Machine-learning based classification of glioblastoma using delta-radiomic features derived from dynamic susceptibility contrast enhanced magnetic resonance images: Introduction

Hemodynamic property incorporated brain tumor segmentation by deep learning and density-based analysis of dynamic susceptibility contrast-enhanced magnetic resonance imaging (MRI)

Background

Magnetic resonance imaging (MRI) is a primary non-invasive imaging modality for tumor segmentation, leveraging its exceptional soft tissue contrast and high resolution. Current segmentation methods typically focus on structural MRI, such as T1-weighted post-contrast-enhanced or fluid-attenuated inversion recovery (FLAIR) sequences. However, these methods overlook the blood perfusion and hemodynamic properties of tumors, readily derived from dynamic susceptibility contrast (DSC) enhanced MRI. This study introduces a novel hybrid method combining density-based analysis of hemodynamic properties in time-dependent perfusion imaging with deep learning spatial segmentation techniques to enhance tumor segmentation.

Methods

First, a U-Net convolutional neural network (CNN) is employed on structural images to delineate a region of interest (ROI). Subsequently, Hierarchical Density-Based Scans (HDBScan) are employed within the ROI to augment segmentation by exploring intratumoral hemodynamic heterogeneity through the investigation of tumor time course profiles unveiled in DSC MRI.

Results

The approach was tested and evaluated using a cohort of 513 patients from the open-source University of Pennsylvania glioblastoma database (UPENN-GBM) dataset, achieving a 74.83% Intersection over Union (IoU) score when compared to structural-only segmentation. The algorithm also exhibited increased precision and localized predictions of heightened segmentation boundary complexity, resulting in a 146.92% increase in contour complexity (ICC) compared to the reference standard provided by the UPENN-GBM dataset. Importantly, segmenting tumors with the developed new approach uncovered a negative correlation of the tumor volume with the scores in the Karnofsky Performance Scale (KPS) clinically used for assessing the functional status of patients (-0.309), which is not observed with the prevailing segmentation standard.

Conclusions

This work demonstrated that including hemodynamic properties of tissues from DSC MRI can improve existing structural or morphological feature-based tumor segmentation techniques with additional information on tumor biology and physiology. This approach can also be applied to other clinical indications that use perfusion MRI for diagnosis or treatment monitoring.

Prioritization of genes for translation: a computational approach

INTRODUCTION

Gene identification for genetic diseases is critical for the development of new diagnostic approaches and personalized treatment options. Prioritization of gene translation is an important consideration in the molecular biology field, allowing researchers to focus on the most promising candidates for further investigation.

AREAS COVERED

In this paper, we discussed different approaches to prioritize genes for translation, including the use of computational tools and machine learning algorithms, as well as experimental techniques such as knockdown and overexpression studies. We also explored the potential biases and limitations of these approaches and proposed strategies to improve the accuracy and reliability of gene prioritization methods. Although numerous computational methods have been developed for this purpose, there is a need for computational methods that incorporate tissue-specific information to enable more accurate prioritization of candidate genes. Such methods should provide tissue-specific predictions, insights into underlying disease mechanisms, and more accurate prioritization of genes.

EXPERT OPINION

Using advanced computational tools and machine learning algorithms to prioritize genes, we can identify potential targets for therapeutic intervention of complex diseases. This represents an up-and-coming method for drug development and personalized medicine.

Application of PET imaging delta radiomics for predicting progression-free survival in rare high-grade glioma

This study assesses the feasibility of using a sample-efficient model to investigate radiomics changes over time for predicting progression-free survival in rare diseases. Eighteen high-grade glioma patients underwent two L-3,4-dihydroxy-6-[18F]-fluoro-phenylalanine positron emission tomography (PET) dynamic scans: the first during treatment and the second at temozolomide chemotherapy discontinuation. Radiomics features from static/dynamic parametric images, alongside conventional features, were extracted. After excluding highly correlated features, 16 different models were trained by combining various feature selection methods and time-to-event survival algorithms. Performance was assessed using cross-validation. To evaluate model robustness, an additional dataset including 35 patients with a single PET scan at therapy discontinuation was used. Model performance was compared with a strategy extracting informative features from the set of 35 patients and applying them to the 18 patients with 2 PET scans. Delta-absolute radiomics achieved the highest performance when the pipeline was directly applied to the 18-patient subset (support vector machine (SVM) and recursive feature elimination (RFE): C-index = 0.783 [0.744-0.818]). This result remained consistent when transferring informative features from 35 patients (SVM + RFE: C-index = 0.751 [0.716-0.784], p = 0.06). In addition, it significantly outperformed delta-absolute conventional (C-index = 0.584 [0.548-0.620], p < 0.001) and single-time-point radiomics features (C-index = 0.546 [0.512-0.580], p < 0.001), highlighting the considerable potential of delta radiomics in rare cancer cohorts.

A clinically relevant computed tomography (CT) radiomics strategy for intracranial rodent brain tumour monitoring

Here, we establish a CT-radiomics based method for application in invasive, orthotopic rodent brain tumour models. Twenty four NOD/SCID mice were implanted with U87R-Luc2 GBM cells and longitudinally imaged via contrast enhanced (CE-CT) imaging. Pyradiomics was employed to extract CT-radiomic features from the tumour-implanted hemisphere and non-tumour-implanted hemisphere of acquired CT-scans. Inter-correlated features were removed (Spearman correlation > 0.85) and remaining features underwent predictive analysis (recursive feature elimination or Boruta algorithm). An area under the curve of the receiver operating characteristic curve was implemented to evaluate radiomic features for their capacity to predict defined outcomes. Firstly, we identified a subset of radiomic features which distinguish the tumour-implanted hemisphere and non- tumour-implanted hemisphere (i.e, tumour presence from normal tissue). Secondly, we successfully translate preclinical CT-radiomic pipelines to GBM patient CT scans (n = 10), identifying similar trends in tumour-specific feature intensities (E.g. 'glszm Zone Entropy'), thereby suggesting a mouse-to-human species conservation (a conservation of radiomic features across species). Thirdly, comparison of features across timepoints identify features which support preclinical tumour detection earlier than is possible by visual assessment of CT scans. This work establishes robust, preclinical CT-radiomic pipelines and describes the application of CE-CT for in-depth orthotopic brain tumour monitoring. Overall we provide evidence for the role of pre-clinical 'discovery' radiomics in the neuro-oncology space.

Untangling the web of glioblastoma treatment resistance using a multi-omic and multidisciplinary approach

Glioblastoma (GBM), the most common human brain tumor, has been notoriously resistant to treatment. As a result, the dismal overall survival of GBM patients has not changed over the past three decades. GBM has been stubbornly resistant to checkpoint inhibitor immunotherapies, which have been remarkably effective in the treatment of other tumors. It is clear that GBM resistance to therapy is multifactorial. Although therapeutic transport into brain tumors is inhibited by the blood brain barrier, there is evolving evidence that overcoming this barrier is not the predominant factor. GBMs generally have a low mutation burden, exist in an immunosuppressed environment and they are inherently resistant to immune stimulation, all of which contribute to treatment resistance. In this review, we evaluate the contribution of multi-omic approaches (genomic and metabolomic) along with analyzing immune cell populations and tumor biophysical characteristics to better understand and overcome GBM multifactorial resistance to treatment.

MRI-based radiomics model for distinguishing Stage I endometrial carcinoma from endometrial polyp: a multicenter study

BACKGROUND

Patients with early endometrial carcinoma (EC) have a good prognosis, but it is difficult to distinguish from endometrial polyps (EPs).

PURPOSE

To develop and assess magnetic resonance imaging (MRI)-based radiomics models for discriminating Stage I EC from EP in a multicenter setting.

MATERIAL AND METHODS

Patients with Stage I EC (n = 202) and EP (n = 99) who underwent preoperative MRI scans were collected in three centers (seven devices). The images from devices 1-3 were utilized for training and validation, and the images from devices 4-7 were utilized for testing, leading to three models. They were evaluated by the area under the receiver operating characteristic curve (AUC) and metrics including accuracy, sensitivity, and specificity. Two radiologists evaluated the endometrial lesions and compared them with the three models.

RESULTS

The AUCs of device 1, 2_ada, device 1, 3_ada, and device 2, 3_ada for discriminating Stage I EC from EP were 0.951, 0.912, and 0.896 for the training set, 0.755, 0.928, and 1.000 for the validation set, and 0.883, 0.956, and 0.878 for the external validation set, respectively. The specificity of the three models was higher, but the accuracy and sensitivity were lower than those of radiologists.

CONCLUSION

Our MRI-based models showed good potential in differentiating Stage I EC from EP and had been validated in multiple centers. Their specificity was higher than that of radiologists and may be used for computer-aided diagnosis in the future to assist clinical diagnosis.

Radiomics characterization of tissues in an animal brain tumor model imaged using dynamic contrast enhanced (DCE) MRI

Here, we investigate radiomics-based characterization of tumor vascular and microenvironmental properties in an orthotopic rat brain tumor model measured using dynamic-contrast-enhanced (DCE) MRI. Thirty-two immune compromised-RNU rats implanted with human U-251N cancer cells were imaged using DCE-MRI (7Tesla, Dual-Gradient-Echo). The aim was to perform pharmacokinetic analysis using a nested model (NM) selection technique to classify brain regions according to vasculature properties considered as the source of truth. A two-dimensional convolutional-based radiomics analysis was performed on the raw-DCE-MRI of the rat brains to generate dynamic radiomics maps. The raw-DCE-MRI and respective radiomics maps were used to build 28 unsupervised Kohonen self-organizing-maps (K-SOMs). A Silhouette-Coefficient (SC), k-fold Nested-Cross-Validation (k-fold-NCV), and feature engineering analyses were performed on the K-SOMs' feature spaces to quantify the distinction power of radiomics features compared to raw-DCE-MRI for classification of different Nested Models. Results showed that eight radiomics features outperformed respective raw-DCE-MRI in prediction of the three nested models. The average percent difference in SCs between radiomics features and raw-DCE-MRI was: 29.875% ± 12.922%, p < 0.001. This work establishes an important first step toward spatiotemporal characterization of brain regions using radiomics signatures, which is fundamental toward staging of tumors and evaluation of tumor response to different treatments.

Predicting Histopathological Grading of Adult Gliomas Based On Preoperative Conventional Multimodal MRI Radiomics: A Machine Learning Model

PURPOSE

The accurate preoperative histopathological grade diagnosis of adult gliomas is of great significance for the formulation of a surgical plan and the implementation of a subsequent treatment. The aim of this study is to establish a predictive model for classifying adult gliomas into grades 2-4 based on preoperative conventional multimodal MRI radiomics.

PATIENTS AND METHODS

Patients with pathologically confirmed gliomas at Huashan Hospital, Fudan University, between February 2017 and July 2019 were retrospectively analyzed. Two regions of interest (ROIs), called the maximum anomaly region (ROI1) and the tumor region (ROI2), were delineated on the patients' preoperative MRIs utilizing the tool ITK-SNAP, and Pyradiomics 3.0 was applied to execute feature extraction. Feature selection was performed utilizing a least absolute shrinkage and selection operator (LASSO) filter. Six classifiers, including Gaussian naive Bayes (GNB), random forest (RF), K-nearest neighbor (KNN), support vector machine (SVM) with a linear kernel, adaptive boosting (AB), and multilayer perceptron (MLP) were used to establish predictive models, and the predictive performance of the six classifiers was evaluated through five-fold cross-validation. The performance of the predictive models was evaluated using the AUC and other metrics. After that, the model with the best predictive performance was tested using the external data from The Cancer Imaging Archive (TCIA).

RESULTS

According to the inclusion and exclusion criteria, 240 patients with gliomas were identified for inclusion in the study, including 106 grade 2, 68 grade 3, and 66 grade 4 gliomas. A total of 150 features was selected, and the MLP classifier had the best predictive performance among the six classifiers based on T2-FLAIR (mean AUC of 0.80 ± 0.07). The SVM classifier had the best predictive performance among the six classifiers based on DWI (mean AUC of 0.84 ± 0.05); the SVM classifier had the best predictive performance among the six classifiers based on CE-T1WI (mean AUC of 0.85 ± 0.06). Among the six classifiers, based on ROI1, the MLP classifier had the best prediction performance (mean AUC of 0.78 ± 0.07); among the six classifiers, based on ROI2, the SVM classifier had the best prediction performance (mean AUC of 0.82 ± 0.07). Among the six classifiers, based on the multimodal MRI of all the ROIs, the SVM classifier had the best prediction performance (average AUC of 0.85 ± 0.04). The SVM classifier, based on the multimodal MRI of all the ROIs, achieved an AUC of 0.81 using the external data from TCIA.

CONCLUSIONS

The prediction model, based on preoperative conventional multimodal MRI radiomics, established in this study can conveniently, accurately, and noninvasively classify adult gliomas into grades 2-4, providing certain assistance for the precise diagnosis and treatment of patients and optimizing their clinical management.

Applying machine learning classifiers to automate quality assessment of paediatric dynamic susceptibility contrast (DSC-) MRI data

OBJECTIVE

Investigate the performance of qualitative review (QR) for assessing dynamic susceptibility contrast (DSC-) MRI data quality in paediatric normal brain and develop an automated alternative to QR.

METHODS

1027 signal-time courses were assessed by Reviewer 1 using QR. 243 were additionally assessed by Reviewer 2 and % disagreements and Cohen's κ (κ) were calculated. The signal drop-to-noise ratio (SDNR), root mean square error (RMSE), full width half maximum (FWHM) and percentage signal recovery (PSR) were calculated for the 1027 signal-time courses. Data quality thresholds for each measure were determined using QR results. The measures and QR results trained machine learning classifiers. Sensitivity, specificity, precision, classification error and area under the curve from a receiver operating characteristic curve were calculated for each threshold and classifier.

RESULTS

Comparing reviewers gave 7% disagreements and κ = 0.83. Data quality thresholds of: 7.6 for SDNR; 0.019 for RMSE; 3 s and 19 s for FWHM; and 42.9 and 130.4% for PSR were produced. SDNR gave the best sensitivity, specificity, precision, classification error and area under the curve values of 0.86, 0.86, 0.93, 14.2% and 0.83. Random forest was the best machine learning classifier, giving sensitivity, specificity, precision, classification error and area under the curve of 0.94, 0.83, 0.93, 9.3% and 0.89.

CONCLUSION

The reviewers showed good agreement. Machine learning classifiers trained on signal-time course measures and QR can assess quality. Combining multiple measures reduces misclassification.

ADVANCES IN KNOWLEDGE

A new automated quality control method was developed, which trained machine learning classifiers using QR results.

AGGN: Attention-based glioma grading network with multi-scale feature extraction and multi-modal information fusion

In this paper, a magnetic resonance imaging (MRI) oriented novel attention-based glioma grading network (AGGN) is proposed. By applying the dual-domain attention mechanism, both channel and spatial information can be considered to assign weights, which benefits highlighting the key modalities and locations in the feature maps. Multi-branch convolution and pooling operations are applied in a multi-scale feature extraction module to separately obtain shallow and deep features on each modality, and a multi-modal information fusion module is adopted to sufficiently merge low-level detailed and high-level semantic features, which promotes the synergistic interaction among different modality information. The proposed AGGN is comprehensively evaluated through extensive experiments, and the results have demonstrated the effectiveness and superiority of the proposed AGGN in comparison to other advanced models, which also presents high generalization ability and strong robustness. In addition, even without the manually labeled tumor masks, AGGN can present considerable performance as other state-of-the-art algorithms, which alleviates the excessive reliance on supervised information in the end-to-end learning paradigm.

Machine Learning in the Classification of Pediatric Posterior Fossa Tumors: A Systematic Review

Background: Posterior fossa tumors (PFTs) are a morbid group of central nervous system tumors that most often present in childhood. While early diagnosis is critical to drive appropriate treatment, definitive diagnosis is currently only achievable through invasive tissue collection and histopathological analyses. Machine learning has been investigated as an alternative means of diagnosis. In this systematic review and meta-analysis, we evaluated the primary literature to identify all machine learning algorithms developed to classify and diagnose pediatric PFTs using imaging or molecular data. Methods: Of the 433 primary papers identified in PubMed, EMBASE, and Web of Science, 25 ultimately met the inclusion criteria. The included papers were extracted for algorithm architecture, study parameters, performance, strengths, and limitations. Results: The algorithms exhibited variable performance based on sample size, classifier(s) used, and individual tumor types being investigated. Ependymoma, medulloblastoma, and pilocytic astrocytoma were the most studied tumors with algorithm accuracies ranging from 37.5% to 94.5%. A minority of studies compared the developed algorithm to a trained neuroradiologist, with three imaging-based algorithms yielding superior performance. Common algorithm and study limitations included small sample sizes, uneven representation of individual tumor types, inconsistent performance reporting, and a lack of application in the clinical environment. Conclusions: Artificial intelligence has the potential to improve the speed and accuracy of diagnosis in this field if the right algorithm is applied to the right scenario. Work is needed to standardize outcome reporting and facilitate additional trials to allow for clinical uptake.

Radiomics features of DSC-PWI in time dimension may provide a new chance to identify ischemic stroke

Ischemic stroke has become a severe disease endangering human life. However, few studies have analyzed the radiomics features that are of great clinical significance for the diagnosis, treatment, and prognosis of patients with ischemic stroke. Due to sufficient cerebral blood flow information in dynamic susceptibility contrast perfusion-weighted imaging (DSC-PWI) images, this study aims to find the critical features hidden in DSC-PWI images to characterize hypoperfusion areas (HA) and normal areas (NA). This study retrospectively analyzed 80 DSC-PWI data of 56 patients with ischemic stroke from 2013 to 2016. For exploring features in HA and NA,13 feature sets (Fmethod) were obtained from different feature selection algorithms. Furthermore, these 13 Fmethod were validated in identifying HA and NA and distinguishing the proportion of ischemic lesions in brain tissue. In identifying HA and NA, the composite score (CS) of the 13 Fmethod ranged from 0.624 to 0.925. FLasso in the 13 Fmethod achieved the best performance with mAcc of 0.958, mPre of 0.96, mAuc of 0.982, mF1 of 0.959, and mRecall of 0.96. As to classifying the proportion of the ischemic region, the best CS was 0.786, with Acc of 0.888 and Pre of 0.863. The classification ability was relatively stable when the reference threshold (RT) was &lt;0.25. Otherwise, when RT was &gt;0.25, the performance will gradually decrease as its increases. These results showed that radiomics features extracted from the Lasso algorithms could accurately reflect cerebral blood flow changes and classify HA and NA. Besides, In the event of ischemic stroke, the ability of radiomics features to distinguish the proportion of ischemic areas needs to be improved. Further research should be conducted on feature engineering, model optimization, and the universality of the algorithms in the future.

A hierarchical machine learning model based on Glioblastoma patients' clinical, biomedical, and image data to analyze their treatment plans

AIM OF STUDY

Glioblastoma Multiforme (GBM) is an aggressive brain cancer in adults that kills most patients in the first year due to ineffective treatment. Different clinical, biomedical, and image data features are needed to analyze GBM, increasing complexities. Besides, they lead to weak performances for machine learning models due to ignoring physicians' knowledge. Therefore, this paper proposes a hierarchical model based on Fuzzy C-mean (FCM) clustering, Wrapper feature selection, and twelve classifiers to analyze treatment plans.

METHODOLOGY/APPROACH

The proposed method finds the effectiveness of previous and current treatment plans, hierarchically determining the best decision for future treatment plans for GBM patients using clinical data, biomedical data, and different image data. A case study is presented based on the Cancer Genome Atlas Glioblastoma Multiforme dataset to prove the effectiveness of the proposed model. This dataset is analyzed using data preprocessing, experts' knowledge, and a feature reduction method based on the Principal Component Analysis. Then, the FCM clustering method is utilized to reinforce classifier learning.

OUTCOMES OF STUDY

The proposed model finds the best combination of Wrapper feature selection and classifier for each cluster based on different measures, including accuracy, sensitivity, specificity, precision, F-score, and G-mean according to a hierarchical structure. It has the best performance among other reinforced classifiers. Besides, this model is compatible with real-world medical processes for GBM patients based on clinical, biomedical, and image data.

Predicting IDH subtype of grade 4 astrocytoma and glioblastoma from tumor radiomic patterns extracted from multiparametric magnetic resonance images using a machine learning approach

Background and purpose

Semantic imaging features have been used for molecular subclassification of high-grade gliomas. Radiomics-based prediction of molecular subgroups has the potential to strategize and individualize therapy. Using MRI texture features, we propose to distinguish between IDH wild type and IDH mutant type high grade gliomas.

Methods

Between 2013 and 2020, 100 patients were retrospectively analyzed for the radiomics study. Immunohistochemistry of the pathological specimen was used to initially identify patients for the IDH mutant/wild phenotype and was then confirmed by Sanger’s sequencing. Image texture analysis was performed on contrast-enhanced T1 (T1C) and T2 weighted (T2W) MR images. Manual segmentation was performed on MR image slices followed by single-slice multiple sampling image augmentation. Both whole tumor multislice segmentation and single-slice multiple sampling approaches were used to arrive at the best model. Radiomic features were extracted, which included first-order features, second-order (GLCM—Grey level co-occurrence matrix), and shape features. Feature enrichment was done using LASSO (Least Absolute Shrinkage and Selection Operator) regression, followed by radiomic classification using Support Vector Machine (SVM) and a 10-fold cross-validation strategy for model development. The area under the Receiver Operator Characteristic (ROC) curve and predictive accuracy were used as diagnostic metrics to evaluate the model to classify IDH mutant and wild-type subgroups.

Results

Multislice analysis resulted in a better model compared to the single-slice multiple-sampling approach. A total of 164 MR-based texture features were extracted, out of which LASSO regression identified 14 distinctive GLCM features for the endpoint, which were used for further model development. The best model was achieved by using combined T1C and T2W MR images using a Quadratic Support Vector Machine Classifier and a 10-fold internal cross-validation approach, which demonstrated a predictive accuracy of 89% with an AUC of 0.89 for each IDH mutant and IDH wild subgroup.

Conclusion

A machine learning classifier of radiomic features extracted from multiparametric MRI images (T1C and T2w) provides important diagnostic information for the non-invasive prediction of the IDH mutant or wild-type phenotype of high-grade gliomas and may have potential use in either escalating or de-escalating adjuvant therapy for gliomas or for using targeted agents in the future.

Novel Survival Features Generated by Clinical Text Information and Radiomics Features May Improve the Prediction of Ischemic Stroke Outcome

Background: Accurate outcome prediction is of great clinical significance in customizing personalized treatment plans, reducing the situation of poor recovery, and objectively and accurately evaluating the treatment effect. This study intended to evaluate the performance of clinical text information (CTI), radiomics features, and survival features (SurvF) for predicting functional outcomes of patients with ischemic stroke. Methods: SurvF was constructed based on CTI and mRS radiomics features (mRSRF) to improve the prediction of the functional outcome in 3 months (90-day mRS). Ten machine learning models predicted functional outcomes in three situations (2-category, 4-category, and 7-category) using seven feature groups constructed by CTI, mRSRF, and SurvF. Results: For 2-category, ALL (CTI + mRSRF+ SurvF) performed best, with an mAUC of 0.884, mAcc of 0.864, mPre of 0.877, mF1 of 0.86, and mRecall of 0.864. For 4-category, ALL also achieved the best mAuc of 0.787, while CTI + SurvF achieved the best score with mAcc = 0.611, mPre = 0.622, mF1 = 0.595, and mRe-call = 0.611. For 7-category, CTI + SurvF performed best, with an mAuc of 0.788, mPre of 0.519, mAcc of 0.529, mF1 of 0.495, and mRecall of 0.47. Conclusions: The above results indicate that mRSRF + CTI can accurately predict functional outcomes in ischemic stroke patients with proper machine learning models. Moreover, combining SurvF will improve the prediction effect compared with the original features. However, limited by the small sample size, further validation on larger and more varied datasets is necessary.

Artificial Intelligence for Survival Prediction in Brain Tumors on Neuroimaging

Survival prediction of patients affected by brain tumors provides essential information to guide surgical planning, adjuvant treatment selection, and patient counseling. Current reliance on clinical factors, such as Karnofsky Performance Status Scale, and simplistic radiological characteristics are, however, inadequate for survival prediction in tumors such as glioma that demonstrate molecular and clinical heterogeneity with variable survival outcomes. Advances in the domain of artificial intelligence have afforded powerful tools to capture a large number of hidden high-dimensional imaging features that reflect abundant information about tumor structure and physiology. Here, we provide an overview of current literature that apply computational analysis tools such as radiomics and machine learning methods to the pipeline of image preprocessing, tumor segmentation, feature extraction, and construction of classifiers to establish survival prediction models based on neuroimaging. We also discuss challenges relating to the development and evaluation of such models and explore ethical issues surrounding the future use of machine learning predictions.

A Survey of Radiomics in Precision Diagnosis and Treatment of Adult Gliomas

Glioma is the most common primary malignant tumor of the adult central nervous system (CNS), which mostly shows invasive growth. In most cases, surgery is often difficult to completely remove, and the recurrence rate and mortality of patients are high. With the continuous development of molecular genetics and the great progress of molecular biology technology, more and more molecular biomarkers have been proved to have important guiding significance in the individualized diagnosis, treatment, and prognosis evaluation of glioma. With the updates of the World Health Organization (WHO) classification of tumors of the CNS in 2021, the diagnosis and treatment of glioma has entered the era of precision medicine in the true sense. Due to its ability to non-invasively achieve accurate identification of glioma from other intracranial tumors, and to predict the grade, genotyping, treatment response, and prognosis of glioma, which provides a scientific basis for the clinical application of individualized diagnosis and treatment model of glioma, radiomics has become a research hotspot in the field of precision medicine. This paper reviewed the research related to radiomics of adult gliomas published in recent years and summarized the research proceedings of radiomics in differential diagnosis, preoperative grading and genotyping, treatment and efficacy evaluation, and survival prediction of adult gliomas.

Cascaded mutual enhancing networks for brain tumor subregion segmentation in multiparametric MRI

Accurate segmentation of glioma and its subregions plays an important role in radiotherapy treatment planning. Due to a very populated multiparameter magnetic resonance imaging image, manual segmentation tasks can be very time-consuming, meticulous, and prone to subjective errors. Here, we propose a novel deep learning framework based on mutual enhancing networks to automatically segment brain tumor subregions. The proposed framework is suitable for the segmentation of brain tumor subregions owing to the contribution of Retina U-Net followed by the implementation of a mutual enhancing strategy between the classification localization map (CLM) module and segmentation module. Retina U-Net is trained to accurately identify view-of-interest and feature maps of the whole tumor (WT), which are then transferred to the CLM module and segmentation module. Subsequently, CLM generated by the CLM module is integrated with the segmentation module to bring forth a mutual enhancing strategy. In this way, our proposed framework first focuses on WT through Retina U-Net, and since WT consists of subregions, a mutual enhancing strategy then further aims to classify and segment subregions embedded within WT. We implemented and evaluated our proposed framework on the BraTS 2020 dataset consisting of 369 cases. We performed a 5-fold cross-validation on 200 datasets and a hold-out test on the remaining 169 cases. To demonstrate the effectiveness of our network design, we compared our method against the networks without Retina U-Net, mutual enhancing strategy, and a recently published Cascaded U-Net architecture. Results of all four methods were compared to the ground truth for segmentation and localization accuracies. Our method yielded significantly (P < 0.01) better values of dice-similarity-coefficient, center-of-mass-distance, and volume difference compared to all three competing methods across all tumor labels (necrosis and non-enhancing, edema, enhancing tumor, WT, tumor core) on both validation and hold-out dataset. Overall quantitative and statistical results of this work demonstrate the ability of our method to both accurately and automatically segment brain tumor subregions.

Hemodynamic Imaging in Cerebral Diffuse Glioma—Part B: Molecular Correlates, Treatment Effect Monitoring, Prognosis, and Future Directions

Gliomas, and glioblastoma in particular, exhibit an extensive intra- and inter-tumoral molecular heterogeneity which represents complex biological features correlating to the efficacy of treatment response and survival. From a neuroimaging point of view, these specific molecular and histopathological features may be used to yield imaging biomarkers as surrogates for distinct tumor genotypes and phenotypes. The development of comprehensive glioma imaging markers has potential for improved glioma characterization that would assist in the clinical work-up of preoperative treatment planning and treatment effect monitoring. In particular, the differentiation of tumor recurrence or true progression from pseudoprogression, pseudoresponse, and radiation-induced necrosis can still not reliably be made through standard neuroimaging only. Given the abundant vascular and hemodynamic alterations present in diffuse glioma, advanced hemodynamic imaging approaches constitute an attractive area of clinical imaging development. In this context, the inclusion of objective measurable glioma imaging features may have the potential to enhance the individualized care of diffuse glioma patients, better informing of standard-of-care treatment efficacy and of novel therapies, such as the immunotherapies that are currently increasingly investigated. In Part B of this two-review series, we assess the available evidence pertaining to hemodynamic imaging for molecular feature prediction, in particular focusing on isocitrate dehydrogenase (IDH) mutation status, MGMT promoter methylation, 1p19q codeletion, and EGFR alterations. The results for the differentiation of tumor progression/recurrence from treatment effects have also been the focus of active research and are presented together with the prognostic correlations identified by advanced hemodynamic imaging studies. Finally, the state-of-the-art concepts and advancements of hemodynamic imaging modalities are reviewed together with the advantages derived from the implementation of radiomics and machine learning analyses pipelines.

Radiomics for precision medicine in glioblastoma

INTRODUCTION

Being the most common primary brain tumor, glioblastoma presents as an extremely challenging malignancy to treat with dismal outcomes despite treatment. Varying molecular epidemiology of glioblastoma between patients and intra-tumoral heterogeneity explains the failure of current one-size-fits-all treatment modalities. Radiomics uses machine learning to identify salient features of the tumor on brain imaging and promises patient-specific management in glioblastoma patients.

METHODS

We performed a comprehensive review of the available literature on studies investigating the role of radiomics and radiogenomics models for the diagnosis, stratification, prognostication as well as treatment planning and monitoring of glioblastoma.

RESULTS

Classifiers based on a combination of various MRI sequences, genetic information and clinical data can predict non-invasive tumor diagnosis, overall survival and treatment response with reasonable accuracy. However, the use of radiomics for glioblastoma treatment remains in infancy as larger sample sizes, standardized image acquisition and data extraction techniques are needed to develop machine learning models that can be translated effectively into clinical practice.

CONCLUSION

Radiomics has the potential to transform the scope of glioblastoma management through personalized medicine.

A review of deep learning-based three-dimensional medical image registration methods

Medical image registration is a vital component of many medical procedures, such as image-guided radiotherapy (IGRT), as it allows for more accurate dose-delivery and better management of side effects. Recently, the successful implementation of deep learning (DL) in various fields has prompted many research groups to apply DL to three-dimensional (3D) medical image registration. Several of these efforts have led to promising results. This review summarized the progress made in DL-based 3D image registration over the past 5 years and identify existing challenges and potential avenues for further research. The collected studies were statistically analyzed based on the region of interest (ROI), image modality, supervision method, and registration evaluation metrics. The studies were classified into three categories: deep iterative registration, supervised registration, and unsupervised registration. The studies are thoroughly reviewed and their unique contributions are highlighted. A summary is presented following a review of each category of study, discussing its advantages, challenges, and trends. Finally, the common challenges for all categories are discussed, and potential future research topics are identified.

Chromatin insulation dynamics in glioblastoma: challenges and future perspectives of precision oncology

Glioblastoma (GBM) is the most aggressive primary brain tumor, having a poor prognosis and a median overall survival of less than two years. Over the last decade, numerous findings regarding the distinct molecular and genetic profiles of GBM have led to the emergence of several therapeutic approaches. Unfortunately, none of them has proven to be effective against GBM progression and recurrence. Epigenetic mechanisms underlying GBM tumor biology, including histone modifications, DNA methylation, and chromatin architecture, have become an attractive target for novel drug discovery strategies. Alterations on chromatin insulator elements (IEs) might lead to aberrant chromatin remodeling via DNA loop formation, causing oncogene reactivation in several types of cancer, including GBM. Importantly, it is shown that mutations affecting the isocitrate dehydrogenase (IDH) 1 and 2 genes, one of the most frequent genetic alterations in gliomas, lead to genome-wide DNA hypermethylation and the consequent IE dysfunction. The relevance of IEs has also been observed in a small population of cancer stem cells known as glioma stem cells (GSCs), which are thought to participate in GBM tumor initiation and drug resistance. Recent studies revealed that epigenomic alterations, specifically chromatin insulation and DNA loop formation, play a crucial role in establishing and maintaining the GSC transcriptional program. This review focuses on the relevance of IEs in GBM biology and their implementation as a potential theranostic target to stratify GBM patients and develop novel therapeutic approaches. We will also discuss the state-of-the-art emerging technologies using big data analysis and how they will settle the bases on future diagnosis and treatment strategies in GBM patients.

Exploring diagnostic performance of T2 mapping in diffuse glioma grading

Background

To evaluate the diagnostic performance of T2 mapping in differentiating WHO grade II glioma from high-grade glioma (HGG).

Methods

We conducted a single-center, retrospective diagnostic study. Confirmed diffuse glioma (WHO grade II-IV) patients who underwent post-contrast T1-weighted imaging, T2-weighted imaging, and T2 mapping were included. All diagnoses were based on histological and molecular tests. Seventy-five percent of cases were subsampled to generate receiver operating characteristic (ROC) curves and areas under the curve (AUC), while the remaining cases were used to test the accuracy of T2 mapping. Subsampling was repeated four times. Age, T2 relaxation time, and contrast-enhancement status were used to generate a multivariable ROC curve. T2 relaxation time was also used to generate ROC curves to predict the isocitrate dehydrogenase (IDH) status.

Results

A total of 159 patients were included in the study. After four repeats of subsampling, the AUCs of the T2 mapping ROC curve were 0.801 (95% CI: 0.724-0.879), 0.795 (95% CI: 0.714-0.875), 0.803 (95% CI: 0.723-0.884), and 0.801 (95% CI: 0.716-0.886), with an average sensitivity of 0.753 and an average specificity of 0.767. When applied to the remaining 25% of cases, the accuracy was 75%, 93.75%, 82.50%, and 71.74%. The AUC of the multivariable ROC was 0.927 (95% CI: 0.882-0.971). IDH-mutant and IDH-wildtype gliomas have significantly different T2 relaxation times (146.28 and 124.10 ms, respectively; P=0.001), and the AUC of IDH-mutant prediction was 0.687 (95% CI: 0.585-0.789).

Conclusions

Quantitative T2 mapping differentiated WHO grade II glioma from HGG with moderate sensitivity and specificity. Given the advantages of short acquisition times and the absence of a contrast agent, our study suggests the application of T2 mapping in pre-operative glioma grading is feasible.

Radiomics Prediction of EGFR Status in Lung Cancer-Our Experience in Using Multiple Feature Extractors and The Cancer Imaging Archive Data

We investigated the performance of multiple radiomics feature extractors/software on predicting epidermal growth factor receptor mutation status in 228 patients with non–small cell lung cancer from publicly available data sets in The Cancer Imaging Archive. The imaging and clinical data were split into training (n = 105) and validation cohorts (n = 123). Two of the most cited open-source feature extractors, IBEX (1563 features) and Pyradiomics (1319 features), and our in-house software, Columbia Image Feature Extractor (CIFE) (1160 features), were used to extract radiomics features. Univariate and multivariate analyses were performed sequentially to predict EGFR mutation status using each individual feature extractor. Our univariate analysis integrated an unsupervised clustering method to identify nonredundant and informative candidate features for the creation of prediction models by multivariate analyses. In training, unsupervised clustering-based univariate analysis identified 5, 6, and 4 features from IBEX, Pyradiomics, and CIFE as candidate features, respectively. Multivariate prediction models using these features from IBEX, Pyradiomics, and CIFE yielded similar areas under the receiver operating characteristic curve of 0.68, 0.67, and 0.69. However, in validation, areas under the receiver operating characteristic curve of multivariate prediction models from IBEX, Pyradiomics, and CIFE decreased to 0.54, 0.56 and 0.64, respectively. Different feature extractors select different radiomics features, which leads to prediction models with varying performance. However, correlation between those selected features from different extractors may indicate these features measure similar imaging phenotypes associated with similar biological characteristics. Overall, attention should be paid to the generalizability of individual radiomics features and radiomics prediction models.

Application of intraoperative B-mode ultrasound and shear wave elastography for glioma grading

Background

To evaluate the value of intraoperative B-mode ultrasound and shear wave elastography (SWE) in differentiating low-grade and high-grade gliomas.

Methods

A total of 172 patients with glioma were examined by B-mode ultrasound to obtain a tumor sonogram. Intraoperative SWE was performed on 52 patients to obtain Young's modulus values of peritumor tissue and tumor tissue, and the differences in conventional B-mode signs and Young's modulus values of gliomas of different grades were then compared. The diagnostic performance of SWE in glioma grading was assessed by receiver operating characteristic (ROC) curve analysis, and the intra- and interobserver reliability of SWE was analyzed by the intraclass correlation coefficient (ICC).

Results

For B-mode ultrasound, patient age, cystic degeneration, and peritumor edema were independent risk factors for high-grade glioma (P<0.05, OR >1). For SWE, Young's modulus values of peritumor tissue, low-grade glioma, and high-grade glioma tissues were 8.20 (7.50, 9.70) kPa, 19.65 (15.30, 24.75) kPa, and 9.55 (8.50, 13.80) kPa, respectively. The area under the ROC curve for the diagnosis of high-grade glioma by SWE was 0.859 (95% CI: 0.758-0.961, P<0.05), and the optimal cutoff value was 12.1 kPa, with 89.3% sensitivity and 75.0% specificity. The intra- and interobserver reliability of SWE in grading gliomas was excellent, with ICCs ranging from 0.921 to 0.965.

Conclusions

High-grade glioma is associated with significantly more severe necrotic cystic degeneration and peritumoral edema on B-mode ultrasound and lower stiffness on SWE. Further, SWE exhibits excellent intra- and interobserver reliability. Intraoperative B-mode ultrasound combined with SWE helps differentiate different grades of gliomas.

Machine learning applications to neuroimaging for glioma detection and classification: An artificial intelligence augmented systematic review

Glioma is the most common primary intraparenchymal tumor of the brain and the 5-year survival rate of high-grade glioma is poor. Magnetic resonance imaging (MRI) is essential for detecting, characterizing and monitoring brain tumors but definitive diagnosis still relies on surgical pathology. Machine learning has been applied to the analysis of MRI data in glioma research and has the potential to change clinical practice and improve patient outcomes. This systematic review synthesizes and analyzes the current state of machine learning applications to glioma MRI data and explores the use of machine learning for systematic review automation. Various datapoints were extracted from the 153 studies that met inclusion criteria and analyzed. Natural language processing (NLP) analysis involved keyword extraction, topic modeling and document classification. Machine learning has been applied to tumor grading and diagnosis, tumor segmentation, non-invasive genomic biomarker identification, detection of progression and patient survival prediction. Model performance was generally strong (AUC = 0.87 ± 0.09; sensitivity = 0.87 ± 0.10; specificity = 0.0.86 ± 0.10; precision = 0.88 ± 0.11). Convolutional neural network, support vector machine and random forest algorithms were top performers. Deep learning document classifiers yielded acceptable performance (mean 5-fold cross-validation AUC = 0.71). Machine learning tools and data resources were synthesized and summarized to facilitate future research. Machine learning has been widely applied to the processing of MRI data in glioma research and has demonstrated substantial utility. NLP and transfer learning resources enabled the successful development of a replicable method for automating the systematic review article screening process, which has potential for shortening the time from discovery to clinical application in medicine.

MRI radiomic signature predicts intracranial progression-free survival in patients with brain metastases of ALK-positive non-small cell lung cancer

Background

Intracranial progression is considered an important cause of treatment failure in anaplastic lymphoma kinase (ALK)-positive non-small cell lung cancer (NSCLC) patients. Recent advances in targeted therapy and radiomics have generated considerable interest for the exploration of prognostic imaging biomarkers to predict the clinical course. Here, we developed a magnetic resonance imaging (MRI) radiomic signature that can stratify survival and intracranial progression.

Methods

We analyzed 87 brain metastatic lesions in 24 ALK-positive NSCLC patients undergoing ALK-inhibitor ensartinib therapy and divided them into training (n=61) and validation (n=26) sets. Radiomic features were extracted and screened from contrast-enhanced MR images. Combined with these selected features, the Rad-score was calculated with multivariate logistic regression. The predictive model and Rad-score performance were assessed in the training set and validated in the validation set; decision curve analysis was performed with the combined training and validation sets to estimate Rad-score’s patient-stratification ability.

Results

The prediction model constructed with nine selected radiomic features could predict intracranial progression within 51 weeks (AUC =0.84 and 0.85 in the training and validation sets, respectively), while clinical and regular MRI characteristics were independent of progression (P>0.05). The decision-curve analysis showed that the radiomic prediction model was clinically useful. The Kaplan-Meier analysis showed that the progression-free survival (PFS) difference between the high- and low-risk groups distinguished by the Rad-score was significant (P=0.017).

Conclusions

Radiomics may provide prognostic information and improve pretreatment risk stratification in ALK-positive NSCLC patients with brain metastases undergoing ensartinib treatment, allowing follow-up and treatment to be tailored to the patient’s individual risk profile.

Multimodal MRI synthesis using unified generative adversarial networks

PURPOSE

Complementary information obtained from multiple contrasts of tissue facilitates physicians assessing, diagnosing and planning treatment of a variety of diseases. However, acquiring multiple contrasts magnetic resonance images (MRI) for every patient using multiple pulse sequences is time-consuming and expensive, where, medical image synthesis has been demonstrated as an effective alternative. The purpose of this study is to develop a unified framework for multimodal MR image synthesis.

METHODS

A unified generative adversarial network consisting of only a single generator and a single discriminator was developed to learn the mappings among images of four different modalities. The generator took an image and its modality label as inputs and learned to synthesize the image in the target modality, while the discriminator was trained to distinguish between real and synthesized images and classify them to their corresponding modalities. The network was trained and tested using multimodal brain MRI consisting of four different contrasts which are T1-weighted (T1), T1-weighted and contrast-enhanced (T1c), T2-weighted (T2), and fluid-attenuated inversion recovery (Flair). Quantitative assessments of our proposed method were made through computing normalized mean absolute error (NMAE), peak signal-to-noise ratio (PSNR), structural similarity index measurement (SSIM), visual information fidelity (VIF), and naturalness image quality evaluator (NIQE).

RESULTS

The proposed model was trained and tested on a cohort of 274 glioma patients with well-aligned multi-types of MRI scans. After the model was trained, tests were conducted by using each of T1, T1c, T2, Flair as a single input modality to generate its respective rest modalities. Our proposed method shows high accuracy and robustness for image synthesis with arbitrary MRI modality that is available in the database as input. For example, with T1 as input modality, the NMAEs for the generated T1c, T2, Flair respectively are 0.034 ± 0.005, 0.041 ± 0.006, and 0.041 ± 0.006, the PSNRs respectively are 32.353 ± 2.525 dB, 30.016 ± 2.577 dB, and 29.091 ± 2.795 dB, the SSIMs are 0.974 ± 0.059, 0.969 ± 0.059, and 0.959 ± 0.059, the VIF are 0.750 ± 0.087, 0.706 ± 0.097, and 0.654 ± 0.062, and NIQE are 1.396 ± 0.401, 1.511 ± 0.460, and 1.259 ± 0.358, respectively.

CONCLUSIONS

We proposed a novel multimodal MR image synthesis method based on a unified generative adversarial network. The network takes an image and its modality label as inputs and synthesizes multimodal images in a single forward pass. The results demonstrate that the proposed method is able to accurately synthesize multimodal MR images from a single MR image.

Artificial intelligence for brain diseases: A systematic review

Artificial intelligence (AI) is a major branch of computer science that is fruitfully used for analyzing complex medical data and extracting meaningful relationships in datasets, for several clinical aims. Specifically, in the brain care domain, several innovative approaches have achieved remarkable results and open new perspectives in terms of diagnosis, planning, and outcome prediction. In this work, we present an overview of different artificial intelligent techniques used in the brain care domain, along with a review of important clinical applications. A systematic and careful literature search in major databases such as Pubmed, Scopus, and Web of Science was carried out using "artificial intelligence" and "brain" as main keywords. Further references were integrated by cross-referencing from key articles. 155 studies out of 2696 were identified, which actually made use of AI algorithms for different purposes (diagnosis, surgical treatment, intra-operative assistance, and postoperative assessment). Artificial neural networks have risen to prominent positions among the most widely used analytical tools. Classic machine learning approaches such as support vector machine and random forest are still widely used. Task-specific algorithms are designed for solving specific problems. Brain images are one of the most used data types. AI has the possibility to improve clinicians' decision-making ability in neuroscience applications. However, major issues still need to be addressed for a better practical use of AI in the brain. To this aim, it is important to both gather comprehensive data and build explainable AI algorithms.

Brain tumor segmentation using 3D Mask R-CNN for dynamic susceptibility contrast enhanced perfusion imaging

The segmentation of neoplasms is an important part of radiotherapy treatment planning, monitoring disease progression, and predicting patient outcome. In the brain, functional magnetic resonance imaging (MRI) like dynamic susceptibility contrast enhanced (DSCE) or T1-weighted dynamic contrast enhanced (DCE) perfusion MRI are important tools for diagnosis. They play a crucial role in providing pre-operative assessment of tumor histology, grading, and biopsy guidance. However, the manual contouring of these neoplasms is tedious, expensive, time-consuming, and vulnerable to inter-observer variability. In this work, we propose a 3D mask region-based convolutional neural network (R-CNN) method to automatically segment brain tumors in DSCE MRI perfusion images. As our goal is to simultaneously localize and segment the tumor, our training process contained both a region-of-interest (ROI) localization and regression with voxel-wise segmentation. The combination of classification loss, ROI location and size regression loss, and segmentation loss were used to supervise the proposed network. We retrospectively investigated 21 patients' perfusion images, with between 50 and 70 perfusion time point volumes, a total of 1260 3D volumes. Tumor contours were automatically segmented by our proposed method and compared against other state-of-the-art methods and those delineated by physicians as the ground truth. The results of our method demonstrated good agreement with the ground truth contours. The average DSC, precision, recall, Hausdorff distance, mean surface distance (MSD), root MSD, and center of mass distance were 0.90 ± 0.04, 0.91 ± 0.04, 0.90 ± 0.06, 7.16 ± 5.78 mm, 0.45 ± 0.34 mm, 1.03 ± 0.72 mm, and 0.86 ± 0.91 mm, respectively. These results support the feasibility of our method in accurately localizing and segmenting brain tumors in DSCE perfusion MRI. Our 3D Mask R-CNN segmentation method in DSCE perfusion imaging has great promise for future clinical use.

A Novel Intelligent System for Brain Tumor Diagnosis Based on a Composite Neutrosophic-Slantlet Transform Domain for Statistical Texture Feature Extraction

Discrete wavelet transform (DWT) is often implemented by an iterative filter bank; hence, a lake of optimization of a discrete time basis is observed with respect to time localization for a constant number of zero moments. This paper discusses and presents an improved form of DWT for feature extraction, called Slantlet transform (SLT) along with neutrosophy, a generalization of fuzzy logic, which is a relatively new logic. Thus, a novel composite NS-SLT model has been suggested as a source to derive statistical texture features that used to identify the malignancy of brain tumor. The MR images in the neutrosophic domain are defined using three membership sets, true (T), false (F), and indeterminate (I); then, SLT was applied to each membership set. Three statistical measurement-based methods are used to extract texture features from images of brain MRI. One-way ANOVA has been applied as a method of reducing the number of extracted features for the classifiers; then, the extracted features are subsequently provided to the four neural network classification techniques, Support Vector Machine Neural Network (SVM-NN), Decision Tree Neural Network (DT-NN), K-Nearest Neighbor Neural Network (KNN-NN), and Naive Bayes Neural Networks (NB-NN), to predict the type of the brain tumor. Meanwhile, the performance of the proposed model is assessed by calculating average accuracy, precision, sensitivity, specificity, and Area Under the Curve (AUC) of the Receiver Operating Characteristic (ROC) curve. The experimental results demonstrate that the proposed approach is quite accurate and efficient for diagnosing brain tumors when the Gray Level Run Length Matrix (GLRLM) features derived from the composite NS-SLT technique is used.

4D-CT deformable image registration using multiscale unsupervised deep learning

Deformable image registration (DIR) of 4D-CT images is important in multiple radiation therapy applications including motion tracking of soft tissue or fiducial markers, target definition, image fusion, dose accumulation and treatment response evaluations. It is very challenging to accurately and quickly register 4D-CT abdominal images due to its large appearance variances and bulky sizes. In this study, we proposed an accurate and fast multi-scale DIR network (MS-DIRNet) for abdominal 4D-CT registration. MS-DIRNet consists of a global network (GlobalNet) and local network (LocalNet). GlobalNet was trained using down-sampled whole image volumes while LocalNet was trained using sampled image patches. MS-DIRNet consists of a generator and a discriminator. The generator was trained to directly predict a deformation vector field (DVF) based on the moving and target images. The generator was implemented using convolutional neural networks with multiple attention gates. The discriminator was trained to differentiate the deformed images from the target images to provide additional DVF regularization. The loss function of MS-DIRNet includes three parts which are image similarity loss, adversarial loss and DVF regularization loss. The MS-DIRNet was trained in a completely unsupervised manner meaning that ground truth DVFs are not needed. Different from traditional DIRs that calculate DVF iteratively, MS-DIRNet is able to calculate the final DVF in a single forward prediction which could significantly expedite the DIR process. The MS-DIRNet was trained and tested on 25 patients' 4D-CT datasets using five-fold cross validation. For registration accuracy evaluation, target registration errors (TREs) of MS-DIRNet were compared to clinically used software. Our results showed that the MS-DIRNet with an average TRE of 1.2 ± 0.8 mm outperformed the commercial software with an average TRE of 2.5 ± 0.8 mm in 4D-CT abdominal DIR, demonstrating the superior performance of our method in fiducial marker tracking and overall soft tissue alignment.