DISCUSSION
One of the main challenges after curative treatment of childhood cancer is maintaining an effective follow up to assess various health related complications. There are a wide range of late effects as a result of cancer and its therapeutic exposure. Studies have shown that the cumulative number of health conditions in survivors of ALL are actually twice that identified among matched controls, particulary cardiovascular, endocrine, reproductive and neurological sequelae10. At a mean duration of around 5 years from diagnosis, 50% of our cohort of ALL survivors showed significant disturbance in neurocognitive function. Around 75% of our cohort showed a deficit in atleast one domain of neurocognition, which includes verbal comprehension, working memory, processing speed and perceptual reasoning. This is comparable to a large series of adult survivors of ALL, where neurocognitive impairment ranged between 28.6% for self-reported behavior problems to 58.9% for direct assessment of executive function4. The mean FSIQ of ALL survivors from developed countries are almost 20-30 points higher than in our cohort11,12. This could be due to regional variation in IQ and socioeconomic status between the two cohorts. However it also signifies the greater need to implement more stringent long term follow up guidelines in LMICs as the baseline prevalence of neurocognitive impairment is high.
Data on neurocognitive outcome in survivors of ALL in LMICs is scarce. According to a review of neurocognitive impairment in Asian Cancer survivors, mild-to-moderate impairment was reported in 10.0–42.8% of survivors13. A large series of cancer survivors from India, showed scholastic problems and psychosocial problems in around 43% and 57% of the cohort, respectively14. Racial, regional and ethnic variations of cancer and its treatment are known, which increases the burden of chronic health condition in this growing survivor population in LMICs. Hence, the importance of country specific databases of childhood cancer survivors.
As early as 1988, neurobehavioural effect of CRT in survivors of ALL have been demonstrated15. These late effects include deficits in attention, memory, processing speed, and visual spatial skills16. Demyelination with loss of white matter, microvascular damage, loss of neuronal precursors in hippocampus, basal ganglia and thalamus are all proposed mechanisms for long term radiation induced cognitive decline17. This neurocognitive morbidity in survivors of childhood ALL has led to the phasing out of radiation, almost completely from the management of ALL18,19. Acute changes in the brain occurring during chemotherapy are also associated with long term neuro behavioural problems20. Despite elimination of CRT, modern treatment protocols including high dose methotrexate and IT methotrexate, have been implicated in long term problems in memory, executive function, attention and processing speed4,21,22. Methotrexate depletes the folate pool which in turn depletes the availability of methyl-groups and produces an excess of homocysteine and alternative one-carbon sources such as choline and betaine. This leads to an impairment in myelin formation and stabilisation and is a likely explanation for methotrexate induced long lasting neurodevelopmental impairment23. Higher plasma methotrexate levels have been shown to be associated with poorer executive function scores4. In one study methotrexate decreased the processing speed by 3% for each 1g/m2of methotrexate4. A study comparing IT methotrexate with triple IT revealed minor reduction in processing speed in survivors of ALL. CRT, high dose methotrexate and IT therapy have a neurotoxic effect in the early course of therapy. As time passes neurocognitive problems are a result of other therapies like glucocorticoids. The cerebellum and hippocampi have higher number of glucocorticoid receptors and are more affected by exposure to prolonged and high doses of steroids24. Comparisons between different glucocorticoids have also been made. Neurocognitive outcome measures evaluated in survivors receiving dexamethasone or prednisolone during induction of ALL, found no difference in outcome measures in either group25. Even though almost half of our cohort had neurocognitive dysfunction, the scores were not different in groups of patients receiveing higher dose of CRT compared to lower doses(24Gy vs 18Gy). Higher dose of methotrexate or use of IT methotrexate so did not appear to have adverse effect on neurocognition in our group of survivors. Other factors that can be evaluated to assess the risk of neurotoxicity include markers of oxidative stress26, cytokines like tumor necrosis factor(TNF) alpha and interleukins27 and cerebrospinal fluid biomarkers28.
Neurocognitive dysfunction is seen to be more significant as age increases29. Attention problems detected at the end of therapy in survivors of ALL predicted decreased academic performance 2 years later, especially in females and children who had a younger age at diagnosis30. It is also believed that patients diagnosed at a younger age and those who received CRT, perform poorly in neurocognitive domains13,16. A younger age at diagnosis correlated with poor working memory scores in our cohort. Other domains were not affected by either the age at diagnosis or assessment. Nevertheless, assessment of neurocognition in survivors of ALL should start at a younger age, especially in children in LMICs, where focus is more on survival than survivorship, to allow them to reach their full potential at school and in the workplace.
Additional factors which put a survivor at increased risk of neurocognitive impairment, like polymorphisms in genes related to neurodevelopment, genes affecting the folate pathway and genes that regulate oxidative stress, have been studied31. This can not only help us in stratifying patients according to their genetic risk but also help in tailoring treatment and prophylaxis early in development32. Since regional variations affect the genetic makeup of an individual, country specific multicentre collaborative studies are in much need.
There is limited research examining how a childs neurocognitive function following treatment for cancer is associated with parenting factors. A recent study reported that parents respond to child executive function difficulties with greater over protection. Greater years of maternal education were related to less parental perceptions of child vulnerability and less over protection33. A univariate analysis in our cohort, showed a significantly lower Intelligence quotient(IQ) in survivors whos’ parents had fewer years of education. Other factors like lower socioeconomic status and increased birth order are also associated with lower IQ scores. These are predictable and understandable findings and may also be seen in children who are not cancer survivors. However, since other factors including radiation dose, intensity of chemotherapy or age at primary diagnosis did not alter outcome, focus of neurocognitive behavioural assessment and intervention should be more focussed towards this more vulnerable group of children and adolescents to mitigate the impact of therapy on neurocognitive outcomes.
Poor neurocognition in ALL survivors, has been associated with a marked decrease in quality of life due to increased chronic health morbidities34. Survivors of ALL have been believed to perform poorly in school because of probable dysfunction in executive function abilities(4(p20). Compared to their siblings, they have greater degrees of inattention-hyperactivity, anxiety, depression, social withdrawal and learning problems. Adolescent survivors with neurocognitive impairment are also less likely to graduate from college than survivors without any impairment30. The St Judes Lifetime Cohort study concluded that though survivors of ALL had significant neurocognitive impairment, overall their educational attainment and employment status was similar to controls4. Delay in schooling is a major concern in survivors of ALL from India. Our cohort showed an average of 27 months of delay in schooling despite our best efforts. LMICs, which have their own cohort of young survivors, with poorer overall exposure to basic education, must focus on routine longitudinal screening in childhood survivors of ALL, optimizing their educational, vocational, and social outcomes based on local, regional and cultural practices.
Following a good assessment, survivors of ALL with neurocognitive impairment, should should be taught deficit specific compensatory interventions. This may be a computerised, home, school or even community based training31. Pharmacological agents have also been studied. Methylphenidate resulted in a statistically significant improvement in measures related to attention35. Future research should focus on interventions to improve neurocognition in ALL survivors.
Several limitations should be considered in interpretation of these results. Firstly, this was a cross sectional study with a small sample size restricted by geographical coverage. Secondly baseline neurocognitive functions of these ALL survivors were not known and there was no healthy control group for comparison. Assessment at time of initiation of treatment may have yieldied falsely low scores associated with anxiety and stress at diagnosis. Thirdly, due to a lack of region specific assessment tools for neurocognitive function, we had to use region specific adaptations of the assessment tool which allowed certain substitutions. Limitations notwithstanding, comparisons were made across treatment groups and patients with or without significant neurocognitive impairment. In addition, the primary focus of this study was to assess the neurocognitive impairment in a group of survivors from a LMIC where baseline educational and vocational opportuinites are challenging even for children without cancer. To the best of our knowledge this was the first study in India, a LMIC, in which WISC-IV scale was used for the assessment of neurocognitive functions in the survivors of childhood ALL. We looked extensively into various factors associated with neurocognitive deficits including host, disease and treatment related factors.
In conclusion, data on neurocognitive outcomes in survivors of ALL are extremely scarce, this being particularly true for our country. Ethnic, geographic and genetic variations predispose survivors differently to adverse chronic neurological toxicities. This can explain the difference in prevalence, severity, and presentation of neurocognitive impairment in survivors of ALL in LMICs. More data from LMICs are needed to assess the time to initiate screening and ways to identify potentially remediable risk factors which in turn will enable us to develop culturally appropriate preventive and rehabilitative interventions.