High-risk features in retinoblastoma: the association between histopathology and MRI.
AI interpretation is pending for this paper.
Open original publication →What the AI sees
Not AI summarized yet.
Research significance
Pending deeper interpretation.
Source abstract
BACKGROUND: Identification of high-risk features is essential for improving treatment outcomes of retinoblastoma. MRI is one of the most used modalities in treatment decision-making; however, its accuracy in detecting extension into different ocular structures is variable. This study aims to evaluate high-risk features in retinoblastoma by comparing MRI with histopathology results. METHODOLOGY: This was a retrospective chart review of patients enrolled in the retinoblastoma programme at the American University of Beirut Medical Center from November 2012 to December 2022. We included enucleated eyes, classified as group D or E per the International Intraocular Retinoblastoma Classification. Data pertaining to high-risk features (including postlaminar optic nerve (ON) and massive choroidal invasion) were collected from preoperative MRI and histopathological examination following enucleation. Data on scleral and anterior segment involvement (not considered high-risk features) were also collected. Morphometric properties were calculated and analysed. RESULTS: Analysis included 48 eyes from 44 patients. Males constituted 52.1% (n=25) of the sample and females 47.9% (n=23). 58% had unilateral retinoblastoma. The median age at diagnosis was 14.5 months (10-34.5). The median follow-up period from the first visit to data collection was 27.5 months (17.3-49). On tissue examination, 3 patients had postlaminar ON involvement and 11 had massive choroidal involvement. For postlaminar ON involvement, the MRI showed a moderate sensitivity (66.7%) and specificity (71.1%), with a high negative predictive value (NPV) of 96.9% but a low positive predictive value (PPV) of 13.3%. In cases of massive choroidal invasion, MRI demonstrated moderate sensitivity (45.5%) and specificity (75.7%), with a good NPV of 82.4% but a low PPV of 35.7%. Additionally, histopathology identified anterior segment involvement in six cases and scleral involvement in two cases. MRI showed high accuracy in detecting invasion of the anterior chamber (79.2%) and moderate accuracy in detecting invasion of the sclera (62.5%). CONCLUSION: MRI showed high levels of variability in detecting high-risk features in advanced retinoblastoma. MRI results ought to be interpreted with caution when making treatment decisions in advanced retinoblastoma. Limitations of our study included a relatively small sample size, retrospective design and possible technical constraints of MRI imaging.