A prospective neurocognitive evaluation of children treated with additional chemotherapy and craniospinal irradiation following isolated central nervous system relapse in acute lymphoblastic leukemia

Abstract

A prospective assessment of neurocognitive performance was conducted in children with acute lymphoblastic leukemia (ALL) following isolated central nervous system (CNS) relapse to evaluate the impact of additional systemic/intrathecal (IT) chemotherapy and craniospinal irradiation (CSI) upon long-term intellectual function. Twenty-one children with ALL manifesting an isolated CNS relapse between 1984 through 1989 underwent serial evaluations of intellectual function. Neurocognitive function was measured by the full-scale intelligence quotient (FSIQ) as determined by the age-appropriate Wechsler Intelligence Scale and by achievement in reading, math, and spelling as assessed by the Wide Range Achievement Test (WRAT). Intelligence testing was initiated following isolated CNS relapse after clearance of cerebrospinal fluid (CSF) cytology but prior to CSI and continued at annual intervals for a minimum of 4 years postmeningeal failure. Protocol treatment for isolated CNS relapse consisted of reinduction and maintenance systemic therapy, intrathecal (IT) triple-agent chemotherapy, and early CSI (cranium to 24 Gy and spine to 15 Gy at 1.5 Gy/fraction) as outlined on the institutional "Total XI" trial. All 21 children attained secondary CNS remission and underwent the planned additional systemic/IT chemotherapy and CSI. Fourteen of the 21 children remain in secondary continuous remission, while the remaining 7 experienced a second relapse and were removed from further neurocognitive assessment. For the eight female and six male long-term survivors, mean ages at original diagnosis and at CSI were 5.7 years (range = 0.6-16.2) and 7.0 years (range = 1.8-17.0), respectively. At a median follow-up interval of 4.6 years (ranges 1.7-6.8) post-CNS relapse, comparison of group mean initial to final FSIQs revealed no statistically significant difference between the two measures (94.5 vs. 95.9, respectively, n = 11, p = 0.52). None of the children are functioning in the mentally retarded range. Final FSIQ outcome directly correlated with initial FSIQ (p = 0.00005, n = 11), age at diagnosis (p = 0.009, n = 14), and age at CSI (p = 0.011, n = 14). In addition, change between initial and final FSIQ scores inversely correlated with age at diagnosis (p = 0.009) and age at CSI (p = 0.018) but not with baseline IQ score (p = 0.41). Initial FSIQ scores were not influenced by either age at diagnosis (p = 0.12) or age at CSI (p = 0.14). Final group mean (range) WRAT scores in reading, math, and spelling were measured to be 94.7 (68-132), 95.6 (69-126), and 93.7 (77-122), respectively (n = 13). Final reading and math scores directly correlated with both age at diagnosis (p = 0.01) and age at CSI (p = 0.009). Spelling outcome did not correlate with either age at diagnosis (p = 0.25) or age at CSI (p = 0.24). Nine children are placed in regular classrooms, while the remaining five require a mixed classroom environment. In our study, children with ALL experiencing an isolated CNS relapse tolerated additional systemic/IT chemotherapy and CSI without apparent deterioration of group serial intellectual scores. However, longitudinal analyses of group intellectual measures obscured the independent impact of initial FSIQ, age at original diagnosis, and age at CSI upon individual neurocognitive outcome. Early and more intensive psychoeducational stimulation may be needed in very young children presenting with low initial FSIQ scores who are treated with CSI and additional chemotherapy following isolated CNS relapse in ALL.