Quantitative Multi-Parameter MRI Evaluation of Hashimoto’s Thyroiditis Changes After Dietary Interventions

Introduction

Hashimoto’s thyroiditis, also known as chronic lymphocytic thyroiditis, is a common autoimmune disease and the leading cause of hypothyroidism. It is characterized by thyroid enlargement, lymphocytic infiltration, and thyroid-specific antibodies [1]. The precise mechanisms that trigger autoimmunity remain under investigation. Epidemiological data suggest that a combination of genetic susceptibility and environmental factors plays a critical role in the breakdown of immune tolerance and the onset of the disease. Identified environmental risk factors include radiation, iodine, selenium, smoking, infections, stress, and certain medications [2]. Hashimoto’s thyroiditis typically has a slow, insidious onset with non-specific early symptoms. As thyroid damage progresses, hypothyroidism develops. Elevated thyroglobulin antibody (TGAb) and thyroid peroxidase antibody (TPOAb) levels are key markers of Hashimoto’s thyroiditis, which often precede hypothyroidism. A firm, diffusely enlarged goiter with isthmus or conus involvement may suggest Hashimoto’s thyroiditis, with diagnosis further supported by elevated antibodies and fine-needle aspiration cytology [3].

Ultrasonography, typically performed following serological tests, is the preferred method for detecting thyroid lesions. If necessary, radionuclide scanning can be used as a supplementary diagnostic tool. However, ultrasonography diagnosis may vary due to the skill and experience of the examiner, the choice of probe, and the similarity between thyroid abnormalities and adjacent structures. Despite advancements in ultrasonography technology, this diagnostic instability persists, and the subjective nature of ultrasonography cannot be overlooked. Additionally, ultrasonography lacks sensitive quantitative measurement techniques, making it challenging to accurately assess the progression of Hashimoto’s thyroiditis [4].

Although the diagnosis and evaluation of Hashimoto’s thyroiditis largely depend on serological tests, including assessments of thyroid function and autoantibodies, and ultrasonography, these methods are limited in their ability to accurately track disease progression compared with in vivo tissue alterations. There remains an unmet need for a tool capable of providing sensitive and visual assessments of both structural and inflammatory changes in the thyroid. Studies have reported that magnetic resonance imaging (MRI) can distinguish signal intensity changes in normal thyroid parenchyma from diffuse thyroid lesions [5]. The Dixon method, introduced by Dixon in 1984, is a water-fat separation imaging technique based on assessment of chemical shift differences. It produces water-only and fat-only images by calculating the phase shift caused by the resonance frequency difference between fat and water. While this method is widely used in abdominal and limb imaging, its application in soft-tissue imaging of the neck remains relatively rare. Diffusion-weighted imaging (DWI) is a functional MRI technique with broad applications. It uses the diffusion of water molecules within tissues to perform quantitative lesion analysis, with its primary quantitative metric being the apparent diffusion coefficient (ADC). ADC can be quantitatively calculated by performing the sequence under different intensities of diffusion gradient magnetic field, largely influenced by the presence of diffusion barriers within the water microenvironment [6–8]. Some researchers have reported promising results. For instance, Zhu et al [9] studied the application of DWI in differentiating between benign and malignant thyroid nodules, while Liu et al [10] explored the application of DWI in differentiating benign from malignant thyroid nodules, and demonstrated that quantitative T1-mapping using 3.0 T MRI is effective in diagnosing autoimmune thyroiditis. Such tools can detect subtle microscopic changes in vivo, providing information that can help to tailor treatment. In addition, nuclear medicine is useful in the diagnosis of thyroid disorders. The study by Fischer et al showed that combining visual with semi-quantitative MIBI uptake assessment significantly improved the classification accuracy of amiodarone-induced thyrotoxicosis, particularly in distinguishing the mixed type [11].

This study focused on evaluating the utility of multi-parametric MRI in assessing the progression of Hashimoto’s thyroiditis, with the aim of providing intuitive, sensitive, and accurate imaging tools for clinical evaluation of this disease.

Material and Methods

ETHICS STATEMENT:

All participants in this study gave their written, signed informed consent. The study protocol was approved by the Ethics Review Committee of the Second Affiliated Hospital of Zhejiang Chinese Medical University (approval number 202400601).

STUDY DESIGN AND PARTICIPANTS:

The study collected and analyzed the imaging and clinical data of 40 patients who underwent thyroid MRI examinations in the radiology department of the hospital where the author is affiliated from March 10, 2024 to September 15, 2024. Forty patients were equally randomized into intervention and control groups. Inclusion criteria were patients meeting the diagnostic criteria for Hashimoto’s thyroiditis, voluntarily undergoing thyroid MRI examination, and age 18–65 years. Exclusion criteria were patients with MRI images of poor quality due to artifacts, a history of acute or chronic infectious diseases, other autoimmune diseases, trauma or surgery of the thyroid gland, claustrophobia, metal implants in the neck, severe cardiovascular or cerebrovascular diseases, tumors, significant organ insufficiency, or individuals who were pregnant or lactating.

Referring to the consensus research report of the National Academies Press in Washington, D.C. [12], and the study by Osowiecka et al [13], we developed 3 dietary management programs for patients in the intervention group. The first program was a gluten-free diet, which involved avoiding gluten-containing grains and supplementing with 100 μg of vitamin D per day. The second program was a low-calorie diet with a nutrient composition of 25% protein, 30% fat, and 45% carbohydrates, designed to meet daily micronutrient and macronutrient needs. Depending on the individual’s resting metabolic rate and energy expenditure, this diet provided 1400–1600 kcal per day, resulting in a calorie deficit of approximately 1000 kcal per day, and included supplementation with 200 μg of 1-selenomethionine and 30 μg of zinc gluconate. The third program was a low-carbohydrate diet with 50–60% protein, 25–30% fat, and 12–15% carbohydrates, emphasizing increased consumption of leafy green vegetables and lean meats while limiting eggs, dairy products, legumes, fruits, bread, pasta, thyroid hormone-containing foods, and rice. Participants in the intervention group were allowed to choose and adhere to one of these dietary programs, and all were evaluated at a follow-up visit after 6 months. Eventually, 5 patients adhered to a gluten-free regimen, 8 were maintained on calorie-restricted protocols, and 7 followed carbohydrate-controlled nutritional plans. In contrast, patients in the control group did not undergo dietary management.

The diagnosis of Hashimoto’s thyroiditis is established based on the following criteria: (1) markedly elevated serum levels of thyroid peroxidase antibody (TPOAb) and/or thyroglobulin antibody (TGAb); (2) characteristic thyroid ultrasound findings demonstrating diffuse heterogeneous changes with thyroid enlargement, heterogeneous echogenicity, and frequently accompanied by multiple hypoechoic areas or thyroid nodules; (3) thyroid dysfunction may or may not be present (not required as an essential diagnostic criterion).

The diagnostic criteria for abnormal thyroid antibodies were defined as TPOAb levels greater than 16.83 IU/mL and TGAb levels greater than 12.33 IU/mL, both of which are 3 times higher than the upper limit of the reference range. All participants underwent testing for thyroid function indicators free triiodothyronine (FT3), free thyroxine (FT4), thyroid-stimulating hormone (TSH), and thyroid autoantibodies TPOAb and TGAb prior to the examination. The reference ranges for thyroid function indices are: FT3: 2.76–6.45 pmol/L; FT4: 11.20–23.81 pmol/L; and TSH: 0.35–5.10 mIU/L. Body mass index (BMI) was calculated using the following formula:

MRI MACHINE SPECIFICATIONS:

This study employed multi-parametric magnetic resonance imaging (MRI) sequences, including T1WI, T2WI, Dixon, and DWI, with post-processing of Dixon and DWI generating in-phase, out-of-phase, water-only, fat-only, and ADC maps.

MRI scans were performed using the SIEMENS MAGNETOM Avanto 1.5 T MRI system. Patients were positioned supine with the scanning center at the 5^th cervical vertebra. The transverse scanning parameters were: (1) T1-weighted imaging (T1WI): TR/TE: 524 ms/11 ms; field of view (FOV): 220 mm; slice thickness: 4 mm; inter-slice gap: 1.2 mm; number of slices: 20; number of excitations: 2. (2) T2-weighted imaging (T2WI) Dixon: TR/TE: 2410 ms/75 ms; FOV: 220 mm; slice thickness: 4 mm; inter-slice gap: 1.2 mm; number of slices: 20; number of excitations: 2. After conventional MRI scanning, DWI was obtained in axial plane by single-shot spin-echo echo-planar imaging on b values of 50 and 800 (DWI: TR/TE: 3370 ms/53 ms; FOV: 260 mm; slice thickness: 4 mm; inter-slice gap: 1.2 mm; number of slices: 20; number of excitations: 2).

MRI ANALYSIS:

MRI parameter measurements were conducted by 2 experienced radiologists on 2 separate occasions, with at least a 2-week interval between each measurement. The radiologists were blinded to the study design and the clinical information of the subjects. The transverse diameter, anteroposterior diameter, and isthmus thickness of the largest cross-sections of the right and left thyroid lobes were measured on T1WI and T2WI serial images. The sizes of the right and left lobes were expressed as the product of the transverse and anteroposterior diameters in mm², and the isthmus thickness was recorded in mm. The regions of interest (ROIs) were delineated with irregular lines on the axial T1WI, T2WI, water-only, fat-only, and ADC images to measure signal intensity. To ensure that the ROIs were kept approximately 2 mm smaller than the thyroid gland edge, signal intensity of the sternocleidomastoid muscle at the same level was measured using a circular ROI. Since the signal intensities of T1WI and T2WI cannot be directly compared between different time points for the same examiner or between different examiners, the ratio of the signal intensities of the thyroid gland to the sternocleidomastoid at the corresponding level was used as an indicator of signal intensity, T1-weighted imaging relative signal intensity (T1WI RSI) and T2-weighted imaging relative signal intensity (T2WI RSI). This method effectively minimizes signal intensity differences caused by scanning conditions and individual variations, thus improving the consistency and reliability of the imaging data. Thyroid water-only and fat-only images were obtained by processing the Dixon sequence using a Siemens post-processing system. At the same anatomical level, the water fraction and fat fraction of the thyroid were measured by outlining ROIs and expressed as percentages. Signal intensity ratio calculation formulas:

STATISTICAL ANALYSIS:

Statistical calculations were performed using SPSS 26.0 (SPSS, Inc., Chicago, Illinois, USA) on a PC. The Shapiro-Wilk test was used to assess the normality of measurement data. Data were expressed as mean±standard deviation. An independent samples t test or Mann-Whitney U test was used to analyze differences between the 2 groups. Paired t tests or paired rank-sum tests were used to analyze differences within the same group before and after the intervention. The differences between multiple groups were compared using analysis of variance (ANOVA) or Kruskal-Wallis H test. A P value of less than 0.05 was considered statistically significant.

Some laboratory test parameters exceeded the measurable range, resulting in over-limit values. Given the skewed data distribution, we opted for non-parametric rank-sum tests for baseline and intra-individual comparisons. Unlike tests based on means and variances, rank-sum tests are more resistant to outliers as they use data ranks instead of raw values. We assigned these over-limit results a value based on the detection-provided thresholds, which did not affect the sample ranks, thus minimally impacting our statistical analysis.

Results

BASELINE DATA COMPARISON:

Baseline data comparison between the intervention and control groups revealed no statistically significant differences in age, sex, BMI, laboratory parameters, or imaging-related indices (Table 1). Before the intervention, we conducted a between-group comparison among patients on gluten-free (n=5), low-calorie (n=8), and low-carbohydrate (n=7) diets within the intervention group. The results demonstrated no statistically significant differences among the 3 groups in age, BMI, laboratory parameters, or imaging indices (Table 2). FT3, FT4, and TSH levels were all within their respective normal ranges.

The FT3, FT4, and TSH levels of all patients were within the normal range. Therefore, it can be assumed that the thyroid function of the study population was essentially normal or within the compensatory range.

LABORATORY AND MRI PARAMETERS COMPARISON IN THE CONTROL GROUP:

The control group received no intervention. When comparing baseline and 6-month follow-up laboratory and imaging parameters in the control group, no statistically significant differences were found except for thyroid isthmus thickness, which was increased (P=0.019) (Table 3).

LABORATORY AND MRI PARAMETER COMPARISON IN THE INTERVENTION GROUP:

In the intervention group, thyroid autoantibody levels significantly decreased (TPOAb, P=0.016; TGAb, P=0.024), and FT3 levels also slightly decreased (P=0.077), remaining within normal ranges. Additionally, the T2-weighted imaging relative signal intensity (T2WI RSI) and water fraction of the thyroid gland significantly decreased (T2WI RSI mean, P<0.001; water fraction mean, P<0.001), while fat fraction increased (fat fraction mean, P=0.030). However, there were no statistically significant changes in thyroid size (including right and left lobes and isthmus) or T1-weighted imaging relative signal intensity (T1WI RSI) (Table 4, Figures 1–3). Subsequently, we analyzed the change values (Δ) between pre- and post-intervention measurements across different dietary management regimens to evaluate treatment efficacy. The results showed no significant differences in the changes (Δ) of these indicators (age, BMI, thyroid function parameters, thyroid autoantibodies, or MRI parameters) among the dietary regimens (P>0.05) (Table 5).

Discussion

STUDY LIMITATIONS AND FUTURE PROSPECTS:

First, this study was a single-center study with a relatively small sample size, which limits the generalizability of these results. Large-scale, multi-center studies are needed to confirm these results. Second, thyroid pathology was not performed in this study due to ethical constraints, and puncture biopsy only reflects localized changes in the sampled area. In contrast, MRI allows a comprehensive, noninvasive assessment of the entire thyroid gland and enables precise MRI-guided puncture and case examination when high-risk or difficult-to-interpret changes are detected. Third, while the sternocleidomastoid muscle was used to standardize signal intensity in this study, measurement errors were unavoidable due to the complex magnetic environment of the neck. Future advancements in thyroid-specific detection coils and homogenization techniques could enhance multi-parametric quantitative MRI (eg, T1- and T2-mapping), enabling more precise assessments and deeper insights into Hashimoto’s thyroiditis mechanisms.

Conclusions

Our findings indicate that dietary management effectively alleviates Hashimoto’s thyroiditis. Additionally, combining multi-parametric MRI sequences offers a valuable quantitative approach for assessing the condition. This approach bridges the gap in traditional testing methods, offering an intuitive, sensitive, and accurate assessment tool that enhances understanding and management of disease progression.