Explainable hybrid vision transformers and convolutional network for multimodal glioma segmentation in brain MRI.

Accurate localization of gliomas, the most common malignant primary brain cancer, and its different sub-region from multimodal magnetic resonance imaging (MRI) volumes are highly important for interventional procedures. Recently, deep learning models have been applied widely to assist automatic lesion segmentation tasks for neurosurgical interventions. However, these models are often complex and represented as "black box" models which limit their applicability in clinical practice. This article introduces new hybrid vision Transformers and convolutional neural networks for accurate and robust glioma segmentation in Brain MRI scans. Our proposed method, TransXAI, provides surgeon-understandable heatmaps to make the neural networks transparent. TransXAI employs a post-hoc explanation technique that provides visual interpretation after the brain tumor localization is made without any network architecture modifications or accuracy tradeoffs. Our experimental findings showed that TransXAI achieves competitive performance in extracting both local and global contexts in addition to generating explainable saliency maps to help understand the prediction of the deep network. Further, visualization maps are obtained to realize the flow of information in the internal layers of the encoder-decoder network and understand the contribution of MRI modalities in the final prediction. The explainability process could provide medical professionals with additional information about the tumor segmentation results and therefore aid in understanding how the deep learning model is capable of processing MRI data successfully. Thus, it enables the physicians' trust in such deep learning systems towards applying them clinically. To facilitate TransXAI model development and results reproducibility, we will share the source code and the pre-trained models after acceptance at https://github.com/razeineldin/TransXAI .

Evaluation of a Navigated 3D Ultrasound Integration for Brain Tumor Surgery: First Results of an Ongoing Prospective Study.

The aim of the study was to assess the quality, accuracy and benefit of navigated 2D and 3D ultrasound for intra-axial tumor surgery in a prospective study. Patients intended for gross total resection were consecutively enrolled. Intraoperatively, a 2D and 3D iUS-based resection was performed. During surgery, the image quality, clinical benefit and navigation accuracy were recorded based on a standardized protocol using Likert's scales. A total of 16 consecutive patients were included. Mean ratings of image quality in 2D iUS were significantly higher than in 3D iUS (p < 0.001). There was no relevant decrease in rating during the surgery in 2D and 3D iUS (p > 0.46). The benefit was rated 2.2 in 2D iUS and 2.6 in 3D iUS (p = 0.08). The benefit remained stable in 2D, while there was a slight decrease in the benefit in 3D after complete tumor resection (p = 0.09). The accuracy was similar in both (mean 2.2 p = 0.88). Seven patients had a small tumor remnant in intraoperative MRT (mean 0.98 cm3) that was not appreciated with iUS. Crucially, 3D iUS allows for an accurate intraoperative update of imaging with slightly lower image quality than 2D iUS. Our preliminary data suggest that the benefit and accuracy of 2D and 3D iUS navigation do not undergo significant variations during tumor resection.

Explainability of deep neural networks for MRI analysis of brain tumors.

PURPOSE: Artificial intelligence (AI), in particular deep neural networks, has achieved remarkable results for medical image analysis in several applications. Yet the lack of explainability of deep neural models is considered the principal restriction before applying these methods in clinical practice. METHODS: In this study, we propose a NeuroXAI framework for explainable AI of deep learning networks to increase the trust of medical experts. NeuroXAI implements seven state-of-the-art explanation methods providing visualization maps to help make deep learning models transparent. RESULTS: NeuroXAI has been applied to two applications of the most widely investigated problems in brain imaging analysis, i.e., image classification and segmentation using magnetic resonance (MR) modality. Visual attention maps of multiple XAI methods have been generated and compared for both applications. Another experiment demonstrated that NeuroXAI can provide information flow visualization on internal layers of a segmentation CNN. CONCLUSION: Due to its open architecture, ease of implementation, and scalability to new XAI methods, NeuroXAI could be utilized to assist radiologists and medical professionals in the detection and diagnosis of brain tumors in the clinical routine of cancer patients. The code of NeuroXAI is publicly accessible at https://github.com/razeineldin/NeuroXAI .

A phase Ib/IIa trial of 9 repurposed drugs combined with temozolomide for the treatment of recurrent glioblastoma: CUSP9v3.

BACKGROUND: The dismal prognosis of glioblastoma (GBM) may be related to the ability of GBM cells to develop mechanisms of treatment resistance. We designed a protocol called Coordinated Undermining of Survival Paths combining 9 repurposed non-oncological drugs with metronomic temozolomide-version 3-(CUSP9v3) to address this issue. The aim of this phase Ib/IIa trial was to assess the safety of CUSP9v3. METHODS: Ten adults with histologically confirmed GBM and recurrent or progressive disease were included. Treatment consisted of aprepitant, auranofin, celecoxib, captopril, disulfiram, itraconazole, minocycline, ritonavir, and sertraline added to metronomic low-dose temozolomide. Treatment was continued until toxicity or progression. Primary endpoint was dose-limiting toxicity defined as either any unmanageable grade 3-4 toxicity or inability to receive at least 7 of the 10 drugs at ≥ 50% of the per-protocol doses at the end of the second treatment cycle. RESULTS: One patient was not evaluable for the primary endpoint (safety). All 9 evaluable patients met the primary endpoint. Ritonavir, temozolomide, captopril, and itraconazole were the drugs most frequently requiring dose modification or pausing. The most common adverse events were nausea, headache, fatigue, diarrhea, and ataxia. Progression-free survival at 12 months was 50%. CONCLUSIONS: CUSP9v3 can be safely administered in patients with recurrent GBM under careful monitoring. A randomized phase II trial is in preparation to assess the efficacy of the CUSP9v3 regimen in GBM.