Related Experiment Video
Updated: Apr 23, 2026
Radiation Planning Assistant - A Streamlined, Fully Automated Radiotherapy Treatment Planning System
Published on: April 11, 2018
Response Evaluation and Surveillance of Head and Neck Tumours Using the NI-RADS Algorithm
Sara Shahid1, Rashed Nazir1, Ahmed Moqeet1
1Department of Radiology, Shifa International Hospital, Islamabad, Pakistan.
Objective:
To evaluate the effectiveness of the American College of Radiology (ACR) Neck Imaging Reporting and Data System (NI-RADS) in the response evaluation of head and neck neoplasms following therapy, utilising contrast-enhanced CT or MRI.
Study Design:
A descriptive study. Place and Duration of the Study: Department of Radiology, Shifa International Hospital, Islamabad, Pakistan, from December 2023 until July 2024.
Methodology:
Post-treatment CT or MRI scans from 70 patients with head and neck cancer were reviewed. Each scan was assessed according to the NI-RADS lexicon and categorised accordingly. Diagnostic accuracy was determined by comparing NI-RADS findings with histopathological results and/or clinical follow-up at a three-month interval, and positron emission tomography (PET), where available. Cases were classified as true negatives (NI-RADS 1 and 2) or true positives (NI-RADS 3 and 4) when imaging findings were concordant with the reference standards, whereas discordant findings were classified as false negatives (NI-RADS 1 and 2) or false positives (NI-RADS 3 and 4). Sensitivity, specificity, and accuracy were calculated using SPSS.
Results:
Recurrence was confirmed in 25 patients either histopathologically or through clinical evaluation and PET, where available. The overall accuracy of NI-RADS categorisation was 82.86% after taking into account all concordant cases (i.e., true positives reported as NI-RADS 3, NI-RADS 4, and true negatives reported as NI-RADS 1 and NI-RADS 2) and discordant cases (i.e., false positives reported as NI-RADS 3, NI-RADS 4, and false negatives reported as NI-RADS 1 and NI-RADS 2) of the primary tumour site. The negative predictive value (NPV) was 92% for NI-RADS 1 and 76.92% for NI-RADS 2, while the positive predictive value (PPV) was 69.23% for NI-RADS 3 and 100% for NI-RADS 4. For neck lesions, the NI-RADS algorithm demonstrated a sensitivity of 80%, specificity of 100%, and an overall accuracy of 98.59%.
Conclusion:
The ACR NI-RADS system demonstrates high diagnostic accuracy and reliable risk stratification for post-treatment surveillance of head and neck cancer patients. Its structured reporting format facilitates a consistent evaluation and aids in guiding clinical decision-making.
Key Words:
Head, Neck, Neoplasms, Recurrence, Risk, NI-RADS lesion.
