Alzahraa Mustafa Alawami, Abdulaziz Al-Sharydah, Mishael M AlObaid, Zahra AlShabeeb, Jomana Alnasser, Maryam Aldeen, Fatimah Abbas A Aldawood, Mohammed Ahmed Al Mulhim, Mohammed S AlAftan, Hussain Al-Arfaj, Ahmad M Al-Essa
Within this cohort, ultrasound demonstrated high diagnostic performance, with the highest estimates of agreement and accuracy; CT showed moderate performance; and MRI results were limited by the small sample size. Given the distinct modality‑specific subgroups, these findings should be interpreted as descriptive and hypothesis‑generating rather than as definitive evidence of modality superiority.
BACKGROUND: Accurate detection of metastatic cervical lymph nodes in thyroid cancer is needed for treatment planning. The present study evaluated the diagnostic performance of ultrasound (US), computed tomography (CT), and magnetic resonance imaging (MRI) for detecting metastatic cervical lymph nodes in a Saudi cohort.
METHODS: This retrospective cohort study reviewed patients with thyroid nodules and cervical lymphadenopathy treated at King Fahd Hospital of the University, Saudi Arabia, between 2015 and 2024. This retrospective cohort study reviewed 150 retrieved records, of which 111 patients underwent detailed assessment; 10 were excluded for missing pathology or incomplete diagnostic data, leaving 101 patients in the final analysis. Conventional imaging features were assessed using predefined modality-specific criteria. US-guided fine needle aspiration cytology was used as the reference standard. Sensitivity, specificity, positive predictive value, negative predictive value, accuracy, and 95% confidence intervals were calculated. Agreement and discordance were assessed using Cohen's kappa and McNemar's exact test, respectively.
RESULTS: A total of 101 patients were included. US was performed in 59 cases, CT in 34, and MRI in 8. FNA cytology was positive in 48/101 cases. Overall radiological classification showed sensitivity of 89.6%, specificity of 88.7%, and accuracy of 89.1%, with substantial agreement with FNA cytology (κ=0.782). In modality-stratified descriptive analyses, US showed the highest observed agreement with FNA cytology, CT showed moderate agreement, and MRI findings were limited by the limited sample size.
CONCLUSION: Within this cohort, ultrasound demonstrated high diagnostic performance, with the highest estimates of agreement and accuracy; CT showed moderate performance; and MRI results were limited by the small sample size. Given the distinct modality‑specific subgroups, these findings should be interpreted as descriptive and hypothesis‑generating rather than as definitive evidence of modality superiority.