In our research, we evaluated the efficacy of DW-MRI in differentiating between benign and malignant thyroid nodules. The results demonstrated that the ADC values at both b 500 and b 800 were significantly lower in malignant nodules compared to benign ones. Interestingly, no significant difference was observed in ADC measurements between these two b values in the normal thyroid parenchyma.

Except for one study [12], all other research has reported lower ADC values for malignant nodules when compared to benign ones, a finding consistent with studies involving various tissues in the literature [13,14,15,16,17,18]. Generally, malignant tumors exhibit hypercellularity, which leads to a reduction in the extracellular matrix and, consequently, a decrease in ADC values [6,7,8]. In the context of the thyroid gland, this characteristic correlates with the presence of large, oval, irregularly nucleated malignant thyroid nodules, accompanied by multiple micronucleoli, a thin chromatin structure, and intranuclear pseudoinclusions [13, 14]. Similarly, our study identified lower ADC values for malignant nodules in comparison to benign ones. Razek et al. observed a broad range of ADC values for benign nodules, attributing this variability to diverse components within the nodules, such as colloid, microcystic necrosis, hemorrhage, fibrous tissue, and calcification [13]. In our research, by aiming to minimize the measurement of cystic components within nodules, we excluded cystic and hemorrhagic nodules from our analysis, thereby achieving a narrower range of ADC values.

In the existing literature, ADC measurements have been conducted using a variety of b values [12,13,14,15, 17, 18]. Schuller-Weiderkamm et al. utilized a b value of 800, whereas Razek et al. employed b values of 250 and 500; Bozgeyik et al. conducted measurements with b values of 100, 200, and 300; Erdem et al. utilized a b value of 1000; Aghaghazvini et al. implemented b values of 50, 500, and 1000; and Zhu et al. explored b values ranging from 0 to 1000 and from 0 to 2000 [12,13,14,15, 17, 18]. The selection of the b value is a critical factor influencing both image quality and the accuracy of ADC measurements. It has been observed that ADC values obtained at lower b values may be affected by the microcapillary perfusion of water molecules in the blood, leading to artificially higher ADC readings [14]. Conversely, higher b values mitigate the influence of perfusion on ADC measurements but can diminish the Signal-to-Noise Ratio (SNR), potentially introducing errors into the measurements [8,9,10]. In previous studies, Aghaghazvini et al. stated in their comparison that the b 1000 value had higher sensitivity and specificity, while Bozgeyik et al. stated that the b 300 value had higher sensitivity and specificity [11,12,13,14,15]. In a study conducted using high b values [14], similar diagnostic performance was found between b 1000 and b 2000 values. Consequently, the diverse range of b values employed across studies has precluded the establishment of an optimal cutoff ADC value. A comparative analysis of the sensitivity and specificity values for ADC500 and ADC800 reveals that ADC500 exhibits superior performance metrics. Specifically, ADC500 demonstrates a higher sensitivity (83.3% compared to 71.4%) and a slightly elevated specificity (90.0% compared to 89.7%) at the identical cutoff point of 1.1 × 10− 3.

The enhanced sensitivity and specificity values of ADC500 indicate its greater utility in diagnostic applications. Consequently, the ROC analysis supports the robustness of both ADC500 and ADC800 as diagnostic tools. However, ADC500, with its higher sensitivity and specificity values, seems to offer a more reliable diagnostic performance compared to ADC800. Therefore, ADC500 is be considered the more useful marker for diagnostic applications.

In recent years, studies employing magnetic field strengths and devices with varied b values have consistently reported lower ADC values in malignant nodules [17, 18]. However, the consistency of these findings has been called into question due to variations in magnetic field strengths, b values, and DW-MRI parameters. As a result, establishing a universally applicable cutoff ADC value presents significant challenges [10, 11].

In our evaluation, two nodules demonstrated discrepancies between their histopathological results and DW-MRI values. For instance, a benign nodule exhibited ADC values akin to those typically seen in malignant nodules, recording b values of 500 and 800 at 0.94 × 10− 3 mm2/s and 0.93 × 10− 3 mm2/s, respectively. Upon pathological examination, extensive fibrotic and calcified areas were identified within the nodule. These findings suggest that the reduced extracellular fluid and the consequent restriction in diffusion could have influenced the ADC measurements. Additionally, the presence of calcifications may lead to inaccuracies in ADC value calculations. Hence, a pre-MRI evaluation of nodules for calcification, fibrosis, and other internal characteristics could mitigate the risk of misinterpretation.

In a separate case involving a malignant nodule, the ADC values were notably higher, with measurements of 1.9 × 10− 3 mm2/s for b 500 and 1.26 × 10− 3 mm2/s for b 800. Pathological analysis confirmed the diagnosis of a macrofollicular variant of papillary carcinoma. This variant is characterized by an increased colloid volume within the follicles, which elevates the extracellular water content, thus enhancing water mobility and diffusion capacity, and resulting in higher ADC values. This observation is consistent with the interpretations of Weidekamm et al. regarding malignant nodules [12]. However, the rarity of this variant in our study, with only a single patient represented, precluded statistical analysis. Consequently, further research involving a larger cohort of patients is essential to deepen our understanding of the variability in ADC values among malignant nodules.

In our study, measurements obtained from normal-looking parenchymal areas devoid of nodules did not show any significant differences across the two b values. This finding is in line with previous research, which has also reported no discernible differences in ADC values between the parenchyma of healthy individuals and those with nodules, suggesting that the pathology does not impact the normal parenchyma.

Our research methodology involved comparing ADC values against patients’ postoperative histopathological findings, providing a direct correlation with the definitive diagnostic outcome [14]. Contrastingly, some studies have opted to compare ADC values with the results from FNAB [12, 14, 15]. However, the reliability of FNAB can sometimes be compromised due to reports of “inadequate” sample collection. Furthermore, FNAB alone is insufficient for the evaluation of follicular neoplasms, as it cannot definitively distinguish between follicular adenoma and follicular carcinoma. Accurate diagnosis in such cases necessitates a thorough examination of the entire nodule, with specific attention to vascular and capsular invasion, to determine the nature of the neoplasm. This highlights the limitations of FNAB and underscores the importance of comprehensive histopathological examination for accurate diagnosis [5,6,7,8,9,10,11,12,13,14,15,16,17,18,19].

Our study acknowledges several limitations. Primarily, the resolution offered by MRI technology falls short when it comes to nodules smaller than one centimeter in size. Due to this limitation, such nodules cannot be distinctly differentiated from the surrounding parenchyma, rendering ADC measurement unfeasible. Furthermore, despite efforts to exclude cystic portions from the ROI, the potential inclusion of cystic areas within small nodules might have inadvertently resulted in elevated ADC values.

Another significant constraint is the limited number and diversity of malignant nodules examined in our study. This limitation restricts the generalizability and applicability of our findings across the broader spectrum of thyroid nodules with varying etiologies.

To address these challenges and enhance the robustness of future research, studies designed with suitable parameters that encompass a more extensive assortment of nodules, inclusive of diverse thyroid nodule etiologies, are imperative. Such studies will not only contribute to overcoming the limitations identified but also aid in refining the diagnostic accuracy and reliability of ADC measurements for thyroid nodules.