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.

Imaging in patients with glioblastoma: A national cohort study.

Background: Glioblastoma is the most common malignant brain tumor in adults and has a poor prognosis. This cohort of patients is diverse and imaging is vital to formulate treatment plans. Despite this, there is relatively little data on patterns of use of imaging and imaging workload in routine practice. Methods: We examined imaging patterns for all patients aged 15-99 years resident in England who were diagnosed with a glioblastoma between 1st January 2013 and 31st December 2014. Patients without imaging and death-certificate-only registrations were excluded. Results: The analytical cohort contained 4,307 patients. There was no significant variation in pre- or postdiagnostic imaging practice by sex or deprivation quintile. Postdiagnostic imaging practice was varied. In the group of patients who were treated most aggressively (surgical debulking and chemoradiation) and were MRI compatible, only 51% had a postoperative MRI within 72 hours of surgery. In patients undergoing surgery who subsequently received radiotherapy, only 61% had a postsurgery and preradiotherapy MRI. Conclusions: Prediagnostic imaging practice is uniform. Postdiagnostic imaging practice was variable. With increasing evidence and clearer recommendations regarding debulking surgery and planning radiotherapy imaging, the reason for this is unclear and will form the basis of further work.

Moving towards vertically integrated artificial intelligence development.

Substantial interest and investment in clinical artificial intelligence (AI) research has not resulted in widespread translation to deployed AI solutions. Current attention has focused on bias and explainability in AI algorithm development, external validity and model generalisability, and lack of equity and representation in existing data. While of great importance, these considerations also reflect a model-centric approach seen in published clinical AI research, which focuses on optimising architecture and performance of an AI model on best available datasets. However, even robustly built models using state-of-the-art algorithms may fail once tested in realistic environments due to unpredictability of real-world conditions, out-of-dataset scenarios, characteristics of deployment infrastructure, and lack of added value to clinical workflows relative to cost and potential clinical risks. In this perspective, we define a vertically integrated approach to AI development that incorporates early, cross-disciplinary, consideration of impact evaluation, data lifecycles, and AI production, and explore its implementation in two contrasting AI development pipelines: a scalable "AI factory" (Mayo Clinic, Rochester, United States), and an end-to-end cervical cancer screening platform for resource poor settings (Paps AI, Mbarara, Uganda). We provide practical recommendations for implementers, and discuss future challenges and novel approaches (including a decentralised federated architecture being developed in the NHS (AI4VBH, London, UK)). Growth in global clinical AI research continues unabated, and introduction of vertically integrated teams and development practices can increase the translational potential of future clinical AI projects.

Imaging Biomarkers of Glioblastoma Treatment Response: A Systematic Review and Meta-Analysis of Recent Machine Learning Studies.

OBJECTIVE: Monitoring biomarkers using machine learning (ML) may determine glioblastoma treatment response. We systematically reviewed quality and performance accuracy of recently published studies. METHODS: Following Preferred Reporting Items for Systematic Reviews and Meta-Analysis: Diagnostic Test Accuracy, we extracted articles from MEDLINE, EMBASE and Cochrane Register between 09/2018-01/2021. Included study participants were adults with glioblastoma having undergone standard treatment (maximal resection, radiotherapy with concomitant and adjuvant temozolomide), and follow-up imaging to determine treatment response status (specifically, distinguishing progression/recurrence from progression/recurrence mimics, the target condition). Using Quality Assessment of Diagnostic Accuracy Studies Two/Checklist for Artificial Intelligence in Medical Imaging, we assessed bias risk and applicability concerns. We determined test set performance accuracy (sensitivity, specificity, precision, F1-score, balanced accuracy). We used a bivariate random-effect model to determine pooled sensitivity, specificity, area-under the receiver operator characteristic curve (ROC-AUC). Pooled measures of balanced accuracy, positive/negative likelihood ratios (PLR/NLR) and diagnostic odds ratio (DOR) were calculated. PROSPERO registered (CRD42021261965). RESULTS: Eighteen studies were included (1335/384 patients for training/testing respectively). Small patient numbers, high bias risk, applicability concerns (particularly confounding in reference standard and patient selection) and low level of evidence, allow limited conclusions from studies. Ten studies (10/18, 56%) included in meta-analysis gave 0.769 (0.649-0.858) sensitivity [pooled (95% CI)]; 0.648 (0.749-0.532) specificity; 0.706 (0.623-0.779) balanced accuracy; 2.220 (1.560-3.140) PLR; 0.366 (0.213-0.572) NLR; 6.670 (2.800-13.500) DOR; 0.765 ROC-AUC. CONCLUSION: ML models using MRI features to distinguish between progression and mimics appear to demonstrate good diagnostic performance. However, study quality and design require improvement.

Glioblastoma post-operative imaging in neuro-oncology: current UK practice (GIN CUP study).

OBJECTIVES: MRI remains the preferred imaging investigation for glioblastoma. Appropriate and timely neuroimaging in the follow-up period is considered to be important in making management decisions. There is a paucity of evidence-based information in current UK, European and international guidelines regarding the optimal timing and type of neuroimaging following initial neurosurgical treatment. This study assessed the current imaging practices amongst UK neuro-oncology centres, thus providing baseline data and informing future practice. METHODS: The lead neuro-oncologist, neuroradiologist and neurosurgeon from every UK neuro-oncology centre were invited to complete an online survey. Participants were asked about current and ideal imaging practices following initial treatment. RESULTS: Ninety-two participants from all 31 neuro-oncology centres completed the survey (100% response rate). Most centres routinely performed an early post-operative MRI (87%, 27/31), whereas only a third performed a pre-radiotherapy MRI (32%, 10/31). The number and timing of scans routinely performed during adjuvant TMZ treatment varied widely between centres. At the end of the adjuvant period, most centres performed an MRI (71%, 22/31), followed by monitoring scans at 3 monthly intervals (81%, 25/31). Additional short-interval imaging was carried out in cases of possible pseudoprogression in most centres (71%, 22/31). Routine use of advanced imaging was infrequent; however, the addition of advanced sequences was the most popular suggestion for ideal imaging practice, followed by changes in the timing of EPMRI. CONCLUSION: Variations in neuroimaging practices exist after initial glioblastoma treatment within the UK. Multicentre, longitudinal, prospective trials are needed to define the optimal imaging schedule for assessment. KEY POINTS: • Variations in imaging practices exist in the frequency, timing and type of interval neuroimaging after initial treatment of glioblastoma within the UK. • Large, multicentre, longitudinal, prospective trials are needed to define the optimal imaging schedule for assessment.