Commentary: study of the neurobehavioral consequences of childhood cancer: entering the genomic era?

The outlook for children diagnosed with cancer has significantly improved and continued to change since the late 1960s. Due to advances in therapy and diagnosis, the focus of psychological research and service has progressively changed from death and dying to surviving and cure, as well as possible neurobehavioral late effects, which is addressed in the accompanying papers (Moore; Butler & Mulhern, this issue). The first papers addressing neurobehavioral issues of survivors of pediatric cancer focused on childhood acute lymphoblastic leukemia (ALL) and reported quite contrasting results. In the first study, Soni, Marten, Pitner, Duenas, and Powazek (1975) found that survivors of childhood ALL who had received central nervous system (CNS) preventative treatment, including 2400 cGy of cranial radiation therapy (CRT), did not have cognitive/intellectual deficits compared with a matched disease control group and were functioning in the average range. In another investigation, which studied similarly treated children, it was found that the majority had evidence of computed tomography brain scan abnormalities such as cortical atrophy, intracerebral calcifications, and white matter translucency (Peylan-Ramu, Poplack, Pizzo, Adornato, & Di Chiro, 1978). Subsequent papers identified reliable intellectual deficits in long-term survivors of childhood ALL compared with siblings or a control group (Eiser, 1978; Meadows et al., 1981; Moss, Nannis, & Poplack, 1981). Although early on it was suggested that the etiology of the deficits seen in survivors of childhood ALL could be mostly psychosocial (Koocher & O’Malley, 1981), the biologic etiology of these effects is now well established. A number of studies have related the neurobehavioral deficits to brain imaging abnormalities (Brouwers, Riccardi, Fedio, & Poplack, 1985; Ciesielski et al., 1994). Secondly, in controlled clinical trials, children who received CNS preventative treatment that included CRT have been shown to have lower levels of neurobehavioral performance than children treated with CNS preventative therapy that did not include CRT (Brouwers, Moss, & Poplack, 1992; Waber, Bernstein, Kammerer, Tarbell, & Sallan, 1992). Subsequent studies also started to better characterize neuropsychological deficits and identify domains at higher risk for late effects (Anderson, Smibert, Ekert, & Godber, 1994; Brouwers, Riccardi, Poplack, & Fedio, 1984; Saykin, Ahles, & McDonald, 2003). An increasing interest in the neurobehavioral consequences of pediatric brain tumors in survivors started (Ellenberg, McComb, Siegel, & Stowe, 1987) when some of the typical neurotoxicities of therapy became better known, the prognosis for pediatric brain tumors improved, and the incidence of pediatric brain tumors increased. As noted in the accompanying papers, a lot of similarities in neurobehavioral profiles and their progression over time have been noted in children with brain tumors and ALL. It is clear that numerous factors determine neurobehavioral outcome in survivors of childhood cancer. Conceptually, these factors can be subdivided into mediators and moderators (Baron & Kenny, 1986). Mediators are factors that specify how or by which mechanisms an effect occurs, while moderators are factors that affect the direction and/or strength of the relation between a mediator and an outcome variable but are not by themselves pathogenic. Whether factors are mediators or moderators has played an important role in the attribution of late effects in the past, as noted before. In the recent past, most research emphasis was on relating the neurobehavioral sequelae to biologic factors, with diseaseand/or treatment-related factors as mediators, and gender, age at diagnosis, time since diagnosis, and age at testing as common moderators. Although the mediators of the sequelae for the various diseases may be very different, the factors that moderate outcome seem to be rather similar for all cases in which, as a consequence of