OBJECTIVETo observe the treatment and prognosis of optic nerve invasion in retinoblastoma (RB).

METHODThe children who had been diagnosed with unilateral RB and had received enucleation from January 2006 to December 2013 in our hospital were recruited. Tumor extension into the optic nerve were disclosed. Optic nerve involvement was classified into four grades according to the degree of invasion. Grade I is superficial invasion of the optic nerve head only, grade II is involvement up to and including the lamina cribrosa, grade III is involvement beyond the lamina cribrosa, and grade IV is involvement up to and including the surgical margin. Grade I and II are called invasion of the optic nerve before the sieve plate. Grade III and IV are called invasion of the optic nerve after the sieve plate. Other high-risk factors included extensive invasions of the choroid, sclera, anterior chamber, iris, and ciliary body. They were divided into two groups according to whether the merger of other high histopathologic risk factors. Treatment was delivered accordingly, and the prognosis of different degrees of optic nerve invasion was observed. The subjects were followed up for 6 months to 7 years (average: 43 months).

RESULTThere were 465 subjects in this study, including 279 boys and 186 girls. The right eye was affected in 260 patients and the left eye in 205 patients. The average time from onset of symptoms to visit was 2.7 months (range 1 day-24 months). Twenty-five patients died, resulting in an overall survival rate of 94.6%. The mortality rate of patients with optic nerve involvement with grade I was 0.4%, grade II was 1.0%, grade III was 8.7% was and grade IV was 60.9%. Of the 338 with invasion of the optic nerve before the sieve plate, two died of recurrence, with a survival rates of 99.4% (336/338). Of the 127 patients who had invasion of the optic nerve after the sieve plate, twenty-three died of recurrence, with a survival rate of 81.9% (104/127) , the difference was statistically significant (χ² = 52.299, P = 0.000). A total of 379 patients did not have any other merged pathology high-risk factors, 8 died, the mortality rate was 2.1%. Of the 86 patients who had complicated with other high-risk factors, 17 died, the mortality rate was 19.8%, the difference was statistically significant (χ² = 42.955, P = 0.000). Of the 338 patients, 304 had invasion of the optic nerve before the sieve plate had not merged other pathology high-risk factors, none died, of the 34 patients who had complicated with other pathology high-risk factors, 2 had died, the mortality rate was 5.9%, the difference was statistically significant (P = 0.010). Of the 127 patients with invasion of the optic nerve after the sieve plate, 76 had not complicated with other pathological high-risk factors, 9 of whom had died, the mortality rate was 11.8%, 51 had complicated with other pathological high-risk factors, 14 of whom had died, the mortality rate was 27.5%, outcomes did significantly differ between the two subgroups (χ² = 5.014, P = 0.025). Cox multivariate analysis showed that invasion of the retrolaminar optic nerve, surgical margin of the optic nerve and sclera were influential factors of colorectal cancer.

CONCLUSIONPatients with optic nerve invasion have an excellent outcome with current therapy. But for those whose resection margin was invaded, which has a high incidence of recurrence, chemotherapy is recommended for patients with postlaminar optic nerve involvement.