Comparison of dosimetric parameters for predicting radiation-induced cataract in paediatric patients

INTRODUCTION: This study compared the predictive ability of radiation-induced cataract between maximum point dose of the lens (Lens Dmax) ≥7 Gy, mean lens dose (Lens Dmean) ≥7 Gy, Lens Dmax ≥10 Gy, and Lens Dmean ≥10 Gy.

METHODS: Patients aged 3-18 years received cranial irradiation or radiation therapy at head and neck area between January 2010 and December 2019 at our institute were included. Patients without baseline and/or follow-up eye examination were excluded. Receiver operating characteristic (ROC) curves identified potential predictors and Cox regression analysed correlations between potential factors and cataract occurrence.

RESULTS: Sixty-three patients (122 eyes) were analysed. Cataracts were detected in 14 eyes (11.5%). Median follow-up time was 4 years (range 0.5-10 years), with cataract developing in a median of 2.5 years (range 0.3-7 years). Three patients (21.4%) developed grade ≥3 cataract. Lens Dmean ≥10 Gy was associated with cataract formation.

CONCLUSION: Lens Dmean ≥10 Gy showed the highest ability for predicting radiation-induced cataract in paediatric patients. Net reclassification improvement (NRI) suggested that changing lens dose constraint from Dmax <7 Gy to Dmean <10 Gy would miss 7% of cataract cases but avoid 28% of unnecessary restrictions. Adopting a mean lens dose <10 Gy was suggested as a constraint for lens dose.