Robust Deep Learning-based Segmentation of Glioblastoma on Routine Clinical MRI Scans Using Sparsified Training

Improved Robustness for Deep Learning-based Segmentation of Multi-Center Myocardial Perfusion MRI Datasets Using Data Adaptive Uncertainty-guided Space-time Analysis

Background

Fully automatic analysis of myocardial perfusion MRI datasets enables rapid and objective reporting of stress/rest studies in patients with suspected ischemic heart disease. Developing deep learning techniques that can analyze multi-center datasets despite limited training data and variations in software (pulse sequence) and hardware (scanner vendor) is an ongoing challenge.

Methods

Datasets from 3 medical centers acquired at 3T (n = 150 subjects; 21,150 first-pass images) were included: an internal dataset (inD; n = 95) and two external datasets (exDs; n = 55) used for evaluating the robustness of the trained deep neural network (DNN) models against differences in pulse sequence (exD-1) and scanner vendor (exD-2). A subset of inD (n = 85) was used for training/validation of a pool of DNNs for segmentation, all using the same spatiotemporal U-Net architecture and hyperparameters but with different parameter initializations. We employed a space-time sliding-patch analysis approach that automatically yields a pixel-wise "uncertainty map" as a byproduct of the segmentation process. In our approach, dubbed Data Adaptive Uncertainty-Guided Space-time (DAUGS) analysis, a given test case is segmented by all members of the DNN pool and the resulting uncertainty maps are leveraged to automatically select the "best" one among the pool of solutions. For comparison, we also trained a DNN using the established approach with the same settings (hyperparameters, data augmentation, etc.).

Results

The proposed DAUGS analysis approach performed similarly to the established approach on the internal dataset (Dice score for the testing subset of inD: 0.896 ± 0.050 vs. 0.890 ± 0.049; p = n.s.) whereas it significantly outperformed on the external datasets (Dice for exD-1: 0.885 ± 0.040 vs. 0.849 ± 0.065, p < 0.005; Dice for exD-2: 0.811 ± 0.070 vs. 0.728 ± 0.149, p < 0.005). Moreover, the number of image series with "failed" segmentation (defined as having myocardial contours that include bloodpool or are noncontiguous in ≥1 segment) was significantly lower for the proposed vs. the established approach (4.3% vs. 17.1%, p < 0.0005).

Conclusions

The proposed DAUGS analysis approach has the potential to improve the robustness of deep learning methods for segmentation of multi-center stress perfusion datasets with variations in the choice of pulse sequence, site location or scanner vendor.

Multi-modal fusion and feature enhancement U-Net coupling with stem cell niches proximity estimation for voxel-wise GBM recurrence prediction. Phys Med Biol 2024;69:10.1088/1361-6560/ad64b8. [PMID: 39019073 PMCID: PMC11308744 DOI: 10.1088/1361-6560/ad64b8] [Citation(s) in RCA: 0] [Impact Index Per Article: 0]

Objective.We aim to develop a Multi-modal Fusion and Feature Enhancement U-Net (MFFE U-Net) coupling with stem cell niche proximity estimation to improve voxel-wise Glioblastoma (GBM) recurrence prediction.Approach.57 patients with pre- and post-surgery magnetic resonance (MR) scans were retrospectively solicited from 4 databases. Post-surgery MR scans included two months before the clinical diagnosis of recurrence and the day of the radiologicaly confirmed recurrence. The recurrences were manually annotated on the T1ce. The high-risk recurrence region was first determined. Then, a sparse multi-modal feature fusion U-Net was developed. The 50 patients from 3 databases were divided into 70% training, 10% validation, and 20% testing. 7 patients from the 4th institution were used as external testing with transfer learning. Model performance was evaluated by recall, precision, F1-score, and Hausdorff Distance at the 95% percentile (HD95). The proposed MFFE U-Net was compared to the support vector machine (SVM) model and two state-of-the-art neural networks. An ablation study was performed.Main results.The MFFE U-Net achieved a precision of 0.79 ± 0.08, a recall of 0.85 ± 0.11, and an F1-score of 0.82 ± 0.09. Statistically significant improvement was observed when comparing MFFE U-Net with proximity estimation couple SVM (SVMPE), mU-Net, and Deeplabv3. The HD95 was 2.75 ± 0.44 mm and 3.91 ± 0.83 mm for the 10 patients used in the model construction and 7 patients used for external testing, respectively. The ablation test showed that all five MR sequences contributed to the performance of the final model, with T1ce contributing the most. Convergence analysis, time efficiency analysis, and visualization of the intermediate results further discovered the characteristics of the proposed method.Significance. We present an advanced MFFE learning framework, MFFE U-Net, for effective voxel-wise GBM recurrence prediction. MFFE U-Net performs significantly better than the state-of-the-art networks and can potentially guide early RT intervention of the disease recurrence.

Marker-controlled watershed algorithm and fuzzy C-means clustering machine learning: automated segmentation of glioblastoma from MRI images in a case series

Introduction and importance

Automated segmentation of glioblastoma multiforme (GBM) from MRI images is crucial for accurate diagnosis and treatment planning. This paper presents a new and innovative approach for automating the segmentation of GBM from MRI images using the marker-controlled watershed segmentation (MCWS) algorithm.

Case presentation and methods

The technique involves several image processing techniques, including adaptive thresholding, morphological filtering, gradient magnitude calculation, and regional maxima identification. The MCWS algorithm efficiently segments images based on local intensity structures using the watershed transform, and fuzzy c-means (FCM) clustering improves segmentation accuracy. The presented approach achieved improved segmentation accuracy in detecting and segmenting GBM tumours from axial T2-weighted (T2-w) MRI images, as demonstrated by the mean characteristics performance metrics for GBM segmentation (sensitivity: 0.9905, specificity: 0.9483, accuracy: 0.9508, precision: 0.5481, F_measure: 0.7052, and jaccard: 0.9340).

Clinical discussion

The results of this study underline the importance of reliable and accurate image segmentation for effective diagnosis and treatment planning of GBM tumours.

Conclusion

The MCWS technique provides an effective and efficient approach for the segmentation of challenging medical images.

MRI-based automatic identification and segmentation of extrahepatic cholangiocarcinoma using deep learning network

BACKGROUND

Accurate identification of extrahepatic cholangiocarcinoma (ECC) from an image is challenging because of the small size and complex background structure. Therefore, considering the limitation of manual delineation, it's necessary to develop automated identification and segmentation methods for ECC. The aim of this study was to develop a deep learning approach for automatic identification and segmentation of ECC using MRI.

METHODS

We recruited 137 ECC patients from our hospital as the main dataset (C1) and an additional 40 patients from other hospitals as the external validation set (C2). All patients underwent axial T1-weighted imaging (T1WI), T2-weighted imaging (T2WI), and diffusion-weighted imaging (DWI). Manual delineations were performed and served as the ground truth. Next, we used 3D VB-Net to establish single-mode automatic identification and segmentation models based on T1WI (model 1), T2WI (model 2), and DWI (model 3) in the training cohort (80% of C1), and compared them with the combined model (model 4). Subsequently, the generalization capability of the best models was evaluated using the testing set (20% of C1) and the external validation set (C2). Finally, the performance of the developed models was further evaluated.

RESULTS

Model 3 showed the best identification performance in the training, testing, and external validation cohorts with success rates of 0.980, 0.786, and 0.725, respectively. Furthermore, model 3 yielded an average Dice similarity coefficient (DSC) of 0.922, 0.495, and 0.466 to segment ECC automatically in the training, testing, and external validation cohorts, respectively.

CONCLUSION

The DWI-based model performed better in automatically identifying and segmenting ECC compared to T1WI and T2WI, which may guide clinical decisions and help determine prognosis.

Multi-class glioma segmentation on real-world data with missing MRI sequences: comparison of three deep learning algorithms

This study tests the generalisability of three Brain Tumor Segmentation (BraTS) challenge models using a multi-center dataset of varying image quality and incomplete MRI datasets. In this retrospective study, DeepMedic, no-new-Unet (nn-Unet), and NVIDIA-net (nv-Net) were trained and tested using manual segmentations from preoperative MRI of glioblastoma (GBM) and low-grade gliomas (LGG) from the BraTS 2021 dataset (1251 in total), in addition to 275 GBM and 205 LGG acquired clinically across 12 hospitals worldwide. Data was split into 80% training, 5% validation, and 15% internal test data. An additional external test-set of 158 GBM and 69 LGG was used to assess generalisability to other hospitals' data. All models' median Dice similarity coefficient (DSC) for both test sets were within, or higher than, previously reported human inter-rater agreement (range of 0.74-0.85). For both test sets, nn-Unet achieved the highest DSC (internal = 0.86, external = 0.93) and the lowest Hausdorff distances (10.07, 13.87 mm, respectively) for all tumor classes (p < 0.001). By applying Sparsified training, missing MRI sequences did not statistically affect the performance. nn-Unet achieves accurate segmentations in clinical settings even in the presence of incomplete MRI datasets. This facilitates future clinical adoption of automated glioma segmentation, which could help inform treatment planning and glioma monitoring.

Deep learning-based algorithm for postoperative glioblastoma MRI segmentation: a promising new tool for tumor burden assessment

OBJECTIVE

Clinical and surgical decisions for glioblastoma patients depend on a tumor imaging-based evaluation. Artificial Intelligence (AI) can be applied to magnetic resonance imaging (MRI) assessment to support clinical practice, surgery planning and prognostic predictions. In a real-world context, the current obstacles for AI are low-quality imaging and postoperative reliability. The aim of this study is to train an automatic algorithm for glioblastoma segmentation on a clinical MRI dataset and to obtain reliable results both pre- and post-operatively.

METHODS

The dataset used for this study comprises 237 (71 preoperative and 166 postoperative) MRIs from 71 patients affected by a histologically confirmed Grade IV Glioma. The implemented U-Net architecture was trained by transfer learning to perform the segmentation task on postoperative MRIs. The training was carried out first on BraTS2021 dataset for preoperative segmentation. Performance is evaluated using DICE score (DS) and Hausdorff 95% (H95).

RESULTS

In preoperative scenario, overall DS is 91.09 (± 0.60) and H95 is 8.35 (± 1.12), considering tumor core, enhancing tumor and whole tumor (ET and edema). In postoperative context, overall DS is 72.31 (± 2.88) and H95 is 23.43 (± 7.24), considering resection cavity (RC), gross tumor volume (GTV) and whole tumor (WT). Remarkably, the RC segmentation obtained a mean DS of 63.52 (± 8.90) in postoperative MRIs.

CONCLUSIONS

The performances achieved by the algorithm are consistent with previous literature for both pre-operative and post-operative glioblastoma's MRI evaluation. Through the proposed algorithm, it is possible to reduce the impact of low-quality images and missing sequences.

Multiparametric MRI: From Simultaneous Rapid Acquisition Methods and Analysis Techniques Using Scoring, Machine Learning, Radiomics, and Deep Learning to the Generation of Novel Metrics

ABSTRACT

With the recent advancements in rapid imaging methods, higher numbers of contrasts and quantitative parameters can be acquired in less and less time. Some acquisition models simultaneously obtain multiparametric images and quantitative maps to reduce scan times and avoid potential issues associated with the registration of different images. Multiparametric magnetic resonance imaging (MRI) has the potential to provide complementary information on a target lesion and thus overcome the limitations of individual techniques. In this review, we introduce methods to acquire multiparametric MRI data in a clinically feasible scan time with a particular focus on simultaneous acquisition techniques, and we discuss how multiparametric MRI data can be analyzed as a whole rather than each parameter separately. Such data analysis approaches include clinical scoring systems, machine learning, radiomics, and deep learning. Other techniques combine multiple images to create new quantitative maps associated with meaningful aspects of human biology. They include the magnetic resonance g-ratio, the inner to the outer diameter of a nerve fiber, and the aerobic glycolytic index, which captures the metabolic status of tumor tissues.

A neural ordinary differential equation model for visualizing deep neural network behaviors in multi-parametric MRI-based glioma segmentation

PURPOSE

To develop a neural ordinary differential equation (ODE) model for visualizing deep neural network behavior during multi-parametric MRI-based glioma segmentation as a method to enhance deep learning explainability.

METHODS

By hypothesizing that deep feature extraction can be modeled as a spatiotemporally continuous process, we implemented a novel deep learning model, Neural ODE, in which deep feature extraction was governed by an ODE parameterized by a neural network. The dynamics of (1) MR images after interactions with the deep neural network and (2) segmentation formation can thus be visualized after solving the ODE. An accumulative contribution curve (ACC) was designed to quantitatively evaluate each MR image's utilization by the deep neural network toward the final segmentation results. The proposed Neural ODE model was demonstrated using 369 glioma patients with a 4-modality multi-parametric MRI protocol: T1, contrast-enhanced T1 (T1-Ce), T2, and FLAIR. Three Neural ODE models were trained to segment enhancing tumor (ET), tumor core (TC), and whole tumor (WT), respectively. The key MRI modalities with significant utilization by deep neural networks were identified based on ACC analysis. Segmentation results by deep neural networks using only the key MRI modalities were compared to those using all four MRI modalities in terms of Dice coefficient, accuracy, sensitivity, and specificity.

RESULTS

All Neural ODE models successfully illustrated image dynamics as expected. ACC analysis identified T1-Ce as the only key modality in ET and TC segmentations, while both FLAIR and T2 were key modalities in WT segmentation. Compared to the U-Net results using all four MRI modalities, the Dice coefficient of ET (0.784→0.775), TC (0.760→0.758), and WT (0.841→0.837) using the key modalities only had minimal differences without significance. Accuracy, sensitivity, and specificity results demonstrated the same patterns.

CONCLUSION

The Neural ODE model offers a new tool for optimizing the deep learning model inputs with enhanced explainability. The presented methodology can be generalized to other medical image-related deep-learning applications.

Deep Learning in MRI-guided Radiation Therapy: A Systematic Review

MRI-guided radiation therapy (MRgRT) offers a precise and adaptive approach to treatment planning. Deep learning applications which augment the capabilities of MRgRT are systematically reviewed. MRI-guided radiation therapy offers a precise, adaptive approach to treatment planning. Deep learning applications which augment the capabilities of MRgRT are systematically reviewed with emphasis placed on underlying methods. Studies are further categorized into the areas of segmentation, synthesis, radiomics, and real time MRI. Finally, clinical implications, current challenges, and future directions are discussed.

Enhancing medical-imaging artificial intelligence through holistic use of time-tested key imaging and clinical parameters: Future insights

Much of the published literature in Radiology-related Artificial Intelligence (AI) focuses on single tasks, such as identifying the presence or absence or severity of specific lesions. Progress comparable to that achieved for general-purpose computer vision has been hampered by the unavailability of large and diverse radiology datasets containing different types of lesions with possibly multiple kinds of abnormalities in the same image. Also, since a diagnosis is rarely achieved through an image alone, radiology AI must be able to employ diverse strategies that consider all available evidence, not just imaging information. Using key imaging and clinical signs will help improve their accuracy and utility tremendously. Employing strategies that consider all available evidence will be a formidable task; we believe that the combination of human and computer intelligence will be superior to either one alone. Further, unless an AI application is explainable, radiologists will not trust it to be either reliable or bias-free; we discuss some approaches aimed at providing better explanations, as well as regulatory concerns regarding explainability (“transparency”). Finally, we look at federated learning, which allows pooling data from multiple locales while maintaining data privacy to create more generalizable and reliable models, and quantum computing, still prototypical but potentially revolutionary in its computing impact.

Predictors for time to awake in patients undergoing awake craniotomies

OBJECTIVE

Awake craniotomies are often characterized by alternating asleep-awake-asleep periods. Preceding the awake phase, patients are weaned from anesthesia and mechanical ventilation. Although clinicians aim to minimize the time to awake for patient safety and operating room efficiency, in some patients, the time to awake exceeds 20 minutes. The goal of this study was to determine the average time to awake and the factors associated with prolonged time to awake (> 20 minutes) in patients undergoing awake craniotomy.

METHODS

Records of patients who underwent awake craniotomy between 2003 and 2020 were evaluated. Time to awake was defined as the time between discontinuation of propofol and remifentanil infusion and the time of extubation. Patient and perioperative characteristics were explored as predictors for time to awake using logistic regression analyses.

RESULTS

Data of 307 patients were analyzed. The median (IQR) time to awake was 13 (10-20) minutes and exceeded 20 minutes in 17% (95% CI 13%-21%) of the patients. In both univariate and multivariable analyses, increased age, nonsmoker status, and American Society of Anesthesiologists (ASA) class III versus II were associated with a time to awake exceeding 20 minutes. BMI, as well as the use of alcohol, drugs, dexamethasone, or antiepileptic agents, was not significantly associated with the time to awake.

CONCLUSIONS

While most patients undergoing awake craniotomy are awake within a reasonable time frame after discontinuation of propofol and remifentanil infusion, time to awake exceeded 20 minutes in 17% of the patients. Increasing age, nonsmoker status, and higher ASA classification were found to be associated with a prolonged time to awake.

Glioblastoma Surgery Imaging-Reporting and Data System: Standardized Reporting of Tumor Volume, Location, and Resectability Based on Automated Segmentations

Treatment decisions for patients with presumed glioblastoma are based on tumor characteristics available from a preoperative MR scan. Tumor characteristics, including volume, location, and resectability, are often estimated or manually delineated. This process is time consuming and subjective. Hence, comparison across cohorts, trials, or registries are subject to assessment bias. In this study, we propose a standardized Glioblastoma Surgery Imaging Reporting and Data System (GSI-RADS) based on an automated method of tumor segmentation that provides standard reports on tumor features that are potentially relevant for glioblastoma surgery. As clinical validation, we determine the agreement in extracted tumor features between the automated method and the current standard of manual segmentations from routine clinical MR scans before treatment. In an observational consecutive cohort of 1596 adult patients with a first time surgery of a glioblastoma from 13 institutions, we segmented gadolinium-enhanced tumor parts both by a human rater and by an automated algorithm. Tumor features were extracted from segmentations of both methods and compared to assess differences, concordance, and equivalence. The laterality, contralateral infiltration, and the laterality indices were in excellent agreement. The native and normalized tumor volumes had excellent agreement, consistency, and equivalence. Multifocality, but not the number of foci, had good agreement and equivalence. The location profiles of cortical and subcortical structures were in excellent agreement. The expected residual tumor volumes and resectability indices had excellent agreement, consistency, and equivalence. Tumor probability maps were in good agreement. In conclusion, automated segmentations are in excellent agreement with manual segmentations and practically equivalent regarding tumor features that are potentially relevant for neurosurgical purposes. Standard GSI-RADS reports can be generated by open access software.

Sensitivity analysis for interpretation of machine learning based segmentation models in cardiac MRI

BACKGROUND

Image segmentation is a common task in medical imaging e.g., for volumetry analysis in cardiac MRI. Artificial neural networks are used to automate this task with performance similar to manual operators. However, this performance is only achieved in the narrow tasks networks are trained on. Performance drops dramatically when data characteristics differ from the training set properties. Moreover, neural networks are commonly considered black boxes, because it is hard to understand how they make decisions and why they fail. Therefore, it is also hard to predict whether they will generalize and work well with new data. Here we present a generic method for segmentation model interpretation. Sensitivity analysis is an approach where model input is modified in a controlled manner and the effect of these modifications on the model output is evaluated. This method yields insights into the sensitivity of the model to these alterations and therefore to the importance of certain features on segmentation performance.

RESULTS

We present an open-source Python library (misas), that facilitates the use of sensitivity analysis with arbitrary data and models. We show that this method is a suitable approach to answer practical questions regarding use and functionality of segmentation models. We demonstrate this in two case studies on cardiac magnetic resonance imaging. The first case study explores the suitability of a published network for use on a public dataset the network has not been trained on. The second case study demonstrates how sensitivity analysis can be used to evaluate the robustness of a newly trained model.

CONCLUSIONS

Sensitivity analysis is a useful tool for deep learning developers as well as users such as clinicians. It extends their toolbox, enabling and improving interpretability of segmentation models. Enhancing our understanding of neural networks through sensitivity analysis also assists in decision making. Although demonstrated only on cardiac magnetic resonance images this approach and software are much more broadly applicable.

A neuro evolutionary algorithm for patient calibrated prediction of survival in Glioblastoma patients

BACKGROUND AND OBJECTIVES

Glioblastoma multiforme (GBM) is the most common and malignant type of primary brain tumors. Radiation therapy (RT) plus concomitant and adjuvant Temozolomide (TMZ) constitute standard treatment of GBM. Existing models for GBM growth do not consider the effect of different schedules on tumor growth and patient survival. However, clinical trials show that treatment schedule and drug dosage significantly affect patient survival. The goal is to provide a patient calibrated model for predicting survival according to the treatment schedule.

METHODS

We propose a top-down method based on artificial neural networks (ANN) and genetic algorithm (GA) to predict survival of GBM patients. A feed forward undercomplete Autoencoder network is integrated with the neuro-evolutionary (NE) algorithm in order to extract a compressed representation of input clinical data. The proposed NE algorithm uses GA to obtain optimal architecture of a multi-layer perceptron (MLP). Taguchi L₁₆ orthogonal design of experiments is used to tune parameters of the proposed NE algorithm. Finally, the optimal MLP is used to predict survival of GBM patients.

RESULTS

Data from 8 related clinical trials have been collected and integrated to train the model. From 847 evaluable cases, 719 were used for train and validation and the remaining 128 cases were used to test the model. Mean absolute error of the predictions on the test data is 0.087 months which shows excellent performance of the proposed model in predicting survival of the patients. Also, the results show that the proposed NE algorithm is superior to other existing models in both the mean and variability of the prediction error.