Role of radiation therapy to the brain in leukemic patients with cranial nerve palsies in the absence of radiological findings.

The value of brain radiotherapy for leukemic patients with cranial nerve palsies in the absence of radiological evidence of leukemic infiltration is not well defined. This retrospective study was undertaken to evaluate the effectiveness of brain irradiation in reversing the cranial nerve palsies in leukemic patients with no radiological evidence of intracranial leukemic infiltration. Records of leukemic patients who received brain radiotherapy between June 1980 and December 1993 were reviewed. Criteria for inclusion were 1) no evidence of intracranial leukemic infiltration by computed axial tomography (CT) or magnetic resonance imaging scan (MRI), 2) no evidence of leukemic infiltration on ophthalmologic examination, and 3) no previous radiotherapy to the brain. Actuarial survival rates were calculated using the Kaplan-Meier method. Pearson's chi-squared test was used to compare responses. Twenty-eight patients met these criteria. The median age was 38 years (range 3-75 years): Seventeen patients had acute lymphoblastic leukemia, nine had acute myelogenous leukemia, and two had chronic myelogenous leukemia. Four patients had initial presentation with leukemia, and 24 presented with relapse. Twenty-six patients had cerebrospinal fluid cytology that was positive for leukemic cells. Fifteen patients had involvement of more than one cranial nerve, and nine had bilateral involvement. The most commonly involved nerves were the facial (n = 18), oculomotor (n = 9), and abducens nerves (n = 8). Twenty-six patients received whole-brain radiotherapy. Two received radiation to the base of the skull only. The median radiation dose was 24 Gy (range 16-30 Gy) at 2-3 Gy per fraction. Every patient had either concomitant intrathecal (n = 6) or systemic (n = 5) chemotherapy or both (n = 17) with radiation. Fourteen patients had complete reversal of the cranial nerve deficit, eight had partial recovery, and four had no response or progression of the disease. The response was unknown in two patients. Factors associated with complete response were unilateral versus bilateral involvement (72% vs. 13%, P = 0.005) and single versus multiple nerve involvement (75% vs. 36%, P = 0.045). In conclusion, radiation therapy to whole brain was effective in reversing cranial nerve deficits from leukemia, although the leukemic infiltration may not be visualized by CT or MRI. No dose-response relationship was observed in the range we examined.