Radiation tolerance of the optic apparatus

Abstract

Many clinicians limit the dose to the optic apparatus (chiasm, optic nerve, and retinas) to 45–50 Gy during treatment planning for targets in the brain and certain head and neck malignancies. Such a restriction may compromise adequate coverage of the planning target volume and lead to suboptimal tumor control, especially for malignant tumors. We sought to determine the incidence of radiation-induced optic injury in patients who received a dose of 50 Gy or more to the optic apparatus. We analyzed dosimetric data for 359 patients treated at the University of Alabama - Birmingham between 1989 and 2003 that were felt to be at risk for receiving dose to the optic apparatus. We determined the maximum dose and mean dose received by the optic chiasm, right and left optic nerves, and right and left retinas. The data were derived from dose-volume histograms (40%) and isodose curves in the rest. A group of 99 patients who received at least 50 Gy to one of the optic structures was identified. Excluding patients with less than 1 year of follow-up due to death (21 patients), lack of sufficient time (13), or being lost to follow-up (17), clinical follow-up of at least 1 year was available for 47 of these patients. The majority of the patients in this cohort had meningiomas (51%) and high-grade gliomas (23%), with the rest consisting of various head and neck or intracranial neoplasms. The primary end point was development of decreased visual acuity (other than cataracts) or a visual field defect attributed to radiation-induced injury. With a median follow-up of 26.2 months [range: 12 - 156], 3 of 47 patients (6.4%) developed clinical symptoms of decreased visual acuity or a visual field defect that could not be attributed to other causes such as tumor progression. Another patient suffered a retinal detachment and vitreous hemorrhage of uncertain etiology. The time from the completion of radiotherapy to manifestation of symptoms ranged from 4.2 to 36.2 months. For the 3 patients with injury to the optic apparatus, the maximum dose to the chiasm was 5030 - 5629 cGy [3 of 3 pts with > 50 Gy], to the ipsilateral optic nerve, 5030 cGy [1 of 3 pts > 50 Gy], and to the ipsilateral retina, 2969 cGy. In 2 of these cases, vision loss was due to optic nerve injury, with enhancement of the nerve seen on MRI [dose to nerve: 4200 - 5030 cGy]. In the other case, vision loss was presumed to be due to chiasmal injury, as the chiasm received 5629 cGy, and the nerve received only 1769 cGy. For the patient with a retinal detachment, the highest dose to the retina was 2067 cGy, and the optic nerve and chiasm received a maximum of 5257 cGy. Of the 44 patients who did not develop toxicity, the maximum dose was greater than 50 Gy for the chiasm in 39 of 44 patients [of these, median dose: 5221 cGy, max: 6660], for 65 optic nerves in 38 patients [median: 5400, max: 7180] and for 18 retinas in 12 patients [median: 5276, max: 6010]. Of 7 patients who had at least 60 Gy to an optic structure, none developed vision loss (median follow-up: 24.8 months). The dose per fraction at the optic apparatus was similar for patients with and without optic injury, with a median of 180 cGy and 183 cGy, respectively. In one patient, toxicity may have been due to a high dose gradient near the optic nerve such that small set-up errors may have caused the actual dose to be much higher than the calculated dose. In addition, this patient was treated using serial tomotherapy, and a small field matching or couch translation error may have introduced additional hotspots. Radiation-induced optic injury occurs in a minority of patients [3 of 47 patients (6.4%); 1 of 42 chiasms (2.3%); 2 of 66 optic nerves (3.0%)] who receive 50 Gy to portions of the optic apparatus. Although our data lack sufficient numbers to demonstrate a dose response, our findings are consistent with a TD 5/5 of 50 Gy or slightly higher. In fact, no patient who received greater than 60 Gy had a complication. For patients with malignant tumors where tumor control improves with doses above 50 Gy, concern over a relatively low chance of optic nerve injury should not compromise treatment of the planning target volume