Reliability of Ultrasound Features in Predicting Thyroid Malignancy

Area of Science:

• Radiology and Imaging
• Oncology
• Endocrinology

Background:

• Increasing incidence of thyroid nodules necessitates accurate diagnostic tools.
• Ultrasound (US) is a preferred initial investigation for thyroid nodules due to safety and accessibility.
• Fine needle aspiration biopsy (FNAB) is invasive and costly, with associated risks.

Purpose of the Study:

• To evaluate the diagnostic accuracy of a predefined five-feature ultrasound composite for identifying malignant thyroid nodules.
• To compare the sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and AUC of the US composite against histopathology.
• To identify optimal combinations of US features for improved diagnostic performance using ROC analysis.

Main Methods:

• A cohort of 313 patients with thyroid nodules was screened; 146 underwent surgery.
• Preoperative ultrasound (US) and fine needle aspiration cytology (FNAC) were performed.
• Diagnostic accuracy was assessed by comparing US features (hypoechogenicity, irregular margin, microcalcification, increased vascularity, suspicious lymph node) and FNAC results with histopathology of surgically excised nodules. A rule of ≥2 positive US features was used.

Main Results:

• The five-feature US composite rule classified 112 patients as positive, with 84 confirmed malignancies.
• The rule demonstrated a sensitivity of 95.45%, specificity of 51%, and PPV of 75%.
• ROC analysis indicated that the combination of irregular margin and microcalcification yielded a balanced sensitivity (74.4%) and specificity (73.5%) for malignancy.

Conclusions:

• The ≥2-feature ultrasound composite rule exhibits high sensitivity for screening malignant thyroid nodules.
• The combination of irregular margins and microcalcifications offers a more balanced diagnostic trade-off, suitable for risk stratification.
• Multidisciplinary assessment integrating US, cytology, and clinical factors is crucial for definitive management decisions.