Two Cases of Remimazolam Anesthesia Managed With Pharmacokinetic Simulations in an Awake Craniotomy of Patients With Obesity.

Anesthetic management for awake craniotomy (AC) often poses problems in patients with obesity including respiratory management. The Japan Awake Craniotomy Guidelines state that the indication for surgery should be carefully considered, especially in patients with obesity (body mass index (BMI) > 30). Patients with obesity often have comorbidities such as dyslipidemia, hypertension, and type 2 diabetes, and this is a risk factor for perioperative morbidity and mortality. Remimazolam, a new intravenous anesthetic, has been reported to be useful in anesthesia management during AC, but its use in obese patients has not been reported. Herein, we report two cases case series in which remimazolam was used in patients with obesity and safely managed under anesthesia with the pharmacokinetic simulations.

Fertility preservation in men with hypogonadotropic hypogonadism secondary to germinoma: two cases of gonadotropin replacement therapy before induction of anticancer chemotherapy.

Hypogonadotropic hypogonadism can be caused by brain tumors. For a malignancy such as a germ cell tumor, chemotherapy combined with radiation is administered. In patients who wish for children, the inability to undergo sperm cryopreservation before treatment because of impaired spermatogenesis and/or ejaculation dysfunction can be problematic. We herein present two cases involving a 26-year-old man and a 30-year-old man with hypogonadotropic hypogonadism due to an intracranial germinoma and both wished to have children. Gonadotropin replacement therapy prior to anticancer chemotherapy resulted in subsequent spontaneous pregnancy or assisted reproductive therapy. Subsequent treatment of the tumor resulted in no recurrence for 9 and 2 years, respectively. Close consultation with an oncologist is mandatory in such cases. Depending on the tumor prognosis, however, it may be possible to delay tumor treatment and prioritize fertility because there is a possibility of impaired spermatogenesis due to additional chemotherapy.

The transformative potential of mRNA vaccines for glioblastoma and human cancer: technological advances and translation to clinical trials.

Originally devised for cancer control, mRNA vaccines have risen to the forefront of medicine as effective instruments for control of infectious disease, notably their pivotal role in combating the COVID-19 pandemic. This review focuses on fundamental aspects of the development of mRNA vaccines, e.g., tumor antigens, vector design, and precise delivery methodologies, - highlighting key technological advances. The recent, promising success of personalized mRNA vaccines against pancreatic cancer and melanoma illustrates the potential value for other intractable, immunologically resistant, solid tumors, such as glioblastoma, as well as the potential for synergies with a combinatorial, immunotherapeutic approach. The impact and progress in human cancer, including pancreatic cancer, head and neck cancer, bladder cancer are reviewed, as are lessons learned from first-in-human CAR-T cell, DNA and dendritic cell vaccines targeting glioblastoma. Going forward, a roadmap is provided for the transformative potential of mRNA vaccines to advance cancer immunotherapy, with a particular focus on the opportunities and challenges of glioblastoma. The current landscape of glioblastoma immunotherapy and gene therapy is reviewed with an eye to combinatorial approaches harnessing RNA science. Preliminary preclinical and clinical data supports the concept that mRNA vaccines could be a viable, novel approach to prolong survival in patients with glioblastoma.

Mesenchymal-type genetic mutations are likely pre-requisite for glioblastoma multiforme to metastasize outside the central nervous system: an original case series and systematic review of the literature.

BACKGROUND: Glioblastoma multiforme (GBM) is the most aggressive and prevalent type of malignant brain tumor, yet they metastasize outside of the central nervous system (CNS) in 0.4% of all cases. Little is known about what enables this subset of GBMs to take root outside the CNS, but genetic mutations likely play a role. METHODS: We conducted a PRISMA-compliant systematic review of metastatic GBM wherein we reviewed 3579 search results and 1080 abstracts, ultimately analyzing data from 139 studies and 211 unique patients. Additionally, we describe four cases of patients with pathologically confirmed GBM metastases outside the CNS treated at our institution. RESULTS: We found that metastases were discovered near previous surgical sites in at least 36.9% of cases. Other sites of metastasis included bone (47.9%), lung (25.6%), lymph nodes (25.1%), scalp (19.2%), and liver (14.2%). On average, metastases were diagnosed 12.1 months after the most recent resection, and the mean survival from discovery was 5.7 months. In our patients, primary GBM lesions revealed mutations in NF1, TERT, TP53, CDK4, and RB1/PTEN genes. Unique to the metastatic lesions were amplifications in genes such as p53 and PDGFRA/KIT, as well as increased vimentin and Ki-67 expression. CONCLUSIONS: In sum, there is strong evidence that GBMs acquire novel mutations to survive outside the CNS. In some cases, tumor cells likely mutate after seeding scalp tissue during surgery, and in others, they mutate and spread without surgery. Future studies and genetic profiling of primary and metastatic lesions may help uncover the mechanisms of spread.

Arterial to jugular-bulb lactate difference in patients undergoing elective brain tumor craniotomy.

Hyperlactatemia is common during tumor craniotomy, but the underlying pathophysiology is unclear. This study measured simultaneous arterial and jugular-bulb lactate concentrations in patients undergoing brain tumor craniotomy to investigate the hypothesis that hyperlactatemia was associated with a net cerebrovascular lactate input. In 20 patients, arterial and jugular-bulb blood was collected hourly from the start of surgery to 6 h postoperatively for measurement of lactate, glucose, and oxygen concentration. For each marker, data were analyzed using a linear mixed-effects model with jugular-bulb concentration as dependent variable, arterial concentration as fixed effect, and patient as random effect. Furthermore, we generated regression lines between arterial and jugular-bulb concentrations. The slope of the regression line between arterial and jugular-bulb lactate was 0.95 (95% CI 0.93-0.97, R2 = 0.98), indicating that increasing arterial lactate levels were associated with an increasingly positive net cerebrovascular balance (net input). The line crossed the identity line at 2.86 (95% CI 0.57-5.16) mmol/L, indicating that lower levels of lactate were associated with a negative net cerebrovascular balance (net output). This suggests a switch from net lactate output during normolactatemia towards net input during hyperlactatemia. Hyperlactatemia in tumor-craniotomy patients probably does not originate from the brain.

Dilated multi-scale residual attention (DMRA) U-Net: three-dimensional (3D) dilated multi-scale residual attention U-Net for brain tumor segmentation.

Background: The precise identification of the position and form of a tumor mass can improve early diagnosis and treatment. However, due to the complicated tumor categories and varying sizes and forms, the segregation of brain gliomas and their internal sub-regions is still very challenging. This study sought to design a new deep-learning network based on three-dimensional (3D) U-Net to address its shortcomings in brain tumor segmentation (BraTS) tasks. Methods: We developed a 3D dilated multi-scale residual attention U-Net (DMRA-U-Net) model for magnetic resonance imaging (MRI) BraTS. It used dilated convolution residual (DCR) modules to better process shallow features, multi-scale convolution residual (MCR) modules in the bottom encoding path to create richer and more comprehensive feature expression while reducing overall information loss or blurring, and a channel attention (CA) module between the encoding and decoding paths to address the problem of retrieving and preserving important features during the processing of deep feature maps. Results: The BraTS 2018-2021 datasets served as the training and evaluation datasets for this study. Further, the proposed architecture was assessed using metrics such as the dice similarity coefficient (DSC), Hausdorff distance (HD), and sensitivity (Sens). The DMRA U-Net model segments the whole tumor (WT), and the tumor core (TC), and the enhancing tumor (ET) regions of brain tumors. Using the suggested architecture, the DSCs were 0.9012, 0.8867, and 0.8813, the HDs were 28.86, 13.34, and 10.88 mm, and the Sens was 0.9429, 0.9452, and 0.9303 for the WT, TC, and ET regions, respectively. Compared to the traditional 3D U-Net, the DSC of the DMRA U-Net increased by 4.5%, 2.5%, and 0.8%, the HD of the DMRA U-Net decreased by 21.83, 16.42, and 10.00, the Sens of the DMRA U-Net increased by 0.4%, 0.7%, and 1.4% for the WT, TC, and ET regions, respectively. Further, the results of the statistical comparison of the performance indicators revealed that our model performed well generally in the segmentation of the WT, TC, and ET regions. Conclusions: We developed a promising tumor segmentation model. Our solution is open sourced and is available at: https://github.com/Gold3nk/dmra-unet.

Preoperative nTMS analysis: a sensitive tool to detect imminent motor deficits in brain tumor patients.

BACKGROUND: One of the challenges in surgery of tumors in motor eloquent areas is the individual risk assessment for postoperative motor disorder. Previously a regression model was developed that permits estimation of the risk prior to surgery based on topographical and neurophysiological data derived from investigation with nTMS (navigated Transcranial Magnetic Stimulation). This study aims to analyze the impact of including additional neurophysiological TMS parameters into the established risk stratification model for motor outcome after brain tumor surgery. METHODS: Biometric and clinical data of 170 patients with glioma in motor eloquent areas were collected prospectively. In addition, the following nTMS parameters were collected bihemispherically prior to surgery: resting motor threshold (RMT), recruitment curve (RC), cortical silent period (CSP) and a nTMS based fibertracking to measure the tumor tract distance (TTD). Motor function was quantified by Medical Research Council Scale (MRCS) preoperatively, seven days and three months postoperatively. Association between nTMS parameters and postoperative motor outcome was investigated in bivariate and multivariable analyses. RESULTS: The bivariate analysis confirmed the association of RMT ratio with the postoperative motor outcome after seven days with higher rates of worsening in patients with RMT ratio > 1.1 compared to patients with RMT ratio ≤ 1.1 (31.6% vs. 15.1%, p = 0.009). Similarly, an association between a pathological CSP ratio and a higher risk of new postoperative motor deficits after seven days was observed (35.3% vs. 16.7% worsening, p = 0.025). A pathological RC Ratio was associated postoperative deterioration of motor function after three months (42.9% vs. 16.2% worsening, p = 0.004). In multiple regression analysis, none of these associations were statistically robust. CONCLUSIONS: The current results suggest that the RC ratio, CSP ratio and RMT ratio individually are sensitive markers associated with the motor outcome 7 days and 3 months after tumor resection in a presumed motor eloquent location. They can therefore supply valuable information during preoperative risk-benefit-balancing. However, underlying neurophysiological mechanisms might be too similar to make the parameters meaningful in a combined model.

MALT1 protease inhibition restrains glioblastoma progression by reversing tumor-associated macrophage-dependent immunosuppression.

MALT1 protease is an intracellular signaling molecule that promotes tumor progression via cancer cell-intrinsic and cancer cell-extrinsic mechanisms. MALT1 has been mostly studied in lymphocytes, and little is known about its role in tumor-associated macrophages. Here, we show that MALT1 plays a key role in glioblastoma (GBM)-associated macrophages. Mechanistically, GBM tumor cells induce a MALT1-NF-κB signaling axis within macrophages, leading to macrophage migration and polarization toward an immunosuppressive phenotype. Inactivation of MALT1 protease promotes transcriptional reprogramming that reduces migration and restores a macrophage "M1-like" phenotype. Preclinical in vivo analysis shows that MALT1 inhibitor treatment results in increased immuno-reactivity of GBM-associated macrophages and reduced GBM tumor growth. Further, the addition of MALT1 inhibitor to temozolomide reduces immunosuppression in the tumor microenvironment, which may enhance the efficacy of this standard-of-care chemotherapeutic. Together, our findings suggest that MALT1 protease inhibition represents a promising macrophage-targeted immunotherapeutic strategy for the treatment of GBM.

Recent advances in Tumor Treating Fields (TTFields) therapy for glioblastoma.

Tumor Treating Fields (TTFields) therapy is a locoregional, anticancer treatment consisting of a noninvasive, portable device that delivers alternating electric fields to tumors through arrays placed on the skin. Based on efficacy and safety data from global pivotal (randomized phase III) clinical studies, TTFields therapy (Optune Gio) is US Food and Drug Administration-approved for newly diagnosed (nd) and recurrent glioblastoma (GBM) and Conformité Européenne-marked for grade 4 glioma. Here we review data on the multimodal TTFields mechanism of action that includes disruption of cancer cell mitosis, inhibition of DNA replication and damage response, interference with cell motility, and enhancement of systemic antitumor immunity (adaptive immunity). We describe new data showing that TTFields therapy has efficacy in a broad range of patients, with a tolerable safety profile extending to high-risk subpopulations. New analyses of clinical study data also confirmed that overall and progression-free survival positively correlated with increased usage of the device and dose of TTFields at the tumor site. Additionally, pilot/early phase clinical studies evaluating TTFields therapy in ndGBM concomitant with immunotherapy as well as radiotherapy have shown promise, and new pivotal studies will explore TTFields therapy in these settings. Finally, we review recent and ongoing studies in patients in pediatric care, other central nervous system tumors and brain metastases, as well as other advanced-stage solid tumors (ie, lung, ovarian, pancreatic, gastric, and hepatic cancers), that highlight the broad potential of TTFields therapy as an adjuvant treatment in oncology.

Radiation therapy communication: Retreatment of multiple histologically confirmed oligodendroglioma drop metastasis with stereotactic radiotherapy in a dog.

A 10-year-old female spayed boxer was treated with stereotactic radiotherapy (SRT) for a suspected glioma in the left piriform lobe. The intra-axial lesion was T2 hyperintense, T2 FLAIR hyperintense, T1 hypointense, and did not uptake contrast. Imaging was performed with an MRI every 3 months, and at the 6-month recheck, new lesions in the left hippocampus and right piriform lobe were evident without clinically apparent neurological progression. A second course of SRT was prescribed for the new lesions. Euthanasia was elected 14 months after the first course of SRT, and necropsy confirmed oligodendroglioma with drop metastasis.

Relationship between radiation dose and cerebral microbleed formation in dogs with intracranial tumors.

BACKGROUND: Cerebral microbleeds (CMBs) are a possible sequela in human brain tumor patients treated with radiation therapy (RT). No such association is reported in dogs. OBJECTIVES: To investigate whether CMBs occur in dogs after radiotherapy, and if there is an association between number and dose, and an increase over time. ANIMALS: Thirty-four client-owned dogs irradiated for primary intracranial neoplasia. ≥2 magnetic resonance imaging (MRI) scans including susceptibility-weighted imaging (SWI) were required. METHODS: Retrospective, observational, single-center study. Cerebral microbleeds identified on 3 T SWI were counted within the entire brain, and within low- (<20 Gy), intermediate- (20-30 Gy), and high- (>30 Gy) dose regions. A generalized linear mixed-effects model was used to analyze the relationship between the CMBs count and the predictor variables (irradiation dose, time after treatment). RESULTS: Median follow-up time was 12.6 months (range, 1.8-37.6 months). Eighty-three MR scans were performed. In 4/15 dogs (27%, 95% CI, 10%-52%) CMBs were present at baseline. ≥1 CMBs after RT were identified in 21/34 dogs (62%, 95% CI, 45%-77%). With each month, the number of CMBs increased by 14% (95% CI, 11%-16%; P < .001). The odds of developing CMBs in the high-dose region are 4.7 times (95% CI, 3.9-5.6; P < .001) greater compared with the low-dose region. CONCLUSION AND CLINICAL IMPORTANCE: RT is 1 possible cause of CMBs formation in dogs. Cerebral microbleeds are most likely to occur in the peritumoral high-dose volume, to be chronic, and to increase in number over time. Their clinical relevance remains unknown.

The roles of glioma-associated macrophages/microglia and potential targets for anti-glioma therapy.

Glioblastoma (GBM) is the central nervous system tumor with the most aggressive behavior, and no definitive therapy has yet been found. The tumor microenvironment of GBM is immunosuppressive and is considered a 'cold tumor' with low lymphocytic infiltration, but is characterized by a high proportion of glioma-associated macrophages/microglia (GAMs). GAMs promote tumor growth and also affect treatment resistance in GBM. In this review, we describe the origin and classification of GAMs in humans and describe the mechanisms of their activation and the cell-cell interactions between tumor cells and GAMs. We also describe the history of GAM detection methods, especially immunohistochemistry, and discusses the merits and limitations of these techniques. In addition, we summarized chemotactic factors for GAMs and the therapies targeting these factors. Recent single-cell RNA analysis and spatial analysis add new insights to our previous knowledge of GAMs. Based on these studies, GBM therapies targeting GAMs are expected to be further developed.

Biophysical and Biological Mechanisms of Tumor Treating Fields in Glioblastoma.

Glioblastoma (GBM) is one of the most aggressive forms of brain cancer that presents with a median survival rate of 14-30 months and along with a discouraging five-year survival rate of 4-5%. Standard treatment of newly diagnosed GBM, also known as the Stupp protocol, includes a maximally safe surgical resection followed by radiation and chemotherapy. Despite these treatment regimens, recurrence is almost inevitable, emphasizing the need for new therapies to combat the aggressive nature of GBMs. Tumor Treating Fields (TTFs) are a relatively new application to the treatment of GBMs, and results have been promising with both progression-free survival and overall survival when TTFs have been used in combination with temozolomide. This article critically reviews the biophysical and biological mechanisms of TTFs, their clinical efficacy, and discusses the results in clinical trials, including EF-11 and EF-14. Both trials have demonstrated that TTFs can enhance progression free survival and overall survival without compromising quality of life or causing severe adverse effects. Despite the high cost associated with TTFs and the need for further analysis to determine the most effective ways to integrate TTFs into GBM treatments, TTFs represent a significant advancement in GBM therapy and offer hope for improved patient prognosis.

Obstructive hydrocephalus due to choroid plexus carcinoma of third ventricle in pediatric: A rare case report.

Choroid plexus carcinoma (CPC) is an uncommon tumor that accounts for less than 1% of all pediatric brain tumors. CPC usually originates in the lateral ventricle, followed by the fourth ventricle; the incidence in the third ventricle is only 5% of all CPC cases (children and adults). We report an extremely rare tumor arising from the choroid plexus of the third ventricle in a 6-year-old child with progressive headache, macrocephaly, left hemiparesis, and sunset eyes. The imaging found a well-defined, lobulated mass with strong enhancement in the posterior part of the third ventricle, resulting in obstructive hydrocephalus. The patient underwent an endoscopic biopsy and histopathological examination, which resulted in choroid plexus carcinoma.

Personalized mRNA vaccines in glioblastoma therapy: from rational design to clinical trials.

Glioblastomas (GBMs) are the most common and aggressive malignant brain tumors, presenting significant challenges for treatment due to their invasive nature and localization in critical brain regions. Standard treatment includes surgical resection followed by radiation and adjuvant chemotherapy with temozolomide (TMZ). Recent advances in immunotherapy, including the use of mRNA vaccines, offer promising alternatives. This review focuses on the emerging use of mRNA vaccines for GBM treatment. We summarize recent advancements, evaluate current obstacles, and discuss notable successes in this field. Our analysis highlights that while mRNA vaccines have shown potential, their use in GBM treatment is still experimental. Ongoing research and clinical trials are essential to fully understand their therapeutic potential. Future developments in mRNA vaccine technology and insights into GBM-specific immune responses may lead to more targeted and effective treatments. Despite the promise, further research is crucial to validate and optimize the effectiveness of mRNA vaccines in combating GBM.

Brain tumor detection and classification in MRI using hybrid ViT and GRU model with explainable AI in Southern Bangladesh.

Brain tumor, a leading cause of uncontrolled cell growth in the central nervous system, presents substantial challenges in medical diagnosis and treatment. Early and accurate detection is essential for effective intervention. This study aims to enhance the detection and classification of brain tumors in Magnetic Resonance Imaging (MRI) scans using an innovative framework combining Vision Transformer (ViT) and Gated Recurrent Unit (GRU) models. We utilized primary MRI data from Bangabandhu Sheikh Mujib Medical College Hospital (BSMMCH) in Faridpur, Bangladesh. Our hybrid ViT-GRU model extracts essential features via ViT and identifies relationships between these features using GRU, addressing class imbalance and outperforming existing diagnostic methods. We extensively processed the dataset, and then trained the model using various optimizers (SGD, Adam, AdamW) and evaluated through rigorous 10-fold cross-validation. Additionally, we incorporated Explainable Artificial Intelligence (XAI) techniques-Attention Map, SHAP, and LIME-to enhance the interpretability of the model's predictions. For the primary dataset BrTMHD-2023, the ViT-GRU model achieved precision, recall, and F1-score metrics of 97%. The highest accuracies obtained with SGD, Adam, and AdamW optimizers were 81.66%, 96.56%, and 98.97%, respectively. Our model outperformed existing Transfer Learning models by 1.26%, as validated through comparative analysis and cross-validation. The proposed model also shows excellent performances with another Brain Tumor Kaggle Dataset outperforming the existing research done on the same dataset with 96.08% accuracy. The proposed ViT-GRU framework significantly improves the detection and classification of brain tumors in MRI scans. The integration of XAI techniques enhances the model's transparency and reliability, fostering trust among clinicians and facilitating clinical application. Future work will expand the dataset and apply findings to real-time diagnostic devices, advancing the field.

Heterogeneous biomechanical/mathematical modeling of spatial prediction of glioblastoma progression using magnetic resonance imaging-based finite element method.

BACKGROUND AND OBJECTIVE: Brain tumors are one of the most common diseases and causes of death in humans. Since the growth of brain tumors has irreparable risks for the patient, predicting the growth of the tumor and knowing its effect on the brain tissue will increase the efficiency of treatment strategies. METHODS: This study examines brain tumor growth using mathematical modeling based on the Reaction-Diffusion equation and the biomechanical model based on continuum mechanics principles. With the help of the image threshold technique of magnetic resonance images, a heterogeneous and close-to-reality environment of the brain has been modeled and experimental data validated the results to achieve maximum accuracy in predicting growth. RESULTS: The obtained results have been compared with the reported conventional models to evaluate the presented model. In addition to incorporating the chemotherapy effects in governing equations, the real-time finite element analysis of the stress tensors of the surrounding tissue of tumor cells and considering its role in changing the shape and growth of the tumor has added to the importance and accuracy of the current model. CONCLUSIONS: The comparison of the obtained results with conventional models shows that the heterogeneous model has higher reliability due to the consideration of the appropriate properties for the different regions of the brain. The presented model can contribute to personalized medicine, aid in understanding the dynamics of tumor growth, optimize treatment regimens, and develop adaptive therapy strategies.

Association between hypertension requiring medication and postoperative 30-day mortality in adult patients with tumor craniotomy: an analysis of data using propensity score matching.

Background: Reliable quantification of the association between hypertension requiring medication and postoperative 30-day mortality in adult patients who undergo craniotomy for tumor resection is limited. We aimed to explore the associations between these factors. Materials and methods: This work was a retrospective cohort study that used propensity score matching (PSM) among 18,642 participants from the American College of Surgeons National Surgical Quality Improvement Program database between 2012 and 2015. Hypertension requiring medication and postoperative 30-day mortality were the independent and dependent target variables, respectively. PSM was conducted via nonparsimonious multivariate logistic regression to balance the confounders. Robust estimation methods were used to investigate the association between hypertension requiring medication and postoperative 30-day mortality. Results: A total of 18,642 participants (52.6% male and 47.4% female) met our inclusion criteria; 7,116 (38.17%) participants with hypertension required medication and had a 3.74% mortality rate versus an overall mortality rate of 2.46% in the adult cohort of patients who underwent craniotomy for tumor resection. In the PSM cohort, the risk of postoperative 30-day mortality significantly increased by 39.0% among patients with hypertension who required medication (OR = 1.390, 95% confidence interval (CI): 1.071-1.804, p = 0.01324) after adjusting for the full covariates. Compared with participants without hypertension requiring medication, those with hypertension requiring medication had a 34.0% greater risk of postoperative 30-day mortality after adjusting for the propensity score (OR = 1.340, 95% CI: 1.040-1.727, p = 0.02366) and a 37.6% greater risk of postoperative 30-day mortality in the inverse probability of treatment weights (IPTW) cohort (OR = 1.376, 95% CI: 1.202, 1.576, p < 0.00001). Conclusion: Among U.S. adult patients undergoing craniotomy for tumor resection, hypertension requiring medication is a notable contributor to 30-day mortality after surgery, with odds ratios ranging from 1.34 to 1.39.

Multimodal brain tumor segmentation and classification from MRI scans based on optimized DeepLabV3+ and interpreted networks information fusion empowered with explainable AI.

Explainable artificial intelligence (XAI) aims to offer machine learning (ML) methods that enable people to comprehend, properly trust, and create more explainable models. In medical imaging, XAI has been adopted to interpret deep learning black box models to demonstrate the trustworthiness of machine decisions and predictions. In this work, we proposed a deep learning and explainable AI-based framework for segmenting and classifying brain tumors. The proposed framework consists of two parts. The first part, encoder-decoder-based DeepLabv3+ architecture, is implemented with Bayesian Optimization (BO) based hyperparameter initialization. The different scales are performed, and features are extracted through the Atrous Spatial Pyramid Pooling (ASPP) technique. The extracted features are passed to the output layer for tumor segmentation. In the second part of the proposed framework, two customized models have been proposed named Inverted Residual Bottleneck 96 layers (IRB-96) and Inverted Residual Bottleneck Self-Attention (IRB-Self). Both models are trained on the selected brain tumor datasets and extracted features from the global average pooling and self-attention layers. Features are fused using a serial approach, and classification is performed. The BO-based hyperparameters optimization of the neural network classifiers is performed and the classification results have been optimized. An XAI method named LIME is implemented to check the interpretability of the proposed models. The experimental process of the proposed framework was performed on the Figshare dataset, and an average segmentation accuracy of 92.68 % and classification accuracy of 95.42 % were obtained, respectively. Compared with state-of-the-art techniques, the proposed framework shows improved accuracy.

Brain tumor grade classification using the ConvNext architecture.

Objective: Brain tumor grade is an important aspect of brain tumor diagnosis and helps to plan for treatment. Traditional methods of diagnosis, including biopsy and manual examination of medical images, are either invasive or may result in inaccurate diagnoses. This study proposes a brain tumor grade classification technique using a modern convolutional neural network (CNN) architecture called ConvNext that inputs magnetic resonance imaging (MRI) data. Methods: Deep learning-based techniques are replacing invasive procedures for consistent, accurate, and non-invasive diagnosis of brain tumors. A well-known challenge of using deep learning architectures in medical imaging is data scarcity. Modern-day architectures have huge trainable parameters and require massive datasets to achieve the desired accuracy and avoid overfitting. Therefore, transfer learning is popular among researchers using medical imaging data. Recently, transformer-based architectures have surpassed CNNs for image data. However, recently proposed CNNs have achieved superior accuracy by introducing some tweaks inspired by vision transformers. This study proposed a technique to extract features from the ConvNext architecture and feed these features to a fully connected neural network for final classification. Results: The proposed study achieved state-of-the-art performance on the BraTS 2019 dataset using pre-trained ConvNext. The best accuracy of 99.5% was achieved when three MRI sequences were input as three channels of the pre-trained CNN. Conclusion: The study demonstrated the efficacy of the representations learned by a modern CNN architecture, which has a higher inductive bias for the image data than vision transformers for brain tumor grade classification.