Applying machine learning classifiers to automate quality assessment of paediatric dynamic susceptibility contrast (DSC-) MRI data.

OBJECTIVE: Investigate the performance of qualitative review (QR) for assessing dynamic susceptibility contrast (DSC-) MRI data quality in paediatric normal brain and develop an automated alternative to QR. METHODS: 1027 signal-time courses were assessed by Reviewer 1 using QR. 243 were additionally assessed by Reviewer 2 and % disagreements and Cohen's κ (κ) were calculated. The signal drop-to-noise ratio (SDNR), root mean square error (RMSE), full width half maximum (FWHM) and percentage signal recovery (PSR) were calculated for the 1027 signal-time courses. Data quality thresholds for each measure were determined using QR results. The measures and QR results trained machine learning classifiers. Sensitivity, specificity, precision, classification error and area under the curve from a receiver operating characteristic curve were calculated for each threshold and classifier. RESULTS: Comparing reviewers gave 7% disagreements and κ = 0.83. Data quality thresholds of: 7.6 for SDNR; 0.019 for RMSE; 3 s and 19 s for FWHM; and 42.9 and 130.4% for PSR were produced. SDNR gave the best sensitivity, specificity, precision, classification error and area under the curve values of 0.86, 0.86, 0.93, 14.2% and 0.83. Random forest was the best machine learning classifier, giving sensitivity, specificity, precision, classification error and area under the curve of 0.94, 0.83, 0.93, 9.3% and 0.89. CONCLUSION: The reviewers showed good agreement. Machine learning classifiers trained on signal-time course measures and QR can assess quality. Combining multiple measures reduces misclassification. ADVANCES IN KNOWLEDGE: A new automated quality control method was developed, which trained machine learning classifiers using QR results.

Response assessment in paediatric intracranial ependymoma: recommendations from the Response Assessment in Pediatric Neuro-Oncology (RAPNO) working group.

Response criteria for paediatric intracranial ependymoma vary historically and across different international cooperative groups. The Response Assessment in the Pediatric Neuro-Oncology (RAPNO) working group, consisting of an international panel of paediatric and adult neuro-oncologists, neuro-radiologists, radiation oncologists, and neurosurgeons, was established to address both the issues and the unique challenges in assessing the response in children with CNS tumours. We established a subcommittee to develop response assessment criteria for paediatric ependymoma. Current practice and literature were reviewed to identify major challenges in assessing the response of paediatric ependymoma to clinical trial therapy. For areas in which data were scarce or unavailable, consensus was reached through an iterative process. RAPNO response assessment recommendations include assessing disease response on the basis of changes in tumour volume, and using event-free survival as a study endpoint for patients entering clinical trials without bulky disease. Our recommendations for response assessment include the use of brain and spine MRI, cerebral spinal fluid cytology, neurological examination, and steroid use. Baseline postoperative imaging to assess for residual tumour should be obtained 24-48 h after surgery. Our consensus recommendations and response definitions should be prospectively validated in clinical trials.

Assessment of Cannabidiol and Δ9-Tetrahydrocannabiol in Mouse Models of Medulloblastoma and Ependymoma.

Children with medulloblastoma and ependymoma are treated with a multidisciplinary approach that incorporates surgery, radiotherapy, and chemotherapy; however, overall survival rates for patients with high-risk disease remain unsatisfactory. Data indicate that plant-derived cannabinoids are effective against adult glioblastoma; however, preclinical evidence supporting their use in pediatric brain cancers is lacking. Here we investigated the potential role for Δ9-tetrahydrocannabinol (THC) and cannabidiol (CBD) in medulloblastoma and ependymoma. Dose-dependent cytotoxicity of medulloblastoma and ependymoma cells was induced by THC and CBD in vitro, and a synergistic reduction in viability was observed when both drugs were combined. Mechanistically, cannabinoids induced cell cycle arrest, in part by the production of reactive oxygen species, autophagy, and apoptosis; however, this did not translate to increased survival in orthotopic transplant models despite being well tolerated. We also tested the combination of cannabinoids with the medulloblastoma drug cyclophosphamide, and despite some in vitro synergism, no survival advantage was observed in vivo. Consequently, clinical benefit from the use of cannabinoids in the treatment of high-grade medulloblastoma and ependymoma is expected to be limited. This study emphasizes the importance of preclinical models in validating therapeutic agent efficacy prior to clinical trials, ensuring that enrolled patients are afforded the most promising therapies available.

The diagnosis of brain tumours in children: a guideline to assist healthcare professionals in the assessment of children who may have a brain tumour.

BACKGROUND: Brain tumours are the commonest solid tumour in children. Children with brain tumours are frequently unwell for months prior to diagnosis. A prolonged period between symptom onset and diagnosis is associated with increased morbidity. OBJECTIVE: To develop an evidence-based clinical guideline to support healthcare professionals in the identification, assessment and investigation of children presenting with symptoms and signs that could be due to a brain tumour. METHODS: A systematic literature review with a meta-analysis and cohort study provided the guideline evidence base. A multi-disciplinary workshop and Delphi consensus voting were used to translate the evidence into a clinical guideline. The results of the literature review and cohort study have been previously published. RESULTS: 20 healthcare professionals and parents participated in the workshop. 77 statements were generated detailing the presenting features of childhood brain tumours, factors that could be used to discriminate brain tumours from other less serious conditions and possible referral pathways for children with brain tumours. 156 healthcare professionals agreed to participate in the Delphi process; 112 completed the first round and 88 completed all three rounds (attrition rate 21%). 64 statements reached consensus. The final guideline comprises 76 recommendations advising on the symptomatology of childhood brain tumours, assessment of children who may have a brain tumour and recommendations for selection for and timing of central nervous system imaging. CONCLUSION: Implementation of this guideline may support clinicians in the identification and timely imaging of children with brain tumours. This may reduce the morbidity currently experienced by many children with brain tumours.

Magnetic resonance spectroscopy in the assessment of pilocytic astrocytomas.

BACKGROUND: Pilocytic astrocytomas (PA) are common childhood brain tumours whose management and prognosis vary widely depending on location. (1)H magnetic resonance spectroscopy (MRS) measures biochemistry in vivo and shows promise for characterising brain tumours and aiding management. METHODS: Single voxel MRS (1.5 Tesla, TE 30 ms, TR 1500 ms) was performed on 27 children with PAs. Cases were designated 'progressors' if tumour progression led to their clinical management plan being altered. RESULTS: Prior to treatment, supratentorial tumours had significantly higher myo-inositol (p<0.01, t-test) and glutamate plus glutamine (p=0.02, t-test) than cerebellar tumours. Optic pathway or thalamic tumours that progressed had a significantly (p=0.04, t-test) lower myo-inositol at initial MRS than those with stable disease. Myo-inositol levels decreased significantly in progressors between the initial and subsequent MRS (p=0.03, paired t-test). Changes in myo-inositol occurred before clinical and radiological progression. CONCLUSIONS: MRS identifies differences with anatomical location in PAs and yields potential non-invasive biomarkers of prognosis.