Accuracy of Computed Tomography and Magnetic Resonance Imaging for Detection of Pathologic Risk Factors in Patients Diagnosed with Retinoblastoma

ABSTRACT

Aim: To assess and compare the diagnostic accuracy of magnetic resonance imaging (MRI) and computed tomography (CT) for detecting the extent of tumor invasion in eyes with advanced retinoblastoma prior to enucleation using histopathologic analysis as a reference.

Methods: A total of 56 patients (68 eyes) enucleated for retinoblastoma were reviewed. Eyes with complete data (CT 28, MRI 16) were included for analysis. Imaging data were reviewed by a pediatric radiologist, blinded to histopathology results. Four high-risk factors which included scleral, choroidal, anterior eye segment invasion, and postlaminar optic nerve invasion were assessed.

Results: For CT, the diagnostic odds ratio (DOR), sensitivity, and specificity for detecting postlaminar optic nerve invasion were 1.7%, 25%, and 83.3%. Choroidal invasion was correctly identified in only 2 of 10 eyes (DOR 0.5, sensitivity 20%, specificity 66.7%). For scleral invasion, CT showed a diagnostic ratio of 1.5%, sensitivity 40%, specificity 69.6% whereas for anterior segment invasion DOR was 9.2%, sensitivity 100%, and specificity 65.4%. MRI showed a higher DOR for detecting postlaminar optic nerve invasion of 39%, sensitivity 77.8%, and specificity 100%. Choroidal and scleral invasion on MRI had a DOR of 6.5% (sensitivity 25%, specificity 100%) and 6% (sensitivity 33.3%, specificity 92.3%) respectively while DOR for anterior segment invasion was 1.3% (sensitivity 50%, specificity 57.1%).

Conclusion: CT showed a poor diagnostic accuracy for all four high-risk factors in advanced retinoblastoma and therefore may not be a useful tool for assessment of tumor extension. For MRI, detection of postlaminar optic nerve invasion was moderately accurate although less accurate for detection of other risk factors. Decision-making for each child should not be based on any single parameter but rather on consideration of clinical factors in combination with radiologic findings.

KEYWORDS:

• Retinoblastoma
• computerized tomography (CT)
• magnetic resonance imaging (MRI)
• optic nerve invasion
• choroidal invasion
• pathologic high-risk factors

Declaration of interest

The authors have no conflict of interest to declare.