There was no significant difference in sex distribution between PTMCs and BMNs in both HT group (P = 0.376) and non-HT group (P = 0.239). The incidences of females with PTMC and BMN were also significantly higher in HT group than in non-HT group (both P < 0.001)HT is the most common autoimmune thyroid disease and the major cause of hypothyroidism. The incidence rate of women is 4–10 times higher than that in men [15, 16]. The disease mainly occurs in patients aged 30–60 years [11, 15–18]. TPO-Ab is positive in 90% – 95% of patients and TG-Ab is positive in 60% – 80% of patients [11, 15, 16]. In the present study, the incidence of females was significantly higher than that of males in both groups, and the incidence of females was also significantly higher in HT group than in non-HT group, which were higher than that reported by the previous studies[15, 16, 19]. The age of the patients with PTMCs was lower than the age of patients with BMNs in both groups, which could be associated with the fact that patients with malignant nodules tended to receive surgical treatment earlier. Besides, our study indicated that the TG-Ab levels in BMNs were lower than those of PTMCs in the non-HT group. We speculated that such differences could be associated with the selection of the samples. Consistent with our previous findings [9], the average size of the PTMCs in both groups was significantly lower than that of the BMNs, which might be related to the fact that the relatively small cystic BMNs were excluded from this study for not meeting the inclusion criteria.
According to the traditional 4–5 grade method of ultrasonic echogenicity intensity, most researchers believed that the echogenicity intensity was of almost equal importance in differentiating benign and malignant thyroid nodules regardless of whether HT is present or not [16, 20–23], of which the sensitivity and specificity of hypoechogenicity for diagnosing malignant nodules were 62% – 87.2% and 42.9% – 58.5%, respectively [13, 14, 24]. The present study demonstrated that UGSR was statistically significant in differentiating PTMCs and BMNs in both HT and non-HT groups, and the AUC was highly consistent (0.890 and 0.901, respectively). The optimal UGSR threshold was 0.687 and 0.727 and the corresponding sensitivity and specificity were 77.46% and 91.23%, 82.05% and 90.67%, respectively, suggesting that the diagnostic efficacy of the UGSR was substantially higher than that of the traditional 4–5 grade method [2, 13, 14]. The UGSRs of the PTMCs in the HT group were lower than in the non-HT group, while the UGSRs of the BMNs were not significantly different between the two groups.
In 2015, Grani et al. [5] discriminated the thyroid nodules in patients receiving FNAC by the UGSR of nodule to surrounding thyroid tissue and the UGSR of nodule to neck strap muscle, demonstrating that the former had higher diagnostic power than the latter. However, in their study, the sample size of malignant nodules was small and the nodules were not classified by pathological subtype and nodule size. In addition, all the nodules were confirmed by FNAC rather than histopathology. In 2018, we performed a controlled study on the UGSR of PTMCs and micronodular goiters in a single medical center [7], and further conducted a two center study in 2021 [9]. The findings of both studies demonstrated that UGSR had important value in differentiating PTMCs and micronodular goiters, and the AUC, optimal UGSR threshold, sensitivity, and specificity were highly consistent (0.895 vs. 0.918, 0.691 vs. 0.721, 86.8% vs. 88.1% and 80.4% vs. 83.3%, respectively). In 2019, Chen et al. [6] classified the papillary carcinomas and nodular goiters based on the size of the tumors, the AUC, optimal UGSR threshold, sensitivity and specificity for diagnosing smaller-sized PTCs were 0.919, 0.692, 97.5% and 72.4%, respectively. The diagnostic performance was in agreement with our previous findings[7, 9]. Although our previous studies and the study by Chen et al. improved the study by Grani et al. [5], these studies were conducted in patients with normal thyroid glands. In the present study, the AUC in the HT group, as well as the AUC and optimal UGSR threshold in the non-HT group were highly consistent with our previous studies, while the optimal UGSR threshold in the HT group was lower than that of the non-HT group and our previous findings [7, 9]. We speculated that the following two factors led to the lower UGSR in the HT group: 1) infiltration of lymphocytes resulted in decreased echogenicity intensity of PTMCs, notably in PTMCs interspersing and growing in HT background and in those with prominent lymphocyte infiltration surrounding the tumor; 2) when measuring the gray scale of thyroid gland with HT, the ROI contained the decreased echogenicity regions caused by abundant lymphocyte infiltration, as well as the increased echogenicity regions due to fibrosis formation. Finally, the average echogenicity intensity of the whole ROI was close to that of the normal thyroid gland. Although the UGSR of the HT group was lower than that of the non-HT group, the UGSR exhibited very high specificity and relatively high sensitivity in differentiating PTMCs and BMNs in the two groups, suggesting that UGSR was also suitable for patients with HT and was of great significance for improving the diagnostic efficiency of malignant nodules in these patients.
There were several limitations in this study. First, for nodules with heterogeneous echo, or thyroid tissue with heterogeneous echo due to HT, the selection and measurement of ROI had uncertainties. In the study, the selection and measurement of the ROI was performed twice by a senior imaging specialist with 19 years’ working experience, which would greatly reduce this deviation. Second, nodular goiters, nodular HT and adenomas were included as BMNs in the present study. Although we did not include all the types of BMNs, these three types could represent the majority of BMNs [25]. Our future studies will further investigate whether UGSR is also suitable for other types of BMNs. Third, the judgment of the nature of thyroid nodules requires the combination of multiple ultrasound features, such as gray level, morphology, taller-than-wide shape, and microcalcification, while the aim of this study was to investigate the value of UGSR for the differential diagnosis of PTMCs and BMNs in patients with HT. Thus, combination of UGSR with other ultrasound signs for identifying benign and malignant thyroid nodules needs to be studied further. Finally, the present study was a single-center retrospective study, while prospective studies with multiple medical centers are required to provide additional objective evidence on the values and robustness of UGSR in the two groups.
In conclusion, UGSR has important values in the differential diagnosis of PTMCs and BMNs and the optimal UGSR threshold is lower in the HT group than in the non-HT group. Sufficient understanding of these characteristics can improve the accuracy of preoperative diagnosis and provide important evidence for clinicians in selecting individualized treatment strategies.