ARRS 2022 Abstracts


1433. MRI Detection of Suspected Nasopharyngeal Carcinoma: A Systematic Review and Meta-Analysis
Authors * Denotes Presenting Author
  1. Vineet Gorolay *; Royal Prince Alfred Hospital; University of Sydney
  2. Naomi Niles; Concord Repatriation General Hospital
  3. Ya Huo; Concord Repatriation General Hospital; University of Sydney
  4. Navid Ahmadi; Concord Repatriation General Hospital; University of New South Wales
  5. Kate Hanneman; Peter Munk Cardiac Centre; University Health Network, University of Toronto
  6. Elizabeth Thompson; Royal Prince Alfred Hospital
  7. Michael Chan; Concord Repatriation General Hospital; University of Sydney
Nasopharyngeal carcinoma (NPC) is a squamous epithelial malignancy typically arising from the lateral pharyngeal recess. Although endoscopic biopsy is recommended for initial diagnosis, a proportion of these lesions are hidden from endoscopic view due to challenging anatomy or submucosal location. Magnetic resonance imaging (MRI) has an established role in locoregional staging and has also been shown to detect endoscopically occult tumors. This meta-analysis investigates the diagnostic performance of MRI for primary detection of NPC.

Materials and Methods:
A systematic review was performed in May 2021 of studies that examined the diagnostic performance of MRI for detection of NPC in patients at-risk or suspected of having NPC. Studies comparing the diagnostic accuracy of the index test (MRI) to a reference standard (histopathology) were included. The primary outcome was accuracy for the detection of NPC. Random effects models were used to pool outcomes for sensitivity, specificity, positive likelihood ratio (LR), and negative LR. Bias and applicability were assessed using the modified QUADAS-2 tool.

Eight studies were included involving 1092 patients of whom 333 were diagnosed with NPC. MRI demonstrated a pooled sensitivity of 97.8% (95% CI 93.5% - 99.3%), specificity of 94.2% (95% CI 89.8% - 96.8%), negative LR of 0.02 (95% CI 0.01-0.06), and positive LR of 17.00 (95% CI 9.56–30.25) for detection of NPC. No statistically significant difference was demonstrated between prospective and retrospective studies for sensitivity (p = 0.705) or specificity (p = 0.638).

This study demonstrates that MRI has a high pooled sensitivity, specificity, and negative predictive value for detection of NPC. MRI may be an important diagnostic tool for high-risk patients prior to endoscopic biopsy and to avoid biopsy in patients with a low post-test probability of disease.