Deformable registration of magnetic resonance images using unsupervised deep learning in neuro-/radiation oncology.

PURPOSE: Accurate deformable registration of magnetic resonance imaging (MRI) scans containing pathologies is challenging due to changes in tissue appearance. In this paper, we developed a novel automated three-dimensional (3D) convolutional U-Net based deformable image registration (ConvUNet-DIR) method using unsupervised learning to establish correspondence between baseline pre-operative and follow-up MRI scans of patients with brain glioma. METHODS: This study involved multi-parametric brain MRI scans (T1, T1-contrast enhanced, T2, FLAIR) acquired at pre-operative and follow-up time for 160 patients diagnosed with glioma, representing the BraTS-Reg 2022 challenge dataset. ConvUNet-DIR, a deep learning-based deformable registration workflow using 3D U-Net style architecture as a core, was developed to establish correspondence between the MRI scans. The workflow consists of three components: (1) the U-Net learns features from pairs of MRI scans and estimates a mapping between them, (2) the grid generator computes the sampling grid based on the derived transformation parameters, and (3) the spatial transformation layer generates a warped image by applying the sampling operation using interpolation. A similarity measure was used as a loss function for the network with a regularization parameter limiting the deformation. The model was trained via unsupervised learning using pairs of MRI scans on a training data set (n = 102) and validated on a validation data set (n = 26) to assess its generalizability. Its performance was evaluated on a test set (n = 32) by computing the Dice score and structural similarity index (SSIM) quantitative metrics. The model's performance also was compared with the baseline state-of-the-art VoxelMorph (VM1 and VM2) learning-based algorithms. RESULTS: The ConvUNet-DIR model showed promising competency in performing accurate 3D deformable registration. It achieved a mean Dice score of 0.975 ± 0.003 and SSIM of 0.908 ± 0.011 on the test set (n = 32). Experimental results also demonstrated that ConvUNet-DIR outperformed the VoxelMorph algorithms concerning Dice (VM1: 0.969 ± 0.006 and VM2: 0.957 ± 0.008) and SSIM (VM1: 0.893 ± 0.012 and VM2: 0.857 ± 0.017) metrics. The time required to perform a registration for a pair of MRI scans is about 1 s on the CPU. CONCLUSIONS: The developed deep learning-based model can perform an end-to-end deformable registration of a pair of 3D MRI scans for glioma patients without human intervention. The model could provide accurate, efficient, and robust deformable registration without needing pre-alignment and labeling. It outperformed the state-of-the-art VoxelMorph learning-based deformable registration algorithms and other supervised/unsupervised deep learning-based methods reported in the literature.

MR-based radiomics predictive modelling of EGFR mutation and HER2 overexpression in metastatic brain adenocarcinoma: a two-centre study.

OBJECTIVES: Magnetic resonance (MR)-based radiomics features of brain metastases are utilised to predict epidermal growth factor receptor (EGFR) mutation and human epidermal growth factor receptor 2 (HER2) overexpression in adenocarcinoma, with the aim to identify the most predictive MR sequence. METHODS: A retrospective inclusion of 268 individuals with brain metastases from adenocarcinoma across two institutions was conducted. Utilising T1-weighted imaging (T1 contrast-enhanced [T1-CE]) and T2 fluid-attenuated inversion recovery (T2-FLAIR) sequences, 1,409 radiomics features were extracted. These sequences were randomly divided into training and test sets at a 7:3 ratio. The selection of relevant features was done using the least absolute shrinkage selection operator, and the training cohort's support vector classifier model was employed to generate the predictive model. The performance of the radiomics features was evaluated using a separate test set. RESULTS: For contrast-enhanced T1-CE cohorts, the radiomics features based on 19 selected characteristics exhibited excellent discrimination. No significant differences in age, sex, and time to metastasis were observed between the groups with EGFR mutations or HER2 + and those with wild-type EGFR or HER2 (p > 0.05). Radiomics feature analysis for T1-CE revealed an area under the curve (AUC) of 0.98, classification accuracy of 0.93, sensitivity of 0.92, and specificity of 0.93 in the training cohort. In the test set, the AUC was 0.82. The 19 radiomics features for the T2-FLAIR sequence showed AUCs of 0.86 in the training set and 0.70 in the test set. CONCLUSIONS: This study developed a T1-CE signature that could serve as a non-invasive adjunctive tool to determine the presence of EGFR mutations and HER2 + status in adenocarcinoma, aiding in the direction of treatment plans. CLINICAL RELEVANCE STATEMENT: We propose radiomics features based on T1-CE brain MR sequences that are both evidence-based and non-invasive. These can be employed to guide clinical treatment planning in patients with brain metastases from adenocarcinoma.

Refining neural network algorithms for accurate brain tumor classification in MRI imagery.

Brain tumor diagnosis using MRI scans poses significant challenges due to the complex nature of tumor appearances and variations. Traditional methods often require extensive manual intervention and are prone to human error, leading to misdiagnosis and delayed treatment. Current approaches primarily include manual examination by radiologists and conventional machine learning techniques. These methods rely heavily on feature extraction and classification algorithms, which may not capture the intricate patterns present in brain MRI images. Conventional techniques often suffer from limited accuracy and generalizability, mainly due to the high variability in tumor appearance and the subjective nature of manual interpretation. Additionally, traditional machine learning models may struggle with the high-dimensional data inherent in MRI images. To address these limitations, our research introduces a deep learning-based model utilizing convolutional neural networks (CNNs).Our model employs a sequential CNN architecture with multiple convolutional, max-pooling, and dropout layers, followed by dense layers for classification. The proposed model demonstrates a significant improvement in diagnostic accuracy, achieving an overall accuracy of 98% on the test dataset. The proposed model demonstrates a significant improvement in diagnostic accuracy, achieving an overall accuracy of 98% on the test dataset. The precision, recall, and F1-scores ranging from 97 to 98% with a roc-auc ranging from 99 to 100% for each tumor category further substantiate the model's effectiveness. Additionally, the utilization of Grad-CAM visualizations provides insights into the model's decision-making process, enhancing interpretability. This research addresses the pressing need for enhanced diagnostic accuracy in identifying brain tumors through MRI imaging, tackling challenges such as variability in tumor appearance and the need for rapid, reliable diagnostic tools.

Enhancing brain tumor detection in MRI images through explainable AI using Grad-CAM with Resnet 50.

This study addresses the critical challenge of detecting brain tumors using MRI images, a pivotal task in medical diagnostics that demands high accuracy and interpretability. While deep learning has shown remarkable success in medical image analysis, there remains a substantial need for models that are not only accurate but also interpretable to healthcare professionals. The existing methodologies, predominantly deep learning-based, often act as black boxes, providing little insight into their decision-making process. This research introduces an integrated approach using ResNet50, a deep learning model, combined with Gradient-weighted Class Activation Mapping (Grad-CAM) to offer a transparent and explainable framework for brain tumor detection. We employed a dataset of MRI images, enhanced through data augmentation, to train and validate our model. The results demonstrate a significant improvement in model performance, with a testing accuracy of 98.52% and precision-recall metrics exceeding 98%, showcasing the model's effectiveness in distinguishing tumor presence. The application of Grad-CAM provides insightful visual explanations, illustrating the model's focus areas in making predictions. This fusion of high accuracy and explainability holds profound implications for medical diagnostics, offering a pathway towards more reliable and interpretable brain tumor detection tools.

Extraneural metastatic ependymoma: distant metastasis to the pleura, lungs, lymph nodes and bone.

Ependymomas are neuroepithelial tumours arising from ependymal cells surrounding the cerebral ventricles that rarely metastasise to extraneural structures. This spread has been reported to occur to the lungs, lymph nodes, liver and bone. We describe the case of a patient with recurrent CNS WHO grade 3 ependymoma with extraneural metastatic disease. He was treated with multiple surgical resections, radiation therapy and salvage chemotherapy for his extraneural metastasis to the lungs, bone, pleural space and lymph nodes.

Association of 5-aminolevulinic acid fluorescence guided resection with photodynamic therapy in recurrent glioblastoma: a matched cohort study.

PURPOSE: Glioblastoma is a malignant and aggressive brain tumour that, although there have been improvements in the first line treatment, there is still no consensus regarding the best standard of care (SOC) upon its inevitable recurrence. There are novel adjuvant therapies that aim to improve local disease control. Nowadays, the association of intraoperative photodynamic therapy (PDT) immediately after a 5-aminolevulinic acid (5-ALA) fluorescence-guided resection (FGR) in malignant gliomas surgery has emerged as a potential and feasible strategy to increase the extent of safe resection and destroy residual tumour in the surgical cavity borders, respectively. OBJECTIVES: To assess the survival rates and safety of the association of intraoperative PDT with 5-ALA FGR, in comparison with a 5-ALA FGR alone, in patients with recurrent glioblastoma. METHODS: This article describes a matched-pair cohort study with two groups of patients submitted to 5-ALA FGR for recurrent glioblastoma. Group 1 was a prospective series of 11 consecutive cases submitted to 5-ALA FGR plus intraoperative PDT; group 2 was a historical series of 11 consecutive cases submitted to 5-ALA FGR alone. Age, sex, Karnofsky performance scale (KPS), 5-ALA post-resection status, T1-contrast-enhanced extent of resection (EOR), previous and post pathology, IDH (Isocitrate dehydrogenase), Ki67, previous and post treatment, brain magnetic resonance imaging (MRI) controls and surgical complications were documented. RESULTS: The Mantel-Cox test showed a significant difference between the survival rates (p = 0.008) of both groups. 4 postoperative complications occurred (36.6%) in each group. As of the last follow-up (January 2024), 7/11 patients in group 1, and 0/11 patients in group 2 were still alive. 6- and 12-months post-treatment, a survival proportion of 71,59% and 57,27% is expected in group 1, versus 45,45% and 9,09% in group 2, respectively. 6 months post-treatment, a progression free survival (PFS) of 61,36% and 18,18% is expected in group 1 and group 2, respectively. CONCLUSION: The association of PDT immediately after 5-ALA FGR for recurrent malignant glioma seems to be associated with better survival without additional or severe morbidity. Despite the need for larger, randomized series, the proposed treatment is a feasible and safe addition to the reoperation.

Artificial intelligence- image learning and its applications in neurooncology: a review.

Image learning involves using artificial intelligence (AI) to analyse radiological images. Various machine and deeplearning- based techniques have been employed to process images and extract relevant features. These can later be used to detect tumours early and predict their survival based on their grading and classification. Radiomics is now also used to predict genetic mutations and differentiate between tumour progression and treatment-related side effects. These were once completely dependent on invasive procedures like biopsy and histopathology. The use and feasibility of these techniques are now widely being explored in neurooncology to devise more accurate management plans and limit morbidity and mortality. Hence, the future of oncology lies in the exploration of AI-based image learning techniques, which can be applied to formulate management plans based on less invasive diagnostic techniques, earlier detection of tumours, and prediction of prognosis based on radiomic features. In this review, we discuss some of these applications of image learning in current medical dynamics.

Quantification of attenuation and speckle features from endoscopic OCT images for the diagnosis of human brain glioma.

Application of optical coherence tomography (OCT) in neurosurgery mostly includes the discrimination between intact and malignant tissues aimed at the detection of brain tumor margins. For particular tissue types, the existing approaches demonstrate low performance, which stimulates the further research for their improvement. The analysis of speckle patterns of brain OCT images is proposed to be taken into account for the discrimination between human brain glioma tissue and intact cortex and white matter. The speckle properties provide additional information of tissue structure, which could help to increase the efficiency of tissue differentiation. The wavelet analysis of OCT speckle patterns was applied to extract the power of local brightness fluctuations in speckle and its standard deviation. The speckle properties are analysed together with attenuation ones using a set of ex vivo brain tissue samples, including glioma of different grades. Various combinations of these features are considered to perform linear discriminant analysis for tissue differentiation. The results reveal that it is reasonable to include the local brightness fluctuations at first two wavelet decomposition levels in the analysis of OCT brain images aimed at neurosurgical diagnosis.

Brain tumor detection with integrating traditional and computational intelligence approaches across diverse imaging modalities - Challenges and future directions.

Brain tumor segmentation and classification play a crucial role in the diagnosis and treatment planning of brain tumors. Accurate and efficient methods for identifying tumor regions and classifying different tumor types are essential for guiding medical interventions. This study comprehensively reviews brain tumor segmentation and classification techniques, exploring various approaches based on image processing, machine learning, and deep learning. Furthermore, our study aims to review existing methodologies, discuss their advantages and limitations, and highlight recent advancements in this field. The impact of existing segmentation and classification techniques for automated brain tumor detection is also critically examined using various open-source datasets of Magnetic Resonance Images (MRI) of different modalities. Moreover, our proposed study highlights the challenges related to segmentation and classification techniques and datasets having various MRI modalities to enable researchers to develop innovative and robust solutions for automated brain tumor detection. The results of this study contribute to the development of automated and robust solutions for analyzing brain tumors, ultimately aiding medical professionals in making informed decisions and providing better patient care.

Structural- and DTI- MRI enable automated prediction of IDH Mutation Status in CNS WHO Grade 2-4 glioma patients: a deep Radiomics Approach.

BACKGROUND: The role of isocitrate dehydrogenase (IDH) mutation status for glioma stratification and prognosis is established. While structural magnetic resonance image (MRI) is a promising biomarker, it may not be sufficient for non-invasive characterisation of IDH mutation status. We investigated the diagnostic value of combined diffusion tensor imaging (DTI) and structural MRI enhanced by a deep radiomics approach based on convolutional neural networks (CNNs) and support vector machine (SVM), to determine the IDH mutation status in Central Nervous System World Health Organization (CNS WHO) grade 2-4 gliomas. METHODS: This retrospective study analyzed the DTI-derived fractional anisotropy (FA) and mean diffusivity (MD) images and structural images including fluid attenuated inversion recovery (FLAIR), non-enhanced T1-, and T2-weighted images of 206 treatment-naïve gliomas, including 146 IDH mutant and 60 IDH-wildtype ones. The lesions were manually segmented by experienced neuroradiologists and the masks were applied to the FA and MD maps. Deep radiomics features were extracted from each subject by applying a pre-trained CNN and statistical description. An SVM classifier was applied to predict IDH status using imaging features in combination with demographic data. RESULTS: We comparatively assessed the CNN-SVM classifier performance in predicting IDH mutation status using standalone and combined structural and DTI-based imaging features. Combined imaging features surpassed stand-alone modalities for the prediction of IDH mutation status [area under the curve (AUC) = 0.846; sensitivity = 0.925; and specificity = 0.567]. Importantly, optimal model performance was noted following the addition of demographic data (patients' age) to structural and DTI imaging features [area under the curve (AUC) = 0.847; sensitivity = 0.911; and specificity = 0.617]. CONCLUSIONS: Imaging features derived from DTI-based FA and MD maps combined with structural MRI, have superior diagnostic value to that provided by standalone structural or DTI sequences. In combination with demographic information, this CNN-SVM model offers a further enhanced non-invasive prediction of IDH mutation status in gliomas.

Comparison of two region-of-interest placement methods for histogram analysis of apparent diffusion coefficient maps for glioma grading.

Objectives: We assessed the value of histogram analysis (HA) of apparent diffusion coefficient (ADC) maps for grading low-grade (LGG) and high-grade (HGG) gliomas. Methods: We compared the diagnostic performance of two region-of-interest (ROI) placement methods (ROI 1: the entire tumor; ROI 2: the tumor excluding cystic and necrotic portions). We retrospectively evaluated 54 patients with supratentorial gliomas (18 LGG and 36 HGG). All subjects underwent standard 3T contrast-enhanced magnetic resonance imaging. Histogram parameters of ADC maps calculated with the two segmentation methods comprised mean, median, maxi-mum, minimum, kurtosis, skewness, entropy, standard deviation (sd), mean of positive pixels (mpp), uniformity of positive pixels, and their ratios (r) between lesion and normal white matter. They were compared using the independent t-test, chi-square test, or Mann-Whitney U test. For statistically significant results, receiver operating characteristic curves were constructed, and the optimal cutoff value, sensitivity, and specificity were determined by maximizing Youden's index. Results: The ROI 1 method resulted in significantly higher rADC mean, rADC median, and rADC mpp for LGG than for HGG; these parameters had value for predicting the histological glioma grade with a cutoff (sensitivity, specificity) of 1.88 (77.8%, 61.1%), 2.25 (44.4%, 97.2%), and 1.88 (77.8%, 63.9%), respectively. The ROI 2 method resulted in significantly higher ADC mean, ADC median, ADC mpp, ADC sd, ADC max, rADC median, rADC mpp, rADC mean, rADC sd, and rADC max for LGG than for HGG, while skewness was lower for LGG than for HGG (0.27 [0.98] vs 0.91 [0.81], p = 0.014). In ROI 2, ADC median, ADC mpp, ADC mean, rADC median, rADC mpp, and rADC mean performed well in differentiating glioma grade with cutoffs (sensitivity, specificity) of 1.28 (77.8%, 88.9%), 1.28 (77.8%, 88.9%), 1.25 (77.8%, 91.7%), 1.81 (83.3%, 91.7%), 1.74 (83.3%, 91.7%), and 1.81 (83.3%, 91.7%), respectively. Conclusions: HA parameters had value for grading gliomas. Ex-cluding cystic and necrotic portions of the tumor for measuring HA parameters was preferable to using the entire tumor as the ROI. In this segmentation, rADC median showed the highest performance in predicting histological glioma grade, followed by rADC mpp, rADC mean, ADC median, ADC mpp, and ADC mean.

ReMIND: The Brain Resection Multimodal Imaging Database.

The standard of care for brain tumors is maximal safe surgical resection. Neuronavigation augments the surgeon's ability to achieve this but loses validity as surgery progresses due to brain shift. Moreover, gliomas are often indistinguishable from surrounding healthy brain tissue. Intraoperative magnetic resonance imaging (iMRI) and ultrasound (iUS) help visualize the tumor and brain shift. iUS is faster and easier to incorporate into surgical workflows but offers a lower contrast between tumorous and healthy tissues than iMRI. With the success of data-hungry Artificial Intelligence algorithms in medical image analysis, the benefits of sharing well-curated data cannot be overstated. To this end, we provide the largest publicly available MRI and iUS database of surgically treated brain tumors, including gliomas (n = 92), metastases (n = 11), and others (n = 11). This collection contains 369 preoperative MRI series, 320 3D iUS series, 301 iMRI series, and 356 segmentations collected from 114 consecutive patients at a single institution. This database is expected to help brain shift and image analysis research and neurosurgical training in interpreting iUS and iMRI.

Intracranial residual lesions following early intensification in a patient with T-cell acute lymphoblastic leukemia: a case report.

BACKGROUND: T-cell acute lymphoblastic leukemia (T-ALL) tends to involve central nervous system (CNS) infiltration at diagnosis. However, cases of residual CNS lesions detected at the end of induction and post early intensification have not been recorded in patients with T-ALL. Also, the ratio and prognosis of patients with residual intracranial lesions have not been defined. CASE PRESENTATION: A 9-year-old boy with T-ALL had multiple intracranial tumors, which were still detected post early intensification. To investigate residual CNS lesions, we used 11C-methionine (MET)-positron emission tomography. Negative MET uptake in CNS lesions and excellent MRD status in bone marrow allowed continuing therapies without hematopoietic cell transplantation. CONCLUSIONS: In cases with residual lesions on imaging studies, treatment strategies should be considered by the systemic response, direct assessment of spinal fluid, along with further development of noninvasive imaging methods in CNS. Further retrospective or prospective studies are required to determine the prognosis and frequency of cases with residual intracranial lesions after induction therapy.

PDZK1 is correlated with DCE-MRI perfusion parameters in high-grade glioma.

OBJECTIVE: This study investigated the relationship between PDZK1 expression and Dynamic Contrast-Enhanced MRI (DCE-MRI) perfusion parameters in High-Grade Glioma (HGG). METHODS: Preoperative DCE-MRI scanning was performed on 80 patients with HGG to obtain DCE perfusion transfer coefficient (Ktrans), vascular plasma volume fraction (vp), extracellular volume fraction (ve), and reverse transfer constant (kep). PDZK1 in HGG patients was detected, and its correlation with DCE-MRI perfusion parameters was assessed by the Pearson method. An analysis of Cox regression was performed to determine the risk factors affecting survival, while Kaplan-Meier and log-rank tests to evaluate PDZK1's prognostic significance, and ROC curve analysis to assess its diagnostic value. RESULTS: PDZK1 was upregulated in HGG patients and predicted poor overall survival and progression-free survival. Moreover, PDZK1 expression distinguished grade III from grade IV HGG. PDZK1 expression was positively correlated with Ktrans 90, and ve_90, and negatively correlated with kep_max, and kep_90. CONCLUSION: PDZK1 is upregulated in HGG, predicts poor survival, and differentiates tumor grading in HGG patients. PDZK1 expression is correlated with DCE-MRI perfusion parameters.

Epidemiology and Management Trends of Neuro-Oncology in Nigeria: A Systematic Review and Pooled Analysis.

OBJECTIVE: Access to neuro-oncologic care in Nigeria has grown exponentially since the first reported cases in the mid-1960s. In this systematic review and pooled analysis, we characterize the growth of neurosurgical oncology in Nigeria and build a reference paper to direct efforts to expand this field. METHODS: We performed an initial literature search of several article databases and gray literature sources. We included and subsequently screened articles published between 1962 and 2021. Several variables were extracted from each study, including the affiliated hospital, the number of patients treated, patient sex, tumor pathology, the types of imaging modalities used for diagnosis, and the interventions used for each individual. Change in these variables was assessed using Chi-squared independence tests and univariate linear regression when appropriate. RESULTS: A total of 147 studies were identified, corresponding to 5,760 patients. Over 4000 cases were reported in the past 2 decades from 21 different Nigerian institutions. The types of tumors reported have increased over time, with increasingly more patients being evaluated via computed tomography (CT) and magnetic resonance imaging (MRI). There is also a prevalent use of radiotherapy, though chemotherapy remains an underreported treatment modality. CONCLUSIONS: This study highlights key trends regarding the prevalence and management of neuro-oncologic pathologies within Nigeria. Further studies are needed to continue to learn and guide the future growth of this field in Nigeria.

Indirect functional connectivity does not predict overall survival in glioblastoma.

BACKGROUND: Lesion network mapping (LNM) is a popular framework to assess clinical syndromes following brain injury. The classical approach involves embedding lesions from patients into a normative functional connectome and using the corresponding functional maps as proxies for disconnections. However, previous studies indicated limited predictive power of this approach in behavioral deficits. We hypothesized similarly low predictiveness for overall survival (OS) in glioblastoma (GBM). METHODS: A retrospective dataset of patients with GBM was included (n = 99). Lesion masks were registered in the normative space to compute disconnectivity maps. The brain functional normative connectome consisted in data from 173 healthy subjects obtained from the Human Connectome Project. A modified version of the LNM was then applied to core regions of GBM masks. Linear regression, classification, and principal component (PCA) analyses were conducted to explore the relationship between disconnectivity and OS. OS was considered both as continuous and categorical (low, intermediate, and high survival) variable. RESULTS: The results revealed no significant associations between OS and network disconnection strength when analyzed at both voxel-wise and classification levels. Moreover, patients stratified into different OS groups did not exhibit significant differences in network connectivity patterns. The spatial similarity among the first PCA of network maps for each OS group suggested a lack of distinctive network patterns associated with survival duration. CONCLUSIONS: Compared with indirect structural measures, functional indirect mapping does not provide significant predictive power for OS in patients with GBM. These findings are consistent with previous research that demonstrated the limitations of indirect functional measures in predicting clinical outcomes, underscoring the need for more comprehensive methodologies and a deeper understanding of the factors influencing clinical outcomes in this challenging disease.

Prognostic relevance of radiological findings on early postoperative MRI for 187 consecutive glioblastoma patients receiving standard therapy.

Several prognostic factors are known to influence survival for patients treated with IDH-wildtype glioblastoma, but unknown factors may remain. We aimed to investigate the prognostic implications of early postoperative MRI findings. A total of 187 glioblastoma patients treated with standard therapy were consecutively included. Patients either underwent a biopsy or surgery followed by an early postoperative MRI. Progression-free survival (PFS) and overall survival (OS) were analysed for known prognostic factors and MRI-derived candidate factors: resection status as defined by the response assessment in neuro-oncology (RANO)-working group (no contrast-enhancing residual tumour, non-measurable contrast-enhancing residual tumour, or measurable contrast-enhancing residual tumour) with biopsy as reference, contrast enhancement patterns (no enhancement, thin linear, thick linear, diffuse, nodular), and the presence of distant tumours. In the multivariate analysis, patients with no contrast-enhancing residual tumour or non-measurable contrast-enhancing residual tumour on the early postoperative MRI displayed a significantly improved progression-free survival compared with patients receiving only a biopsy. Only patients with non-measurable contrast-enhancing residual tumour showed improved overall survival in the multivariate analysis. Contrast enhancement patterns were not associated with survival. The presence of distant tumours was significantly associated with both poor progression-free survival and overall survival and should be considered incorporated into prognostic models.

Prediction of treatment response after stereotactic radiosurgery of brain metastasis using deep learning and radiomics on longitudinal MRI data.

We developed artificial intelligence models to predict the brain metastasis (BM) treatment response after stereotactic radiosurgery (SRS) using longitudinal magnetic resonance imaging (MRI) data and evaluated prediction accuracy changes according to the number of sequential MRI scans. We included four sequential MRI scans for 194 patients with BM and 369 target lesions for the Developmental dataset. The data were randomly split (8:2 ratio) for training and testing. For external validation, 172 MRI scans from 43 patients with BM and 62 target lesions were additionally enrolled. The maximum axial diameter (Dmax), radiomics, and deep learning (DL) models were generated for comparison. We evaluated the simple convolutional neural network (CNN) model and a gated recurrent unit (Conv-GRU)-based CNN model in the DL arm. The Conv-GRU model performed superior to the simple CNN models. For both datasets, the area under the curve (AUC) was significantly higher for the two-dimensional (2D) Conv-GRU model than for the 3D Conv-GRU, Dmax, and radiomics models. The accuracy of the 2D Conv-GRU model increased with the number of follow-up studies. In conclusion, using longitudinal MRI data, the 2D Conv-GRU model outperformed all other models in predicting the treatment response after SRS of BM.