This was a hospital-based descriptive cross-sectional study conducted at the ultrasound unit of the Department of Radiology and Department of Pathology of MNRH, Kampala, Uganda, between November 2020 and March 2021. The radiology department provides wide imaging services, including US, computed tomography, plain radiography, and interventional radiology. It has eleven radiographers, five consultant radiologists, and four nurses. The department of pathology is situated at the School of Biomedical Science and serves the roles of teaching and research, as well as offering diagnostic histopathology/cytopathology and autopsy services. Most the patients referred for US-guided FNA are from the endocrine-surgical outpatient clinic at MNRH. The clinic runs every Wednesday from 8:00 AM to 2:00 PM and receives approximately 15–17 patients weekly.

All consenting participants with thyroid nodules ≥ 1 cm on B-mode ultrasound scans were scheduled for US-guided FNA. All participants were at the age of 18 and above during the study period. Participants with prolonged bleeding time, extensively calcified nodules, emphysema, and clinical and laboratory features of thyrotoxicosis were excluded from the study.

The sample size was determined using Kish Leslie’s formula.

$$N=\frac^\alpha /2 p\left(1-p\right)}^}$$

where:

N = desired sample size.

Z = Z score corresponding to 1.96 for 95% confidence level.

p = the estimated proportion of people with suspicious nodular thyroid sonography findings; 50% was used to obtain the maximum sample size

d = margin of error at the 95% level of significance, which is 0.05

Taking the prevalence of suspicious nodular thyroid sonographic findings as 50%, we obtained an approximate sample size of 385 participants.

The finite population correction formula was used to adjust the sample size before data collection since the accessible population was 200 patients for the duration of the study. The 200 participants were estimated by multiplying the number of FNA Referrals received per week [10] by 20 weeks (anticipated duration of the study).

Using the finite correction formula, the final adjusted sample size was 132 participants.

$$S=\frac^\!\left/ \!_\right.}=s=385/\left(1+385/200\right)=132$$

where:

S is the adjusted sample size.

Population size was the expected number of participants within the 5-month study period.

All patients referred from the endocrine-surgical outpatient clinic with TNs in a B-mode US scan scheduled for US-guided FNA during the study period were screened for nodules using US. Those with TNs ≥ 1 cm in B-mode US and who consented were recruited into the study. Under the supervision of a radiologist, The US evaluation was performed on an SIUI machine, model Apogee 5300, manufactured January 2015 by Hamburg Germany. Ultrasound machine with high-frequency linear probes of 7.5 MHz for obese patients or large thyroid lesions, a 5 MHz transducer was used for greater penetration. US was performed with the patient in the supine position and the neck hyperextended, and the entire gland was examined. Hyperextension of the neck was obtained by placing a pillow under the shoulders. The neck was scanned in sagittal, transverse, and oblique sections to optimally visualize both lobes of the thyroid and isthmus. Color Doppler imaging was utilized. Imaging of the lower poles of the thyroid was obtained by making the patient swallow as this tends to raise the thyroid gland in the neck.

Thyroid nodules sonographic characteristics, such as composition, shape, echogenicity, margins and echogenic foci, were recorded, and points were assigned to each nodule for separate categories according to ACR TI-RADS guidelines [11]. The sum of the points in each category determined the TI-RADS level assigned to each nodule, with TR1 indicating 0 points; TR2 – 2 points; TR3 – 3 points; TR4 – 4–6 points; and TR5 – 7 or more points (Fig. 1). The final sonographic diagnosis was reached with the help of a consultant radiologist. The data obtained from the ACR point table were used to correlate with cytology results. A 23-gauge needle with a clear hub and clear syringe (5–10 ml) was used to obtain a sample from each nodule, and a maximum of two nodules were sampled per patient. The sample was gently expelled onto the surface of a labeled microscope slide from the needle tip. A smearing slide was then slid over the specimen, ensuring that both slides were smeared. One slide underwent wet fixation with alcohol, which was stained using the Papanicolaou method, while the other slide was air-dried at room temperature and stained with the Diff quick method. After this, an experienced pathologist evaluated all samples according to the TBSRTC [13]. The categories and their risk of malignancy were recorded as follows in (Table 1).

Nomenclature of categorization of thyroid nodule features per the five lexicon categories [22]

The ACR TI-RADS level, which ranges from TR1 (benign) to TR5 (high suspicion of malignancy), was also used to categorize the nodules. Coded US images were stored and printed.

The data were entered into EpiData version 3.1 and then exported into Stata statistical software version 14 for analysis. To describe patient characteristics, categorical variables were summarized using frequencies and percentages, while continuous variables used the mean and standard deviation. While 132 patients were enrolled in the study, 29 of them had 2 or more nodules but only two nodules were biopsied, resulting in an analytic sample size of 161. Pairwise analysis was not performed to control for clustering because the patient identifiers were replaced with study numbers for the study purpose. The ACR TI-RADS classification criteria were used to classify nodules and then presented as frequencies and percentages. The difference in the proportions of the ACR TI-RADS sonographic criteria was tested using Fischer’s exact test. Spearman’s correlation coefficient was used to establish the correlation between ACRTI-RADS and cytology findings.

To determine the diagnostic effectiveness of ACR TI-RADS in characterizing thyroid nodules and predicting cytological findings, sensitivity, specificity, positive and negative predictive values, and positive and negative likelihood ratios with corresponding 95% confidence levels were calculated using the Bethesda system of thyroid classification as a gold standard. ACR TI-RADS was dichotomized by considering TR4 and TR5 as a positive screen for malignancy and TR1 to 3 as screen negative. The Bethesda System was also dichotomized by classifying 4 to 6 as malignancy and 1–3 as no malignancy.