Risk factors for treatment related mortality in childhood acute lymphoblastic leukaemia

Cancer remains the leading cause of disease-related mortality among children aged 1-14 years in developed countries. Acute lymphoblastic leukaemia (ALL) is the most common malignancy in childhood and accounts for about 25% of all childhood cancers. In Norway about 30-40 children are diagnosed with ALL each year. Treatment comprises of a combination of 6-12 different chemotherapy drugs administered over a period of 2.5 years with additional supportive care. Side effects remain a great challenge for both patients and physicians. One of the most common side effects is immunosuppression which increases the risk of infection. Today, the overall survival rate in childhood ALL is 80-85%. About 75% of deaths related to ALL are caused by the disease itself, most after relapse, but about 25% of deaths are caused by treatment and are classified as treatment-related deaths (TRDs). In this study, the incidence, risk factors and causes of death were investigated for 88 TRDs among 2,700 patients in the Nordic countries Sweden, Denmark, Finland, Iceland and Norway. The incidence of TRD was 3.2%, which was stable in the two protocols, the Nordic Society of Paediatric Haematology and Oncology (NOPHO) ALL 1992 and ALL 2000 protocols. The risk factors identified were female gender, treatment risk-group, T-cell disease, Down syndrome and stem cell transplantation. Seventy-five percent of the deaths were related to infection, 10% to bleeding or thrombosis, 10% to organ failure and 5% to the tumour burden. We have no good explanation as to why some patients experience more severe infections than others, even when receiving the same drugs, drug dosage and supportive care. It is reasonable to believe that genetic factors play a role. In accordance with the increasing understanding of host genetic variation the past 10 years, we investigated 34,000 single nucleotide polymorphisms (SNPs) relevant in childhood ALL in a Danish ALL cohort including 69 children and detected 24 SNPs which associated with infections during induction treatment. Through CART-analysis (classification and regression tree) a four-SNP risk profile were identified as highly predictive of risk of infections during the 50 days induction treatment. The four SNPs belong to the genes R51F1, CBR1, POLDIP3 and CCL11 which regulate drug metabolism, cell-growth and inflammation. If these findings are replicated in larger studies, such knowledge may be useful for developing personalized medicine with more tailored therapy and supportive care, which will hopefully reduce the severity of side effects and increase overall survival. Studies involving multiple biological markers, such as SNPs and short tandem repeats (STRs), often use high-quality DNA extracted from blood samples. However, blood samples are not always easy to obtain from study participants, especially if the patient has died. Archival bone marrow samples from patients with leukaemia are available and represent a potential DNA source. We were interested in exploring whether such archival material is usable for the analysis and identification of multiple markers. DNA from 21 bone marrow smears and 13 bone marrow biopsies were extracted and quantified. Because of small amounts of DNA from each sample whole genome amplification (WGA) was applied. Ten different STR-markers and 34,000 SNPs were analysed and compared with corresponding blood samples. For the STR markers 90% detection rates were obtained. Shorter markers (107bp-242bp) from samples stored 0-3 years gave better results compared with longer markers (219bp-317bp) stored 4-10 years. Multiple SNP analysis was more complicated and only seven of 34 archival samples gave acceptable results (SNP call-rates above 50%). However, in two samples, nearly 100% of SNP-markers were detected. Although increasing the total amount of DNA, WGA reduced the analysis quality. In conclusion, DNA from archival bone-marrow samples might be used in multiple marker analysis, but adjustments of the laboratory set-up are essential to optimize this method.