Oncology and haematology
G38 CHANGE OF CEREBRAL BLOOD FLOW OVER TIME IN CHILDREN WITH SICKLE CELL DISEASE
K. Kotecha, D. E. Saunders, F. J. Kirkham. Institute of Child Health & Great Ormond Street Hospital, London
Around 10% of children with sickle cell disease (SCD) have a stroke by the age of 20 and a further 24% have silent infarction (SI). Non-invasive methods of imaging such as transcranial Doppler (TCD) have been used to diagnose vasculopathy in SCD. Cerebral blood flow velocities (CBFV) >170 cm/sec are associated with higher risk of subsequent stroke, potentially preventable by blood transfusion. In adults, low CBFV is associated with SI.
This study has examined the natural history of CBFV as measured by TCD in a cohort of 64 patients who were not on a regular transfusion programme. Highest recorded average CBFV was measured on both sides of the internal carotid and middle cerebral arteries in 1991–1993 (time point 1) and again in 1996–1999 (time point 2). A score was given for each TCD measurement as follows: 0 = normal CBFV, 1 = CBFV>200 cm/sec, 2 = CBFV<50 cm/sec. The median age at time point 1 was 7 (range 1.4–15 years).
Thirty nine patients remained normal throughout and 15 with initially normal velocities changed to velocity <50 cm/s at time point 2. Of the 5 (8%) patients with velocities >200 cm/s at time point 1, 4 had velocities <50 cm/s at time point 2 and 4/5 patients with velocities <50 cm/s at time point 1 remained low at time point 2. Patients with abnormally high or low CBFV at time point 1 were more likely to have abnormally low CBFV at time point 2 (χ2, p = 0.007). Although clinical stroke was commoner in those with abnormal TCD (χ2 = 0.003) and SI, the latter relationship was not significant (χ2, p = 0.397).
There is progressive CVD in children with SCD over time as evidenced by an increase in the number of children with abnormally low TCD. There is also an increase in the prevalence of SI and incidence of clinical stroke in children with velocities >200 cm/sec and <50 cm/sec compared with the group with normal velocities. Studies of the predictors of CVD in a larger cohort are underway.
G39 PROSPECTIVE AUDIT OF EFFICACY OF A REGIONAL SUPPORTIVE CARE POLICY FOR FEBRILE NEUTROPENIA
C. Duncan, S. Freeman, J. Rabbs, J. Chisholm, A. Shankar, M. Michelagnoli, V. Grant, J. Hartley, E. Price, U. Riley, M. Sharland, K. Pritchard-Jones. Thames Paediatric Oncology Centres
Aims: To monitor efficacy and safety of guidelines for management of febrile neutropenia (FN) and identify antibiotic resistance among blood culture isolates.
Methods: Prospective audit of FN admissions to paediatric “shared care” units (POSCUs) of childhood cancer patients managed jointly with the four tertiary paediatric oncology centres in the Thames region (TPOC) over an 18 month period (7/01–12/02).
Results: 32/61 POSCUs returned forms on 433 FN episodes in 212 patients. 82% of admissions used recommended first line antibiotics (piptazobactam and gentamicin). Blood cultures were positive in 129 episodes (34%). Among 149 isolates, 81% were Gram positive (96 coagulase negative staphylococci (CNS), 10 Staph aureus (3 MRSA)), 18% were gram negative isolates, and <1% fungal. Low resistance was documented to first line antibiotics: among Gram positive isolates 2.5% resistant to piptazobactam and gentamicin, 12% to gentamicin alone, and 2.5% to piptazobactam alone; among Gram negative isolates none was resistant to piptazobactam and gentamicin, 11% to piptazobactam alone, and none to gentamicicn. Median length of hospital stay was 5 days. 3% of children presented in septic shock, 36% had respiratory symptoms, and 27% had fever only. 8/12 in septic shock required transfer to PICU, 1 died pre transfer (E coli sepsis). 75% had positive blood cultures, 50% were Gram negative isolates and 25% Gram positive isolates. All the isolates were sensitive to first line antibiotics.
Conclusion: The guidelines were well followed and appeared safe with a low mortality and morbidity. The current first line antibiotics appear appropriate with low resistance documented. It is planned to have risk assessment on admission and at 48 hours to identify a low risk group of children who could be considered for earlier discharge on oral antibiotics.
G40 EARLY INTENSIVE INDUCTION THERAPY IN PHILADELPHIA POSITIVE CHILDHOOD ACUTE LYMPHOBLASTIC LEUKAEMIA (PH+ ALL) INFLUENCES TREATMENT RESPONSE: LESSONS FROM THE MRC ALL97/01 TRIAL
A. Roy, M. Bradburn, A. Moorman, J. Burrett, C. Mitchell, S. Kinsey, I. Hann, V. Saha on behalf of the MRC ALL 97 Trial Coordinators and the UK Children’s Leukaemia Working PartyCancer Research UK Children’s Cancer Group, Royal London Hospital
Aims: To compare the outcome of children with Ph+ ALL treated on the UK protocol for childhood ALL, MRC ALL 97/01 with that obtained for children with Ph+ ALL treated on BFM/AIEOP protocol.
Methods: For the MRC trial, a diagnosis of Ph+ ALL was made by karyotyping and/or fluorescent in situ hybridisation. Outcome was evaluated according to NCI risk criteria and treatment response. These data were compared with published results of the BFM/AIEOP group. Statistical methods included Kaplan Meier analysis, log rank, and stratified log rank tests.
Results: See tables.
Conclusions: Although the overall outcome in the two study groups are comparable; high risk patients in the MRC group responded better to induction chemotherapy and had a significantly better event free survival when compared with those in the BFM/AIEOP group. We believe that this is because of the early introduction of intensive chemotherapy in the MRC trial.
G41 GENE EXPRESSION PROFILING IN CHILDHOOD ACUTE LEUKAEMIA
F. W. Van Delft, T. Bellotti, L. K. Jones, N. Patel, M. Skinner, V. Chauhan, S. Senderovich, M. Hubank, I. Hann, D. Fletcher, J. Sargent, C. Taylor, A. Gammerman, V. CR-UK, Children’s Cancer Group, Department of Paediatric Haematology and Oncology, Royal London Hospital, London, UK
Aims: Successful treatment of patients with leukaemia depends on correct diagnosis and early identification of patients at risk for disease relapse. We have prospectively analysed the gene expression pattern of leukaemic blasts at diagnosis with the use of oligonucleotide microarrays for classification purposes and to unravel genes involved in leukaemogenesis. We hypothesise that correlation with risk parameters will help identify patients at risk for disease relapse at the outset of treatment and develop more accurate stratification.
Methods: Children diagnosed with acute leukaemia at the Royal London and Great Ormond Street Hospitals have been included in this study. Ethical approval and written consent were obtained. The expression profile of leukaemic cells was determined using oligonucleotide microarrays. Data were analysed with Affymetrix Micro Array Suite 5.0.
Results: Analysis of the microarray data created statistically significant gene lists discriminating acute lymphoblastic (ALL) from acute myeloid leukaemia (AML). These gene lists were used to predict disease phenotype in three patients, hitherto unclassified. All were classified as AML. Therapy was adjusted accordingly and remission achieved. We identified genes differentially expressed in ALL subgroups defined by the presence of E2A-PBX1, TEL-AML1, hyperdiploid karyotype, or T cell immunophenotype. Analysis even identified the presence of chromosomal abnormalities undetected by conventional methods. Comparison of profiles between patients with standard and intermediate risk revealed prognostic genes.
Conclusion: We find microarray analysis a powerful diagnostic tool for tumour classification and it promises to refine risk stratification in the near future.
G42 AN EVALUATION OF LISA, A COMPUTERISED DECISION SUPPORT SYSTEM FOR CHEMOTHERAPY DOSING DURING MAINTENANCE THERAPY IN CHILDHOOD ACUTE LYMPHOBLASTIC LEUKEMIA
J. Bury1, A. Roy2, C. Hurt3, M. Bradburn4, S. Cross1, J. Fox3, V. Saha2. 1Division of Genomic Medicine, University of Sheffield Medical School, UK; 2Children’s Cancer group, Cancer Research UK, Royal London Hospital, UK; 3Advanced Computation Laboratory, Cancer Research UK, London, UK; 4Statistics Unit, Cancer Research UK, London, UK
Aims: To assess the acceptability to clinicians of a web based decision support system (“LISA”) designed to assist with dosage adjustments during maintenance therapy for childhood acute lymphoblastic leukemia (ALL), and to evaluate the potential impact of the system on accuracy and speed of decision making.
Methods: We performed a balanced block crossover experiment, in which 36 clinicians with varying degrees of experience in making dosage adjustments during maintenance (classified as novices, intermediates, and experts) were asked to decide on appropriate oral chemotherapy dosages for eight simulated cases: four using LISA and four using conventional paper based records and guidance.
Outcome Measures: Number of protocol consistent dosage decisions made; time taken to manage each case; accuracy of dosage calculations; and clinicians’ opinions about the utility and usability of the system and whether or not they would use it in practice. Qualitative data on issues subjects felt would impact upon the successful deployment of the system was also gathered.
Key Results: LISA significantly reduced the number of errors made in interpreting the protocol and/or calculating correct drug dosages (0/144 with LISA v 51/144 without, p<0.01). However, using LISA did not affect the number of times subjects (usually experts) deliberately overrode the protocol, reporting that they felt strict adherence was inappropriate (7/144 times using LISA, 6/144 without). Using LISA significantly reduced the time taken by novices to reach a decision and carry out the dosage calculation for each case (125.4 sec v 156.5 sec, p = 0.015), but increased the time taken by experts (133.6 sec v 110.8 sec, p = 0.017). 35/36 subjects said they would be likely to use the system if it were available.
Conclusions: A system similar to LISA is likely to be acceptable to clinicians. The system is likely to lead to an increase in protocol compliance and a decrease in prescribing errors, while allowing clinicians to override the protocol in specific cases where sound reasons exist for doing so.
G43 RITUXIMAB IN THE TREATMENT OF LYMPHOPROLIFERATIVE DISEASE IN CHILDREN. RESULTS FROM THE UNITED KINGDOM CHILDREN’S CANCER STUDY GROUP (UKCCSG) LPD REGISTRY
B. Messahel1, M. Taj1, R. Hobson2, I. Hann3, C. R. Pinkerton1. 1Royal Marsden Hospital, Downs Road, Sutton, Surrey, UK; 2UKCCSG, Leicester; 3Hospital for Sick Children, Great Ormond Street, London
Introduction: Survival of children with lymphoproliferative disease (LPD) still remains poor, particularly in non-post transplant lymphoproliferative disease (PTLD) cases. Anti CD20 (rituximab), a highly specific mouse/human chimaeric antibody, has shown promise in this condition. Response and tolerance of rixuximab in 18 children with LPD was reviewed.
Methods: A retrospective study of data in children with LPD treated with rituximab in UKCCSG centres is presented.
Results: A total of 18 children (13 males, 5 females) aged 11 months to 18 years were studied. Three had primary immunodeficiency, 2 secondary LPD due to prolonged immunosuppression, and 13 children had PTLD. Transplant types were 3 heart, 1 kidney, 4 liver, and 5 bone marrow. The mean time from diagnosis to LPD was 2 years (range 0.3–5 years) in the non-PTLD group and 1.6 years from transplant (range 0.1–10 years) in the PTLD group. There was multi-organ involvement in 7 patients. The remaining had single site disease (4 nodal, 2 lung, 2 GI tract, 1 bone, 1 cutaneous, and 1 liver). In the PTLD group all patients had a reduction in immunosuppression and 6 had COP chemotherapy. Rituximab was used as second/third line therapy in all patients and given in a dose of 375 mg/m2 IV weekly infusion. One patient had a dose of 200 mg/m2 due to renal dysfunction. 8 patients received 4 courses and 10, 1–3 courses. Fever was the main side effect in 4 patients. Eleven patients responded to rituximab, 6 CR (complete response), 3 PR (partial response), and 2 MR (minimal response). Mean duration of follow up was 8.2 months (range 1–30 months). The overall response rate was 11 (61%) and survival 50% (2 died of unrelated causes).
Conclusion: In children with LPD not responding to standard treatment rituximab showed a high response rate and should be evaluated in early stage disease.