Molecular subtyping of diffuse gliomas using magnetic resonance imaging: comparison and correlation between radiomics and deep learning

Machine learning decision support model construction for craniotomy approach of pineal region tumors based on MRI images

BACKGROUND

Pineal region tumors (PRTs) are rare but deep-seated brain tumors, and complete surgical resection is crucial for effective tumor treatment. The choice of surgical approach is often challenging due to the low incidence and deep location. This study aims to combine machine learning and deep learning algorithms with pre-operative MRI images to build a model for PRTs surgical approaches recommendation, striving to model clinical experience for practical reference and education.

METHODS

This study was a retrospective study which enrolled a total of 173 patients diagnosed with PRTs radiologically from our hospital. Three traditional surgical approaches of were recorded for prediction label. Clinical and VASARI related radiological information were selected for machine learning prediction model construction. And MRI images from axial, sagittal and coronal views of orientation were also used for deep learning craniotomy approach prediction model establishment and evaluation.

RESULTS

5 machine learning methods were applied to construct the predictive classifiers with the clinical and VASARI features and all methods could achieve area under the ROC (Receiver operating characteristic) curve (AUC) values over than 0.7. And also, 3 deep learning algorithms (ResNet-50, EfficientNetV2-m and ViT) were applied based on MRI images from different orientations. EfficientNetV2-m achieved the highest AUC value of 0.89, demonstrating a significant high performance of prediction. And class activation mapping was used to reveal that the tumor itself and its surrounding relations are crucial areas for model decision-making.

CONCLUSION

In our study, we used machine learning and deep learning to construct surgical approach recommendation models. Deep learning could achieve high performance of prediction and provide efficient and personalized decision support tools for PRTs surgical approach.

CLINICAL TRIAL NUMBER

Not applicable.

Development and validation of a multi-omics hemorrhagic transformation model based on hyperattenuated imaging markers following mechanical thrombectomy

This study aimed to develop a predictive model integrating clinical, radiomics, and deep learning (DL) features of hyperattenuated imaging markers (HIM) from computed tomography scans immediately following mechanical thrombectomy (MT) to predict hemorrhagic transformation (HT). A total of 239 patients with HIM who underwent MT were enrolled, with 191 patients (80%) in the training cohort and 48 patients (20%) in the validation cohort. Additionally, the model was tested on an internal prospective cohort of 49 patients. A total of 1834 radiomics features and 2048 DL features were extracted from HIM images. Statistical methods, such as analysis of variance, Pearson's correlation coefficient, principal component analysis, and least absolute shrinkage and selection operator, were used to select the most significant features. A K-Nearest Neighbor classifier was employed to develop a combined model integrating clinical, radiomics, and DL features for HT prediction. Model performance was evaluated using metrics such as accuracy, sensitivity, specificity, receiver operating characteristic curves, and area under curve (AUC). In the training, validation, and test cohorts, the combined model achieved AUCs of 0.926, 0.923, and 0.887, respectively, outperforming other models, including clinical, radiomics, and DL models, as well as hybrid models combining subsets of features (Clinical + Radiomics, DL + Radiomics, and Clinical + DL) in predicting HT. The combined model, which integrates clinical, radiomics, and DL features derived from HIM, demonstrated efficacy in noninvasively predicting HT. These findings suggest its potential utility in guiding clinical decision-making for patients with MT.

Optimizing Deep Learning Models for Luminal and Nonluminal Breast Cancer Classification Using Multidimensional ROI in DCE-MRI-A Multicenter Study

OBJECTIVES

Previous deep learning studies have not explored the synergistic effects of ROI dimensions (2D/2.5D/3D), peritumoral expansion levels (0-8 mm), and segmentation scenarios (ROI only vs. ROI original). Our study aims to evaluate the performance of multidimensional deep transfer learning models in distinguishing molecular subtypes of breast cancer (luminal vs. nonluminal) using DCE-MRI. Under two segmentation scenarios, we systematically compare the effects of ROI dimensions and peritumoral expansion levels to optimize multidimensional deep learning models via transfer learning for distinguishing luminal from nonluminal breast cancers in DCE-MRI-based analysis.

MATERIALS AND METHODS

From October 2020 to October 2023, data from 426 patients with primary invasive breast cancer were retrospectively collected. Patients were divided into three cohorts: (1) training cohort, n = 108, from SYSU Hospital (Zhuhai, China); (2) validation cohort 1, n = 165, from HZ Hospital (Huizhou, China); and (3) validation cohort 2, n = 153, from LY Hospital (Linyi, China). ROIs were delineated, and expansions of 2, 4, 6, and 8 mm beyond the lesion boundary were performed. We assessed the performance of various deep transfer learning models, considering precise segmentation (ROI only and ROI original) and varying peritumoral regions, using ROC curves and decision curve analysis.

RESULTS

The 2.5D1-based deep learning model (ROI original, 4 mm expansion) demonstrated optimal performance, achieving an AUC of 0.808 (95% CI 0.715-0.901) in the training cohort, 0.766 (95% CI 0.682-0.850) in validation cohort 1, and 0.799 (95% CI 0.725-0.874) in validation cohort 2.

CONCLUSION

The study highlights that the 2.5D1-based deep learning model utilizing the three principal slices of the minimum bounding box (ROI original) with a 4 mm peritumoral region is effective in distinguishing between luminal and nonluminal breast cancer tumors, serving as a potential diagnostic tool.

MRI transformer deep learning and radiomics for predicting IDH wild type TERT promoter mutant gliomas

This study aims to predict IDH wt with TERTp-mut gliomas using multiparametric MRI sequences through a novel fusion model, while matching model classification metrics with patient risk stratification aids in crafting personalized diagnostic and prognosis evaluations.Preoperative T1CE and T2FLAIR sequences from 1185 glioma patients were analyzed. A MultiChannel_2.5D_DL model and a 2D DL model, both based on the cross-scale attention vision transformer (CrossFormer) neural network, along with a Radiomics model, were developed. These were integrated via ensemble learning into a stacking model. The MultiChannel_2.5D_DL model outperformed the 2D_DL and Radiomics models, with AUCs of 0.806-0.870. The stacking model achieved the highest AUC (0.855-0.904) across validation sets. Patients were stratified into high-risk and low-risk groups based on stacking model scores, with significant survival differences observed via Kaplan-Meier analysis and log-rank tests. The stacking model effectively identifies IDH wt TERTp-mutant gliomas and stratifies patient risk, aiding personalized prognosis.

Deep learning based on intratumoral heterogeneity predicts histopathologic grade of hepatocellular carcinoma

OBJECTIVES

The potential of medical imaging to non-invasively assess intratumoral heterogeneity (ITH) is increasingly being recognized. This study aimed to investigate the value of the ITH-based deep learning model for preoperative prediction of histopathologic grade in hepatocellular carcinoma (HCC).

MATERIALS AND METHODS

A total of 858 patients from primary cohort and two external cohorts were included. 3.0T or 1.5T axial portal venous phase MRI images were collected. We conducted radiomics feature-driven K-means clustering for automatic partition to reveal ITH. 2.5D and 3D deep learning models based on ResNet architecture were trained to extract deep learning hidden features of each subregion. The selected features were used to train Random Forest classifier, which constructed the feature-fusion model.

RESULTS

The extracted voxel-level radiomics features were unsupervised clustered by K-means to generate three subregions. In the 2.5D deep learning, the feature-fusion model based on ITH had superior predictive efficacy than the whole-tumor model (AUC: 0.82 vs. 0.72; p = 0.004). Even in the validation and external test sets, this model maintained a high AUC of 0.78-0.83, and net reclassification indices indicated that it could improve prediction by 25-28%. Regarding the prognostic value, overall survival (OS) and recurrence-free survival (RFS) could be significantly stratified by the 2.5D feature-fusion model, and multivariable Cox regressions indicated its signature was identified as a risk predictor for OS and RFS (p < 0.05).

CONCLUSION

The ITH-based feature-fusion model provided a non-invasive method for classifying tumor differentiation in HCC, which may serve as a promising strategy for stratification management.

Convolutional neural network-assisted Raman spectroscopy for high-precision diagnosis of glioblastoma

Glioblastoma multiforme (GBM) is the most lethal intracranial tumor with a median survival of approximately 15 months. Due to its highly invasive properties, it is particularly difficult to accurately identify the tumor margins intraoperatively. The current gold standard for diagnosing GBM during surgery is pathology, but it is time-consuming. Under these circumstances, we developed a method combining Raman spectroscopy (RS) with convolutional neural networks (CNN) to distinguish GBM. Analysis of the spectra of normal brain samples (478 spectra) and GBM samples (462 spectra) from 29 in situ intracranial tumor-bearing mice showed that this method identified GBM tissue with 96.8 % accuracy. Subsequently, spectral analysis of 23 normal human brain tissues (223 spectra) versus 21 tissues from patients with pathologically diagnosed GBM (267 spectra) revealed that the accuracy of this method was 93.9 %. Most importantly, for the difference peaks in the spectra of GBM and normal brain tissue, the common difference peaks in the mouse and human spectra were at 750 cm-1, 1440 cm-1, and 1586 cm-1, which emphasized the differences in cytochrome C and lipids between GBM samples and normal brain samples in both mice and human. The preliminary results showed that CNN-assisted RS is simple to operate and can rapidly and accurately identify whether it is GBM tissue or normal brain tissue.

Interpretable multimodal transformer for prediction of molecular subtypes and grades in adult-type diffuse gliomas

Molecular subtyping and grading of adult-type diffuse gliomas are essential for treatment decisions and patient prognosis. We introduce GlioMT, an interpretable multimodal transformer that integrates imaging and clinical data to predict the molecular subtype and grade of adult-type diffuse gliomas according to the 2021 WHO classification. GlioMT is trained on multiparametric MRI data from an institutional set of 1053 patients with adult-type diffuse gliomas to predict the IDH mutation status, 1p/19q codeletion status, and tumor grade. External validation on the TCGA (200 patients) and UCSF (477 patients) shows that GlioMT outperforms conventional CNNs and visual transformers, achieving AUCs of 0.915 (TCGA) and 0.981 (UCSF) for IDH mutation, 0.854 (TCGA) and 0.806 (UCSF) for 1p/19q codeletion, and 0.862 (TCGA) and 0.960 (UCSF) for grade prediction. GlioMT enhances the reliability of clinical decision-making by offering interpretability through attention maps and contributions of imaging and clinical data.

Multiparametric MRI-based radiomics combined with 3D deep transfer learning to predict cervical stromal invasion in patients with endometrial carcinoma

OBJECTIVE

To develop and compare various preoperative cervical stromal invasion (CSI) prediction models, including radiomics, three-dimensional (3D) deep transfer learning (DTL), and integrated models, using single-sequence and multiparametric MRI.

METHODS

Data from 466 early-stage endometrial carcinoma (EC) patients from three centers were collected. Radiomics models were constructed based on T2-weighted imaging (T2WI), diffusion-weighted imaging (DWI), apparent diffusion coefficient (ADC) mapping, contrast-enhanced T1-weighted imaging (CE-T1WI), and four combined sequences as well as 3D DTL models. Two integrated models were created using ensemble and stacking algorithms based on optimal radiomics and DTL models. Model performance and clinical benefits were assessed using area under the curve (AUC), decision curve analysis (DCA), net reclassification index (NRI), integrated discrimination index (IDI), and the Delong test for model comparisons.

RESULTS

Multiparametric MRI models were superior to single-sequence models for radiomics or DTL models. Ensemble and stacking integrated models displayed excellent performance. The stacking model had the highest average AUC (0.908) and accuracy (0.883) in external validation groups 1 and 2 (AUC = 0.965 and 0.851, respectively) and emerged as the best predictive model for CSI. All models significantly outperformed the radiologist (P < 0.05). In terms of net benefits, all models demonstrated favorable outcomes in DCA, NRI, and IDI, with the stacking model yielding the highest net benefit.

CONCLUSION

Multiparametric MRI-based radiomics combined with 3D DTL can be used to noninvasively predict CSI in EC patients with greater diagnostic accuracy than the radiologist. Stacking integrated models showed significant potential utility in predicting CSI. Which helps to provide new treatment strategy for clinicians to treat early-stage EC patients.

Radiomics in glioma: emerging trends and challenges

Radiomics is a promising neuroimaging technique for extracting and analyzing quantitative glioma features. This review discusses the application, emerging trends, and challenges associated with using radiomics in glioma. Integrating deep learning algorithms enhances various radiomics components, including image normalization, region of interest segmentation, feature extraction, feature selection, and model construction and can potentially improve model accuracy and performance. Moreover, investigating specific tumor habitats of glioblastomas aids in a better understanding of glioblastoma aggressiveness and the development of effective treatment strategies. Additionally, advanced imaging techniques, such as diffusion-weighted imaging, perfusion-weighted imaging, magnetic resonance spectroscopy, magnetic resonance fingerprinting, functional MRI, and positron emission tomography, can provide supplementary information for tumor characterization and classification. Furthermore, radiomics analysis helps understand the glioma immune microenvironment by predicting immune-related biomarkers and characterizing immune responses within tumors. Integrating multi-omics data, such as genomics, transcriptomics, proteomics, and pathomics, with radiomics, aids the understanding of the biological significance of the underlying radiomics features and improves the prediction of genetic mutations, prognosis, and treatment response in patients with glioma. Addressing challenges, such as model reproducibility, model generalizability, model interpretability, and multi-omics data integration, is crucial for the clinical translation of radiomics in glioma.

Deep learning radiomics model based on contrast-enhanced MRI for distinguishing between tuberculous spondylitis and pyogenic spondylitis

PURPOSE

This study aimed to develop and validate a deep learning radiomics nomogram (DLRN) to differentiate between tuberculous spondylitis (TS) and pyogenic spondylitis (PS) using contrast-enhanced MRI (CE-MRI).

METHODS

A retrospective approach was employed, enrolling patients diagnosed with TS or PS based on pathological examination at two centers. Clinical features were evaluated to establish a clinical model. Radiomics and deep learning (DL) features were extracted from contrast-enhanced T1-weighted images and subsequently fused. Following feature selection, radiomics, DL, combined DL-radiomics (DLR), and a deep learning radiomics nomogram (DLRN) were developed to differentiate TS from PS. Performance was assessed using metrics including the area under the curve (AUC), calibration curves, and decision curve analysis (DCA).

RESULTS

A total of 147 patients met the study criteria. Center 1 comprised the training cohort with 102 patients (52 TS and 50 PS), while Center 2 served as the external test cohort with 45 patients (17 TS and 28 PS). The DLRN model exhibited the highest diagnostic accuracy, achieving an AUC of 0.994 (95% CI: 0.983-1.000) in the training cohort and 0.859 (95% CI: 0.744-0.975) in the external test cohort. Calibration curves indicated good agreement for DLRN, and decision curve analysis (DCA) demonstrated it provided the greatest clinical benefit.

CONCLUSION

The CE-MRI-based DLRN showed robust diagnostic capability for distinguishing between TS and PS in clinical practice.

Virtual Biopsy for the Prediction of MGMT Promoter Methylation in Gliomas: A Comprehensive Review of Radiomics and Deep Learning Approaches Applied to MRI

Background/Objectives: The methylation status of the O6-methylguanine-DNA methyltransferase (MGMT) promoter in gliomas has emerged as a critical biomarker for prognosis and treatment response. Conventional methods for assessing MGMT promoter methylation, such as methylation-specific PCR, are invasive and require tissue sampling. Methods: A comprehensive literature search was performed in compliance with the updated PRISMA 2020 guidelines within electronic databases MEDLINE/PubMed, Scopus, and IEEE Xplore. Search terms, including "MGMT", "methylation", "glioma", "glioblastoma", "machine learning", "deep learning", and "radiomics", were adopted in various MeSH combinations. Original studies in the English, Italian, German, and French languages were considered for inclusion. Results: This review analyzed 34 studies conducted in the last six years, focusing on assessing MGMT methylation status using radiomics (RD), deep learning (DL), or combined approaches. These studies utilized radiological data from the public (e.g., BraTS, TCGA) and private institutional datasets. Sixteen studies focused exclusively on glioblastoma (GBM), while others included low- and high-grade gliomas. Twenty-seven studies reported diagnostic accuracy, with fourteen achieving values above 80%. The combined use of DL and RD generally resulted in higher accuracy, sensitivity, and specificity, although some studies reported lower minimum accuracy compared to studies using a single model. Conclusions: The integration of RD and DL offers a powerful, non-invasive tool for precisely recognizing MGMT promoter methylation status in gliomas, paving the way for enhanced personalized medicine in neuro-oncology. The heterogeneity of study populations, data sources, and methodologies reflected the complexity of the pipeline and machine learning algorithms, which may require general standardization to be implemented in clinical practice.

Magnetic resonance imaging radiomics based on artificial intelligence is helpful to evaluate the prognosis of single hepatocellular carcinoma

Background

Previous studies mostly use single-type features to establish a prediction model. We aim to develop a comprehensive prediction model that effectively identify patients with poor prognosis for single hepatocellular carcinoma (HCC) based on artificial intelligence (AI). Patients and methods: 236 single HCC patients were studied to establish a comprehensive prediction model. We collected the basic information of patients and used AI to extract the features of magnetic resonance (MR) images.

Results

The clinical model based on linear regression (LR) algorithm (AUC: 0.658, 95%CI: 0.5021-0.8137), the radiomics model and deep transfer learning (DTL) model based on light gradient-boosting machine (Light GBM) algorithm (AUC: 0.761, 95%CI: 0.6326-0.8886 and AUC: 0.784, 95%CI: 0.6587-0.9087, respectively) were the optimal prediction models. A comparison revealed that the integrated nomogram had the largest area under the receiver operating characteristic curve (AUC) (all P < 0.05). In the training cohort, the integrated nomogram was predictive of recurrence-free survival (RFS) as well as overall survival (OS) (C-index: 0.735 and 0.712, P < 0.001). In the test cohort, the integrated nomogram also can predict RFS and OS (C-index: 0.718 and 0.740, P < 0.001) in patients.

Conclusion

The integrated nomogram composed of signatures in the prediction models can not only predict the postoperative recurrence of single HCC patients but also stratify the risk of OS after the operation.

Deep Learning of radiology-genomics integration for computational oncology: A mini review

In the field of computational oncology, patient status is often assessed using radiology-genomics, which includes two key technologies and data, such as radiology and genomics. Recent advances in deep learning have facilitated the integration of radiology-genomics data, and even new omics data, significantly improving the robustness and accuracy of clinical predictions. These factors are driving artificial intelligence (AI) closer to practical clinical applications. In particular, deep learning models are crucial in identifying new radiology-genomics biomarkers and therapeutic targets, supported by explainable AI (xAI) methods. This review focuses on recent developments in deep learning for radiology-genomics integration, highlights current challenges, and outlines some research directions for multimodal integration and biomarker discovery of radiology-genomics or radiology-omics that are urgently needed in computational oncology.

Glioma subtype prediction based on radiomics of tumor and peritumoral edema under automatic segmentation

Comprehensive and non-invasive preoperative molecular diagnosis is important for prognostic and therapy decision-making in adult-type diffuse gliomas. We employed a deep learning method for automatic segmentation of brain gliomas directly from conventional magnetic resonance imaging (MRI) scans of the tumor core and peritumoral edema regions based on available glioma MRI data provided in the BraTS2021. Three-dimensional volumes of interest were segmented from 424 cases of glioma imaging data retrospectively obtained from two medical centers using the segmentation method and radiomic features were extracted. We developed a subtype prediction model based on extracted radiomic features and analyzed significance and correlations between glioma morphological characteristics and pathological features using data from patients with adult-type diffuse glioma. The automated segmentation achieved mean Dice scores of 0.884 and 0.889 for the tumor core and whole tumor, respectively. The area under the receiver operating characteristic curve for the prediction of adult-type diffuse gliomas subtypes was 0.945. "Glioblastoma, IDH-wildtype", "Astrocytoma, IDH-mutant", and "Oligodendroglioma, IDH-mutant, 1p/19q-coded" showed AUCs of 0.96, 0.914, and 0.961, respectively, for subtype prediction. Glioma morphological characteristics, molecular and pathological levels, and clinical data showed significant differences and correlations. An automatic segmentation model for gliomas based on 3D U-Nets was developed, and the prediction model for gliomas built using the parameters obtained from the automatic segmentation model showed high overall performance.

Application of machine learning for the differentiation of thymomas and thymic cysts using deep transfer learning: A multi-center comparison of diagnostic performance based on different dimensional models

OBJECTIVE

This study aimed to evaluate the feasibility and performance of deep transfer learning (DTL) networks with different types and dimensions in differentiating thymomas from thymic cysts in a retrospective cohort.

MATERIALS AND METHODS

Based on chest-enhanced computed tomography (CT), the region of interest was delineated, and the maximum cross section of the lesion was selected as the input image. Five convolutional neural networks (CNNs) and Vision Transformer (ViT) were used to construct a 2D DTL model. The 2D model constructed by the maximum section (n) and the upper and lower layers (n - 1, n + 1) of the lesion was used for feature extraction, and the features were selected. The remaining features were pre-fused to construct a 2.5D model. The whole lesion image was selected for input and constructing a 3D model.

RESULTS

In the 2D model, the area under curve (AUC) of Resnet50 was 0.950 in the training cohort and 0.907 in the internal validation cohort. In the 2.5D model, the AUCs of Vgg11 in the internal validation cohort and external validation cohort 1 were 0.937 and 0.965, respectively. The AUCs of Inception_v3 in the training cohort and external validation cohort 2 were 0.981 and 0.950, respectively. The AUC values of 3D_Resnet50 in the four cohorts were 0.987, 0.937, 0.938, and 0.905.

CONCLUSIONS

The DTL model based on multiple different dimensions can be used as a highly sensitive and specific tool for the non-invasive differential diagnosis of thymomas and thymic cysts to assist clinicians in decision-making.

Two decades of progress in glioma methylation research: the rise of temozolomide resistance and immunotherapy insights

Aims

This study aims to systematically analyze the global trends in glioma methylation research using bibliometric methodologies. We focus on identifying the scholarly trajectory and key research interests, and we utilize these insights to predict future research directions within the epigenetic context of glioma.

Methods

We performed a comprehensive literature search of the Web of Science Core Collection (WoSCC) to identify articles related to glioma methylation published from January 1, 2004, to December 31, 2023. The analysis included full-text publications in the English language and excluded non-research publications. Analysis and visualization were performed using GraphPad Prism, CiteSpace, and VOSviewer software.

Results

The search identified 3,744 publications within the WoSCC database, including 3,124 original research articles and 620 review articles. The research output gradually increased from 2004 to 2007, followed by a significant increase after 2008, which peaked in 2022. A minor decline in publication output was noted during 2020-2021, potentially linked to the coronavirus disease 2019 pandemic. The United States and China were the leading contributors, collectively accounting for 57.85% of the total research output. The Helmholtz Association of Germany, the German Cancer Research Center (DKFZ), and the Ruprecht Karls University of Heidelberg were the most productive institutions. The Journal of Neuro-Oncology led in terms of publication volume, while Neuro-Oncology had the highest Impact Factor. The analysis of publishing authors revealed Michael Weller as the most prolific contributor. The co-citation network analysis identified David N. Louis's article as the most frequently cited. The keyword analysis revealed "temozolomide," "expression," "survival," and "DNA methylation" as the most prominent keywords, while "heterogeneity," "overall survival," and "tumor microenvironment" showed the strongest citation bursts.

Conclusions

The findings of this study illustrate the increasing scholarly interest in glioma methylation, with a notable increase in research output over the past two decades. This study provides a comprehensive overview of the research landscape, highlighting the importance of temozolomide, DNA methylation, and the tumor microenvironment in glioma research. Despite its limitations, this study offers valuable insights into the current research trends and potential future directions, particularly in the realm of immunotherapy and epigenetic editing techniques.

Radiogenomics: bridging the gap between imaging and genomics for precision oncology

Genomics allows the tracing of origin and evolution of cancer at molecular scale and underpin modern cancer diagnosis and treatment systems. Yet, molecular biomarker-guided clinical decision-making encounters major challenges in the realm of individualized medicine, consisting of the invasiveness of procedures and the sampling errors due to high tumor heterogeneity. By contrast, medical imaging enables noninvasive and global characterization of tumors at a low cost. In recent years, radiomics has overcomes the limitations of human visual evaluation by high-throughput quantitative analysis, enabling the comprehensive utilization of the vast amount of information underlying radiological images. The cross-scale integration of radiomics and genomics (hereafter radiogenomics) has the enormous potential to enhance cancer decoding and act as a catalyst for digital precision medicine. Herein, we provide a comprehensive overview of the current framework and potential clinical applications of radiogenomics in patient care. We also highlight recent research advances to illustrate how radiogenomics can address common clinical problems in solid tumors such as breast cancer, lung cancer, and glioma. Finally, we analyze existing literature to outline challenges and propose solutions, while also identifying future research pathways. We believe that the perspectives shared in this survey will provide a valuable guide for researchers in the realm of radiogenomics aiming to advance precision oncology.

Development of an MRI-Based Comprehensive Model Fusing Clinical, Radiomics and Deep Learning Models for Preoperative Histological Stratification in Intracranial Solitary Fibrous Tumor

BACKGROUND

Accurate preoperative histological stratification (HS) of intracranial solitary fibrous tumors (ISFTs) can help predict patient outcomes and develop personalized treatment plans. However, the role of a comprehensive model based on clinical, radiomics and deep learning (CRDL) features in preoperative HS of ISFT remains unclear.

PURPOSE

To investigate the feasibility of a CRDL model based on magnetic resonance imaging (MRI) in preoperative HS in ISFT.

STUDY TYPE

Retrospective.

POPULATION

Three hundred and ninety-eight patients from Beijing Tiantan Hospital, Capital Medical University (primary training cohort) and 49 patients from Lanzhou University Second Hospital (external validation cohort) with ISFT based on histopathological findings (237 World Health Organization [WHO] tumor grade 1 or 2, and 210 WHO tumor grade 3).

FIELD STRENGTH/SEQUENCE

3.0 T/T1-weighted imaging (T1) by using spin echo sequence, T2-weighted imaging (T2) by using fast spin echo sequence, and T1-weighted contrast-enhanced imaging (T1C) by using two-dimensional fast spin echo sequence.

ASSESSMENT

Area under the receiver operating characteristic curve (AUC) was used to assess the performance of the CRDL model and a clinical model (CM) in preoperative HS in the external validation cohort. The decision curve analysis (DCA) was used to evaluate the clinical net benefit provided by the CRDL model.

STATISTICAL TESTS

Cohen's kappa, intra-/inter-class correlation coefficients (ICCs), Chi-square test, Fisher's exact test, Student's t-test, AUC, DCA, calibration curves, DeLong test. A P value <0.05 was considered statistically significant.

RESULTS

The CRDL model had significantly better discrimination ability than the CM (AUC [95% confidence interval, CI]: 0.895 [0.807-0.912] vs. 0.810 [0.745-0.874], respectively) in the external validation cohort. The CRDL model can provide a clinical net benefit for preoperative HS at a threshold probability >20%.

DATA CONCLUSION

The proposed CRDL model holds promise for preoperative HS in ISFT, which is important for predicting patient outcomes and developing personalized treatment plans.

LEVEL OF EVIDENCE

3 TECHNICAL EFFICACY: Stage 2.

CT-Based Super-Resolution Deep Learning Models with Attention Mechanisms for Predicting Spread Through Air Spaces of Solid or Part-Solid Lung Adenocarcinoma

RATIONALE AND OBJECTIVES

Spread through air space (STAS) is a novel invasive pattern of lung adenocarcinoma (LUAD), and preoperative knowledge of STAS status is helpful in choosing an appropriate surgical approach.

MATERIALS AND METHODS

This retrospective study collected and analyzed 602 patients diagnosed with LUAD from two medical centers: center 1 was randomly partitioned into training (n = 358) and validation cohorts (n = 154) at a 7:3 ratio; and center 2 was the external test cohort (n = 90). Super resolution was performed on all images to acquire high-resolution images, which were used to train the SE-ResNet50 model, before creating an equivalent parameter ResNet50 model. Disparities were compared between the two models using receiver operating characteristic curves, area under the curve, accuracy, precision, sensitivity, and specificity.

RESULTS

In this study, 512 and 90 patients with LUAD were enrolled from centers 1 and 2, respectively. The curve values of the SE-ResNet50 and ResNet50 models were compared for training, validation, and test cohorts, resulting in values of 0.933 vs 0.909, 0.783 vs 0.728, and 0.806 vs 0.695, respectively. In the external test cohort, the accuracy of the SE-ResNet50 model demonstrated a 10% improvement over the ResNet50 model (82.2% vs 72.2%).

CONCLUSION

The SE-ResNet50 model based on computed tomography super-resolution has great potential for predicting STAS status in patients with solid or partially solid LUAD, with superior predictive performance compared to traditional deep learning models.

MRI-based intratumoral and peritumoral radiomics for preoperative prediction of glioma grade: a multicenter study

Background

Accurate preoperative prediction of glioma is crucial for developing individualized treatment decisions and assessing prognosis. In this study, we aimed to establish and evaluate the value of integrated models by incorporating the intratumoral and peritumoral features from conventional MRI and clinical characteristics in the prediction of glioma grade.

Methods

A total of 213 glioma patients from two centers were included in the retrospective analysis, among which, 132 patients were classified as the training cohort and internal validation set, and the remaining 81 patients were zoned as the independent external testing cohort. A total of 7728 features were extracted from MRI sequences and various volumes of interest (VOIs). After feature selection, 30 radiomic models depended on five sets of machine learning classifiers, different MRI sequences, and four different combinations of predictive feature sources, including features from the intratumoral region only, features from the peritumoral edema region only, features from the fusion area including intratumoral and peritumoral edema region (VOI-fusion), and features from the intratumoral region with the addition of features from peritumoral edema region (feature-fusion), were established to select the optimal model. A nomogram based on the clinical parameter and optimal radiomic model was constructed for predicting glioma grade in clinical practice.

Results

The intratumoral radiomic models based on contrast-enhanced T1-weighted and T2-flair sequences outperformed those based on a single MRI sequence. Moreover, the internal validation and independent external test underscored that the XGBoost machine learning classifier, incorporating features extracted from VOI-fusion, showed superior predictive efficiency in differentiating between low-grade gliomas (LGG) and high-grade gliomas (HGG), with an AUC of 0.805 in the external test. The radiomic models of VOI-fusion yielded higher prediction efficiency than those of feature-fusion. Additionally, the developed nomogram presented an optimal predictive efficacy with an AUC of 0.825 in the testing cohort.

Conclusion

This study systematically investigated the effect of intratumoral and peritumoral radiomics to predict glioma grading with conventional MRI. The optimal model was the XGBoost classifier coupled radiomic model based on VOI-fusion. The radiomic models that depended on VOI-fusion outperformed those that depended on feature-fusion, suggesting that peritumoral features should be rationally utilized in radiomic studies.

Preoperative Discrimination of CDKN2A/B Homozygous Deletion Status in Isocitrate Dehydrogenase-Mutant Astrocytoma: A Deep Learning-Based Radiomics Model Using MRI

BACKGROUND

Cyclin-dependent kinase inhibitor 2A/B (CDKN2A/B) homozygous deletion has been verified as an independent and critical biomarker of negative prognosis and short survival in isocitrate dehydrogenase (IDH)-mutant astrocytoma. Therefore, noninvasive and accurate discrimination of CDKN2A/B homozygous deletion status is essential for the clinical management of IDH-mutant astrocytoma patients.

PURPOSE

To develop a noninvasive, robust preoperative model based on MR image features for discriminating CDKN2A/B homozygous deletion status of IDH-mutant astrocytoma.

STUDY TYPE

Retrospective.

POPULATION

Two hundred fifty-one patients: 107 patients with CDKN2A/B homozygous deletion and 144 patients without CDKN2A/B homozygous deletion.

FIELD STRENGTH/SEQUENCE

3.0 T/1.5 T: Contrast-enhanced T1-weighted spin-echo inversion recovery sequence (CE-T1WI) and T2-weighted fluid-attenuation spin-echo inversion recovery sequence (T2FLAIR).

ASSESSMENT

A total of 1106 radiomics and 1000 deep learning features extracted from CE-T1WI and T2FLAIR were used to develop models to discriminate the CDKN2A/B homozygous deletion status. Radiomics models, deep learning-based radiomics (DLR) models and the final integrated model combining radiomics features with deep learning features were developed and compared their preoperative discrimination performance.

STATISTICAL TESTING

Pearson chi-square test and Mann Whitney U test were used for assessing the statistical differences in patients' clinical characteristics. The Delong test compared the statistical differences of receiver operating characteristic (ROC) curves and area under the curve (AUC) of different models. The significance threshold is P < 0.05.

RESULTS

The final combined model (training AUC = 0.966; validation AUC = 0.935; test group: AUC = 0.943) outperformed the optimal models based on only radiomics or DLR features (training: AUC = 0.916 and 0.952; validation: AUC = 0.886 and 0.912; test group: AUC = 0.862 and 0.902).

DATA CONCLUSION

Whether based on a single sequence or a combination of two sequences, radiomics and DLR models have achieved promising performance in assessing CDKN2A/B homozygous deletion status. However, the final model combining both deep learning and radiomics features from CE-T1WI and T2FLAIR outperformed the optimal radiomics or DLR model.

EVIDENCE LEVEL

4 TECHNICAL EFFICACY: Stage 2.

Circular RNAs in glioma: Molecular functions and pathological implications

Circular RNAs (circRNAs) are a special class of non-coding RNAs with the ring structure. They are stable, abundant and conservative across mammals. The biogenesis and molecular properties of circRNAs are being elucidated, which exert regulatory functions not only through miRNA and protein sponge, but also via translation and exosomal interaction. Accumulating studies have demonstrated that circRNAs are aberrantly expressed in various diseases, especially in cancer. Glioma is one of the most common malignant cerebral neoplasms with poor prognosis. The accurate diagnosis and effective therapies of glioma have always been challenged, there is an urgent need for developing promising therapeutic intervention. Therefore, exploring novel biomarkers is crucial for diagnosis, treatment and prognosis of the glioma which can provide better assistance in guiding treatment. Recent findings found that circRNAs are systematically altered in glioma and may play critical roles in glioma tumorigenesis, proliferation, invasion and metastasis. Due to their distinct functional properties, they are considered as the potential therapeutic targets, diagnostic and prognostic biomarkers. This review elaborates on current advances towards the biogenesis, translation and interaction of circRNAs in many diseases and focused on the role of their involvement in glioma progression, highlighting the potential value of circRNAs in glioma.

The effect of data resampling methods in radiomics

Radiomic datasets can be class-imbalanced, for instance, when the prevalence of diseases varies notably, meaning that the number of positive samples is much smaller than that of negative samples. In these cases, the majority class may dominate the model's training and thus negatively affect the model's predictive performance, leading to bias. Therefore, resampling methods are often utilized to class-balance the data. However, several resampling methods exist, and neither their relative predictive performance nor their impact on feature selection has been systematically analyzed. In this study, we aimed to measure the impact of nine resampling methods on radiomic models utilizing a set of fifteen publicly available datasets regarding their predictive performance. Furthermore, we evaluated the agreement and similarity of the set of selected features. Our results show that applying resampling methods did not improve the predictive performance on average. On specific datasets, slight improvements in predictive performance (+ 0.015 in AUC) could be seen. A considerable disagreement on the set of selected features was seen (only 28.7% of features agreed), which strongly impedes feature interpretability. However, selected features are similar when considering their correlation (82.9% of features correlated on average).

Automatic Osteoporosis Screening System Using Radiomics and Deep Learning from Low-Dose Chest CT Images

OBJECTIVE

Develop two fully automatic osteoporosis screening systems using deep learning (DL) and radiomics (Rad) techniques based on low-dose chest CT (LDCT) images and evaluate their diagnostic effectiveness.

METHODS

In total, 434 patients who underwent LDCT and bone mineral density (BMD) examination were retrospectively enrolled and divided into the development set (n = 333) and temporal validation set (n = 101). An automatic thoracic vertebra cancellous bone (TVCB) segmentation model was developed. The Dice similarity coefficient (DSC) was used to evaluate the segmentation performance. Furthermore, the three-class Rad and DL models were developed to distinguish osteoporosis, osteopenia, and normal bone mass. The diagnostic performance of these models was evaluated using the receiver operating characteristic (ROC) curve and decision curve analysis (DCA).

RESULTS

The automatic segmentation model achieved excellent segmentation performance, with a mean DSC of 0.96 ± 0.02 in the temporal validation set. The Rad model was used to identify osteoporosis, osteopenia, and normal BMD in the temporal validation set, with respective area under the receiver operating characteristic curve (AUC) values of 0.943, 0.801, and 0.932. The DL model achieved higher AUC values of 0.983, 0.906, and 0.969 for the same categories in the same validation set. The Delong test affirmed that both models performed similarly in BMD assessment. However, the accuracy of the DL model is 81.2%, which is better than the 73.3% accuracy of the Rad model in the temporal validation set. Additionally, DCA indicated that the DL model provided a greater net benefit compared to the Rad model across the majority of the reasonable threshold probabilities Conclusions: The automated segmentation framework we developed can accurately segment cancellous bone on low-dose chest CT images. These predictive models, which are based on deep learning and radiomics, provided comparable diagnostic performance in automatic BMD assessment. Nevertheless, it is important to highlight that the DL model demonstrates higher accuracy and precision than the Rad model.

CT-based intratumoral and peritumoral deep transfer learning features prediction of lymph node metastasis in non-small cell lung cancer

BACKGROUND

The main metastatic route for lung cancer is lymph node metastasis, and studies have shown that non-small cell lung cancer (NSCLC) has a high risk of lymph node infiltration.

OBJECTIVE

This study aimed to compare the performance of handcrafted radiomics (HR) features and deep transfer learning (DTL) features in Computed Tomography (CT) of intratumoral and peritumoral regions in predicting the metastatic status of NSCLC lymph nodes in different machine learning classifier models.

METHODS

We retrospectively collected data of 199 patients with pathologically confirmed NSCLC. All patients were divided into training (n = 159) and validation (n = 40) cohorts, respectively. The best HR and DTL features in the intratumoral and peritumoral regions were extracted and selected, respectively. Support Vector Machine (SVM), k-Nearest Neighbors (KNN), Light Gradient Boosting Machine (Light GBM), Multilayer Perceptron (MLP), and Logistic Regression (LR) models were constructed, and the performance of the models was evaluated.

RESULTS

Among the five models in the training and validation cohorts, the LR classifier model performed best in terms of HR and DTL features. The AUCs of the training cohort were 0.841 (95% CI: 0.776-0.907) and 0.955 (95% CI: 0.926-0.983), and the AUCs of the validation cohort were 0.812 (95% CI: 0.677-0.948) and 0.893 (95% CI: 0.795-0.991), respectively. The DTL signature was superior to the handcrafted radiomics signature.

CONCLUSIONS

Compared with the radiomics signature, the DTL signature constructed based on intratumoral and peritumoral areas in CT can better predict NSCLC lymph node metastasis.

Artificial Intelligence in Andrology: From Semen Analysis to Image Diagnostics

Artificial intelligence (AI) in medicine has gained a lot of momentum in the last decades and has been applied to various fields of medicine. Advances in computer science, medical informatics, robotics, and the need for personalized medicine have facilitated the role of AI in modern healthcare. Similarly, as in other fields, AI applications, such as machine learning, artificial neural networks, and deep learning, have shown great potential in andrology and reproductive medicine. AI-based tools are poised to become valuable assets with abilities to support and aid in diagnosing and treating male infertility, and in improving the accuracy of patient care. These automated, AI-based predictions may offer consistency and efficiency in terms of time and cost in infertility research and clinical management. In andrology and reproductive medicine, AI has been used for objective sperm, oocyte, and embryo selection, prediction of surgical outcomes, cost-effective assessment, development of robotic surgery, and clinical decision-making systems. In the future, better integration and implementation of AI into medicine will undoubtedly lead to pioneering evidence-based breakthroughs and the reshaping of andrology and reproductive medicine.

Noninvasive prediction of IDH mutation status in gliomas using preoperative multiparametric MRI radiomics nomogram: A mutlicenter study

PURPOSE

To establish and validate a radiomics nomogram for preoperative prediction of isocitrate dehydrogenase (IDH) mutation status of gliomas in a multicenter setting.

METHODS

414 gliomas patients were collected (306 from local institution and 108 from TCGA). 851 radiomics features were extracted from contrast-enhanced T1-weighted (CE-T1W) and fluid attenuated inversion recovery (FLAIR) sequence, respectively. The features were refined using least absolute shrinkage and selection operator (LASSO) regression combing 10-fold cross-validation. The optimal radiomics features with age and sex were processed by multivariate logistic regression analysis to construct a prediction model, which was developed in the training dataset and assessed in the test and validation dataset. Receiver operating characteristic (ROC) curve, calibration curve and decision curve analysis were applied in the test and external validation datasets to evaluate the performance of the prediction model.

RESULTS

Ten robust radiomics features were selected from the 1702 features (four CE-T1W features and six FLAIR features). A nomogram was plotted to represent the prediction model. The accuracy and AUC of the radiomics nomogram achieved 86.96% and 0.891(0.809-0.947) in the test dataset and 84.26% and 0.881(0.805-0.936) in the external validation dataset (all p < 0.05). The positive predictive value (PPV) and negative predictive value (NPV) were 83.72% and 87.75% in the test dataset and 87.81% and 82.09% in the external validation dataset.

CONCLUSION

IDH genotypes of gliomas can be identified by preoperative multiparametric MRI radiomics nomogram and might be clinically meaningful for treatment strategy and prognosis stratification of gliomas.

Coordinatized lesion location analysis empowering ROI-based radiomics diagnosis on brain gliomas

OBJECTIVES

To assess the value of coordinatized lesion location analysis (CLLA), in empowering ROI-based imaging diagnosis of gliomas by improving accuracy and generalization performances.

METHODS

In this retrospective study, pre-operative contrasted T1-weighted and T2-weighted MR images were obtained from patients with gliomas from three centers: Jinling Hospital, Tiantan Hospital, and the Cancer Genome Atlas Program. Based on CLLA and ROI-based radiomic analyses, a fusion location-radiomics model was constructed to predict tumor grades, isocitrate dehydrogenase (IDH) status, and overall survival (OS). An inter-site cross-validation strategy was used for assessing the performances of the fusion model on accuracy and generalization with the value of area under the curve (AUC) and delta accuracy (ACC) (ACCtesting-ACCtraining). Comparisons of diagnostic performances were performed between the fusion model and the other two models constructed with location and radiomics analysis using DeLong's test and Wilcoxon signed ranks test.

RESULTS

A total of 679 patients (mean age, 50 years ± 14 [standard deviation]; 388 men) were enrolled. Based on tumor location probabilistic maps, fusion location-radiomics models (averaged AUC values of grade/IDH/OS: 0.756/0.748/0.768) showed the highest accuracy in contrast to radiomics models (0.731/0.686/0.716) and location models (0.706/0.712/0.740). Notably, fusion models ([median Delta ACC: - 0.125, interquartile range: 0.130]) demonstrated improved generalization than that of radiomics model ([- 0.200, 0.195], p = 0.018).

CONCLUSIONS

CLLA could empower ROI-based radiomics diagnosis of gliomas by improving the accuracy and generalization of the models.

CLINICAL RELEVANCE STATEMENT

This study proposed a coordinatized lesion location analysis for glioma diagnosis, which could improve the performances of the conventional ROI-based radiomics model in accuracy and generalization.

KEY POINTS

• Using coordinatized lesion location analysis, we mapped anatomic distribution patterns of gliomas with specific pathological and clinical features and constructed glioma prediction models. • We integrated coordinatized lesion location analysis into ROI-based analysis of radiomics to propose new fusion location-radiomics models. • Fusion location-radiomics models, with the advantages of being less influenced by variabilities, improved accuracy, and generalization performances of ROI-based radiomics models on predicting the diagnosis of gliomas.

A Lightweight Deep Learning Framework for Automatic MRI Data Sorting and Artifacts Detection

The purpose of this study is to develop a lightweight and easily deployable deep learning system for fully automated content-based brain MRI sorting and artifacts detection. 22092 MRI volumes from 4076 patients between 2017 and 2021 were involved in this retrospective study. The dataset mainly contains 4 common contrast (T1-weighted (T1w), contrast-enhanced T1-weighted (T1c), T2-weighted (T2w), fluid-attenuated inversion recovery (FLAIR)) in three perspectives (axial, coronal, and sagittal), and magnetic resonance angiography (MRA), as well as three typical artifacts (motion, aliasing, and metal artifacts). In the proposed architecture, a pre-trained EfficientNetB0 with the fully connected layers removed was used as the feature extractor and a multilayer perceptron (MLP) module with four hidden layers was used as the classifier. Precision, recall, F1_Score, accuracy, the number of trainable parameters, and float-point of operations (FLOPs) were calculated to evaluate the performance of the proposed model. The proposed model was also compared with four other existing CNN-based models in terms of classification performance and model size. The overall precision, recall, F1_Score, and accuracy of the proposed model were 0.983, 0.926, 0.950, and 0.991, respectively. The performance of the proposed model was outperformed the other four CNN-based models. The number of trainable parameters and FLOPs were the smallest among the investigated models. Our proposed model can accurately sort head MRI scans and identify artifacts with minimum computational resources and can be used as a tool to support big medical imaging data research and facilitate large-scale database management.

Development and Validation of Multi-Omics Thymoma Risk Classification Model Based on Transfer Learning

The paper aims to develop prediction model that integrates clinical, radiomics, and deep features using transfer learning to stratifying between high and low risk of thymoma. Our study enrolled 150 patients with thymoma (76 low-risk and 74 high-risk) who underwent surgical resection and pathologically confirmed in Shengjing Hospital of China Medical University from January 2018 to December 2020. The training cohort consisted of 120 patients (80%) and the test cohort consisted of 30 patients (20%). The 2590 radiomics and 192 deep features from non-enhanced, arterial, and venous phase CT images were extracted and ANOVA, Pearson correlation coefficient, PCA, and LASSO were used to select the most significant features. A fusion model that integrated clinical, radiomics, and deep features was developed with SVM classifiers to predict the risk level of thymoma, and accuracy, sensitivity, specificity, ROC curves, and AUC were applied to evaluate the classification model. In both the training and test cohorts, the fusion model demonstrated better performance in stratifying high and low risk of thymoma. It had AUCs of 0.99 and 0.95, and an accuracy of 0.93 and 0.83, respectively. This was compared to the clinical model (AUCs of 0.70 and 0.51, accuracy of 0.68 and 0.47), the radiomics model (AUCs of 0.97 and 0.82, accuracy of 0.93 and 0.80), and the deep model (AUCs of 0.94 and 0.85, accuracy of 0.88 and 0.80). The fusion model integrating clinical, radiomics and deep features based on transfer learning was efficient for noninvasively stratifying high risk and low risk of thymoma. The models could help to determine surgery strategy for thymoma cancer.

Improving Noninvasive Classification of Molecular Subtypes of Adult Gliomas With Diffusion-Weighted MR Imaging: An Externally Validated Machine Learning Algorithm

BACKGROUND

Genetic testing for molecular markers of gliomas sometimes is unavailable because of time-consuming and expensive, even limited tumor specimens or nonsurgery cases.

PURPOSE

To train a three-class radiomic model classifying three molecular subtypes including isocitrate dehydrogenase (IDH) mutations and 1p/19q-noncodeleted (IDHmut-noncodel), IDH wild-type (IDHwt), IDH-mutant and 1p/19q-codeleted (IDHmut-codel) of adult gliomas and investigate whether radiomic features from diffusion-weighted imaging (DWI) could bring additive value.

STUDY TYPE

Retrospective.

POPULATION

A total of 755 patients including 111 IDHmut-noncodel, 571 IDHwt, and 73 IDHmut-codel cases were divided into training (n = 480) and internal validation set (n = 275); 139 patients including 21 IDHmut-noncodel, 104 IDHwt, and 14 IDHmut-codel cases were utilized as external validation set.

FIELD STRENGTH/SEQUENCE

A 1.5 T or 3.0 T/multiparametric MRI, including T1-weighted (T1), T1-weighted gadolinium contrast-enhanced (T1c), T2-weighted (T2), fluid attenuated inversion recovery (FLAIR), and DWI.

ASSESSMENT

The performance of multiparametric radiomic model (random-forest model) using 22 selected features from T1, T2, FLAIR, T1c images and apparent diffusion coefficient (ADC) maps, and conventional radiomic model using 20 selected features from T1, T2, FLAIR, and T1c images was assessed in internal and external validation sets by comparing probability values and actual incidence.

STATISTICAL TESTS

Mann-Whitney U test, Chi-Squared test, Wilcoxon test, receiver operating curve (ROC), and area under the curve (AUC); DeLong analysis. P < 0.05 was statistically significant.

RESULTS

The multiparametric radiomic model achieved AUC values for IDHmut-noncodel, IDHwt, and IDHmut-codel of 0.8181, 0.8524, and 0.8502 in internal validation set and 0.7571, 0.7779, and 0.7491 in external validation set, respectively. Multiparametric radiomic model showed significantly better diagnostic performance after DeLong analysis, especially in classifying IDHwt and IDHmut-noncodel subtypes.

DATA CONCLUSION

Radiomic features from DWI could bring additive value and improve the performance of conventional MRI-based radiomic model for classifying the molecular subtypes especially IDHmut-noncodel and IDHwt of adult gliomas.

LEVEL OF EVIDENCE: 3

TECHNICAL EFFICACY

Stage 2.

Systematic Literature Review of Machine Learning Algorithms Using Pretherapy Radiologic Imaging for Glioma Molecular Subtype Prediction

BACKGROUND

The molecular profile of gliomas is a prognostic indicator for survival, driving clinical decision-making for treatment. Pathology-based molecular diagnosis is challenging because of the invasiveness of the procedure, exclusion from neoadjuvant therapy options, and the heterogeneous nature of the tumor.

PURPOSE

We performed a systematic review of algorithms that predict molecular subtypes of gliomas from MR Imaging.

DATA SOURCES

Data sources were Ovid Embase, Ovid MEDLINE, Cochrane Central Register of Controlled Trials, Web of Science.

STUDY SELECTION

Per the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines, 12,318 abstracts were screened and 1323 underwent full-text review, with 85 articles meeting the inclusion criteria.

DATA ANALYSIS

We compared prediction results from different machine learning approaches for predicting molecular subtypes of gliomas. Bias analysis was conducted for each study, following the Prediction model Risk Of Bias Assessment Tool (PROBAST) guidelines.

DATA SYNTHESIS

Isocitrate dehydrogenase mutation status was reported with an area under the curve and accuracy of 0.88 and 85% in internal validation and 0.86 and 87% in limited external validation data sets, respectively. For the prediction of O6-methylguanine-DNA methyltransferase promoter methylation, the area under the curve and accuracy in internal validation data sets were 0.79 and 77%, and in limited external validation, 0.89 and 83%, respectively. PROBAST scoring demonstrated high bias in all articles.

LIMITATIONS

The low number of external validation and studies with incomplete data resulted in unequal data analysis. Comparing the best prediction pipelines of each study may introduce bias.

CONCLUSIONS

While the high area under the curve and accuracy for the prediction of molecular subtypes of gliomas are reported in internal and external validation data sets, limited use of external validation and the increased risk of bias in all articles may present obstacles for clinical translation of these techniques.

Radiogenomics: a key component of precision cancer medicine

Radiogenomics, focusing on the relationship between genomics and imaging phenotypes, has been widely applied to address tumour heterogeneity and predict immune responsiveness and progression. It is an inevitable consequence of current trends in precision medicine, as radiogenomics costs less than traditional genetic sequencing and provides access to whole-tumour information rather than limited biopsy specimens. By providing voxel-by-voxel genetic information, radiogenomics can allow tailored therapy targeting a complete, heterogeneous tumour or set of tumours. In addition to quantifying lesion characteristics, radiogenomics can also be used to distinguish benign from malignant entities, as well as patient characteristics, to better stratify patients according to disease risk, thereby enabling more precise imaging and screening. Here, we have characterised the radiogenomic application in precision medicine using a multi-omic approach. we outline the main applications of radiogenomics in diagnosis, treatment planning and evaluations in the field of oncology with the aim of developing quantitative and personalised medicine. Finally, we discuss the challenges in the field of radiogenomics and the scope and clinical applicability of these methods.

Deep-learning and conventional radiomics to predict IDH genotyping status based on magnetic resonance imaging data in adult diffuse glioma

Objectives

In adult diffuse glioma, preoperative detection of isocitrate dehydrogenase (IDH) status helps clinicians develop surgical strategies and evaluate patient prognosis. Here, we aim to identify an optimal machine-learning model for prediction of IDH genotyping by combining deep-learning (DL) signatures and conventional radiomics (CR) features as model predictors.

Methods

In this study, a total of 486 patients with adult diffuse gliomas were retrospectively collected from our medical center (n=268) and the public database (TCGA, n=218). All included patients were randomly divided into the training and validation sets by using nested 10-fold cross-validation. A total of 6,736 CR features were extracted from four MRI modalities in each patient, namely T1WI, T1CE, T2WI, and FLAIR. The LASSO algorithm was performed for CR feature selection. In each MRI modality, we applied a CNN+LSTM-based neural network to extract DL features and integrate these features into a DL signature after the fully connected layer with sigmoid activation. Eight classic machine-learning models were analyzed and compared in terms of their prediction performance and stability in IDH genotyping by combining the LASSO-selected CR features and integrated DL signatures as model predictors. In the validation sets, the prediction performance was evaluated by using accuracy and the area under the curve (AUC) of the receiver operating characteristics, while the model stability was analyzed by using the relative standard deviation of the AUC (RSDAUC). Subgroup analyses of DL signatures and CR features were also individually conducted to explore their independent prediction values.

Results

Logistic regression (LR) achieved favorable prediction performance (AUC: 0.920 ± 0.043, accuracy: 0.843 ± 0.044), whereas support vector machine with the linear kernel (l-SVM) displayed low prediction performance (AUC: 0.812 ± 0.052, accuracy: 0.821 ± 0.050). With regard to stability, LR also showed high robustness against data perturbation (RSDAUC: 4.7%). Subgroup analyses showed that DL signatures outperformed CR features (DL, AUC: 0.915 ± 0.054, accuracy: 0.835 ± 0.061, RSDAUC: 5.9%; CR, AUC: 0.830 ± 0.066, accuracy: 0.771 ± 0.051, RSDAUC: 8.0%), while DL and DL+CR achieved similar prediction results.

Conclusion

In IDH genotyping, LR is a promising machine-learning classification model. Compared with CR features, DL signatures exhibit markedly superior prediction values and discriminative capability.

Accurate Differentiation of Spinal Tuberculosis and Spinal Metastases Using MR-Based Deep Learning Algorithms

Purpose

To explore the application of deep learning (DL) methods based on T2 sagittal MR images for discriminating between spinal tuberculosis (STB) and spinal metastases (SM).

Patients and Methods

A total of 121 patients with histologically confirmed STB and SM across four institutions were retrospectively analyzed. Data from two institutions were used for developing deep learning models and internal validation, while the remaining institutions' data were used for external testing. Utilizing MVITV2, EfficientNet-B3, ResNet101, and ResNet34 as backbone networks, we developed four distinct DL models and evaluated their diagnostic performance based on metrics such as accuracy (ACC), area under the receiver operating characteristic curve (AUC), F1 score, and confusion matrix. Furthermore, the external test images were blindly evaluated by two spine surgeons with different levels of experience. We also used Gradient-Class Activation Maps to visualize the high-dimensional features of different DL models.

Results

For the internal validation set, MVITV2 outperformed other models with an accuracy of 98.7%, F1 score of 98.6%, and AUC of 0.98. Other models followed in this order: EfficientNet-B3 (ACC: 96.1%, F1 score: 95.9%, AUC: 0.99), ResNet101 (ACC: 85.5%, F1 score: 84.8%, AUC: 0.90), and ResNet34 (ACC: 81.6%, F1 score: 80.7%, AUC: 0.85). For the external test set, MVITV2 again performed excellently with an accuracy of 91.9%, F1 score of 91.5%, and an AUC of 0.95. EfficientNet-B3 came second (ACC: 85.9, F1 score: 91.5%, AUC: 0.91), followed by ResNet101 (ACC:80.8, F1 score: 80.0%, AUC: 0.87) and ResNet34 (ACC: 78.8, F1 score: 77.9%, AUC: 0.86). Additionally, the diagnostic accuracy of the less experienced spine surgeon was 73.7%, while that of the more experienced surgeon was 88.9%.

Conclusion

Deep learning based on T2WI sagittal images can help discriminate between STB and SM, and can achieve a level of diagnostic performance comparable with that produced by experienced spine surgeons.

Identification of Origin for Spinal Metastases from MR Images: Comparison Between Radiomics and Deep Learning Methods

OBJECTIVE

To determine whether spinal metastatic lesions originated from lung cancer or from other cancers based on spinal contrast-enhanced T1 (CET1) magnetic resonance (MR) images analyzed using radiomics (RAD) and deep learning (DL) methods.

METHODS

We recruited and retrospectively reviewed 173 patients diagnosed with spinal metastases at two different centers between July 2018 and June 2021. Of these, 68 involved lung cancer and 105 were other types of cancer. They were assigned to an internal cohort of 149 patients, randomly divided into a training set and a validation set, and to an external cohort of 24 patients. All patients underwent CET1-MR imaging before surgery or biopsy. We developed two predictive algorithms: a DL model and a RAD model. We compared performance between models, and against human radiological assessment, via accuracy (ACC) and receiver operating characteristic (ROC) analyses. Furthermore, we analyzed the correlation between RAD and DL features.

RESULTS

The DL model outperformed RAD model across the board, with ACC/ area under the receiver operating characteristic curve (AUC) values of 0.93/0.94 (DL) versus 0.84/0.93 (RAD) when applied to the training set from the internal cohort, 0.74/0.76 versus 0.72/0.75 when applied to the validation set, and 0.72/0.76 versus 0.69/0.72 when applied to the external test cohort. For the validation set, it also outperformed expert radiological assessment (ACC: 0.65, AUC: 0.68). We only found weak correlations between DL and RAD features.

CONCLUSION

The DL algorithm successfully identified the origin of spinal metastases from pre-operative CET1-MR images, outperforming both RAD models and expert assessment by trained radiologists.

Performance of deep learning algorithms to distinguish high-grade glioma from low-grade glioma: A systematic review and meta-analysis

This study aims to evaluate deep learning (DL) performance in differentiating low- and high-grade glioma. Search online database for studies continuously published from 1st January 2015 until 16th August 2022. The random-effects model was used for synthesis, based on pooled sensitivity (SE), specificity (SP), and area under the curve (AUC). Heterogeneity was estimated using the Higgins inconsistency index (I²). 33 were ultimately included in the meta-analysis. The overall pooled SE and SP were 94% and 93%, with an AUC of 0.98. There was great heterogeneity in this field. Our evidence-based study shows DL achieves high accuracy in glioma grading. Subgroup analysis reveals several limitations in this field: 1) Diagnostic trials require standard method for data merging for AI; 2) small sample size; 3) poor-quality image preprocessing; 4) not standard algorithm development; 5) not standard data report; 6) different definition of HGG and LGG; and 7) poor extrapolation.

Differential diagnosis of benign and malignant vertebral compression fractures: Comparison and correlation of radiomics and deep learning frameworks based on spinal CT and clinical characteristics

PURPOSE

Differentiating benign from malignant vertebral compression fractures (VCFs) is a diagnostic dilemma in clinical practice. To improve the accuracy and efficiency of diagnosis, we evaluated the performance of deep learning and radiomics methods based on computed tomography (CT) and clinical characteristics in differentiating between Osteoporosis VCFs (OVCFs) and malignant VCFs (MVCFs).

METHODS

We enrolled a total of 280 patients (155 with OVCFs and 125 with MVCFs) and randomly divided them into a training set (80%, n = 224) and a validation set (20%, n = 56). We developed three predictive models: a deep learning (DL) model, a radiomics (Rad) model, and a combined DL_Rad model, using CT and clinical characteristics data. The Inception_V3 served as the backbone of the DL model. The input data for the DL_Rad model consisted of the combined features of Rad and DCNN features. We calculated the receiver operating characteristic curve, area under the curve (AUC), and accuracy (ACC) to assess the performance of the models. Additionally, we calculated the correlation between Rad features and DCNN features.

RESULTS

For the training set, the DL_Rad model achieved the best results, with an AUC of 0.99 and ACC of 0.99, followed by the Rad model (AUC: 0.99, ACC: 0.97) and DL model (AUC: 0.99, ACC: 0.94). For the validation set, the DL_Rad model (with an AUC of 0.97 and ACC of 0.93) outperformed the Rad model (with an AUC: 0.93 and ACC: 0.91) and the DL model (with an AUC: 0.89 and ACC: 0.88). Rad features achieved better classifier performance than the DCNN features, and their general correlations were weak.

CONCLUSIONS

The Deep learnig model, Radiomics model, and Deep learning Radiomics model achieved promising results in discriminating MVCFs from OVCFs, and the DL_Rad model performed the best.

Emerging role of artificial intelligence in diagnosis, classification and clinical management of glioma

Glioma represents a dominant primary intracranial malignancy in the central nervous system. Artificial intelligence that mainly includes machine learning, and deep learning computational approaches, presents a unique opportunity to enhance clinical management of glioma through improving tumor segmentation, diagnosis, differentiation, grading, treatment, prediction of clinical outcomes (prognosis, and recurrence), molecular features, clinical classification, characterization of the tumor microenvironment, and drug discovery. A growing body of recent studies apply artificial intelligence-based models to disparate data sources of glioma, covering imaging modalities, digital pathology, high-throughput multi-omics data (especially emerging single-cell RNA sequencing and spatial transcriptome), etc. While these early findings are promising, future studies are required to normalize artificial intelligence-based models to improve the generalizability and interpretability of the results. Despite prominent issues, targeted clinical application of artificial intelligence approaches in glioma will facilitate the development of precision medicine of this field. If these challenges can be overcome, artificial intelligence has the potential to profoundly change the way patients with or at risk of glioma are provided with more rational care.

Multiparametric magnetic resonance imaging-derived deep learning network to determine ferroptosis-related gene signatures in gliomas

Introduction

Ferroptosis-related gene (FRG) signature is important for assessing novel therapeutic approaches and prognosis in glioma. We trained a deep learning network for determining FRG signatures using multiparametric magnetic resonance imaging (MRI).

Methods

FRGs of patients with glioma were acquired from public databases. FRG-related risk score stratifying prognosis was developed from The Cancer Genome Atlas (TCGA) and validated using the Chinese Glioma Genome Atlas. Multiparametric MRI-derived glioma images and the corresponding genomic information were obtained for 122 cases from TCGA and The Cancer Imaging Archive. The deep learning network was trained using 3D-Resnet, and threefold cross-validation was performed to evaluate the predictive performance.

Results

The FRG-related risk score was associated with poor clinicopathological features and had a high predictive value for glioma prognosis. Based on the FRG-related risk score, patients with glioma were successfully classified into two subgroups (28 and 94 in the high- and low-risk groups, respectively). The deep learning networks TC (enhancing tumor and non-enhancing portion of the tumor core) mask achieved an average cross-validation accuracy of 0.842 and an average AUC of 0.781, while the deep learning networks WT (whole tumor and peritumoral edema) mask achieved an average cross-validation accuracy of 0.825 and an average AUC of 0.781.

Discussion

Our findings indicate that FRG signature is a prognostic indicator of glioma. In addition, we developed a deep learning network that has high classification accuracy in automatically determining FRG signatures, which may be an important step toward the clinical translation of novel therapeutic approaches and prognosis of glioma.

Noninvasive Classification of Glioma Subtypes Using Multiparametric MRI to Improve Deep Learning

Background: Deep learning (DL) methods can noninvasively predict glioma subtypes; however, there is no set paradigm for the selection of network structures and input data, including the image combination method, image processing strategy, type of numeric data, and others. Purpose: To compare different combinations of DL frameworks (ResNet, ConvNext, and vision transformer (VIT)), image preprocessing strategies, magnetic resonance imaging (MRI) sequences, and numerical data for increasing the accuracy of DL models for differentiating glioma subtypes prior to surgery. Methods: Our dataset consisted of 211 patients with newly diagnosed gliomas who underwent preoperative MRI with standard and diffusion-weighted imaging methods. Different data combinations were used as input for the three different DL classifiers. Results: The accuracy of the image preprocessing strategies, including skull stripping, segment addition, and individual treatment of slices, was 5%, 10%, and 12.5% higher, respectively, than that of the other strategies. The accuracy increased by 7.5% and 10% following the addition of ADC and numeric data, respectively. ResNet34 exhibited the best performance, which was 5% and 17.5% higher than that of ConvNext tiny and VIT-base, respectively. Data Conclusions: The findings demonstrated that the addition of quantitatively numeric data, ADC images, and effective image preprocessing strategies improved model accuracy for datasets of similar size. The performance of ResNet was superior for small or medium datasets.

Deep learning model for diagnosing early gastric cancer using preoperative computed tomography images

BACKGROUND

Early gastric cancer (EGC) is defined as a lesion restricted to the mucosa or submucosa, independent of size or evidence of regional lymph node metastases. Although computed tomography (CT) is the main technique for determining the stage of gastric cancer (GC), the accuracy of CT for determining tumor invasion of EGC was still unsatisfactory by radiologists. In this research, we attempted to construct an AI model to discriminate EGC in portal venous phase CT images.

METHODS

We retrospectively collected 658 GC patients from the first affiliated hospital of Nanchang university, and divided them into training and internal validation cohorts with a ratio of 8:2. As the external validation cohort, 93 GC patients were recruited from the second affiliated hospital of Soochow university. We developed several prediction models based on various convolutional neural networks, and compared their predictive performance.

RESULTS

The deep learning model based on the ResNet101 neural network represented sufficient discrimination of EGC. In two validation cohorts, the areas under the curves (AUCs) for the receiver operating characteristic (ROC) curves were 0.993 (95% CI: 0.984-1.000) and 0.968 (95% CI: 0.935-1.000), respectively, and the accuracy was 0.946 and 0.914. Additionally, the deep learning model can also differentiate between mucosa and submucosa tumors of EGC.

CONCLUSIONS

These results suggested that deep learning classifiers have the potential to be used as a screening tool for EGC, which is crucial in the individualized treatment of EGC patients.

Artificial intelligence in multi-parametric magnetic resonance imaging: A review

Multi-parametric magnetic resonance imaging (mpMRI) is an indispensable tool in the clinical workflow for the diagnosis and treatment planning of various diseases. Machine learning-based artificial intelligence (AI) methods, especially those adopting the deep learning technique, have been extensively employed to perform mpMRI image classification, segmentation, registration, detection, reconstruction, and super-resolution. The current availability of increasing computational power and fast-improving AI algorithms have empowered numerous computer-based systems for applying mpMRI to disease diagnosis, imaging-guided radiotherapy, patient risk and overall survival time prediction, and the development of advanced quantitative imaging technology for magnetic resonance fingerprinting. However, the wide application of these developed systems in the clinic is still limited by a number of factors, including robustness, reliability, and interpretability. This survey aims to provide an overview for new researchers in the field as well as radiologists with the hope that they can understand the general concepts, main application scenarios, and remaining challenges of AI in mpMRI. This article is protected by copyright. All rights reserved.

A Novel Deep Learning Radiomics Model to Discriminate AD, MCI and NC: An Exploratory Study Based on Tau PET Scans from ADNI

OBJECTIVE

We explored a novel model based on deep learning radiomics (DLR) to differentiate Alzheimer's disease (AD) patients, mild cognitive impairment (MCI) patients and normal control (NC) subjects. This model was validated in an exploratory study using tau positron emission tomography (tau-PET) scans.

METHODS

In this study, we selected tau-PET scans from the Alzheimer's Disease Neuroimaging Initiative database (ADNI), which included a total of 211 NC, 197 MCI, and 117 AD subjects. The dataset was divided into one training/validation group and one separate external group for testing. The proposed DLR model contained the following three steps: (1) pre-training of candidate deep learning models; (2) extraction and selection of DLR features; (3) classification based on support vector machine (SVM). In the comparative experiments, we compared the DLR model with three traditional models, including the SUVR model, traditional radiomics model, and a clinical model. Ten-fold cross-validation was carried out 200 times in the experiments.

RESULTS

Compared with other models, the DLR model achieved the best classification performance, with an accuracy of 90.76% ± 2.15% in NC vs. MCI, 88.43% ± 2.32% in MCI vs. AD, and 99.92% ± 0.51% in NC vs. AD.

CONCLUSIONS

Our proposed DLR model had the potential clinical value to discriminate AD, MCI and NC.

Predicting MGMT Promoter Methylation in Diffuse Gliomas Using Deep Learning with Radiomics

This study aimed to investigate the feasibility of predicting oxygen 6-methylguanine-DNA methyltransferase (MGMT) promoter methylation in diffuse gliomas by developing a deep learning approach using MRI radiomics. A total of 111 patients with diffuse gliomas participated in the retrospective study (56 patients with MGMT promoter methylation and 55 patients with MGMT promoter unmethylation). The radiomics features of the two regions of interest (ROI) (the whole tumor area and the tumor core area) for four sequences, including T1 weighted image (T1WI), T2 weighted image (T2WI), apparent diffusion coefficient (ADC) maps, and T1 contrast-enhanced (T1CE) MR images were extracted and jointly fed into the residual network. Then the deep learning method was developed and evaluated with a five-fold cross-validation, where in each fold, the dataset was randomly divided into training (80%) and validation (20%) cohorts. We compared the performance of all models using area under the curve (AUC) and average accuracy of validation cohorts and calculated the 10 most important features of the best model via a class activation map. Based on the ROI of the whole tumor, the predictive capacity of the T1CE and ADC model achieved the highest AUC value of 0.85. Based on the ROI of the tumor core, the T1CE and ADC model achieved the highest AUC value of 0.90. After comparison, the T1CE combined with the ADC model based on the ROI of the tumor core exhibited the best performance, with the highest average accuracy (0.91) and AUC (0.90) among all models. The deep learning method using MRI radiomics has excellent diagnostic performance with a high accuracy in predicting MGMT promoter methylation in diffuse gliomas.

Using Deep Learning Radiomics to Distinguish Cognitively Normal Adults at Risk of Alzheimer's Disease From Normal Control: An Exploratory Study Based on Structural MRI

Objectives

We proposed a novel deep learning radiomics (DLR) method to distinguish cognitively normal adults at risk of Alzheimer's disease (AD) from normal control based on T1-weighted structural MRI images.

Methods

In this study, we selected MRI data from the Alzheimer's Disease Neuroimaging Initiative Database (ADNI), which included 417 cognitively normal adults. These subjects were divided into 181 individuals at risk of Alzheimer's disease (preAD group) and 236 normal control individuals (NC group) according to standard uptake ratio >1.18 calculated by amyloid Positron Emission Tomography (PET). We further divided the preaAD group into APOE+ and APOE- subgroups according to whether APOE ε4 was positive or not. All data sets were divided into one training/validation group and one independent test group. The proposed DLR method included three steps: (1) the pre-training of basic deep learning (DL) models, (2) the extraction, selection and fusion of DLR features, and (3) classification. The support vector machine (SVM) was used as the classifier. In the comparative experiments, we compared our proposed DLR method with three existing models: hippocampal model, clinical model, and traditional radiomics model. Ten-fold cross-validation was performed with 100 time repetitions.

Results

The DLR method achieved the best classification performance between preAD and NC than other models with an accuracy of 89.85% ± 1.12%. In comparison, the accuracies of the other three models were 72.44% ± 1.37%, 82.00% ± 4.09% and 79.65% ± 2.21%. In addition, the DLR model also showed the best classification performance (85.45% ± 9.04% and 92.80% ± 2.61%) in the subgroup experiment.

Conclusion

The results showed that the DLR method provided a potentially clinical value to distinguish preAD from NC.

AI-Driven Image Analysis in Central Nervous System Tumors-Traditional Machine Learning, Deep Learning and Hybrid Models

The interpretation of imaging in medicine in general and in oncology specifically remains problematic due to several limitations which include the need to incorporate detailed clinical history, patient and disease-specific history, clinical exam features, previous and ongoing treatment, and account for the dependency on reproducible human interpretation of multiple factors with incomplete data linkage. To standardize reporting, minimize bias, expedite management, and improve outcomes, the use of Artificial Intelligence (AI) has gained significant prominence in imaging analysis. In oncology, AI methods have as a result been explored in most cancer types with ongoing progress in employing AI towards imaging for oncology treatment, assessing treatment response, and understanding and communicating prognosis. Challenges remain with limited available data sets, variability in imaging changes over time augmented by a growing heterogeneity in analysis approaches. We review the imaging analysis workflow and examine how hand-crafted features also referred to as traditional Machine Learning (ML), Deep Learning (DL) approaches, and hybrid analyses, are being employed in AI-driven imaging analysis in central nervous system tumors. ML, DL, and hybrid approaches coexist, and their combination may produce superior results although data in this space is as yet novel, and conclusions and pitfalls have yet to be fully explored. We note the growing technical complexities that may become increasingly separated from the clinic and enforce the acute need for clinician engagement to guide progress and ensure that conclusions derived from AI-driven imaging analysis reflect that same level of scrutiny lent to other avenues of clinical research.

Deep Learning Supplants Visual Analysis by Experienced Operators for the Diagnosis of Cardiac Amyloidosis by Cine-CMR

Background: Diagnosing cardiac amyloidosis (CA) from cine-CMR (cardiac magnetic resonance) alone is not reliable. In this study, we tested if a convolutional neural network (CNN) could outperform the visual diagnosis of experienced operators. Method: 119 patients with cardiac amyloidosis and 122 patients with left ventricular hypertrophy (LVH) of other origins were retrospectively selected. Diastolic and systolic cine-CMR images were preprocessed and labeled. A dual-input visual geometry group (VGG ) model was used for binary image classification. All images belonging to the same patient were distributed in the same set. Accuracy and area under the curve (AUC) were calculated per frame and per patient from a 40% held-out test set. Results were compared to a visual analysis assessed by three experienced operators. Results: frame-based comparisons between humans and a CNN provided an accuracy of 0.605 vs. 0.746 (p < 0.0008) and an AUC of 0.630 vs. 0.824 (p < 0.0001). Patient-based comparisons provided an accuracy of 0.660 vs. 0.825 (p < 0.008) and an AUC of 0.727 vs. 0.895 (p < 0.002). Conclusion: based on cine-CMR images alone, a CNN is able to discriminate cardiac amyloidosis from LVH of other origins better than experienced human operators (15 to 20 points more in absolute value for accuracy and AUC), demonstrating a unique capability to identify what the eyes cannot see through classical radiological analysis.