculated based on the predictive VRR.Some factors influencing VRR have been reported in previous studies, including the volume and component of nodules [18][19][20][21] and the energy of ablation [22,23].However, in the prognosis assessment of benign thyroid nodules after thermal ablation, few studies have combined various characteristics in routine US and CEUS to establish a nomogram to accurately predict VRR, which could affect the personalized decision regarding whether the patients should undergo thermal ablation.
In the present study, the VRR of nodule was 77.0 ± 20.6% (18.4-100%) at the 12-month follow-up, similar to that reported in previous studies [15,18].Seven influencing factors of VRR, including echogenicity, component, calcification, enhancement degree, enhancement defect, ring enhancement in routine and CEUS characteristics, and ablation energy were screened out.Subsequently, a nomogram was established based on the seven identified influencing factors, which first combined characteristics of both routine US and CEUS, which could more precisely predict the probability of VRR !75.0% at the 12-month follow-up.
Most factors influencing VRR screened in the present study were consistent with those of previous reports.However, cystic solid nodules were identified as a negative factor influencing VRR in the present model, which has been reported as a positive factor in previous studies [18][19][20][21].In our experience, if the liquid component inside the nodule was liable to draw out during operation, the volume of the nodule would be significantly reduced immediately after thermal ablation, and the residual ablation zone would be completely absorbed during follow-up with a higher VRR.However, if the liquid component inside the nodule was jelly-like and too sticky to be drawn out, the volume of the nodule would not be reduced significantly during ablation and would not be easily absorbed during follow-up.Therefore, a high VRR is not always achieved for benign thyroid nodules with a large proportion of cystic components, and it primarily depends on the properties of the internal liquid component.
Hyperechoic thyroid nodules were negatively correlated with VRR in the present study, which has not been reported previously.In some studies, thyroid nodules with hyperechoic nodules on routine US contained fewer cells and fibrosis, and more follicles than hypoechoic nodules [24,25].Dissolution and absorption of the ablation zone primarily depend on the immunoreaction initiated by the necrosis of cells after thermal ablation; fewer cells in hyperechoic thyroid nodules might induce fewer immunoreactions.Therefore, hyperechoic nodules generally result in a lower VRR after thermal ablation than hypoechoic nodules, which generally contain abundant cells.
In the present study, the baseline volume of nodules has no correlation with VRR.Some previous studies found that smaller nodules had higher VRR compared with larger nodules [26][27][28].However, the results in the above studies were based on single-factor regression, which neglected the possible influence of other characteristics, such as components, calcification of nodules, and techniques of ablation.Bernardi believed that larger baseline volume was significantly associated with the likelihood of being retreated on logistic regression model analyses in the RFA and laser ablation of benign thyroid nodules [29], which was consistent with our clinical practice.However, the patients who underwent incomplete ablation or secondary ablation were excluded from the present study.
RFA and microwaves are the main techniques for thermal ablation in benign thyroid nodules, however, there is still a  controversy of the difference in clinical efficacy between them.Some previous studies reported that RFA had higher VRR in benign thyroid nodules compared with microwaves, especially in the 12 months follow-up [30,31].But other studies found that there was no difference between these two techniques in efficacy and safety for benign thyroid nodules [22,32].In the present study, there was a significant difference between groups of VRR !75.0% and <75.0% in ablation techniques (p ¼ .011);however, this difference was not found in the further multiple logistics regression (p ¼ .123),and there was a significant negative correlation between energy of ablation with VRR (OR: 0.43, CI: 0.24-0.74,p ¼ .003).
Few studies have examined the correlation between CEUS and VRR in benign thyroid nodules, except for a retrospective study by Fu et al. who believed that nodules with hyper-enhancement on CEUS had a better VRR because hyper-enhancement indicated abundant blood feeding of nodules, which could be beneficial to the absorption of necrotic nodules.Notably, hyper-enhancement was negatively correlated with the VRR in the present study.We considered that hyper-enhancement of thyroid nodules on CEUS accounted for the abundant preoperative blood feeding, which was completely destroyed during thermal ablation and should not be beneficial for the absorption of the ablation zone.However, ring hyper-enhancement surrounding thyroid nodules was positively correlated with VRR in the present study, which indicated a relatively rich vascular network surrounding nodules that may help remove substances derived from the necrotic cells and immune response in the decomposition and absorption of the ablation zone.
In the present study, a nomogram was established to predict the VRR of benign thyroid nodules that underwent thermal ablation, which included routine US and CEUS characteristics as positive or negative influencing factors.This nomogram would help make personalized decisions regarding the maximum diameter that triggers thermal ablation of benign thyroid nodules.For example, if an ideal maximum diameter of the residual ablation zone was set, a suitable maximum diameter triggering thermal ablation would be confirmed according to the nomogram.
This study has a few limitations.First, this was a retrospective study with possible inherent biases, and the sample size was relatively small.Second, the mechanism of the factors influencing VRR has not been verified by pathology or experimental research.Third, owing to the limited number of cases, this model was validated by internal 1000 bootstrap repetitions, and further external validation was required.
In conclusion, seven factors influencing VRR in benign thyroid nodules after thermal ablation were identified through a comprehensive analysis of characteristics in clinic, routine US, and CEUS.A nomogram was further established to predict the probability of a VRR !75% at the 12-month follow-up.To the best of our knowledge, this is the first study to establish a nomogram to make personalized decisions on triggering thermal ablation of benign thyroid nodules based on the relatively accurate prediction of VRR.
); (b) maximum diameter of benign thyroid nodules larger than 2 cm with compression symptoms; (c) exclusively treated by thermal ablation; and (d) follow-up !12 months.The exclusion criteria were as follows: (a) insufficient clinical imaging data before ablation or during follow-up, (b) pure cystic thyroid nodules, (c) history of thyroid surgery, (d) incomplete ablation or unplanned reoperation, (e) multiple nodule ablation, and (f) scheduled staging ablation.

Figure 1 .
Figure 1.Flow chart of patients' inclusion and exclusion criteria.

Figure 2 .
Figure 2. Forest plot of multivariate logistic regression for volume reduction rate !75.0% in benign thyroid nodules after thermal ablation.

Figure 3 .
Figure 3.A nomogram to predict the probability of volume reduction rate !75.0% in benign thyroid nodules after thermal ablation at 12-month follow-up.

Figure 4 .
Figure 4. (a) The receiver-operating characteristics curve of the nomogram, the AUC was 0.926.(b) The calibration curve of the nomogram.(c) The decision curve analysis (DCA) of the nomogram.

Table 1 .
The variables included in logistic regression model and definition.solid component of nodules has lower echogenicity compared with the surrounding thyroid gland.Isoechoic: the solid component of nodules has similar echogenicity compared with surrounding thyroid gland.Hyperechoic: the solid component of nodules has higher echogenicity compared with surrounding thyroid gland.Component of nodules Solid nodules: the nodule is entirely composed of solid tissue.Solid dominant nodules: the solid component account for more than 80% of nodules.Cystic-solid nodules: the solid component account for more than 20% but less than 80% of nodules.Cystic dominant nodules: the nodules have solid component but the solid component account for less than 20% of nodules.Calcification Yes: the nodules with any kind of calcification.No Enhancement degree Hyper-enhancement: the peak enhancement degree of nodules is higher than surrounding thyroid gland.Nohyper-enhancement: the peak enhancement degree of nodules is lower than or similar to surrounding thyroid gland.Enhancement uniformity Homogeneous/heterogeneous Enhancement mode Overall: the contrast agent fills the nodules as a whole.Centripetal: the contrast agent fills the nodules from periphery to central area.Eccentric: the contrast agent fill the nodules from periphery to central area.Enhancement defect Yes: there is some area without contrast agent perfusion in the solid thyroid nodules in contrast-enhanced ultrasound.Watt (W) Â duration of ablation(s)/1000 Â volume of nodules (cm 3 ) and VRR <75% occurred in 93 (39.1%) cases.The nodule completely disappeared in six (2.5%) cases (VRR ¼ 100%).Complications were encountered in nine (3.8%) cases in the present study.Major complications occurred in three cases (1.3%), including hoarseness with choke in two cases, and choke in one case.Minor complications occurred in six cases (2.5%), including perithyroid hematoma in three cases, fever and toothache in one case, xerostomia in one case, and transient hypertension and dizziness in one case.All patients recovered spontaneously.

Table 2 .
Demographic and clinical information of patients.