Glioblastoma and radiotherapy: A multicenter AI study for Survival Predictions from MRI (GRASP study).

BACKGROUND: The aim was to predict survival of glioblastoma at 8 months after radiotherapy (a period allowing for completing a typical course of adjuvant temozolomide), by applying deep learning to the first brain MRI after radiotherapy completion. METHODS: Retrospective and prospective data were collected from 206 consecutive glioblastoma, isocitrate dehydrogenase -wildtype patients diagnosed between March 2014 and February 2022 across 11 UK centers. Models were trained on 158 retrospective patients from 3 centers. Holdout test sets were retrospective (n = 19; internal validation), and prospective (n = 29; external validation from 8 distinct centers). Neural network branches for T2-weighted and contrast-enhanced T1-weighted inputs were concatenated to predict survival. A nonimaging branch (demographics/MGMT/treatment data) was also combined with the imaging model. We investigated the influence of individual MR sequences; nonimaging features; and weighted dense blocks pretrained for abnormality detection. RESULTS: The imaging model outperformed the nonimaging model in all test sets (area under the receiver-operating characteristic curve, AUC P = .038) and performed similarly to a combined imaging/nonimaging model (P > .05). Imaging, nonimaging, and combined models applied to amalgamated test sets gave AUCs of 0.93, 0.79, and 0.91. Initializing the imaging model with pretrained weights from 10 000s of brain MRIs improved performance considerably (amalgamated test sets without pretraining 0.64; P = .003). CONCLUSIONS: A deep learning model using MRI images after radiotherapy reliably and accurately determined survival of glioblastoma. The model serves as a prognostic biomarker identifying patients who will not survive beyond a typical course of adjuvant temozolomide, thereby stratifying patients into those who might require early second-line or clinical trial treatment.

Overcoming challenges of translating deep-learning models for glioblastoma: the ZGBM consortium.

OBJECTIVE: To report imaging protocol and scheduling variance in routine care of glioblastoma patients in order to demonstrate challenges of integrating deep-learning models in glioblastoma care pathways. Additionally, to understand the most common imaging studies and image contrasts to inform the development of potentially robust deep-learning models. METHODS: MR imaging data were analysed from a random sample of five patients from the prospective cohort across five participating sites of the ZGBM consortium. Reported clinical and treatment data alongside DICOM header information were analysed to understand treatment pathway imaging schedules. RESULTS: All sites perform all structural imaging at every stage in the pathway except for the presurgical study, where in some sites only contrast-enhanced T1-weighted imaging is performed. Diffusion MRI is the most common non-structural imaging type, performed at every site. CONCLUSION: The imaging protocol and scheduling varies across the UK, making it challenging to develop machine-learning models that could perform robustly at other centres. Structural imaging is performed most consistently across all centres. ADVANCES IN KNOWLEDGE: Successful translation of deep-learning models will likely be based on structural post-treatment imaging unless there is significant effort made to standardise non-structural or peri-operative imaging protocols and schedules.

Cranial subarachnoid and subdural haemorrhage caused by spinal melanoma metastasis.

A 73-year-old female with a history of metastatic melanoma presented with extensive intracranial subarachnoid haemorrhage. Cranial imaging failed to reveal any vascular anomaly or tumour. MRI of the neuroaxis revealed a melanoma metastasis at T10 with associated subarachnoid haemorrhage. We review the literature on this rare presentation of subarachnoid haemorrhage.

Management of elderly patients with glioblastoma-multiforme-a systematic review.

The management of elderly patients with glioblastoma-multiforme (GBM) remains poorly defined with many experts in the past advocating best supportive care, in view of limited evidence on efficacy of more aggressive treatment protocols. There is randomised evidence (NORDIC and NA-O8 studies) to support the use of surgery followed by adjuvant monotherapy with either radiotherapy (RT) using hypofractionated regimes (e.g. 36 Gy in 6 fractions OR 40 Gy in 15 fractions) or chemotherapy with temozolomide (TMZ) in patients expressing methylation of promoter for O6-methylguanine-DNA methyltransferase enzyme. However, the role of combined-modality therapy involving the use of combined RT and TMZ protocols has remained controversial with data from the EORTC (European Organisation for Research and Treatment of Cancer)-NCIC (National Cancer Institute of Canada) studies indicating that patients more than 65 years of age may not benefit significantly from combining standard RT fractionation using 60 Gy in 30 fractions with concurrent and adjuvant TMZ. More recently, randomised data has emerged on combining hypofractionated RT with concurrent and adjuvant TMZ. We provide a comprehensive review of literature with the aim of defining an evidence-based algorithm for management of elderly glioblastoma-multiforme population.

Imaging assessment of acute ischaemic stroke: a review of radiological methods.

Acute ischaemic stroke is the second largest cause of death worldwide and a cause of major physical and psychological morbidity. Current evidence based treatment includes intravenous thrombolysis (IVT) and mechanical thrombectomy (MT), both requiring careful patient selection and to be administered as quickly as possible within a limited time window from symptom onset. Imaging plays a crucial role identifying patients who may benefit from MT or IVT whilst excluding those that may be harmed. For IVT, imaging must as a minimum exclude haemorrhage, stroke mimics and provide an estimate of non-viable brain. For MT, imaging must in addition detect and characterize intra-arterial thrombus and assess the intra and extracranial arterial architecture. More advanced imaging techniques may be used to assess more accurately the volume of non-viable and potentially salvageable brain tissue. It is highly likely that further research will identify patients who would benefit from treatment beyond currently accepted time windows for IVT (4.5 h) and MT (6 h) and patients with an unknown time of symptom onset. Current evidence indicates that best outcomes are achieved when treatment is instituted as soon as possible after symptom onset. A rapid, efficient imaging pathway including interpretation is fundamental to achieving the best outcomes. This review summarizes current techniques for imaging assessment of acute stroke, highlighting strengths and limitations of each. The optimum pathway is a balance between diagnostic information, local resources, specialization and the time taken to acquire, process and interpret the data. As new evidence emerges, it is likely that the minimum required imaging data will change.

Acute hemiparesis in Sturge-Weber syndrome.

A young woman, known to have Sturge-Weber syndrome, was admitted with headache and left hemiparesis. Neuroimaging showed chronic occlusion of the venous sinuses without evidence of acute thrombus formation, or a recent vascular event.

Mechanism of action of octreotide in acromegalic tumours in vivo using dynamic contrast-enhanced magnetic resonance imaging.

CONTEXT: Octreotide causes significant tumour shrinkage in patients with acromegaly but the exact mechanism of action is unclear in vivo. OBJECTIVE: To determine the mechanism of action of octreotide in vivo using dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI). DESIGN: Five patients with acromegaly were treated with octreotide as primary medical therapy. DCE-MRI was done at baseline and 24 weeks. Local ethical committee approval was granted. SETTING: Study was done in a tertiary care centre. PATIENTS: Five patients with newly diagnosed acromegaly were recruited. INTERVENTION: Patients were started on subcutaneous octreotide and DCE-MRI was done on 0 and 24 weeks. MAIN OUTCOME MEASURES: Amplitude of contrast intake, exchange rate and maximum enhancement index of tumour tissue was compared before and after treatment. RESULTS: Amplitude of contrast intake (9.87 +/- 3.52 vs. 4.97 +/- 1.96 P < or = 0.05) and exchange rate (6.27 +/- 1.57 vs. 1.63 +/- 0.76 P value < or = 0.01) were significantly higher at baseline in adenoma compared to normal pituitary tissue but was comparable to normal pituitary tissue after treatment. There was a significant decrease in amplitude of contrast intake and exchange rate which relates to functional vascularity of adenoma at 24 weeks compared to baseline (P-values 0.026 and 0.002 respectively) but there were no significant changes in the normal pituitary tissue. CONCLUSION: DCE-MRI in acromegalic tumours treated with octreotide showed a significant reduction in functional vascularity after octreotide therapy compared to baseline in pituitary adenomas. This supports the antiangiogenic action of somatostatin analogue therapy in vitro, but it remains unclear if this mechanism is important clinically in analogue pre-treatment reducing the effect of radiotherapy on these pituitary tumours.