High-Risk Retinoblastoma Based on International Classification Systems: Analysis of 1362 Eyes.

Purpose: To determine the predictive value of International Intraocular Retinoblastoma Classification schemes and the American Joint Committee on Cancer (AJCC) classification for histopathological high-risk features (HRFs).

Design: Multicentric international collaborative retrospective case series.

Subjects: One thousand three hundred and sixty-two patients with retinoblastoma from 16 centers and 11 countries.

Intervention: Primary enucleation; adjuvant therapy in patients with HRF.

Main outcome measures: High-risk retinoblastoma defined as 1 or more HRF (anterior segment involvement, massive choroidal invasion, minor choroidal infiltration with prelaminar optic nerve invasion, retrolaminar or resected optic nerve cut end involvement, scleral or microscopic extrascleral infiltration); metastasis-free survival (MFS).

Results: Of the 1362 patients, 751 (55.1%) had HRF. According to the International Classification of Retinoblastoma (ICRB) (Philadelphia vs. Los Angeles [LA]) versus Children's Oncology Group (COG) classification schemes, the positive predictive value (PPV) of group D eyes for HRF was 42.0% versus 35.1% versus 43.2%, respectively, and that for group E eyes was 58.5% versus 59.0% versus 59.5%, respectively. Comparing group D versus group E eyes, there was higher mean number of HRF (standard deviation, range) among group E eyes using the ICRB Philadelphia (0.7 [0.9, 0.0-6.0] vs. 1.3 [1.7, 0.0-9.0], P < 0.001), ICRB LA (0.6 [0.8, 0.0-6.0] vs. 1.3 [1.7, 0.0-9.0], P < 0.001) and COG (0.8 [1.2, 0.0-7.0] vs. 1.3 [1.6, 0.0-8.0], P < 0.001) classifications. The PPV for HRF was above 55% for AJCC clinical tumor (cT) group cT3a with increments through cT3e to 72.3%. An agreement between ICRB Philadelphia versus ICRB LA, ICRB LA versus COG, and ICRB Philadelphia versus COG was 0.9, 0.8, and 0.8, respectively (P < 0.001). Metastasis-free survival rates and overall survival rates were also comparable between all intraocular retinoblastoma classification schemes but better stratified within the AJCC scheme.

Conclusions: All intraocular retinoblastoma classification schemes predict HRF and MFS equally. Group E includes a wide spectrum equivalent to the AJCC group cT3. Uniform grouping with subcategorization of group E might improve risk stratification. We propose that everyone across the retinoblastoma world henceforth adopts the AJCC classification for all reporting and publishing.

Financial disclosure(s): The authors have no proprietary or commercial interest in any materials discussed in this article.