A nodule is defined as any localized enlargement of the thyroid gland [1]. Its frequency is 4% to 7% for palpable nodules and 30% to 50% for nodules discovered by ultrasound [2].

The discovery of a thyroid nodule raises the question of whether it may be cancerous. The advent of fine needle aspiration cytology has greatly helped to resolve this issue. This test is now considered the most effective way of distinguishing between benign and malignant nodules [3][4]. In some medical facilities, however, fine needle aspiration is not available. Clinicians must then rely on a range of clinical and paraclinical arguments to make the best possible diagnosis if they want to avoid unnecessary and costly interventions.

The aim of this study was to demonstrate the role of cervical ultrasound in the investigation and characterization of thyroid nodules and its predictive value in determining malignancy or benignity by searching for a correlation between ultrasound findings and definitive pathological results.

The specific objectives were to:

• Describe the ultrasound features of thyroid nodules;

• Perform EU-TIRADS classification;

• Compare EU-TIRADS classification with histological data.

We conducted a prospective descriptive and analytical study over a ten-month period, from 1 January 2023 to 31 October 2023, in the medical imaging and otorhinolaryngology departments of the Louga Regional Hospital Center.

All patients monitored for a solitary thyroid nodule or within a multinodular goiter, who underwent cervical ultrasound and thyroidectomy (total or partial) during the study period with an available pathological examination result, were included.

The ultrasounds were performed using a Siemens Acuson NX3 Elite machine with a 12 MHz superficial probe for B-mode thyroid examination with color Doppler coupling. We described the ultrasound features of the thyroid nodules, then established the EU-TIRADS classification [5] and finally compared the EU-TIRADS class with the histological data. For patients with heteromultinodular goiter, we considered the nodule with the highest EU-TIRADS score.

The data was entered and sorted using Excel 2019 software and analyzed using SPSS 29, enabling us to correlate the ultrasound and histological characteristics of each nodule examined based on the data collected. We performed a chi-square test with a significance level of 5%.

The average age of patients in our study was 42.35 years, ranging from 18 to 70 years. The frequency of thyroid nodules increases with age. This has been clearly demonstrated by Mortensen et al.’s assessments [6]. In our study, the percentage of malignancy increased with age (p = 0.02) and was 66.7% in the over-60 age group. This was also the case in a Moroccan study by Abousaad [7]. In the latter study, the malignancy rate in patients over 60 years of age was 57.1%. Extreme ages have a higher risk of malignancy in the literature [8]-[10]. According to Belfior et al. [11], this risk is around 50% after 60 years of age.

In our study, there was a predominance of females with a sex ratio of 0.03. According to Mazzaferri et al. [12], females are more represented than males, with a ratio of 3 females to 1 male, thus demonstrating the role played by the presence of sex steroid receptors in follicular cells.

Multinodular thyroid goiter was for a long time considered “benign” compared to solitary nodules [13]-[15]. This notion has been revised by several authors who found that the risk of malignancy was identical for both groups [16]. In our series, this risk was almost identical for solitary nodules and multinodular thyroid goiter (18.8% vs. 16.7%).

The average size of nodules in our series was 26.2 mm in length. These results are comparable to those found in Diourbel (Senegal) [17], with an average size of 28 mm, and those of Abousaad [7], who found 24 mm. Some authors [18][19] find that a size greater than 4 cm is highly suspicious, while others disagree with this notion [20][21]. In Sfar et al.’s study [22], nodules larger than two centimeters were cancerous with a statistically significant relationship (p = 0.001) and, according to Zhao et al.’s study [23], smaller nodules are associated with an increased risk of cancer (with p < 0.01), which is consistent with our study, where the suspicion of malignancy was higher for nodules smaller than 30 mm (p = 0.03).

The orientation of the nodule is a predictive criterion for malignancy, and a non-oval nodule is one of the five strong signs of suspected malignancy [5]. In our study, the malignancy rate of non-oval nodules was 100% with a statistically significant relationship (p = 0.03). This is also consistent with the results of the Moroccan study by Azal [24], which found a significant correlation (p = 0.0001). The contours of the nodules are also important. Indistinct and/or irregular contours are arguments in favour of malignancy. In our study, irregular contours were found in 5.9% of nodules and were malignant in 100% of cases, with a statistically significant relationship (p = 0.002). This is the case in the studies by Abousaad [7] and Bouaity etal. [25], with respective malignancy rates of 80.6% and 75.8%. In several studies, such as that by Azizi et al. [26], irregular contours had a statistically significant relationship in predicting cancer (p = 0.001), as was the case for Sfar et al. [22] (p = 0.01) and Ríos et al. [27] (p = 0.0001).

The echogenicity of thyroid nodules is important to consider. Nodules can be anechoic, hyperechoic, isoechoic, or hypoechoic [28]. Hyperechoic nodules are rarely malignant (1.3% to 4%) [29]. Hypoechoic nodules are more than half the time cancerous. This is the most common form of thyroid cancer. With regard to the echogenicity of nodules, our series is consistent with those of Azal [24] and Abousaad [7].

Calcifications are frequently found. Their presence increases the risk of cancer by a factor of 2.5, which could reach 75% for solitary nodules [30]. In the series by El korbi et al. [31], malignant nodules contained microcalcifications with a statistically significant relationship with cancer prediction (p = 0.02), as in the studies by Hasnaoui etal. [32] (p < 0.01) and Abousaad [7] (p = 0.0001). This is consistent with the results of our series, where microcalcifications had a statistically significant relationship with cancer prediction (p = 0.006).

The EU-TIRADS score of the European Thyroid Association [5] is a 4-point score (from 2 to 5). Each score corresponds to a risk of cancer. The score is based on four highly suspicious signs: high hypoechogenicity, micro-lobulated or spiculated contours, disharmonious shape, and microcalcifications. A single sign is sufficient to classify the nodule as score 5. The results of our study are broadly consistent with this system, as are those of Azal [24] (Table 1).

Table 1. Percentage of malignancy according to the EU-TIRADS classification based on studies.

The main limitation of our study remains the small sample size (n = 34).

The comparison of ultrasound and histological data in our study has enabled us to objectively identify the following criteria for malignancy in thyroid nodules: size, non-oval shape, irregular contours, hypoechoic appearance, and the presence of microcalcifications. A pertinent ultrasound analysis remains highly useful in the pre-surgical identification of factors predictive of malignancy in a thyroid nodule, supplemented by ultrasound-guided fine needle aspiration, which remains, however, rarely available in routine practice.