Abstract
BACKGROUND: Acute lymphoblastic leukemia (ALL) and brain tumor (BT) survivors are at risk for post-treatment IQ declines. The extent to which lower scores represent global cognitive decline versus domain-specific impairment remains unclear. This study examined discrepancies between processing speed and estimated IQ (EIQ) scores and identified clinical characteristics associated with score discrepancies in a sample of pediatric cancer survivors. PROCEDURE: Survivors (50 ALL, 50 BT) ages 12-17 years completed cognitive testing. The Wechsler Abbreviated Scale of Intelligence provided an untimed measure of general reasoning ability (EIQ). The age-appropriate Wechsler Intelligence Scale provided a Processing Speed Index (PSI) score. Scores were examined and compared. RESULTS: Survivors' PSI scores were lower than their EIQ scores (BT t(45) =6.3, p<0.001; ALL t(49) =6.9, p<0.001). For BT survivors, lower PSI scores were associated with history of craniospinal irradiation, t(44) =3.3, p<0.01. For ALL survivors, lower PSI scores were associated with male gender, grade retention, and time since diagnosis, F(3, 46) =10.1, p<0.001. Clinically significant EIQ-PSI score discrepancies were identified in 41.3% of BT and 14.0% of ALL survivors. CONCLUSIONS: Many pediatric BT and ALL survivors exhibit slower processing speed than expected for age, whereas general reasoning ability remains largely intact. Risk factors associated with larger EIQ-PSI discrepancies include the following: BT diagnosis, craniospinal irradiation (BT only), male gender, and younger age at diagnosis (ALL only). Grade retention was frequent and associated with lower EIQ scores (both groups) and PSI scores (ALL only). Describing post-treatment cognitive declines using global measures of intellectual ability may underestimate dysfunction or fail to isolate specific underlying deficits contributing to impairment.