Abstract
Background:Tumour classification, based on histopathology or molecular pathology, is of value to predict tumour behaviour and to select appropriate treatment. In retinoblastoma, pathology information is not available at diagnosis and only exists for enucleated tumours. Alternative methods of tumour classification, using noninvasive techniques such as magnetic resonance spectroscopy, are urgently required to guide treatment decisions at the time of diagnosis.Methods:High-resolution magic-angle spinning magnetic resonance spectroscopy (HR-MAS MRS) was undertaken on enucleated retinoblastomas. Principal component analysis and cluster analysis of the HR-MAS MRS data was used to identify tumour subgroups. Individual metabolite concentrations were determined and were correlated with histopathological risk factors for each group.Results:Multivariate analysis identified three metabolic subgroups of retinoblastoma, with the most discriminatory metabolites being taurine, hypotaurine, total-choline and creatine. Metabolite concentrations correlated with specific histopathological features: taurine was correlated with differentiation, total-choline and phosphocholine with retrolaminar optic nerve invasion, and total lipids with necrosis.Conclusions:We have demonstrated that a metabolite-based classification of retinoblastoma can be obtained using ex vivo magnetic resonance spectroscopy, and that the subgroups identified correlate with histopathological features. This result justifies future studies to validate the clinical relevance of these subgroups and highlights the potential of in vivo MRS as a noninvasive diagnostic tool for retinoblastoma patient stratification.
Highlights
Tumour classification, based on histopathology or molecular pathology, is of value to predict tumour behaviour and to select appropriate treatment
Metabolite concentrations correlated with specific histopathological features: taurine was correlated with differentiation, total-choline and phosphocholine with retrolaminar optic nerve invasion, and total lipids with necrosis
We have demonstrated that a metabolite-based classification of retinoblastoma can be obtained using ex vivo magnetic resonance spectroscopy, and that the subgroups identified correlate with histopathological features
Summary
High-resolution magic-angle spinning magnetic resonance spectroscopy (HR-MAS MRS) was undertaken on enucleated retinoblastomas. Principal component analysis and cluster analysis of the HR-MAS MRS data was used to identify tumour subgroups. Frozen tumour tissue samples from eyes enucleated without prior treatment were obtained from 52 retinoblastoma patients (48 unilateral, 4 bilateral), treated at Birmingham Children’s Hospital from 2004 to 2013. Consent was obtained for tissue banking for ethically approved research. Following removal of a portion of the tumour for HR-MAS MRS, the entire enucleated eye was processed into paraffin as part of standard clinical practice (Sastre et al, 2009), and every 5th–10th section stained with H&E for analysis. Histological analysis was undertaken on the whole paraffin-embedded tumour rather than on the tissue fragment used for HR-MAS MRS. A small subset of post-HR-MAS MRS tumours (n 1⁄4 10) were subsequently embedded in paraffin to confirm that the histological features of the HR-MAS MRS biopsy accurately reflected those of the whole tumour (Supplementary Table 3)
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
