Development of Volumetric MR Spectroscopy for Spatial Metabolic Imaging of the Brain

Development of Volumetric MR Spectroscopy for Spatial Metabolic Imaging of the BrainGjennomført midtveisevaluering i PhD-løp. Er tidligere nesten ferdig med alle PhD-kurs. Innhentet og bearbeidet data. Sendt en artikkel til tidsskrift.Kandidaten var sykmeldt en stor del av vår og sommer, ble mor i høst og er nå i mammapermisjon. Det ble gjennomført halvtidsevaluering av PhD tidlig i 2023. Første del av året: Gjennomført PhD-kurs, er nå enten ferdig eller mangler mindre enn 5 studiepoeng. Midtveisevaluering gjennomført. Mars-oktober sykmeldt. Deretter i mammapermisjon. Det er sendt inn en artikkel basert på arbeidet hun hadde gjort før hun ble sykmeldt.

Develop an Artificial Intelligence (AI)-based algorithm to analyze spectroscopic dataWe have examined deep learning algorithms to analyze magnetic resonance spectroscopic imaging (MRSI) and magnetic resonance (MR) spectroscopy data, focusing on detecting isocitrate dehydrogenase (IDH) mutations.We have continued working on deep learning algorithms for processing and analyzing data from magnetic resonance (MR) spectroscopy, focusing on detecting IDH mutations. This resulted in a conference paper, «Domain-Transformation of MRS-derived Time Series with Deep Learning» which has been submitted for ISMRM 2023. We have collaborated with an AI-based industrial partner in Germany. This collaboration aimed to explore state-of-the-art AI algorithms to identify and classify MR spectra acquired from different subjects and institutions. In particular, we aim to detect cancer-origin spectra from healthy-origin MR spectra. This task resulted in one conference paper, «A Debiased Autoencoder for Glioma Detection in Spectroscopic MRI» for AI Innovations 2022, presented by the Ph.D. candidate at the conference. The Ph.D. candidate performed this study in Berlin during a research stay of six months. In addition, the candidate submitted more updated and refined experiments to two international conferences: ISMRM and OHBM.

Spatial Metabolic Imaging of the BrainDevelop an Artificial Intelligence (AI)-based platform for analyzing magnetic resonance spectroscopic imaging (MRSI) data automatically We have examined deep learning algorithms for processing and analyzing data from magnetic resonance (MR) spectroscopy, focusing on detecting isocitrate dehydrogenase (IDH) mutations.We have examined deep learning algorithms for processing and analyzing data from MR spectroscopy, focusing on detecting IDH mutations. We aimed to develop an automated MRSI analysis platform in the first task. This has resulted in two conference abstracts, “Domain- Transformation of MRI-derived Time Series with Deep Learning” for Nordic AI meet 2021; “A Deep Neural Network for Detection of Glioblastomas in Spectroscopic MRI ” for ISMRM 2022, the first of which have been presented at the conference. We have implemented deep learning algorithms for processing and analyzing data from magnetic resonance (MR) spectroscopy, focusing on detecting IDH mutations. This is part of the main task (to develop automated MRSI analysis). This has resulted in two conference abstracts, “Domain- Transformation of MRI-derived Time Series with Deep Learning” for Nordic AI meet 2021; “A Deep Neural Network for Detection of Glioblastomas in Spectroscopic MRI ” for ISMRM 2022, the first of which have been presented at the conference. We have also worked on the segmentation of brain tumors in MR images using deep learning to detect specific mutations in the tumor. This resulted in one conference abstract, “A Deep-Learning Based Glioma Segmentation with Classification of IDH Mutation Status, “submitted to the human brain mapping 2022. We have started a collaboration with an AI-based industrial partner in Germany. We aim to develop deep learning algorithms with specific focus on medical domains.