Abstract
Purpose:
To analyze the reported success rates for intravenous chemotherapy for the treatment of intraocular retinoblastoma (RB), based on classification group and chemotherapy regimen.
Methods:
PubMed was searched for English articles containing the words “retinoblastoma” and “chemotherapy”, yielding 2329 papers. In vitro/genetic studies (n=1119), animal studies (n=291), studies unrelated to retinoblastoma treatment (n=473) or using a different treatment modality (n=137) were excluded. Case reports/small series (n=114) and editorials (n=108) were excluded. Thus, 87 studies met our inclusion criteria. Success was defined as globe retention without salvage external beam radiotherapy (EBRT). Studies with planned EBRT were excluded. Studies in which eye classification (Reese-Ellsworth [R-E] or International Classification of Retinoblastoma [ICRB]) could not be determined were excluded, as were papers in which it was not possible to identify which eyes avoided both EBRT and enucleation. Duplicate papers reporting on previously published cohorts of patients were excluded. This yielded 23 studies included in this analysis. Weighted point estimates and 95% confidence intervals for success rates by classification group were calculated. We also assessed for different success rates based on chemotherapy regimen and focal treatment and created a model based on these variables.
Results:
Weighted estimates of globe salvage rate with chemoreduction were 83% (95%CI: 67-92%) for R-E Group 1 eyes, 76% (95%CI: 66-84%) for Group 2 eyes, 67% (95%CI: 53-78%) for Group 3 eyes, 49% (95%CI: 36-63%) for Group 4 eyes, and 38% (95%CI: 28-49%) for Group 5 eyes (39% [95%CI: 24-57%] for 5A eyes and 46% [95%CI: 34-59%] for 5B eyes). Using the ICRB, weighted globe salvage rates were 93% (95%CI: 79-98%) for Group A eyes, 83% (95%CI: 71-91%) for Group B, 74% (95%CI: 50-89%) for Group C, 40% (95%CI: 30-52%) for Group D, and 24% (95%CI: 9-52%) for Group E. No statistically-significant difference between chemotherapy regimens was detected.
Conclusions:
Reporting of outcomes in retinoblastoma is not standardized. There is significant selection bias in many studies, as worse eyes are more likely to be enucleated primarily without an attempt at chemoreduction. Both R-E and ICRB predict chemoreduction success. New techniques such as intra-arterial and intravitreal chemotherapy should be compared to these historical success rates of IV chemotherapy.