Predictive values of ultrasound characters associated with malignant thyroid nodules in Yaoundé: a cross-sectional study

Introduction most ultrasound criteria are defined in developed countries and commonly used in practice to assess the malignancy risk of thyroid nodules. This practice does not take into consideration some aspects of our context as delay of consultation and insufficient iodine intake. The objective of this study was to determine the predictive values of ultrasound characters associated with malignant thyroid nodules in our environment. Methods we conducted a cross-sectional, prospective, and analytical study in three hospitals in Yaoundé over a six-month period in 2022. Our sample consisted of thyroid nodules with ultrasound, cytopathological, and histopathological data. The ultrasound characters and histology status of category III thyroid nodules and higher in Bethesda score were analysed in univariate and multivariate statistics to determine their predictive values. Results eighty-nine nodules were obtained according to our inclusion criteria. The sex ratio was 0.46 and the average age of the patients was 46 years (IQR=42-59). The cancer prevalence in our sample was 22.47%. On ultrasound assessment, the characters associated to malignant histology (p<0.05) were nodules count, echogenicity, echostructure, presence or absence of microcalcifications, margins, and type of vascularization. Positive predictive values ranged from 26.15 to 57.14%, while negative predictive values ranged from 12.5 to 33.3%. Conclusion taken alone, the ultrasound characters of suspected thyroid nodules have poor predictive values. There was a high variability in sensitivity but that was generally good (60-95%) while specificity was low. The prediction of malignant thyroid nodules is correlated with the association of at least two ultrasound criteria supported by clinical arguments.


Introduction
The thyroid nodule is any localized hypertrophy of the thyroid gland [1].In countries with sufficient iodine intakes, the clinical prevalence is 5.3 and 6.4% for women, 0.8 and 1.6% for men.The prevalence is about 3 times higher in women and increases with age [1,2].The ultrasound prevalence is about 10 times higher, roughly equal to that of the decade of the subjects examined [1,2].In Cameroon, the ultrasound prevalence of thyroid nodules is 28.3% according to Moifo et al. [3].On ultrasound, several characters are used to describe the thyroid nodule: size, shape, margins, echogenicity, echostructure, type of vascularization, and capsule break-in.Other associated lesions such as the presence or absence of microcalcifications in the thyroid parenchyma may help to histological diagnosis [1].
The Thyroid Imaging Reporting and Data System (TIRADS) classifies thyroid nodules according to their malignancy risk using ultrasound criteria [3,4].However, several TIRADS classifications exist according to world regions [4][5][6][7][8][9].It is therefore apparent that a highly reliable, reproducible, and clinically practical TIRADS classification will greatly improve communication between clinicians and radiologists.This will even be more helpful in settings where Fine Needle Aspiration Biopsy (FNAB) is not readily available and so decisions will therefore be based to a great extent on the ultrasound features of the lesions and TIRADS classification as this implies the potential risk for malignancy [3].In the absence of studies reporting predictive values of nodulerelated ultrasound characters in practice in sub-Saharan Africa, radiologists and clinicians use the different TIRADS classifications (Eu-TIRADS and ACR-TIRADS) to characterize nodules on ultrasound.The aim of our study was to determine ultrasound characters associated with malignant thyroid nodules and their predictive values in Yaoundé.

Methods
Study design and setting: we conducted a crosssectional, prospective, and analytical study during a six-months period (April to September 2022) in three university hospitals in Yaoundé.
Participants: our sampling was not exhaustive.The source population consisted of patients who underwent thyroid surgery and the target population was patients with thyroid nodules on ultrasound and whose cytological examination of the echoguided FNAB showed a category III or higher according to the Bethesda´s score, edition 2017 [10].We excluded patients whose ultrasound did not provide all the descriptive features of thyroid nodules and those whose histology was not recorded in the file.
Variables: variables in ultrasound included nodules count, size, shape, margins, echogenicity, echostructure, type of vascularization, capsular break-in or not, presence of microcalcifications or not [1].In histology, histopathological status was our main variable of interest.
Data sources: data were collected through a structured questionnaire and reported in a database.Nodules count, size, shape, margins, echogenicity, echostructure, type of vascularization, capsular break-in or not, and presence of microcalcifications or not were assessed in ultrasound.The histological status of nodules was given by anatomopathologist.
Bias: to exclude any potential bias, we asked to ultrasonographist to describe nodules in the same way.
Study size: according to an ultrasound prevalence of thyroid nodules at 28.3% [3], an accuracy of 1.5%, and a type I error of 5%, we obtained a sample size of 312 nodules by using OpenEpi software.
Quantitative variables: quantitative data were analyzed by grouping them by interval.Groups were done according to the literature for better comparison.The age was split in 10-year intervals and the nodule size in <7 mm, 7-20 mm, and >20 mm.
Statistical methods: collected data were analysed using the r-studio software for Windows version 4.3.1.Qualitative data were represented by numbers and frequencies.Continuous variables were expressed by their mean (standard deviation) when normally distributed; otherwise, by their median (interquartile interval (IQR)).The ultrasound characters of the nodules were opposed to malignant nodules to determine the association link.The Chi-square test or Fisher´s exact probability was used for the proportion comparison.Continuous variables were compared using the t-student test or its nonparametric equivalent.Logistic regression was used to investigate independent factors associated with malignancy.The explanatory variables associated with malignancy in univariate analysis with p<0.05 were introduced in the same logistic regression model to look for independent factors associated with malignant nodules.The significance level was p<0.05.Predictive values, sensitivity, and specificity were calculated for each ultrasound character frequently associated with malignant nodules and reported in the literature.These were the oval shape, size >7 mm, hypoechogenicity, solid echostructure, irregular margins, and mixed vascularization.
Ethical considerations: permissions to carry out the study was obtained from the Comité Institutionnel d´Ethique et de Recherche of the Faculty of Medicine and Biomedical Sciences at the University of Yaoundé I and from General Managers of Hospitals.Written informed consent was obtained from the study participants before their surgery.This consent includes the possibility to use their informations for scientific purposes.

Results
Participants: the surgery procedure was performed on 112 patients for category III thyroid nodules and higher in Bethesda´s score.We excluded 16 for incomplete data on thyroid ultrasound and 7 for lack of post-operative histology.Thus, our study was based on a sample of 89 thyroid nodules.
Descriptive data: women made up 69% (n=61) of our sample with a sex ratio of 0.46.The median age was 46 years (IQR=42-59) with extremes of 16 and 72 years (Table 1).
Main results: predictive values of ultrasound factors associated with thyroid nodule malignancy are given in Table 4. Positive predictive values on ultrasound ranged from 26.15 to 57.14%, while negative predictive values ranged from 12.5 to 33.3%.
Sensitivity and specificity of ultrasound characters of thyroid nodules: sensitivity ranged from 60 to 95% and specificity 5.88 to 82.85 (Table 4).

Discussion
Our study revealed that the number of nodules, echogenicity, echostructure, mixed vascularization, and microcalcifications were significantly associated with thyroid nodules malignancy (p<0.05).Positive predictive values on ultrasound ranged from 26.15 to 57.14%, while negative predictive values ranged from 12.5 to 33.3%.The main limitation of our study was the sample size; limited to 89 thyroid nodules.
Sex and age of patients: the nodular pathology of the thyroid gland predominates in female subjects with a sex ratio of 1 man to 3 to 4 women according to the series.Its incidence increases with age [2,11].Our findings were similar to the literature with subjects mostly female (1 man for 2.18 women) beyond 40 years (57.5%).
Nodule histology: the risk of malignancy in front of a thyroid nodule increases with the Bethesda category.It ranges from 6-18% for category III to 94-96% for category VI.This risk is even higher in the case of non-invasive follicular thyroid neoplasia with papillary nuclear characters identical to those of carcinoma [10].The risk of malignancy in the different Bethesda categories studied (category III to VI) was not determined in our study.On average, 5% of thyroid nodules are malignant, with rates ranging from 1.5% to 38.1% depending on patient selection [12,13].We report a prevalence of 22.47% based on nodules selected in category III and more than Bethesda and therefore potentially malignant.
Ultrasound characters associated with malignant thyroid nodules and association strength: in the literature, several authors report among the ultrasound characters associated with malignant thyroid nodules those that can be considered major.Thus, Hong et al. in 2010 and Moifo et al. in 2013 identified irregular margins, taller-than-wide shape, presence of microcalcifications, and marked hypoechogenicity as major ultrasound characters associated with malignant nodules [3,14,15].Other ultrasound characters such as lymphadenopathy and local invasion of adjacent structures are associated with nodule malignancy but are rarely found [16].In our study, the US characters associated with the malignant nodule were number of nodules (p=0.043),echogenicity (p=0.008),echostructure (p=0.039),microcalcifications (p<0.001),margins (p<0.001), and type of vascularization (p=0.012).Capsular break-in and nodule shape are recognized as characters associated with malignant histological forms [16] but in our study, this was not the case; certainly because of the small size of our sample.After analysis of the operative parts and comparison with the ultrasound data, the number of nodules (OR=2.83,CI95=1.01-8.05),echogenicity (OR=2.23;CI95=1.14-5.33),echostructure (OR=3.59;CI95=1.44-18.44),mixed vascularization (OR=0.40;CI95=0.19-0.73)and microcalcifications (OR=7.12;CI95=2.45-22.09)had a significant association (p<0.05) with nodules in our series.
Predictive values of ultrasound characters associated with malignant thyroid nodules: the predictive values of the different ultrasound characters were generally poor in our study and very variable according to other series in the literature (Table 5).
Each of these variables alone gives the nodule an increased risk of malignancy by a factor of 1.5 to 3 [1].However, when these variables are associated, the risk increases significantly [17].The solid nodule associated with other characters is the most studied association in the literature [17,18].The size of the nodules was not predictive of malignancy in our series as in that of Besbes et al. in 2007 [13] but remains a prognostic element in case of malignancy if it is greater than 30 mm [1,19].Intranodular vascularization associated or not with peripheral Doppler vascularization signals neoplastic nodular pathology [13,20].Microcalcifications correspond to calcospherite lesions in papillary and medullary cancers but are not pathognomonic [21] despite good specificity; 82.5% in our series, 95.2% for Solbiati et al. [22], and 91% for Chatti et al. [23].

Conclusion
Taken alone, the ultrasound characters of suspected thyroid nodules have poor predictive values.There is a high variability in sensitivity but that is generally good (60-95%) while specificity was low.The prediction of malignant thyroid nodules is correlated with the association of at least 2 ultrasound criteria supported by clinical arguments.According to the literature, the association of solid-oval-hypoechogenic-nodule with irregular contours confers a strong suspicion of malignancy to the nodule.

Table 1 :
sex and age of patients

Table 2 :
ultrasound characters and bivariate analysis

Table 3 :
association strength of ultrasound characters with malignant nodule

Table 4 :
predictive values of ultrasound factors associated with thyroid nodule malignancy

Table 5 :
positive predictive values of ultrasound characters in some studies

Table 3 :
association strength of ultrasound characters with malignant nodules

Table 4 :
predictive values of ultrasound factors associated with thyroid nodule malignancy Sp: specificity; PPV: positive predictive value; NPV: negative predictive value

Table 5 :
positive predictive values of ultrasound characters in some studies Ultrasound