Elastography-Based Evaluation of Treatment Response Following Radiofrequency Ablation of Thyroid Nodules

Summary

A standardized shear-wave and strain elastography follow-up protocol was applied before and after thyroid nodule radiofrequency ablation to quantify stiffness changes longitudinally. Using mixed-effects modeling and multivariable regression, early post-ablation elastography changes were independently associated with 12-month volume reduction response.

Abstract

Radiofrequency ablation (RFA) is an established minimally invasive treatment for benign thyroid nodules, yet response assessment relies largely on delayed volume reduction, limiting early post-procedural decision-making. In this retrospective cohort, participants with a single index thyroid nodule underwent standardized B-mode ultrasound and elastography at baseline, immediately post-ablation, and at 1, 3, 6, and 12 months. Shear-wave elastography (SWE) stiffness and strain ratio were acquired using prespecified regions of interest and quality-control criteria. Twelve-month response was defined by a volume reduction ratio (VRR) ≥ 50%. Longitudinal changes were evaluated using linear mixed-effects models, and immediate post-RFA elastography changes (immediate minus baseline) were related to 12-month VRR using correlation and multivariable regression adjusted for baseline nodule volume, nodule composition, delivered energy, and age. Among the 187 included participants, 146 were responders, and 41 were non-responders. SWE increased immediately after RFA in responders (31.5 ± 5.1 kPa to 41.7 ± 5.6 kPa) and non-responders (35.1 ± 5.3 kPa to 43.4 ± 5.8 kPa), followed by progressive decline with lower 12-month SWE in responders than non-responders (21.6 ± 4.7 vs 31.2 ± 5.2 kPa). VRR separated early, with 12-month VRR of 73.6 ± 6.9% in responders versus 38.5 ± 8.7% in non-responders. In mixed-effects models, the immediate SWE increase was significant (estimate +10.22 kPa, P < 0.001), and late group divergence was evident (non-responder × 12 months estimate +9.41 kPa, P < 0.001). Early elastography change correlated with 12-month VRR (Δstrain ratio: r = 0.445, P < 0.001; ΔSWE: r = 0.302, P < 0.001) and remained independently associated with VRR in adjusted models (Δstrain ratio β = 14.80% per +1.0, P < 0.001; ΔSWE β = 1.80% per +5 kPa, P = 0.011). Immediate post-RFA elastography changes capture early tissue effects and provide independent prognostic information for 12-month volumetric response.

Introduction

Thyroid nodules have become a defining “disease of detection” in modern endocrine practice, not because their biology has abruptly changed, but because imaging has made the thyroid exquisitely visible^1,2,3. The measured prevalence depends strongly on how nodules are sought: classic epidemiologic syntheses estimate 2%–6% by palpation, approximately 19%–35% by ultrasound, and up to about 8%–65% in autopsy series, illustrating how method sensitivity reshapes the apparent epidemiology¹. In iodine-deficient settings and with high-frequency transducers, ultrasound surveys can report even higher detection rates; for example, a population-based investigation using 13-MHz technology identified nodules in 68% of adults, emphasizing both the influence of equipment and the underlying susceptibility of the thyroid to nodular change across the lifespan². China provides a particularly vivid view of scale because ultrasound is frequently incorporated into health examinations. In a national health examination cohort including 6,985,956 participants, the overall prevalence of thyroid nodules was 36.9%, with an age- and sex-standardized prevalence of 38.0%, and a marked female predominance³. Even when most detected lesions are clinically indolent, the aggregate consequences are substantial, including repeated imaging, procedural cascades, heightened symptom vigilance, and a persistent tension between reassurance and intervention. This tension is most acute for benign nodules that are symptomatic, cosmetically conspicuous, or demonstrably enlarging, where “benign” does not equate to “inconsequential”, and where patient-centered outcomes often hinge on whether a therapy can relieve local effects while preserving function and avoiding surgical morbidity.

Over the past decade, ultrasound-guided thermal ablation has moved from niche adoption to an increasingly established alternative to surgery for selected thyroid nodules, particularly benign solid or predominantly solid nodules and, in carefully chosen contexts, recurrent lesions^4,5. Radiofrequency ablation (RFA) has become the flagship technique because it couples real-time sonographic visualization with controllable energy delivery and a complication profile that is favorable in experienced hands^4,5. Beyond the obvious appeal of avoiding a cervical scar and general anesthesia in many cases, RFA aligns with an organ-sparing philosophy in which the lesion is remodeled while thyroid reserve is preserved. Contemporary outcome reporting has converged on the volume reduction ratio (VRR), commonly evaluated at 6 to 12 months, with treatment success in many series and pooled analyses defined using thresholds such as at least 50% volume reduction at 1 year⁶. Yet volume reduction is inherently a late surrogate. The biological effect of ablation begins immediately with protein denaturation and coagulative necrosis, then evolves through inflammation, resorption, and fibrosis over weeks to months⁶. As a result, the earliest clinically decisive question, namely, to provide additional information regarding early post-ablation tissue response, often cannot be answered with confidence using size alone^4,6. Doppler vascularity and grayscale echotexture add useful context, but they remain indirect measures of tissue transformation, and they can be confounded in the early post-procedural window by reactive hyperemia, edema, or heterogeneous necrotic change. An early, objective signal of adequate treatment could rationalize follow-up intensity, help anticipate the likelihood of delayed volume response, and may help identify patients with less favorable longitudinal response patterns before months of watchful waiting consolidate into persistent symptoms or regrowth^2,3.

Elastography is compelling in this setting because it measures a physical property that ablation is designed to change. Ultrasound elastography, encompassing strain-based and shear-wave techniques, characterizes tissue stiffness semi-quantitatively through deformation patterns or quantitatively through shear-wave velocity and derived elastic modulus⁷. In thyroid imaging, elastography has been extensively evaluated as an adjunct for malignancy risk stratification. However, rigorous testing has tempered early enthusiasm regarding its incremental diagnostic value over conventional ultrasound in broad clinical populations, reminding the field that elegant physics does not guarantee clinical gain in every context⁸. The post-ablation problem is different. Here, the target lesion is known and treated. The task is not histologic discrimination but response assessment, and the core biological transition is from viable, deformable tissue toward a less compliant necrotic and fibrotic composite. Early clinical investigations across thermal modalities support this mechanistic premise. In patients undergoing microwave ablation, elastography-based scoring demonstrated measurable structural alteration after therapy and was proposed as a feasible follow-up tool⁹. In RFA cohorts, elastography has been explored as a method to quantify stiffness changes in treated nodules and adjacent reference tissues, offering a potential window into ablation-induced remodeling that may precede macroscopic shrinkage¹⁰. Technical learning-curve work has also incorporated strain elastography as part of sonographic characterization around the procedure, reflecting a growing, although still inconsistent, recognition that stiffness patterns can add information beyond size¹¹. Taken together, these studies suggest that elastography could evolve from a diagnostic adjunct into a response biomarker. However, lack of protocol standardization and variability in acquisition methods remain important barriers to broader clinical implementation^10,11. Existing reports vary in elastography modality, acquisition parameters, region-of-interest strategy, reference tissue selection, and assessment timing, which makes it difficult to translate stiffness changes into clinically robust rules that can guide early decision-making across operators and platforms.

This study was designed to address that translational gap by evaluating an elastography-based approach for assessing treatment response following RFA of thyroid nodules, with explicit attention to practical implementation in routine follow-up. The central premise of the present study is that response after RFA is fundamentally a biomechanical process that begins with thermal injury and progresses through longitudinal structural and stiffness changes after ablation, and that stiffness metrics should therefore change in a directionally consistent and clinically interpretable manner. It was hypothesized that elastography-derived indices would demonstrate systematic post-ablation shifts consistent with treatment effect, track the response trajectory over follow-up, and provide complementary information to conventional endpoints such as VRR and Doppler vascularity, particularly in the early period when size-based assessment is least informative^4,6,7. The novelty of the present study lies in reframing elastography from an additional feature for malignancy discrimination to an objective, quantifiable readout of post-ablation tissue transformation, and in aligning elastography measurements with the clinical questions that drive management after RFA, including longitudinal treatment response assessment and characterization of post-ablation stiffness change during follow-up. If elastography can reliably operationalize these questions, it has the potential to shorten the feedback loop between intervention and outcome, potentially enabling earlier characterization of longitudinal response patterns after RFA while preserving the minimally invasive advantages that have made thyroid RFA an increasingly important option for patients with benign but burdensome nodules^4,6.

Protocol

Ethics statement
All procedures involving human participants were conducted in accordance with the institutional guidelines of Xingtai Hospital and in compliance with the principles of the Declaration of Helsinki and its later amendments. The study was reviewed and approved by the Human Research Ethics Committee of Xingtai Hospital; because all participants provided standard written clinical informed consent for RFA and ultrasound examinations as part of routine care, the committee approved the study with a waiver of additional written research consent and permitted verbal consent for study participation where required. Participant confidentiality was protected through secure, access-restricted data management, with de-identification of all study data prior to analysis.

Study design and overall schematic
This study was designed as a retrospective, single-center clinical cohort to evaluate elastography-based assessment of treatment response following ultrasound-guided RFA of thyroid nodules. Consecutive eligible patients undergoing RFA at our institution were screened and enrolled until a predefined target of 187 participants was reached. To ensure analytic independence and to avoid within-patient clustering, only one index nodule per participant was included, defined a priori as the nodule selected for treatment at the index ablation session. Patients with diffuse thyroid diseases that could substantially affect background thyroid stiffness measurements, including diffuse Hashimoto thyroiditis with marked heterogeneous parenchymal involvement, uncontrolled Graves disease, diffuse thyroiditis, or diffuse goiter, were excluded. Isolated thyroid autoantibody positivity without diffuse structural thyroid abnormality on ultrasound was not considered an exclusion criterion.

The study followed a standardized longitudinal workflow. Baseline clinical evaluation and imaging were performed within 2 weeks before ablation and included conventional ultrasound, Doppler assessment, and elastography acquisition using a prespecified protocol. RFA was conducted under real-time ultrasound guidance according to institutional standards. Immediate post-ablation imaging was obtained within 30 min of energy delivery to document technical success and acute tissue changes. Scheduled follow-up visits were conducted at 1, 3, 6, and 12 months after ablation, with repeat ultrasound and elastography performed at each time point using identical acquisition settings and measurement strategies to minimize technical variability.

Sample size calculation and statistical assumptions
The primary endpoint for sample size estimation was the within-nodule change in quantitative elastography at 1 month after ablation compared with baseline, expressed as the difference in mean stiffness (for example, elastic modulus in kPa) measured using a standardized region of interest strategy. Because the primary analysis compares paired measurements within the same nodule, the sample size was estimated using a paired mean difference framework. Let Δ denote the expected mean change in stiffness from baseline to 1 month, and let σ_d denote the standard deviation of the paired differences. For a two-sided test at significance level α with power 1−β, the required sample size is:

Using α = 0.05 (two-sided) and power 1−β = 0.80, Z1−α/2 = 1.96 and Z1−β = 0.84. Based on pilot measurements and published post-ablation variability in thyroid elastography, this study assumed β_d = 35 kPa and a conservative clinically meaningful change Δ = 8 kPa. This yields a minimum required sample size of at least 150 participants for the primary paired analysis of elastography-derived stiffness before and after RFA.

Baseline clinical assessment prior to ablation
At enrollment, a standardized clinical history was obtained with emphasis on nodule-related symptoms and factors relevant to procedural safety. Age, sex, anthropometric measures, and pertinent comorbidities, including cardiovascular and pulmonary disease, diabetes mellitus, chronic kidney disease, and prior neck surgery or irradiation, were recorded. Current medications were documented, with particular attention to antiplatelet agents and anticoagulants, and any planned peri-procedural adjustments according to institutional policy were recorded. The indication for RFA was confirmed, specifying whether the primary driver was compressive symptoms, cosmetic concern, documented growth on serial imaging, or a combination of these factors.

Baseline symptom burden was established using a prespecified instrument. Local compressive symptoms such as dysphagia, throat pressure, foreign-body sensation, dyspnea, and voice-related complaints were assessed, and their severity was documented using a numeric rating scale or an institutionally adopted symptom score. When cosmetic impact was an indication, the cosmetic appearance was graded using a standardized cosmetic scale, and patient-reported dissatisfaction was recorded. Relevant thyroid history, including duration of nodule awareness, prior fine-needle aspiration results, prior benign cytology confirmation strategy (single versus repeated sampling as required), and any previous medical therapy for thyroid disease, was captured.

Baseline laboratory testing was obtained according to routine clinical practice, including thyroid-stimulating hormone and additional thyroid function parameters where indicated. Cytology and pathology reports confirming benign status for the index nodule, as well as any ultrasound risk stratification documentation if available, were reviewed and recorded. A baseline safety assessment, including vital signs and a focused airway evaluation when the nodule abutted the trachea or caused symptoms suggestive of airway compromise, was established. The planned follow-up schedule was documented, and standardized pre-imaging instructions, including avoidance of swallowing during elastography acquisition, were reinforced to improve measurement consistency across visits.

Conventional ultrasound acquisition and nodule volume measurement
Grayscale ultrasound was performed with the patient supine and the neck gently extended. A standardized imaging preset was used, and key settings were kept consistent across visits, including frequency range, depth, focus, gain, and dynamic range. Frame rate, elastography scale range, persistence, and smoothing settings were kept constant across baseline and follow-up examinations. Ultrasound examinations were performed using a high-frequency linear-array transducer (5–14 MHz) equipped with shear-wave and strain elastography functionality. All ultrasound and elastography examinations were performed using the same ultrasound platform and software version throughout the study period to minimize inter-system variability in stiffness measurements. Transverse and longitudinal views that clearly delineated the nodule margins and adjacent anatomic landmarks were acquired. The maximal orthogonal diameters of the index nodule—anteroposterior (A), transverse (B), and longitudinal (C)—were measured. Electronic calipers were placed at the outer edge of the nodule on frozen images, avoiding inclusion of surrounding parenchyma or perinodular halo. All measurements were recorded in millimeters and, when feasible, each dimension was repeated and the mean value was used. Nodule volume was calculated using the ellipsoid formula:

The same measurement approach was repeated at each follow-up visit. The VRR was computed as:

Elastography acquisition and quality control
Elastography was performed immediately after conventional ultrasound with the patient supine and the neck extended. The patient was asked to avoid swallowing and to briefly suspend breathing during acquisition to minimize motion artifacts. The same elastography mode and settings were used at all visits.

For shear wave elastography, the sampling box was placed to include the solid portion of the index nodule and, when feasible, adjacent normal thyroid tissue at a similar depth. Circular regions of interest were positioned preferentially within the central solid component of the nodule while avoiding cystic degeneration, macrocalcification, peripheral halo regions, and obvious post-ablation cavitary change. After region-of-interest placement, image acquisition was initiated only when the elastography color map remained visually stable for at least 3–5 s. Perinodular stiffness assessment was not systematically incorporated into the predefined acquisition protocol because the primary study objective focused on longitudinal stiffness evolution within the treated nodule itself. At least three technically adequate measurements were obtained, and the median stiffness value was recorded as the primary metric; the interquartile range was documented when available.

For strain elastography, gentle, consistent compression was applied guided by the system’s quality indicator. Compression was applied manually at a slow, rhythmic frequency while monitoring the real-time quality indicator provided by the ultrasound system. Excessive probe pressure, causing visible deformation of the thyroid capsule or unstable elastography color mapping, was avoided. A reference region of interest was defined in adjacent normal thyroid tissue at a similar depth, and the strain ratio was calculated. Each elastography acquisition sequence was maintained for approximately 5–10 s to ensure image stabilization and reproducible sampling. At least three acceptable measurements were acquired, and the median value was recorded.

Measurements were accepted only when the elastography map was stable, region-of-interest placement followed the prespecified rules, and there was no gross motion or compression artifact. Measurements with incomplete color filling, obvious signal voids within the region of interest, or marked frame-to-frame fluctuation were rejected and reacquired. Ultrasound and elastography examinations were performed in a temperature-controlled examination room to minimize environmental influences on tissue stiffness measurements. Representative elastography images with the corresponding B-mode images were stored at each time point for traceability and quality review.

RFA procedure
RFA was performed under continuous ultrasound guidance by operators who had completed institutional training in thyroid ablation techniques, had more than 5 years of experience in thyroid ultrasound-guided interventions, and had each performed more than 200 thyroid RFA procedures before study initiation. The patient was positioned supine with the neck extended. The cervical region was prepared and draped using standard sterile technique. Local anesthesia was administered according to routine practice; conscious sedation was used when clinically indicated, with continuous monitoring of vital signs.

The radiofrequency electrode was introduced into the index nodule under real-time ultrasound guidance, preferentially using a transisthmic approach to improve electrode stability and to minimize unintended displacement. When the index nodule was adjacent to the trachea, carotid sheath, esophagus, or recurrent laryngeal nerve region, hydrodissection was performed by slow injection of sterile fluid under ultrasound guidance to maintain a visible separation between the ablation zone and adjacent critical structures throughout energy delivery.

Radiofrequency energy was delivered using a moving-shot technique, beginning in the deepest portion of the nodule and progressing systematically toward the superficial regions to achieve overlapping ablation zones. Each ablation unit was treated until transient hyperechoic change occupied the targeted area before the electrode tip was repositioned to the adjacent untreated portion. Power output was adjusted in response to real-time echogenic changes and patient tolerance, and ablation was continued until the targeted nodule demonstrated the expected transient hyperechoic change and intranodular vascularity was markedly reduced on Doppler assessment. Adequate ablation was additionally supported by visualization of a continuous hyperechoic zone covering the targeted treatment unit without obvious residual intranodular vascular signals.

Key procedural parameters, including ablation time, power range, estimated delivered energy, use of hydrodissection, and any intraprocedural events, were recorded. After completion of ablation, the patient was observed according to institutional protocol, and standardized post-procedural instructions were provided. Ablation was temporarily interrupted or terminated when patients developed significant pain, voice change, uncontrolled coughing, progressive swelling, or other findings raising concern for thermal injury or procedural complications.

Immediate post-ablation imaging and documentation
Within 30 min after completion of RFA, an immediate post-procedural ultrasound examination was performed using the same imaging presets as baseline. Grayscale images were acquired to document the extent of the treated area and to identify typical post-ablation features, including transient hyperechogenicity and changes in internal architecture. Intranodular and perinodular vascularity were assessed using color or power Doppler to confirm reduction or absence of blood flow within the ablated zone.

Elastography acquisition was repeated after transient procedure-related motion had resolved and the patient was able to maintain stable breath suspension, following the identical acquisition protocol used at baseline. Stiffness values were recorded, and representative elastography maps together with corresponding B-mode images were stored. Any technical limitations encountered during acquisition, such as patient discomfort or motion artifacts, were documented. These immediate post-ablation images served as a procedural reference and were not used as the primary endpoint but provided context for subsequent longitudinal changes.

Follow-up schedule and standardized endpoints
Follow-up visits were scheduled at 1 month, 3 months, 6 months, and 12 months after ablation. At each visit, a standardized evaluation was performed, including symptom assessment, cosmetic grading when applicable, conventional ultrasound, and elastography using the same acquisition parameters and measurement strategy as at baseline. The primary elastography endpoint was the change in stiffness of the index nodule relative to baseline at predefined follow-up time points, with particular emphasis on early change at 1 month. The primary conventional imaging endpoint was the VRR at 12 months. Secondary endpoints included VRR at earlier follow-up visits, longitudinal stiffness trajectories, and the presence of residual or recurrent vascularized tissue within the treated area. Adverse events and any additional interventions during follow-up were recorded systematically.

Data management and blinding
Each participant was assigned a unique study identifier to ensure anonymization. Clinical data and imaging files were stored in a secure institutional database with access restricted to authorized study personnel. Elastography and volumetric measurements are performed by trained readers who are blinded to clinical symptom scores and prior elastography results when feasible. To ensure consistency, maintain a written measurement manual specifying region-of-interest placement rules, image quality criteria, and calculation procedures. In cases of substantial inter-reader discrepancy exceeding a prespecified threshold, measurements are reviewed jointly and resolved by consensus. All changes to the dataset are logged to preserve an audit trail. All ultrasound and elastography acquisitions were performed by operators with prior experience in thyroid ultrasound and RFA imaging assessment who underwent protocol-specific training before study initiation. Stored images were reviewed offline using the same predefined region-of-interest placement rules and quality-control criteria applied during acquisition.

Statistical analysis
Continuous variables were assessed for distributional characteristics using visual inspection of histograms and the Shapiro–Wilk test. Normally distributed variables are presented as mean ± standard deviation, whereas non-normally distributed variables are summarized as median with interquartile range. Categorical variables are reported as counts and percentages. The primary analytical objective was to evaluate within-nodule changes in elastography-derived stiffness following RFA. For the primary paired comparison between baseline and post-ablation time points, differences in stiffness values were analyzed using paired t tests when the distribution of paired differences approximated normality; otherwise, the Wilcoxon signed-rank test was applied. Effect sizes are reported as mean or median changes with corresponding 95% confidence intervals where appropriate. To characterize longitudinal changes in elastography measurements across all follow-up visits (baseline, immediate post-ablation, 1 month, 3 months, 6 months, and 12 months), linear mixed-effects models were constructed with random intercepts at the participant level to account for within-subject correlation. Time was modeled as a categorical fixed effect rather than as a continuous variable because post-ablation biomechanical changes were not expected to follow a linear trajectory across follow-up visits. In particular, the study design anticipated an acute immediate post-ablation stiffness increase followed by progressive remodeling and late decline, making categorical modeling more appropriate for capturing potentially non-linear temporal response patterns without imposing linearity assumptions. Model assumptions were evaluated by inspection of residual plots. When appropriate, stiffness values were log-transformed to improve model fit. Associations between elastography-derived stiffness changes and conventional treatment response metrics were explored in secondary analyses. Correlation between early stiffness change and VRR at subsequent follow-up was assessed using Pearson or Spearman correlation coefficients, depending on data distribution. Multivariable regression models were used to examine whether early elastography changes independently predicted later volumetric response after adjustment for baseline nodule volume, nodule composition (solid versus predominantly solid), delivered energy, and age. Interaction terms between subgroup variables and time were tested within the mixed-effects modeling framework. For longitudinal analyses, mixed-effects models inherently accommodate missing observations under the assumption of missing at random. All statistical tests were two-sided, and a P value < 0.05 was considered to indicate statistical significance.

Results

Study design and overall schematic
The excluded participants did not meet predefined eligibility criteria, had incomplete baseline imaging data, or lacked standardized longitudinal follow-up. Among the 187 included participants, complete 12-month outcome classification was available for all cases included in the primary analysis, although isolated intermittent missing observations at individual follow-up visits were accommodated within the mixed-effects modeling framework (Figure 1). Using the prespecified 12-month response definition based on VRR, 146 nodules were classified as responders (VRR ≥ 50%) and 41 as non-responders (VRR < 50%) (Figure 1). This design enabled paired within-nodule comparisons (baseline as the internal control) and standardized longitudinal assessment across predefined follow-up time points.

Baseline participant and nodule characteristics
Baseline clinical characteristics are summarized in Table 1. Non-responders were older than responders (49.63 ± 8.01 vs 45.94 ± 7.88 years, P = 0.009) and had slightly higher baseline symptom and cosmetic burden (pressure symptom score 3.43 ± 0.57 vs 3.19 ± 0.63, P = 0.029; cosmetic concern score 3.01 ± 0.47 vs 2.79 ± 0.49, P = 0.011). Thyroid function was largely similar between groups, although TSH was modestly higher among non-responders (1.86 ± 0.31 mIU/L vs 1.74 ± 0.33 mIU/L, P = 0.039) (Table 1). These baseline differences indicate a less favorable clinical profile among non-responders prior to intervention. Baseline ultrasound showed significantly larger nodules in non-responders, including greater maximum diameter (33.87 ± 4.03 mm vs 29.31 ± 3.94 mm, P < 0.001) and baseline volume (12.04 ± 2.41 mL vs 8.33 ± 1.90 mL, P < 0.001) (Table 1). Nodule composition also differed, with a shift toward higher cystic component categories in non-responders (P < 0.001) (Table 1). These baseline ultrasound characteristics provide an anatomic context for subsequent volumetric and biomechanical response patterns. No participant demonstrated diffuse thyroid parenchymal abnormalities severe enough to meet the predefined exclusion criteria for diffuse thyroid disease.

Elastography measurements were obtained under standardized conditions with prespecified region-of-interest rules and acceptance criteria to minimize motion and compression artifacts. The use of baseline within-nodule measurements as internal references may help reduce the influence of interindividual variability and support the interpretation of longitudinal stiffness changes. Procedural characteristics are shown in Table 1. Consistent with larger baseline nodule burden, non-responders had longer procedure duration (21.13 ± 3.01 min vs 18.42 ± 2.79 min, P < 0.001) and higher delivered energy (13.18 ± 2.63 kJ vs 11.04 ± 1.98 kJ, P < 0.001) (Table 1). These procedural differences were expected, given baseline size and composition differences and were accounted for in adjusted analyses of response prediction.

Longitudinal elastography trajectories after RFA
Immediately after RFA, both elastography modalities demonstrated a marked acute stiffness increase (Table 2; Figure 2), consistent with the hypothesized immediate biomechanical effect of thermal coagulation. Shear-wave elastography (SWE) increased from 31.5 ± 5.1 to 41.7 ± 5.6 kPa in responders and from 35.1 ± 5.3 to 43.4 ± 5.8 kPa in non-responders (Table 2; Figure 2). Strain ratio increased concordantly (responders 2.41 ± 0.37 to 3.14 ± 0.41; non-responders 2.77 ± 0.40 to 3.35 ± 0.42) (Table 2; Figure 2). In contrast, immediate post-ablation volume remained close to baseline in both groups (Table 2), supporting the study hypothesis that elastography provides an earlier, early imaging indicator of treatment-related tissue change than short-interval volumetry.

Longitudinal volumetric outcomes and VRR
Over follow-up, elastography demonstrated progressive remodeling after the immediate post-ablation rise (Table 2; Figure 2). In responders, SWE fell below baseline by 3 months and continued to decline through 12 months (31.5 ± 5.1 kPa at baseline to 21.6 ± 4.7 kPa at 12 months), whereas non-responders showed an attenuated reduction and remained substantially stiffer at 12 months (31.2 ± 5.2 kPa) (Table 2; Figure 2). Strain ratio followed the same pattern, decreasing to 1.87 ± 0.35 at 12 months in responders versus 2.55 ± 0.41 in non-responders (Table 2; Figure 2).

Volumetric outcomes evolved in parallel with biomechanical changes (Table 2; Figure 3). Nodule volume decreased steadily in both groups, with persistent separation across time points; at 12 months, responders reached 2.20 ± 0.45 mL compared with 7.35 ± 1.80 mL in non-responders (Table 2; Figure 3). VRR also separated early and remained distinct: at 1 month, VRR was 31.5 ± 6.3% in responders versus 16.1 ± 3.8% in non-responders, and at 12 months VRR was 73.6 ± 6.9% versus 38.5 ± 8.7%, respectively (Table 2; Figure 3). Together, these aligned elastography and VRR trajectories support the hypothesis that post-RFA biomechanical remodeling tracks and differentiates durable volumetric response.

Early elastography change as an independent predictor of 12-month volumetric response
Mixed-effects models corroborated the descriptive trajectories and quantified group differences over time (Table 3). For SWE, time effects indicated a significant immediate increase relative to baseline (estimate +10.22 kPa, P < 0.001) followed by progressively lower values at later visits (e.g., 12 months vs baseline estimate -9.02 kPa, P < 0.001). Importantly, time-by-group interactions showed divergence emerging at later follow-up, with non-responders exhibiting higher SWE relative to responders from 3 months onward (non-responder × 3 months estimate +4.20 kPa, P < 0.001; non-responder × 12 months estimate +9.41 kPa, P < 0.001) (Table 3). Strain ratio models demonstrated the same pattern, with significant late divergence (non-responder × 12 months estimate +0.68, P < 0.001) (Table 3).

In correlation analyses, early post-ablation elastography changes (immediate minus baseline) were positively associated with 12-month volumetric response (Figure 4). Δstrain ratio showed a moderate correlation with 12-month VRR (Pearson r = 0.445, P < 0.001), whereas ΔSWE demonstrated a weaker but statistically significant correlation (Pearson r = 0.302, P < 0.001). These findings indicate that larger acute post-ablation stiffness changes were associated with greater subsequent volume reduction, supporting the hypothesis that elastography captures early biomechanical effects that are informative for longer-term treatment response.

In multivariable regression adjusting for baseline nodule volume, composition, delivered energy, and age, early post-ablation elastography changes independently predicted 12-month VRR (Table 4). Early change in strain ratio (immediate minus baseline) was strongly associated with higher 12-month VRR (β = 14.80% per +1.0, P < 0.001), and early SWE change also contributed independently (β = 1.80% per +5 kPa, P = 0.011). Larger baseline volume was associated with lower 12-month VRR (β = −1.05% per +1 mL, P < 0.001), whereas solid composition and higher delivered energy were associated with higher VRR (both P < 0.05) (Table 4).

During follow-up, no treated nodules were subsequently reclassified as malignant or alternative pathologies based on repeat imaging, cytology, or clinical assessment. Although some non-responding nodules demonstrated persistent residual volume or attenuated volume reduction, overt regrowth requiring repeat ablation or surgery was not observed during the 12-month follow-up period.

DATA AVAILABILITY:
The raw data underlying the analyses presented in this study have been prepared and uploaded as Supplementary Table 1 to improve transparency and reproducibility. The dataset includes de-identified participant-level demographic, clinical, ultrasound, elastography, procedural, and longitudinal follow-up variables used in the analyses.

Figure 1: Study design and analytical workflow. Participants screened during the study period were assessed for eligibility and included after predefined exclusions. One index thyroid nodule per participant was analyzed. Response status at 12 months was defined by volume reduction ratio (VRR): responder (VRR ≥ 50%) and non-responder (VRR < 50%). Please click here to view a larger version of this figure.

Figure 2: Longitudinal shear-wave elastography stiffness after radiofrequency ablation. Mean shear-wave elastography (SWE) stiffness (kPa) is shown at baseline, immediately post-ablation, and at 1, 3, 6, and 12 months for the overall cohort and stratified by 12-month response status. Error bars indicate standard deviation. Please click here to view a larger version of this figure.

Figure 3: Longitudinal changes in nodule volume and volume reduction ratio. (A) Mean nodule volume (mL) and (B) mean volume reduction ratio (VRR, %) at baseline, immediately post-ablation, and at 1, 3, 6, and 12 months, stratified by 12-month response status. Error bars indicate standard deviation. Brackets indicate between-group comparisons at each time point. Please click here to view a larger version of this figure.

Figure 4: Relationship between early elastography changes and 12-month volumetric response. Scatter plots show associations between early post-ablation elastography changes (immediate minus baseline) and 12-month VRR: (A) Δstrain ratio and (B) ΔSWE stiffness. Scatter point positions were preserved from the original overall analysis to facilitate visual comparison across groups. Please click here to view a larger version of this figure.

Section	Variable	Responder (n = 146)	Non-responder (n = 41)	Test statistic (t/χ2)	P value
Demographics and baseline clinical status
	Age, years	45.94 ± 7.88	49.63 ± 8.01	-2.640	0.009
	Sex	Female: 119 (81.5%); Male: 27 (18.5%)	Female: 29 (70.7%); Male: 12 (29.3%)	2.252	0.133
	BMI, kg/m2	23.74 ± 1.96	24.74 ± 1.94	-1.914	0.057
	Smoking (current/ever)	No: 126 (86.3%); Yes: 20 (13.7%)	No: 32 (78.0%); Yes: 9 (22.0%)	1.664	0.197
	Hypertension	No: 119 (81.5%); Yes: 27 (18.5%)	No: 30 (73.2%); Yes: 11 (26.8%)	1.374	0.241
	Diabetes mellitus	No: 136 (93.2%); Yes: 10 (6.8%)	No: 37 (90.2%); Yes: 4 (9.8%)	0.391	0.532
	Coronary artery disease	No: 140 (95.9%); Yes: 6 (4.1%)	No: 38 (92.7%); Yes: 3 (7.3%)	0.719	0.397
	Chronic kidney disease	No: 143 (97.9%); Yes: 3 (2.1%)	No: 40 (97.6%); Yes: 1 (2.4%)	0.023	0.881
	COPD/asthma	No: 140 (95.9%); Yes: 6 (4.1%)	No: 39 (95.1%); Yes: 2 (4.9%)	0.046	0.830
	Pressure symptom score, 0–10	3.19 ± 0.63	3.43 ± 0.57	-2.199	0.029
	Cosmetic concern score, 1–4	2.79 ± 0.49	3.01 ± 0.47	-2.562	0.011
Thyroid labs and autoimmunity
	TSH, mIU/L	1.74 ± 0.33	1.86 ± 0.31	-2.084	0.039
	Free T4, pmol/L	15.06 ± 2.03	14.95 ± 1.92	0.310	0.757
	TPOAb positive	No: 129 (88.4%); Yes: 17 (11.6%)	No: 35 (85.4%); Yes: 6 (14.6%)	0.265	0.606
	TgAb positive	No: 132 (90.4%); Yes: 14 (9.6%)	No: 36 (87.8%); Yes: 5 (12.2%)	0.238	0.626
	Levothyroxine use	No: 134 (91.8%); Yes: 12 (8.2%)	No: 36 (87.8%); Yes: 5 (12.2%)	0.612	0.434
	Antiplatelet/anticoagulant use	No: 139 (95.2%); Yes: 7 (4.8%)	No: 38 (92.7%); Yes: 3 (7.3%)	0.402	0.526
Baseline ultrasound features
	Nodule location	Left: 66 (45.2%); Isthmus: 6 (4.1%); Right: 74 (50.7%)	Left: 19 (46.3%); Isthmus: 2 (4.9%); Right: 20 (48.8%)	0.076	0.963
	Maximum diameter, mm	29.31 ± 3.94	33.87 ± 4.03	-6.516	<0.001
	Baseline nodule volume, mL	8.33 ± 1.90	12.04 ± 2.41	-10.385	<0.001
	Composition (solid)	No: 44 (30.1%); Yes: 102 (69.9%)	No: 17 (41.5%); Yes: 24 (58.5%)	1.868	0.172
	Cystic component	<10%: 89 (61.0%); 10–25%: 47 (32.2%); >25%: 10 (6.8%)	<10%: 10 (24.4%); 10–25%: 20 (48.8%); >25%: 11 (26.8%)	21.923	<0.001
	Calcification	Coarse: 10 (6.8%); Micro: 35 (24.0%); None: 101 (69.2%)	Coarse: 3 (7.3%); Micro: 8 (19.5%); None: 30 (73.2%)	0.360	0.835
	Baseline vascularity grade	0: 27 (18.5%); 1: 47 (32.2%); 2: 53 (36.3%); 3: 19 (13.0%)	0: 5 (12.2%); 1: 13 (31.7%); 2: 15 (36.6%); 3: 8 (19.5%)	1.681	0.641
Baseline elastography
	Baseline SWE stiffness, kPa	31.50 ± 5.11	35.08 ± 5.34	-3.925	<0.001
	Baseline strain ratio	2.41 ± 0.37	2.77 ± 0.40	-5.407	<0.001
Procedural parameters & safety
	Procedure time, min	18.42 ± 2.79	21.13 ± 3.01	-5.401	<0.001
	Ablation power, W	35.02 ± 4.78	37.05 ± 4.63	-2.419	0.017
	Delivered energy, kJ	11.04 ± 1.98	13.18 ± 2.63	-5.665	<0.001
	Hydrodissection used	No: 100 (68.5%); Yes: 46 (31.5%)	No: 25 (61.0%); Yes: 16 (39.0%)	0.816	0.366
	Sedation used	No: 115 (78.8%); Yes: 31 (21.2%)	No: 30 (73.2%); Yes: 11 (26.8%)	0.576	0.448
	Intra-procedural pain, VAS 0–10	2.60 ± 0.59	2.78 ± 0.61	-1.713	0.088
	Minor bleeding/hematoma	No: 139 (95.2%); Yes: 7 (4.8%)	No: 37 (90.2%); Yes: 4 (9.8%)	1.423	0.233
	Transient voice change	No: 143 (97.9%); Yes: 3 (2.1%)	No: 41 (100.0%); Yes: 0 (0.0%)	0.856	0.355

Table 1: Baseline demographic, clinical, laboratory, ultrasound, elastography, and procedural characteristics stratified by 12-month volumetric response after radiofrequency ablation (VRR ≥50% vs <50%). Data are shown as mean ± SD (or median [IQR]) and n (%). Between-group comparisons used Student’s t test or Wilcoxon rank-sum test for continuous variables and χ² or Fisher’s exact test for categorical variables, as appropriate. Two-sided P < 0.05 indicated statistical significance. Abbreviations: BMI, body mass index; TSH, thyroid-stimulating hormone; FT4, free thyroxine; TPOAb, thyroid peroxidase antibody; TgAb, thyroglobulin antibody; SWE, shear-wave elastography; ROI, region of interest; VAS, visual analog scale; RFA, radiofrequency ablation; VRR, volume reduction ratio.

Visit	Group	n	Volume (mL), mean ± SD	VRR (%), mean ± SD	SWE (kPa), mean ± SD	Strain ratio, mean ± SD
Baseline	Responder	146	8.33 ± 1.90	—	31.5 ± 5.1	2.41 ± 0.37
	Non-responder	41	12.04 ± 2.41	—	35.1 ± 5.3	2.77 ± 0.40
Immediate	Responder	146	8.75 ± 1.99	—	41.7 ± 5.6	3.14 ± 0.41
	Non-responder	41	12.66 ± 2.54	—	43.4 ± 5.8	3.35 ± 0.42
1 month	Responder	146	5.69 ± 1.35	31.5 ± 6.3	33.6 ± 5.0	2.51 ± 0.39
	Non-responder	41	10.08 ± 2.36	16.1 ± 3.8	36.1 ± 5.3	2.82 ± 0.41
3 months	Responder	146	3.72 ± 0.85	55.3 ± 7.1	26.4 ± 4.8	2.05 ± 0.36
	Non-responder	41	8.70 ± 2.10	28.1 ± 6.8	32.6 ± 5.1	2.62 ± 0.39
6 months	Responder	146	2.99 ± 0.70	64.1 ± 7.0	23.8 ± 4.7	1.91 ± 0.35
	Non-responder	41	7.93 ± 1.95	34.2 ± 8.1	31.6 ± 5.3	2.56 ± 0.40
12 months	Responder	146	2.20 ± 0.45	73.6 ± 6.9	21.6 ± 4.7	1.87 ± 0.35
	Non-responder	41	7.35 ± 1.80	38.5 ± 8.7	31.2 ± 5.2	2.55 ± 0.41

Table 2: Longitudinal changes in nodule volume, volume reduction ratio, and elastography metrics from baseline to 12 months following radiofrequency ablation, stratified by treatment response. Values are summarized at each visit; n denotes available observations. VRR was calculated relative to baseline volume. Longitudinal inference was performed using mixed-effects models. Abbreviations: VRR, volume reduction ratio; SWE, shear-wave elastography; RFA, radiofrequency ablation.

Outcome	Fixed effect term	Estimate	SE	t	P value
SWE stiffness (kPa)
	(Intercept) Baseline, Responder	31.51	0.55	57.273	<0.001
	Group: Non-responder (vs Responder)	3.58	0.92	3.891	<0.001
	Time: Immediate (vs Baseline)	10.22	0.62	16.452	<0.001
	Time: 1 month (vs Baseline)	2.17	0.65	3.502	<0.001
	Time: 3 months (vs Baseline)	-5.14	0.61	-8.361	<0.001
	Time: 6 months (vs Baseline)	-7.73	0.64	-12.419	<0.001
	Time: 12 months (vs Baseline)	-9.02	0.63	-14.286	<0.001
	Group × Time: Non-responder × Immediate	-0.80	0.98	-0.816	0.415
	Group × Time: Non-responder × 1 month	0.45	0.95	0.474	0.636
	Group × Time: Non-responder × 3 months	4.20	0.96	4.375	<0.001
	Group × Time: Non-responder × 6 months	6.10	0.98	6.224	<0.001
	Group × Time: Non-responder × 12 months	9.41	1.02	9.216	<0.001
Strain ratio
	(Intercept) Baseline, Responder	2.41	0.04	60.252	<0.001
	Group: Non-responder (vs Responder)	0.36	0.07	5.143	<0.001
	Time: Immediate (vs Baseline)	0.73	0.05	14.601	<0.001
	Time: 1 month (vs Baseline)	0.10	0.07	2.012	0.046
	Time: 3 months (vs Baseline)	-0.36	0.08	-7.211	<0.001
	Time: 6 months (vs Baseline)	-0.51	0.06	-10.032	<0.001
	Time: 12 months (vs Baseline)	-0.54	0.05	-10.802	<0.001
	Group × Time: Non-responder × Immediate	-0.08	0.08	-1.000	0.318
	Group × Time: Non-responder × 1 month	0.05	0.09	0.625	0.532
	Group × Time: Non-responder × 3 months	0.32	0.04	4.031	<0.001
	Group × Time: Non-responder × 6 months	0.55	0.08	6.875	<0.001
	Group × Time: Non-responder × 12 months	0.68	0.07	8.521	<0.001

Table 3: Linear mixed-effects model results for longitudinal elastography trajectories across follow-up visits, including fixed effects for time and time-by-response interaction with participant-level random intercepts. Linear mixed-effects models included participant-level random intercepts and categorical time effects; response group and time-by-group interaction terms were used to compare trajectories. Model fit was assessed using residual diagnostics; log-transformation was applied when needed. Estimates are reported with 95% CIs. Abbreviations: SWE, shear-wave elastography; CI, confidence interval; VRR, volume reduction ratio.

Outcome	Term	Beta	SE	t	df	P value	95% CI
12-month VRR (%)
	(Intercept)	82.00	4.20	19.524	180	<0.001	73.71 to 90.29
	Δ Strain ratio (Immediate - Baseline), per 1.0	14.80	2.40	6.167	180	<0.001	10.06 to 19.54
	Δ SWE (Immediate - Baseline), per 5 kPa	1.80	0.70	2.571	180	0.011	0.42 to 3.18
	Baseline nodule volume, per 1 mL	-1.05	0.28	-3.752	180	<0.001	-1.60 to -0.50
	Solid composition (Yes vs No)	3.20	1.40	2.286	180	0.023	0.44 to 5.96
	Delivered energy, per 1 kJ	0.55	0.22	2.512	180	0.013	0.12 to 0.98
	Age, per 10 years	-1.10	0.70	-1.571	180	0.118	-2.48 to 0.28

Table 4: Multivariable regression analyses evaluating whether early post-ablation elastography changes independently predict 12-month volumetric response after adjustment for baseline nodule and procedural covariates. Multivariable linear regression evaluated associations between early elastography change (immediate minus baseline) and 12-month VRR after covariate adjustment. Coefficients (β) with SE, t, df, two-sided P, and 95% CIs are reported; df represents residual degrees of freedom. Abbreviations: VRR, volume reduction ratio; SWE, shear-wave elastography; SE, standard error; CI, confidence interval; df, degrees of freedom.

Supplementary File 1: Raw data of this study.Please click here to download this file.

Discussion

This study evaluated an elastography-based approach for assessing treatment response after RFA of thyroid nodules by using elastography as a quantitative imaging biomarker associated with post-ablation stiffness change. The central proposal is that stiffness, measured by SWE and strain elastography, can serve as an early indicator of ablation effect and a longitudinal marker of post-ablation remodeling, complementing conventional volume-based endpoints. The findings support this concept at two levels. First, both SWE stiffness and strain ratio increased immediately after ablation, a pattern consistent with acute thermal coagulation and early tissue reaction. Second, stiffness then declined over follow-up, with responders demonstrating a greater late reduction than non-responders, mirroring the separation observed in volumetric outcomes and VRR. In addition, early post-ablation changes in elastography showed statistically significant associations with 12-month VRR and remained independently associated after adjustment for baseline nodule and procedural covariates. Together, these results support the hypothesis that elastography captures longitudinal stiffness changes associated with RFA treatment response, and that early stiffness shifts carry prognostic information that is not fully explained by baseline size, composition, or energy delivery alone^{12,13,14,15,16}.

The observed trajectories are consistent with previously described post-ablation tissue changes and offer a clinically useful narrative for post-ablation imaging^12,13,14. The immediate stiffness rise likely reflects coagulative necrosis within the ablation zone, along with edema and early inflammatory change, all of which increase resistance to deformation and alter wave propagation^15,16. Over time, the decline in stiffness is consistent with progressive resorption of necrotic tissue, organization of the ablation zone, and remodeling that ultimately yields shrinkage and symptom improvement^15,17,18,19. The fact that responders exhibited a larger late reduction in stiffness suggests that successful ablation is not simply an acute event but a sustained remodeling process, and that incomplete treatment or residual viable tissue may be expressed as a blunted softening trajectory²⁰. This perspective helps reconcile a common limitation of early volume assessment. Immediately after RFA, volume may remain near baseline or transiently fluctuate due to edema, hemorrhage, or reactive change, making early volumetry an imperfect indicator of whether the ablation zone is adequate. Elastography, by contrast, responds to the primary physical consequence of ablation at a time when size-based metrics are least discriminative²¹. The divergence in late stiffness trajectories, quantified in the mixed effects models, further supports the concept that stiffness captures meaningful differences in tissue evolution that align with durable volume reduction.

From the standpoint of scientific advancement, this work contributes by shifting elastography from a predominantly diagnostic role toward a therapeutic monitoring paradigm. In benign thyroid nodule care, the field has increasingly adopted image-guided ablation, yet follow-up strategies still depend heavily on delayed volumetric outcomes such as VRR at 6 to 12 months. A biomarker that provides an objective early readout of treatment effect could provide additional imaging information during the early post-procedural period and support further investigation into risk-adapted follow-up strategies^22,23. Similar elastography-based approaches have also been explored in other thermal ablation settings²⁴. Stiffness mapping is attractive because it is noninvasive, can be integrated into routine ultrasound workflows, and can be repeated without contrast agents or radiation²⁵. From a workflow perspective, elastography can be incorporated into routine post-ablation ultrasound examinations without substantial extension of examination time when standardized acquisition protocols are used. However, broader implementation will require attention to cross-platform variability, operator experience, and harmonization of acquisition parameters across different ultrasound systems. To reduce this source of variability in the present study, all elastography measurements were acquired using the same ultrasound platform and software environment throughout follow-up. Clinical usability may depend less on absolute stiffness values than on consistent longitudinal change patterns obtained under stable acquisition conditions. Accordingly, standardized operator training, predefined quality-control rules, and protocol harmonization remain essential before elastography can be more broadly integrated into routine post-ablation surveillance pathways. The present results suggest that a standardized elastography acquisition protocol may help yield more consistent longitudinal stiffness patterns suitable for clinical research applications. Importantly, the present approach treats baseline within-nodule values as internal controls, thereby emphasizing change metrics that may be more robust across individuals than absolute stiffness alone. An additional consideration is that the present elastography protocol focused primarily on stiffness measurements within the solid component of the treated nodule, whereas perinodular tissue was not systematically evaluated. This represents a potentially important limitation because the peripheral treatment zone may contain biologically meaningful information related to inflammatory response, thermal diffusion, evolving fibrosis, residual viable tissue, and early tissue remodeling after ablation. Emerging imaging literature has increasingly emphasized the diagnostic and biological relevance of lesion periphery and microenvironmental characteristics in treatment-response assessment. Accordingly, future studies incorporating combined intranodular and perinodular stiffness mapping may provide a more comprehensive biomechanical characterization of post-ablation tissue evolution and may further improve early response assessment after thyroid RFA.

In practical implementation, several technical pitfalls should be considered during elastography acquisition and post-ablation assessment. Excessive transducer compression, swallowing motion, unstable breath suspension, heterogeneous cystic degeneration, and incomplete color filling may all introduce measurement instability and reduce reproducibility. Immediate post-ablation assessment may also be influenced by edema, transient hemorrhage, or reactive inflammatory change. Maintaining stable probe positioning, minimizing external compression, using predefined region-of-interest placement rules, and reacquiring measurements when marked frame-to-frame fluctuation or signal voids are present are important troubleshooting strategies for improving acquisition reliability.

Several procedural variables appear particularly important for reproducible longitudinal elastography assessment. These include consistent transducer pressure, stable neck positioning, standardized acquisition timing, avoidance of heavily calcified or cystic regions, and maintenance of comparable region-of-interest depth across serial examinations. Reproducibility may also depend on maintaining identical elastography presets and minimizing inter-operator variation through protocol-specific training and predefined quality-control criteria.

Several specific research applications follow from these findings. First, elastography may help stratify patients according to subsequent response trajectories with suboptimal early biomechanical response who might benefit from closer surveillance, earlier adjunct imaging, or consideration of additional ablation. Accordingly, stiffness thresholds derived from a single protocol or platform should be interpreted cautiously until broader multicenter standardization becomes available. Second, stiffness trajectories may serve as intermediate endpoints in clinical trials of ablation technique, electrode strategy, or adjunctive procedural methods, potentially reducing the time required to detect meaningful differences between approaches. Third, elastography may help interpret ambiguous cases in which volume reduction is delayed yet symptoms change, or in which volume reduction occurs but residual stiffness suggests incomplete remodeling. Fourth, incorporating elastography into predictive models could improve individualized prognostication by combining early change metrics with established determinants such as baseline volume, composition, and delivered energy. The multivariable results in this study support the feasibility of such integration, since early strain ratio change and SWE change contributed independent information beyond conventional covariates. These findings suggest the potential feasibility of incorporating elastography metrics into future predictive or risk-stratification models.

There are also other approaches to studying this hypothesis, and they deserve consideration. Perfusion-based imaging, including contrast-enhanced ultrasound and vascularity assessment, can identify residual viable tissue and may predict longer-term response by estimating the effective treated fraction²⁶. Morphologic metrics such as the initial ablation ratio and early ablation zone volume have also been used to forecast long-term shrinkage and retreatment risk^27,28. Cross-sectional imaging can further characterize ablation zones and complications in selected cases. Patient-reported outcomes, symptom relief, and cosmetic improvement may provide more direct measures of clinical benefit and should be incorporated alongside imaging biomarkers. In this context, elastography should be viewed as complementary rather than exclusive. A strong future direction is head-to-head comparison of elastography against perfusion-based methods and morphologic surrogates within the same cohort, using harmonized follow-up schedules and standardized outcome definitions. Although the present study demonstrated significant associations between early elastography change and later VRR, the current cohort was not specifically designed or powered for the derivation of clinically actionable cutoff values with validated sensitivity and specificity estimates. Prospective multicenter studies with predefined outcome-driven threshold analyses will therefore be important for future clinical translation. Finally, the field would benefit from prospective studies in which early elastography-guided management decisions are tested against standard follow-up strategies, with outcomes including VRR, symptom and cosmetic improvement, and retreatment rates.

Limitations of this study should be considered when interpreting the findings and planning further work. First, elastography is inherently sensitive to acquisition conditions, including transducer pressure, depth, region of interest placement, and device-specific reconstruction, and these factors may introduce measurement variability despite protocolization. To reduce potential confounding from diffuse thyroid stiffness abnormalities, patients with clinically significant diffuse thyroid disease and diffuse heterogeneous parenchymal involvement were excluded during enrollment. Formal interobserver reproducibility analysis was not performed in the present study, although standardized acquisition protocols and operator training procedures were applied to improve measurement consistency. Second, the follow-up schedule captures key milestones but may not fully characterize very early evolution beyond the immediate post-ablation interval, during which edema and reactive change could influence stiffness and complicate separation of acute treatment effect from transient post-procedural physiology. Third, response classification based on a VRR threshold, while widely used, does not encompass all dimensions of clinical success, since symptom improvement, cosmetic outcomes, and long-term regrowth may not perfectly track VRR. In addition, the present follow-up duration was limited to 12 months, whereas treated thyroid nodules may continue to remodel, shrink, or demonstrate delayed regrowth over longer observation periods. Although overt regrowth requiring repeat ablation or surgery was not observed during the current follow-up interval, future studies with extended longitudinal follow-up should incorporate regrowth patterns, retreatment rates, and repeat ablation requirements as additional clinically relevant outcome measures. Fourth, although multivariable models adjust for major baseline and procedural covariates, residual confounding remains possible, including unmeasured technical nuances such as margin completeness, electrode positioning dynamics, and operator experience, which may affect both early stiffness change and long-term response. In addition, because all procedures were performed by experienced thyroid RFA operators at a single center, external generalizability to lower-volume settings requires further evaluation. Fifth, the study did not include direct comparisons with alternative early response tools, such as perfusion-based imaging or formal ablation zone viability metrics, within the same protocol, so the incremental value of elastography relative to those methods cannot be fully established here. Because the study was observational and imaging-based, the identified associations should not be interpreted as proof of direct mechanistic causation between elastography changes and underlying tissue biology. Future studies may further evaluate the independent predictive performance of baseline elastography, immediate post-ablation stiffness measurements, and 1-month follow-up elastography individually, as well as the relative percentage change between baseline and immediate post-RFA values. Such analyses may help clarify whether absolute stiffness measurements or dynamic stiffness-change metrics provide greater prognostic utility for long-term volumetric response assessment after thyroid RFA.

Elastography-based assessment provides a feasible and protocol-standardized approach for longitudinal evaluation after thyroid nodule radiofrequency ablation. Early post-ablation stiffness changes were independently associated with 12-month volumetric response and may provide clinically useful information before substantial volume reduction becomes apparent. These findings support the potential role of standardized elastography protocols as complementary imaging biomarkers for post-ablation response assessment and longitudinal follow-up after thyroid RFA.

Materials

List of materials used in this article

Name	Company	Catalog Number	Comments
3M Tegaderm Film Dressing	Tegaderm	https://www.3m.com/3M/en_US/p/c/medical/bandages-dressings/film/i/health-care/medical/	Adhesive dressing; Post-procedure site coverage.
Disposable syringe	Becton Dickinson	BD 309628	Hydrodissection and local anesthetic administration
Grounding pads for RFA	3M Healthcare	Model 9547	Single-use return electrodes compatible with RF.
High-frequency linear-array transducer (5–14 MHz)	SuperSonic Imagine	SL15-4 linear probe	Thyroid ultrasound and elastography acquisition
High-frequency linear-array ultrasound transducer (7–15 MHz)	Mindray	L11-3U	Linear probe optimized for thyroid imaging and elastography.
IntelliVue	Philips Healthcare	MX450	Patient monitoring system; ECG, SpO2, noninvasive BP monitor during RFA.
IntelliVue	Philips Healthcare	MP50	Vital signs monitor; Continuous peri-procedural monitoring
Lidocaine hydrochloride injection	Hospira Inc.	NDC 0409-4276-17	Local anesthesia before ablation
lme4 package	CRAN	RRID:SCR_015654	Linear mixed-effects modeling in R.
lmerTest package	CRAN	RRID:SCR_015656	Mixed model testing and p-values in R.
Microflex Powder-Free Gloves	Ansell Healthcare		Aseptic barrier for procedural team.
Non-sterile ultrasound gel	EcoGel		Coupling medium for baseline and follow-up imaging.
Picture Archiving and Communication System (PACS)	GE Healthcare	Centricity PACS	Enterprise-grade image storage and retrieval system.
Povidone-Iodine 10%	Betadine		Pre-procedural antiseptic.
Precision Workstation	Dell	T5820	For image review and data analysis.
PrecisionGlide 22G–27G	BD (Becton Dickinson)	https://www.bd.com/en-us/products-and-solutions/products	Needle set for local anesthesia/hydrodissection.
Premixed Saline Bag 500 mL (0.9% NaCl)	Baxter Healthcare		Fluid for hydrodissection and flushing.
R statistical computing software	R Foundation	RRID:SCR_001905	Statistical analysis environment.
RFA percutaneous cooled electrode	STARmed Co., Ltd.	Star RF Electrode 18-10s10F	Active RFA electrode for thyroid ablation (TaeWoong Medical USA)
RStudio integrated development environment	Posit (formerly RStudio)	RRID:SCR_000432	IDE for R statistical analyses.
Shear-wave elastography software module	Mindray	SWE Module (RESONA Series)	Quantitative stiffness mapping in kPa within the US system.
Sterile drape kit	Medline Industries	DYNJP2500	Sterile preparation for thyroid RFA procedure
Sterile gauze and swabs	Covidien	Sterile Sponge Kit	Used for skin prep and hemostasis.
Sterile normal saline	Baxter Healthcare	2B1324X	Hydrodissection during thyroid RFA
Sterile procedure drapes (Blue Drape Pack)	Medline Industries		Standard drape for sterile percutaneous RFA setup.
Sterile ultrasound gel	Aquasonic 100	Conductive Gel 100 mL	Sterile coupling medium for US guidance.
Sterile ultrasound probe cover	Parker Laboratories	Model 1001	Sterile cover for intra-procedural imaging.
Strain elastography software module	Mindray	SE Module (RESONA Series)	Strain ratio calculation with integrated quality indicator.
Syringes (5/10/20 mL)	BD (Becton Dickinson)	Monoject Syringes	For anesthetic and hydrodissection delivery.
Three-way stopcock	BD (Becton Dickinson)	Model 305129	Stopcock and tubing for controlled fluid management.
Thyroid RF Electrode	STARmed Co., Ltd.	https://en.starmed4u.com/	Internally cooled RF electrode; Radiofrequency ablation of thyroid nodules
Ultrasound coupling gel	Parker Laboratories	Aquasonic 100	Ultrasound and elastography acquisition
Ultrasound imaging system with elastography capability	SuperSonic Imagine	Aixplorer	Used for B-mode ultrasound, Doppler imaging, SWE, and strain elastography
VIVA Combo RF Generator	STARmed Co., Ltd.		RF generator for percutaneous thyroid ablation (TaeWoong Medical USA)
VIVA RF Generator	STARmed Co., Ltd.		Ultrasound-guided thyroid radiofrequency ablation
Xylocaine 1% 10 mL	AstraZeneca		Local anesthetic (lidocaine) for procedural analgesia.