Deep Transfer Learning and Radiomics Feature Prediction of Survival of Patients with High-Grade Gliomas

A Stacked Multimodality Model Based on Functional MRI Features and Deep Learning Radiomics for Predicting the Early Response to Radiotherapy in Nasopharyngeal Carcinoma

BACKGROUND

This study aimed to construct and assess a comprehensive model that integrates MRI-derived deep learning radiomics, functional imaging (fMRI), and clinical indicators to predict early efficacy of radiotherapy in nasopharyngeal carcinoma (NPC).

METHODS

This retrospective study recruited NPC patients with radiotherapy from two Chinese hospitals between October 2018 and July 2022, divided into a training set (hospital I, 194 cases), an internal validation set (hospital I, 82 cases), and an external validation set (hospital II, 40 cases). We extracted 3404 radiomic features and 2048 deep learning features from multi-sequence MRI includes T1WI, CE-T1WI, T2WI and T2WI/FS. Additionally, both the Apparent diffusion coefficient (ADC), its maximum (ADCmax) and Tumor blood flow (TBF), its maximum (TBFmax) were obtained by Diffusion-weighted imaging (DWI) and Arterial spin labeling (ASL) respectively. We used four classifiers (LR, XGBoost, SVM and KNN) and stacked algorithm as model construction methods. The area under the receiver operating characteristic curve (AUC) and decision curve analysis was used to assess models.

RESULTS

The manual radiomics model based on XGBoost and the deep learning model based on KNN (the AUCs in the training set: 0.909, 0.823, respectively) showed better predictive efficacy than other machine learning algorithms. The stacked model that integrated MRI-based deep learning radiomics, fMRI, and hematological indicators, has the strongest efficacy prediction ability of AUC in the training set [0.984 (95%CI: 0.972-0.996)], the internal validation set [0.936 (95%CI: 0.885-0.987)], and the external validation set [0.959 (95%CI: 0.901-1.000)].

CONCLUSION

Our research has developed a clinical-radiomics integrated model based on MRI which can predict early radiotherapy response in NPC and provide guidance for personalized treatment.

Leveraging segmentation-guided spatial feature embedding for overall survival prediction in glioblastoma with multimodal magnetic resonance imaging

BACKGROUND AND OBJECTIVE

Patients with glioblastoma have a five-year relative survival rate of less than 5 %. Thus, accurately predicting the overall survival (OS) of patients with glioblastoma is crucial for effective treatment planning.

METHODS

To fully leverage the imaging characteristics of glioblastomas, we propose a segmentation-guided regression method for predicting OS of patients with brain tumors using multimodal magnetic resonance imaging. Specifically, a brain tumor segmentation network was first pre-trained without leveraging survival information. Subsequently, the survival regression network was jointly trained with the guidance of brain tumor segmentation, focusing on tumor voxels and suppressing irrelevant backgrounds.

RESULTS

Our proposed framework, based on the well-known backbone of UNETR++, achieved a Dice score of 0.7910, Spearman correlation of 0.4112, and Harrell's concordance index of 0.6488. The model consistently showed promising results compared with baseline methods on two different datasets (BraTS and UCSF-PDGM). Furthermore, ablation studies on our training configurations demonstrated that both the pre-training segmentation network and contrastive loss significantly improved all metrics for OS prediction.

CONCLUSIONS

In this study, we propose a joint learning framework based on a pre-trained segmentation backbone for OS prediction by leveraging a brain tumor segmentation map. By utilizing a spatial feature map, our model can operate using a sliding-window approach, which can be adopted by varying the matrix sizes and resolutions of the input images.

The application value of support vector machine model based on multimodal MRI in predicting IDH-1mutation and Ki-67 expression in glioma

PURPOSE

To investigate the application value of support vector machine (SVM) model based on diffusion-weighted imaging (DWI), dynamic contrast-enhanced (DCE) and amide proton transfer- weighted (APTW) imaging in predicting isocitrate dehydrogenase 1(IDH-1) mutation and Ki-67 expression in glioma.

METHODS

The DWI, DCE and APTW images of 309 patients with glioma confirmed by pathology were retrospectively analyzed and divided into the IDH-1 group (IDH-1(+) group and IDH-1(-) group) and Ki-67 group (low expression group (Ki-67 ≤ 10%) and high expression group (Ki-67 > 10%)). All cases were divided into the training set, and validation set according to the ratio of 7:3. The training set was used to select features and establish machine learning models. The SVM model was established with the data after feature selection. Four single sequence models and one combined model were established in IDH-1 group and Ki-67 group. The receiver operator characteristic (ROC) curve was used to evaluate the diagnostic performance of the model. Validation set data was used for further validation.

RESULTS

Both in the IDH-1 group and Ki-67 group, the combined model had better predictive efficiency than single sequence model, although the single sequence model had a better predictive efficiency. In the Ki-67 group, the combined model was built from six selected radiomics features, and the AUC were 0.965 and 0.931 in the training and validation sets, respectively. In the IDH-1 group, the combined model was built from four selected radiomics features, and the AUC were 0.997 and 0.967 in the training and validation sets, respectively.

CONCLUSION

The radiomics model established by DWI, DCE and APTW images could be used to detect IDH-1 mutation and Ki-67 expression in glioma patients before surgery. The prediction performance of the radiomics model based on the combination sequence was better than that of the single sequence model.

BTK Expression Level Prediction and the High-Grade Glioma Prognosis Using Radiomic Machine Learning Models

We aimed to study whether the Bruton's tyrosine kinase (BTK) expression is correlated with the prognosis of patients with high-grade gliomas (HGGs) and predict its expression level prior to surgery, by constructing radiomic models. Clinical and gene expression data of 310 patients from The Cancer Genome Atlas (TCGA) were included for gene-based prognostic analysis. Among them, contrast-enhanced T1-weighted imaging (T1WI + C) from The Cancer Imaging Archive (TCIA) with genomic data was selected from 82 patients for radiomic models, including support vector machine (SVM) and logistic regression (LR) models. Furthermore, the nomogram incorporating radiomic signatures was constructed to evaluate its clinical efficacy. BTK was identified as an independent risk factor for HGGs through univariate and multivariate Cox regression analyses. Three radiomic features were selected to construct the SVM and LR models, and the validation set showed area under curve (AUCs) values of 0.711 (95% CI, 0.598-0.824) and 0.736 (95% CI, 0.627-0.844), respectively. The median survival times of the high Rad_score and low-Rad_score groups based on LR model were 15.53 and 23.03 months, respectively. In addition, the total risk score of each patient was used to construct a predictive nomogram, and the AUCs calculated from the corresponding time-dependent ROC curves were 0.533, 0.659, and 0.767 for 1, 3, and 5 years, respectively. BTK is an independent risk factor associated with poor prognosis in patients, and the radiomic model constructed in this study can effectively and non-invasively predict preoperative BTK expression levels and patient prognosis based on T1WI + C.

Serum CD133-Associated Proteins Identified by Machine Learning Are Connected to Neural Development, Cancer Pathways, and 12-Month Survival in Glioblastoma

Glioma is the most prevalent type of primary central nervous system cancer, while glioblastoma (GBM) is its most aggressive variant, with a median survival of only 15 months when treated with maximal surgical resection followed by chemoradiation therapy (CRT). CD133 is a potentially significant GBM biomarker. However, current clinical biomarker studies rely on invasive tissue samples. These make prolonged data acquisition impossible, resulting in increased interest in the use of liquid biopsies. Our study, analyzed 7289 serum proteins from 109 patients with pathology-proven GBM obtained prior to CRT using the aptamer-based SOMAScan® proteomic assay technology. We developed a novel methodology that identified 24 proteins linked to both serum CD133 and 12-month overall survival (OS) through a multi-step machine learning (ML) analysis. These identified proteins were subsequently subjected to survival and clustering evaluations, categorizing patients into five risk groups that accurately predicted 12-month OS based on their protein profiles. Most of these proteins are involved in brain function, neural development, and/or cancer biology signaling, highlighting their significance and potential predictive value. Identifying these proteins provides a valuable foundation for future serum investigations as validation of clinically applicable GBM biomarkers can unlock immense potential for diagnostics and treatment monitoring.

Patient-Specific, Mechanistic Models of Tumor Growth Incorporating Artificial Intelligence and Big Data

Despite the remarkable advances in cancer diagnosis, treatment, and management over the past decade, malignant tumors remain a major public health problem. Further progress in combating cancer may be enabled by personalizing the delivery of therapies according to the predicted response for each individual patient. The design of personalized therapies requires the integration of patient-specific information with an appropriate mathematical model of tumor response. A fundamental barrier to realizing this paradigm is the current lack of a rigorous yet practical mathematical theory of tumor initiation, development, invasion, and response to therapy. We begin this review with an overview of different approaches to modeling tumor growth and treatment, including mechanistic as well as data-driven models based on big data and artificial intelligence. We then present illustrative examples of mathematical models manifesting their utility and discuss the limitations of stand-alone mechanistic and data-driven models. We then discuss the potential of mechanistic models for not only predicting but also optimizing response to therapy on a patient-specific basis. We describe current efforts and future possibilities to integrate mechanistic and data-driven models. We conclude by proposing five fundamental challenges that must be addressed to fully realize personalized care for cancer patients driven by computational models.

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.

Ultrasound-based deep learning radiomics model for differentiating benign, borderline, and malignant ovarian tumours: a multi-class classification exploratory study

BACKGROUND

Accurate preoperative identification of ovarian tumour subtypes is imperative for patients as it enables physicians to custom-tailor precise and individualized management strategies. So, we have developed an ultrasound (US)-based multiclass prediction algorithm for differentiating between benign, borderline, and malignant ovarian tumours.

METHODS

We randomised data from 849 patients with ovarian tumours into training and testing sets in a ratio of 8:2. The regions of interest on the US images were segmented and handcrafted radiomics features were extracted and screened. We applied the one-versus-rest method in multiclass classification. We inputted the best features into machine learning (ML) models and constructed a radiomic signature (Rad_Sig). US images of the maximum trimmed ovarian tumour sections were inputted into a pre-trained convolutional neural network (CNN) model. After internal enhancement and complex algorithms, each sample's predicted probability, known as the deep transfer learning signature (DTL_Sig), was generated. Clinical baseline data were analysed. Statistically significant clinical parameters and US semantic features in the training set were used to construct clinical signatures (Clinic_Sig). The prediction results of Rad_Sig, DTL_Sig, and Clinic_Sig for each sample were fused as new feature sets, to build the combined model, namely, the deep learning radiomic signature (DLR_Sig). We used the receiver operating characteristic (ROC) curve and the area under the ROC curve (AUC) to estimate the performance of the multiclass classification model.

RESULTS

The training set included 440 benign, 44 borderline, and 196 malignant ovarian tumours. The testing set included 109 benign, 11 borderline, and 49 malignant ovarian tumours. DLR_Sig three-class prediction model had the best overall and class-specific classification performance, with micro- and macro-average AUC of 0.90 and 0.84, respectively, on the testing set. Categories of identification AUC were 0.84, 0.85, and 0.83 for benign, borderline, and malignant ovarian tumours, respectively. In the confusion matrix, the classifier models of Clinic_Sig and Rad_Sig could not recognise borderline ovarian tumours. However, the proportions of borderline and malignant ovarian tumours identified by DLR_Sig were the highest at 54.55% and 63.27%, respectively.

CONCLUSIONS

The three-class prediction model of US-based DLR_Sig can discriminate between benign, borderline, and malignant ovarian tumours. Therefore, it may guide clinicians in determining the differential management of patients with ovarian tumours.

Artificial intelligence-based MRI radiomics and radiogenomics in glioma

The specific genetic subtypes that gliomas exhibit result in variable clinical courses and the need to involve multidisciplinary teams of neurologists, epileptologists, neurooncologists and neurosurgeons. Currently, the diagnosis of gliomas pivots mainly around the preliminary radiological findings and the subsequent definitive surgical diagnosis (via surgical sampling). Radiomics and radiogenomics present a potential to precisely diagnose and predict survival and treatment responses, via morphological, textural, and functional features derived from MRI data, as well as genomic data. In spite of their advantages, it is still lacking standardized processes of feature extraction and analysis methodology among different research groups, which have made external validations infeasible. Radiomics and radiogenomics can be used to better understand the genomic basis of gliomas, such as tumor spatial heterogeneity, treatment response, molecular classifications and tumor microenvironment immune infiltration. These novel techniques have also been used to predict histological features, grade or even overall survival in gliomas. In this review, workflows of radiomics and radiogenomics are elucidated, with recent research on machine learning or artificial intelligence in glioma.

Peritumoral Radiomics for Identification of Telomerase Reverse Transcriptase Promoter Mutation in Patients With Glioblastoma Based on Preoperative MRI

Purpose: To evaluate the value of intra- and peritumoral deep learning (DL) features based on multi-parametric magnetic resonance imaging (MRI) for identifying telomerase reverse transcriptase (TERT) promoter mutation in glioblastoma (GBM). Methods: In this study, we included 229 patients with GBM who underwent preoperative MRI in two hospitals between November 2016 and September 2022. We used four 2D Convolutional Neural Networks (GoogLeNet, DenseNet121, VGG16, and MobileNetV3-Large) to extract intra- and peritumoral DL features. The Mann-Whitney U test, Pearson correlation analysis, least absolute shrinkage and selection operator, and logistic regression analysis were used for feature selection and construction of DL radiomics (DLR) signatures in different regions. These multi-parametric and multi-region signatures were combined to identify TERT promoter mutation. The area under the receiver operating characteristic curve (AUC) was used to evaluate the effects of the signatures. Results: The signatures based on the DL features from the peritumoral regions with expansion distances of 2 mm, 8 mm, and 10 mm using the GoogLeNet architecture correlated with the optimal AUC values (test set: .823, .753, and .768) in the T2-weighted, T1-weighted contrast-enhanced, and T1-weighted images. Using the stacking fusion method, DLR with multi-parameter and multi-region fusion achieved the best discrimination with AUC values of .948 and .902 in the training and test sets, respectively. Conclusions: The radiomics model based on the fusion of multi-parameter MRI intra- and peritumoral DLR signatures may help to identify TERT promoter mutation in patients with GBM.

Convolutional Neural Networks for Glioma Segmentation and Prognosis: A Systematic Review

Deep learning (DL) is poised to redefine the way medical images are processed and analyzed. Convolutional neural networks (CNNs), a specific type of DL architecture, are exceptional for high-throughput processing, allowing for the effective extraction of relevant diagnostic patterns from large volumes of complex visual data. This technology has garnered substantial interest in the field of neuro-oncology as a promising tool to enhance medical imaging throughput and analysis. A multitude of methods harnessing MRI-based CNNs have been proposed for brain tumor segmentation, classification, and prognosis prediction. They are often applied to gliomas, the most common primary brain cancer, to classify subtypes with the goal of guiding therapy decisions. Additionally, the difficulty of repeating brain biopsies to evaluate treatment response in the setting of often confusing imaging findings provides a unique niche for CNNs to help distinguish the treatment response to gliomas. For example, glioblastoma, the most aggressive type of brain cancer, can grow due to poor treatment response, can appear to grow acutely due to treatment-related inflammation as the tumor dies (pseudo-progression), or falsely appear to be regrowing after treatment as a result of brain damage from radiation (radiation necrosis). CNNs are being applied to separate this diagnostic dilemma. This review provides a detailed synthesis of recent DL methods and applications for intratumor segmentation, glioma classification, and prognosis prediction. Furthermore, this review discusses the future direction of MRI-based CNN in the field of neuro-oncology and challenges in model interpretability, data availability, and computation efficiency.

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.

Value Proposition of FDA-Approved Artificial Intelligence Algorithms for Neuroimaging

PURPOSE

The number of FDA-cleared artificial intelligence (AI) algorithms for neuroimaging has grown in the past decade. The adoption of these algorithms into clinical practice depends largely on whether this technology provides value in the clinical setting. The objective of this study was to analyze trends in FDA-cleared AI algorithms for neuroimaging and understand their value proposition as advertised by the AI developers and vendors.

METHODS

A list of AI algorithms cleared by the FDA for neuroimaging between May 2008 and August 2022 was extracted from the ACR Data Science Institute AI Central database. Product information for each device was collected from the database. For each device, information on the advertised value as presented on the developer's website was collected.

RESULTS

A total of 59 AI neuroimaging algorithms were cleared by the FDA between May 2008 and August 2022. Most of these algorithms (24 of 59) were compatible with noncontrast CT, 21 with MRI, 9 with CT perfusion, 8 with CT angiography, 3 with MR perfusion, and 2 with PET. Six algorithms were compatible with multiple imaging techniques. Of the 59 algorithms, websites were located that discussed the product value for 55 algorithms. The most widely advertised value proposition was improved quality of care (38 of 55 [69.1%]). A total of 24 algorithms (43.6%) proposed saving user time, 9 (15.7%) advertised decreased costs, and 6 (10.9%) described increased revenue. Product websites for 26 algorithms (43.6%) showed user testimonials advertising the value of the technology.

CONCLUSIONS

The results of this study indicate a wide range of value propositions advertised by developers and vendors of AI algorithms for neuroimaging. Most vendors advertised that their products would improve patient care. Further research is necessary to determine whether the value claimed by developers is actually demonstrated in clinical practice.

Deep Learning Radiomics for the Assessment of Telomerase Reverse Transcriptase Promoter Mutation Status in Patients With Glioblastoma Using Multiparametric MRI

BACKGROUND

Studies have shown that magnetic resonance imaging (MRI)-based deep learning radiomics (DLR) has the potential to assess glioma grade; however, its role in predicting telomerase reverse transcriptase (TERT) promoter mutation status in patients with glioblastoma (GBM) remains unclear.

PURPOSE

To evaluate the value of deep learning (DL) in multiparametric MRI-based radiomics in identifying TERT promoter mutations in patients with GBM preoperatively.

STUDY TYPE

Retrospective.

POPULATION

A total of 274 patients with isocitrate dehydrogenase-wildtype GBM were included in the study. The training and external validation cohorts included 156 (54.3 ± 12.7 years; 96 males) and 118 (54 .2 ± 13.4 years; 73 males) patients, respectively.

FIELD STRENGTH/SEQUENCE

Axial contrast-enhanced T1-weighted spin-echo inversion recovery sequence (T1CE), T1-weighted spin-echo inversion recovery sequence (T1WI), and T2-weighted spin-echo inversion recovery sequence (T2WI) on 1.5-T and 3.0-T scanners were used in this study.

ASSESSMENT

Overall tumor area regions (the tumor core and edema) were segmented, and the radiomics and DL features were extracted from preprocessed multiparameter preoperative brain MRI images-T1WI, T1CE, and T2WI. A model based on the DLR signature, clinical signature, and clinical DLR (CDLR) nomogram was developed and validated to identify TERT promoter mutation status.

STATISTICAL TESTS

The Mann-Whitney U test, Pearson test, least absolute shrinkage and selection operator, and logistic regression analysis were applied for feature selection and construction of radiomics and DL signatures. Results were considered statistically significant at P-value <0.05.

RESULTS

The DLR signature showed the best discriminative power for predicting TERT promoter mutations, yielding an AUC of 0.990 and 0.890 in the training and external validation cohorts, respectively. Furthermore, the DLR signature outperformed CDLR nomogram (P = 0.670) and significantly outperformed clinical models in the validation cohort.

DATA CONCLUSION

The multiparameter MRI-based DLR signature exhibited a promising performance for the assessment of TERT promoter mutations in patients with GBM, which could provide information for individualized treatment.

LEVEL OF EVIDENCE

3 TECHNICAL EFFICACY: Stage 2.

Pretreatment MR-based radiomics in patients with glioblastoma: A systematic review and meta-analysis of prognostic endpoints

PURPOSE

Recent studies have shown promise of MR-based radiomics in predicting the survival of patients with untreated glioblastoma. This study aimed to comprehensively collate evidence to assess the prognostic value of radiomics in glioblastoma.

METHODS

PubMed-MEDLINE, Embase, and Web of Science were searched to find original articles investigating the prognostic value of MR-based radiomics in glioblastoma published up to July 14, 2023. Concordance indexes (C-indexes) and Cox proportional hazards ratios (HRs) of overall survival (OS) and progression-free survival (PFS) were pooled via random-effects modeling. For studies aimed at classifying long-term and short-term PFS, a hierarchical regression model was used to calculate pooled sensitivity and specificity. Between-study heterogeneity was assessed using the Higgin inconsistency index (I2). Subgroup regression analysis was performed to find potential factors contributing to heterogeneity. Publication bias was assessed via funnel plots and the Egger test.

RESULTS

Among 1371 abstracts, 18 and 17 studies were included for qualitative and quantitative data synthesis, respectively. Respective pooled C-indexes and HRs for OS were 0.65 (95 % confidence interval [CI], 0.58-0.72) and 2.88 (95 % CI, 2.28-3.64), whereas those for PFS were 0.61 (95 % CI, 0.55-0.66) and 2.78 (95 % CI, 1.91-4.03). Among 4 studies that predicted short-term PFS, the pooled sensitivity and specificity were 0.77 (95 % CI, 0.58-0.89) and 0.60 (95 % CI, 0.45-0.73), respectively. There was a substantial between-study heterogeneity among studies with the survival endpoint of OS C-index (n = 9, I2 = 83.8 %). Publication bias was not observed overall.

CONCLUSION

Pretreatment MR-based radiomics provided modest prognostic value in both OS and PFS in patients with glioblastoma.

Auto-Segmentation and Classification of Glioma Tumors with the Goals of Treatment Response Assessment Using Deep Learning Based on Magnetic Resonance Imaging

Glioma is the most common primary intracranial neoplasm in adults. Radiotherapy is a treatment approach in glioma patients, and Magnetic Resonance Imaging (MRI) is a beneficial diagnostic tool in treatment planning. Treatment response assessment in glioma patients is usually based on the Response Assessment in Neuro Oncology (RANO) criteria. The limitation of assessment based on RANO is two-dimensional (2D) manual measurements. Deep learning (DL) has great potential in neuro-oncology to improve the accuracy of response assessment. In the current research, firstly, the BraTS 2018 Challenge dataset included 210 HGG and 75 LGG were applied to train a designed U-Net network for automatic tumor and intra-tumoral segmentation, followed by training of the designed classifier with transfer learning for determining grading HGG and LGG. Then, designed networks were employed for the segmentation and classification of local MRI images of 49 glioma patients pre and post-radiotherapy. The results of tumor segmentation and its intra-tumoral regions were utilized to determine the volume of different regions and treatment response assessment. Treatment response assessment demonstrated that radiotherapy is effective on the whole tumor and enhancing region with p-value ≤ 0.05 with a 95% confidence level, while it did not affect necrosis and peri-tumoral edema regions. This work demonstrated the potential of using deep learning in MRI images to provide a beneficial tool in the automated treatment response assessment so that the patient can obtain the best treatment.

Assessment of brain cancer atlas maps with multimodal imaging features

BACKGROUND

Glioblastoma Multiforme (GBM) is a fast-growing and highly aggressive brain tumor that invades the nearby brain tissue and presents secondary nodular lesions across the whole brain but generally does not spread to distant organs. Without treatment, GBM can result in death in about 6 months. The challenges are known to depend on multiple factors: brain localization, resistance to conventional therapy, disrupted tumor blood supply inhibiting effective drug delivery, complications from peritumoral edema, intracranial hypertension, seizures, and neurotoxicity.

MAIN TEXT

Imaging techniques are routinely used to obtain accurate detections of lesions that localize brain tumors. Especially magnetic resonance imaging (MRI) delivers multimodal images both before and after the administration of contrast, which results in displaying enhancement and describing physiological features as hemodynamic processes. This review considers one possible extension of the use of radiomics in GBM studies, one that recalibrates the analysis of targeted segmentations to the whole organ scale. After identifying critical areas of research, the focus is on illustrating the potential utility of an integrated approach with multimodal imaging, radiomic data processing and brain atlases as the main components. The templates associated with the outcome of straightforward analyses represent promising inference tools able to spatio-temporally inform on the GBM evolution while being generalizable also to other cancers.

CONCLUSIONS

The focus on novel inference strategies applicable to complex cancer systems and based on building radiomic models from multimodal imaging data can be well supported by machine learning and other computational tools potentially able to translate suitably processed information into more accurate patient stratifications and evaluations of treatment efficacy.

Analysis of Functional Neuroplastic Changes in the Cortical Language System in Relation to Different Growth Patterns of Glioblastoma

The interpretation of fMRI data in glioblastoma (GB) is challenging as these tumors exhibit specific hemodynamic processes which, together with malignancy, tumor volume and proximity to eloquent cortex areas, may lead to misinterpretations of fMRI signals. The aim of this study was to investigate if different radiologically defined GB tumor growth patterns may also influence the fMRI signal, activation pattern and functional connectivity differently. Sixty-four patients with left-hemispheric glioblastoma were included and stratified according to their radiologically defined tumor growth pattern into groups with a uniform (U-TGP) or diffuse tumor growth pattern (D-TGP). Task-based fMRI data were analyzed using SPM12 with the marsbar, LI and CONN toolboxes. The percent signal change and the laterality index were analyzed, as well as functional connectivity between 23 selected ROIs. Comparisons of both patient groups showed only minor non-significant differences, indicating that the tumor growth pattern is not a relevant influencing factor for fMRI signal. In addition to these results, signal reductions were found in areas that were not affected by the tumor underlining that a GB is not a localized but rather a systemic disease affecting the entire brain.

Assessment and prediction of glioblastoma therapy response: challenges and opportunities

Glioblastoma is the most aggressive type of primary adult brain tumour. The median survival of patients with glioblastoma remains approximately 15 months, and the 5-year survival rate is <10%. Current treatment options are limited, and the standard of care has remained relatively constant since 2011. Over the last decade, a range of different treatment regimens have been investigated with very limited success. Tumour recurrence is almost inevitable with the current treatment strategies, as glioblastoma tumours are highly heterogeneous and invasive. Additionally, another challenging issue facing patients with glioblastoma is how to distinguish between tumour progression and treatment effects, especially when relying on routine diagnostic imaging techniques in the clinic. The specificity of routine imaging for identifying tumour progression early or in a timely manner is poor due to the appearance similarity of post-treatment effects. Here, we concisely describe the current status and challenges in the assessment and early prediction of therapy response and the early detection of tumour progression or recurrence. We also summarize and discuss studies of advanced approaches such as quantitative imaging, liquid biomarker discovery and machine intelligence that hold exceptional potential to aid in the therapy monitoring of this malignancy and early prediction of therapy response, which may decisively transform the conventional detection methods in the era of precision medicine.

Deep learning-based harmonization of CT reconstruction kernels towards improved clinical task performance

OBJECTIVES

To develop a deep learning-based harmonization framework, assessing whether it can improve performance of radiomics models given different kernels in different clinical tasks and additionally generalize to mitigate the effects of new/unobserved kernels on radiomics features.

METHODS

Patient data with 2 reconstruction kernels and phantom data with 22 reconstruction kernels were included. Eighty-five patients were studied for lymph node metastasis (LNM) prediction, and 164 patients for differential diagnosis between lung cancer (LC) and pulmonary tuberculosis (TB). Two convolutional neural network (CNN) models were developed to convert images (i) from B70f to B30f (CNNa) and (ii) from B30f to B70f (CNNb). Model performance between the two kernels was evaluated using AUC and compared with other well-known harmonization methods. Patient-normalized feature difference (PNFD) was used to identify the incompatible kernels (i.e., kernel with median PNFD > 1) with baseline (B30f/B70f), and measure the ability of the CNN models to convert the non-comparable kernels.

RESULTS

For LC versus pulmonary TB diagnosis, AUCs of CNNa vs. others were 0.85 vs. 0.54-0.74 (p = 0.0001-0.0003), and for CNNb vs. others: 0.87 vs. 0.54-0.86 (p = 0.0001-0.55). For LNM prediction, AUCs of CNNa vs. others were 0.68 vs. 0.56-0.61 (p = 0.10-0.39), and for CNNb vs. others: 0.78 vs. 0.70-0.73 (p = 0.07-0.40). After CNN harmonization, 17 of 20 (85%) of investigated unknown kernels produced comparable radiomics feature values relative to baseline (median PNFD from 1.10-2.31 to 0.23-1.13).

CONCLUSION

The CNN harmonization effectively improved performance of radiomics models between reconstruction kernels in different clinical tasks, and reduced feature differences between unknown kernels vs. baseline.

KEY POINTS

• The soft (B30f) and sharp (B70f) kernels strongly affect radiomics reproducibility and generalizability. • The convolutional neural network (CNN) harmonization methods performed better than location-scale (ComBat and centering-scaling) and matrix factorization harmonization methods (based on singular value decomposition (SVD) and independent component analysis (ICA)) in both clinical tasks. • The CNN harmonization methods improve feature reproducibility not only between specific kernels (B30f and B70f) from the same scanner, but also between unobserved kernels from different scanners of different vendors.

A Deep Learning Radiomics Nomogram to Predict Response to Neoadjuvant Chemotherapy for Locally Advanced Cervical Cancer: A Two-Center Study

Purpose: This study aimed to establish a deep learning radiomics nomogram (DLRN) based on multiparametric MR images for predicting the response to neoadjuvant chemotherapy (NACT) in patients with locally advanced cervical cancer (LACC). Methods: Patients with LACC (FIGO stage IB-IIIB) who underwent preoperative NACT were enrolled from center 1 (220 cases) and center 2 (independent external validation dataset, 65 cases). Handcrafted and deep learning-based radiomics features were extracted from T2WI, DWI and contrast-enhanced (CE)-T1WI, and radiomics signatures were built based on the optimal features. Two types of radiomics signatures and clinical features were integrated into the DLRN for prediction. The AUC, calibration curve and decision curve analysis (DCA) were employed to illustrate the performance of these models and their clinical utility. In addition, disease-free survival (DFS) was assessed by Kaplan–Meier survival curves based on the DLRN. Results: The DLRN showed favorable predictive values in differentiating responders from nonresponders to NACT with AUCs of 0.963, 0.940 and 0.910 in the three datasets, with good calibration (all p > 0.05). Furthermore, the DLRN performed better than the clinical model and handcrafted radiomics signature in all datasets (all p < 0.05) and slightly higher than the DL-based radiomics signature in the internal validation dataset (p = 0.251). DCA indicated that the DLRN has potential in clinical applications. Furthermore, the DLRN was strongly correlated with the DFS of LACC patients (HR = 0.223; p = 0.004). Conclusion: The DLRN performed well in preoperatively predicting the therapeutic response in LACC and could provide valuable information for individualized treatment.

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.

Differentiation of acute and chronic vertebral compression fractures using conventional CT based on deep transfer learning features and hand-crafted radiomics features

BACKGROUND

We evaluated the diagnostic efficacy of deep learning radiomics (DLR) and hand-crafted radiomics (HCR) features in differentiating acute and chronic vertebral compression fractures (VCFs).

METHODS

A total of 365 patients with VCFs were retrospectively analysed based on their computed tomography (CT) scan data. All patients completed MRI examination within 2 weeks. There were 315 acute VCFs and 205 chronic VCFs. Deep transfer learning (DTL) features and HCR features were extracted from CT images of patients with VCFs using DLR and traditional radiomics, respectively, and feature fusion was performed to establish the least absolute shrinkage and selection operator. The MRI display of vertebral bone marrow oedema was used as the gold standard for acute VCF, and the model performance was evaluated using the receiver operating characteristic (ROC).To separately evaluate the effectiveness of DLR, traditional radiomics and feature fusion in the differential diagnosis of acute and chronic VCFs, we constructed a nomogram based on the clinical baseline data to visualize the classification evaluation. The predictive power of each model was compared using the Delong test, and the clinical value of the nomogram was evaluated using decision curve analysis (DCA).

RESULTS

Fifty DTL features were obtained from DLR, 41 HCR features were obtained from traditional radiomics, and 77 features fusion were obtained after feature screening and fusion of the two. The area under the curve (AUC) of the DLR model in the training cohort and test cohort were 0.992 (95% confidence interval (CI), 0.983-0.999) and 0.871 (95% CI, 0.805-0.938), respectively. While the AUCs of the conventional radiomics model in the training cohort and test cohort were 0.973 (95% CI, 0.955-0.990) and 0.854 (95% CI, 0.773-0.934), respectively. The AUCs of the features fusion model in the training cohort and test cohort were 0.997 (95% CI, 0.994-0.999) and 0.915 (95% CI, 0.855-0.974), respectively. The AUCs of nomogram constructed by the features fusion in combination with clinical baseline data were 0.998 (95% CI, 0.996-0.999) and 0.946 (95% CI, 0.906-0.987) in the training cohort and test cohort, respectively. The Delong test showed that the differences between the features fusion model and the nomogram in the training cohort and the test cohort were not statistically significant (P values were 0.794 and 0.668, respectively), and the differences in the other prediction models in the training cohort and the test cohort were statistically significant (P < 0.05). DCA showed that the nomogram had high clinical value.

CONCLUSION

The features fusion model can be used for the differential diagnosis of acute and chronic VCFs, and its differential diagnosis ability is improved when compared with that when either radiomics is used alone. At the same time, the nomogram has a high predictive value for acute and chronic VCFs and can be a potential decision-making tool to assist clinicians, especially when a patient is unable to undergo spinal MRI examination.

MR Intensity Normalization Methods Impact Sequence Specific Radiomics Prognostic Model Performance in Primary and Recurrent High-Grade Glioma

PURPOSE

This study investigates the impact of different intensity normalization (IN) methods on the overall survival (OS) radiomics models' performance of MR sequences in primary (pHGG) and recurrent high-grade glioma (rHGG).

METHODS

MR scans acquired before radiotherapy were retrieved from two independent cohorts (rHGG C1: 197, pHGG C2: 141) from multiple scanners (15, 14). The sequences are T1 weighted (w), contrast-enhanced T1w (T1wce), T2w, and T2w-FLAIR. Sequence-specific significant features (SF) associated with OS, extracted from the tumour volume, were derived after applying 15 different IN methods. Survival analyses were conducted using Cox proportional hazard (CPH) and Poisson regression (POI) models. A ranking score was assigned based on the 10-fold cross-validated (CV) concordance index (C-I), mean square error (MSE), and the Akaike information criterion (AICs), to evaluate the methods' performance.

RESULTS

Scatter plots of the 10-CV C-I and MSE against the AIC showed an impact on the survival predictions between the IN methods and MR sequences (C1/C2 C-I range: 0.62-0.71/0.61-0.72, MSE range: 0.20-0.42/0.13-0.22). White stripe showed stable results for T1wce (C1/C2 C-I: 0.71/0.65, MSE: 0.21/0.14). Combat (0.68/0.62, 0.22/0.15) and histogram matching (HM, 0.67/0.64, 0.22/0.15) showed consistent prediction results for T2w models. They were also the top-performing methods for T1w in C2 (Combat: 0.67, 0.13; HM: 0.67, 0.13); however, only HM achieved high predictions in C1 (0.66, 0.22). After eliminating IN impacted SF using Spearman's rank-order correlation coefficient, a mean decrease in the C-I and MSE of 0.05 and 0.03 was observed in all four sequences.

CONCLUSION

The IN method impacted the predictive power of survival models; thus, performance is sequence-dependent.

Image-based deep learning identifies glioblastoma risk groups with genomic and transcriptomic heterogeneity: a multi-center study

OBJECTIVES

To develop and validate a deep learning imaging signature (DLIS) for risk stratification in patients with multiforme (GBM), and to investigate the biological pathways and genetic alterations underlying the DLIS.

METHODS

The DLIS was developed from multi-parametric MRI based on a training set (n = 600) and validated on an internal validation set (n = 164), an external test set 1 (n = 100), an external test set 2 (n = 161), and a public TCIA set (n = 88). A co-profiling framework based on a radiogenomics analysis dataset (n = 127) using multiscale high-dimensional data, including imaging, transcriptome, and genome, was established to uncover the biological pathways and genetic alterations underpinning the DLIS.

RESULTS

The DLIS was associated with survival (log-rank p < 0.001) and was an independent predictor (p < 0.001). The integrated nomogram incorporating the DLIS achieved improved C indices than the clinicomolecular nomogram (net reclassification improvement 0.39, p < 0.001). DLIS significantly correlated with core pathways of GBM (apoptosis and cell cycle-related P53 and RB pathways, and cell proliferation-related RTK pathway), as well as key genetic alterations (del_CDNK2A). The prognostic value of DLIS-correlated genes was externally confirmed on TCGA/CGGA sets (p < 0.01).

CONCLUSIONS

Our study offers a biologically interpretable deep learning predictor of survival outcomes in patients with GBM, which is crucial for better understanding GBM patient's prognosis and guiding individualized treatment.

KEY POINTS

• MRI-based deep learning imaging signature (DLIS) stratifies GBM into risk groups with distinct molecular characteristics. • DLIS is associated with P53, RB, and RTK pathways and del_CDNK2A mutation. • The prognostic value of DLIS-correlated pathway genes is externally demonstrated.

Predicting effectiveness of anti-VEGF injection through self-supervised learning in OCT images

Anti-vascular endothelial growth factor (Anti-VEGF) therapy has become a standard way for choroidal neovascularization (CNV) and cystoid macular edema (CME) treatment. However, anti-VEGF injection is a long-term therapy with expensive cost and may be not effective for some patients. Therefore, predicting the effectiveness of anti-VEGF injection before the therapy is necessary. In this study, a new optical coherence tomography (OCT) images based self-supervised learning (OCT-SSL) model for predicting the effectiveness of anti-VEGF injection is developed. In OCT-SSL, we pre-train a deep encoder-decoder network through self-supervised learning to learn the general features using a public OCT image dataset. Then, model fine-tuning is performed on our own OCT dataset to learn the discriminative features to predict the effectiveness of anti-VEGF. Finally, classifier trained by the features from fine-tuned encoder as a feature extractor is built to predict the response. Experimental results on our private OCT dataset demonstrated that the proposed OCT-SSL can achieve an average accuracy, area under the curve (AUC), sensitivity and specificity of 0.93, 0.98, 0.94 and 0.91, respectively. Meanwhile, it is found that not only the lesion region but also the normal region in OCT image is related to the effectiveness of anti-VEGF.

Development and validation of a preoperative MRI-based radiomics nomogram to predict progression-free survival in patients with clival chordomas

Objectives

The aim of this study was to establish and validate a MRI-based radiomics nomogram to predict progression-free survival (PFS) of clival chordoma.

Methods

A total of 174 patients were enrolled in the study (train cohort: 121 cases, test cohort: 53 cases). Radiomic features were extracted from multiparametric MRIs. Intraclass correlation coefficient analysis and a Lasso and Elastic-Net regularized generalized linear model were used for feature selection. Then, a nomogram was established via univariate and multivariate Cox regression analysis in the train cohort. The performance of this nomogram was assessed by area under curve (AUC) and calibration curve.

Results

A total of 3318 radiomic features were extracted from each patient, of which 2563 radiomic features were stable features. After feature selection, seven radiomic features were selected. Cox regression analysis revealed that 2 clinical factors (degree of resection, and presence or absence of primary chordoma) and 4 radiomic features were independent prognostic factors. The AUC of the established nomogram was 0.747, 0.807, and 0.904 for PFS prediction at 1, 3, and 5 years in the train cohort, respectively, compared with 0.582, 0.852, and 0.914 in the test cohort. Calibration and risk score stratified survival curves were satisfactory in the train and test cohort.

Conclusions

The presented nomogram demonstrated a favorable predictive accuracy of PFS, which provided a novel tool to predict prognosis and risk stratification. Our results suggest that radiomic analysis can effectively help neurosurgeons perform individualized evaluations of patients with clival chordomas.

Federated learning enables big data for rare cancer boundary detection

Although machine learning (ML) has shown promise across disciplines, out-of-sample generalizability is concerning. This is currently addressed by sharing multi-site data, but such centralization is challenging/infeasible to scale due to various limitations. Federated ML (FL) provides an alternative paradigm for accurate and generalizable ML, by only sharing numerical model updates. Here we present the largest FL study to-date, involving data from 71 sites across 6 continents, to generate an automatic tumor boundary detector for the rare disease of glioblastoma, reporting the largest such dataset in the literature (n = 6, 314). We demonstrate a 33% delineation improvement for the surgically targetable tumor, and 23% for the complete tumor extent, over a publicly trained model. We anticipate our study to: 1) enable more healthcare studies informed by large diverse data, ensuring meaningful results for rare diseases and underrepresented populations, 2) facilitate further analyses for glioblastoma by releasing our consensus model, and 3) demonstrate the FL effectiveness at such scale and task-complexity as a paradigm shift for multi-site collaborations, alleviating the need for data-sharing.

Current Evidence, Limitations and Future Challenges of Survival Prediction for Glioblastoma Based on Advanced Noninvasive Methods: A Narrative Review

Background and Objectives: Survival estimation for patients diagnosed with Glioblastoma (GBM) is an important information to consider in patient management and communication. Despite some known risk factors, survival estimation remains a major challenge. Novel non-invasive technologies such as radiomics and artificial intelligence (AI) have been implemented to increase the accuracy of these predictions. In this article, we reviewed and discussed the most significant available research on survival estimation for GBM through advanced non-invasive methods. Materials and Methods: PubMed database was queried for articles reporting on survival prognosis for GBM through advanced image and data management methods. Articles including in their title or abstract the following terms were initially screened: ((glioma) AND (survival)) AND ((artificial intelligence) OR (radiomics)). Exclusively English full-text articles, reporting on humans, published as of 1 September 2022 were considered. Articles not reporting on overall survival, evaluating the effects of new therapies or including other tumors were excluded. Research with a radiomics-based methodology were evaluated using the radiomics quality score (RQS). Results: 382 articles were identified. After applying the inclusion criteria, 46 articles remained for further analysis. These articles were thoroughly assessed, summarized and discussed. The results of the RQS revealed some of the limitations of current radiomics investigation on this field. Limitations of analyzed studies included data availability, patient selection and heterogeneity of methodologies. Future challenges on this field are increasing data availability, improving the general understanding of how AI handles data and establishing solid correlations between image features and tumor's biology. Conclusions: Radiomics and AI methods of data processing offer a new paradigm of possibilities to tackle the question of survival prognosis in GBM.

Predicting Survival in Patients with Brain Tumors: Current State-of-the-Art of AI Methods Applied to MRI

Given growing clinical needs, in recent years Artificial Intelligence (AI) techniques have increasingly been used to define the best approaches for survival assessment and prediction in patients with brain tumors. Advances in computational resources, and the collection of (mainly) public databases, have promoted this rapid development. This narrative review of the current state-of-the-art aimed to survey current applications of AI in predicting survival in patients with brain tumors, with a focus on Magnetic Resonance Imaging (MRI). An extensive search was performed on PubMed and Google Scholar using a Boolean research query based on MeSH terms and restricting the search to the period between 2012 and 2022. Fifty studies were selected, mainly based on Machine Learning (ML), Deep Learning (DL), radiomics-based methods, and methods that exploit traditional imaging techniques for survival assessment. In addition, we focused on two distinct tasks related to survival assessment: the first on the classification of subjects into survival classes (short and long-term or eventually short, mid and long-term) to stratify patients in distinct groups. The second focused on quantification, in days or months, of the individual survival interval. Our survey showed excellent state-of-the-art methods for the first, with accuracy up to ∼98%. The latter task appears to be the most challenging, but state-of-the-art techniques showed promising results, albeit with limitations, with C-Index up to ∼0.91. In conclusion, according to the specific task, the available computational methods perform differently, and the choice of the best one to use is non-univocal and dependent on many aspects. Unequivocally, the use of features derived from quantitative imaging has been shown to be advantageous for AI applications, including survival prediction. This evidence from the literature motivates further research in the field of AI-powered methods for survival prediction in patients with brain tumors, in particular, using the wealth of information provided by quantitative MRI techniques.

Deep learning models for predicting the survival of patients with chondrosarcoma based on a surveillance, epidemiology, and end results analysis

Background

Accurate prediction of prognosis is critical for therapeutic decisions in chondrosarcoma patients. Several prognostic models have been created utilizing multivariate Cox regression or binary classification-based machine learning approaches to predict the 3- and 5-year survival of patients with chondrosarcoma, but few studies have investigated the results of combining deep learning with time-to-event prediction. Compared with simplifying the prediction as a binary classification problem, modeling the probability of an event as a function of time by combining it with deep learning can provide better accuracy and flexibility.

Materials and methods

Patients with the diagnosis of chondrosarcoma between 2000 and 2018 were extracted from the Surveillance, Epidemiology, and End Results (SEER) registry. Three algorithms—two based on neural networks (DeepSurv, neural multi-task logistic regression [NMTLR]) and one on ensemble learning (random survival forest [RSF])—were selected for training. Meanwhile, a multivariate Cox proportional hazards (CoxPH) model was also constructed for comparison. The dataset was randomly divided into training and testing datasets at a ratio of 7:3. Hyperparameter tuning was conducted through a 1000-repeated random search with 5-fold cross-validation on the training dataset. The model performance was assessed using the concordance index (C-index), Brier score, and Integrated Brier Score (IBS). The accuracy of predicting 1-, 3-, 5- and 10-year survival was evaluated using receiver operating characteristic curves (ROC), calibration curves, and the area under the ROC curves (AUC).

Results

A total of 3145 patients were finally enrolled in our study. The mean age at diagnosis was 52 ± 18 years, 1662 of the 3145 patients were male (53%), and mean survival time was 83 ± 67 months. Two deep learning models outperformed the RSF and classical CoxPH models, with the C-index on test datasets achieving values of 0.832 (DeepSurv) and 0.821 (NMTLR). The DeepSurv model produced better accuracy and calibrated survival estimates in predicting 1-, 3- 5- and 10-year survival (AUC:0.895-0.937). We deployed the DeepSurv model as a web application for use in clinical practice; it can be accessed through https://share.streamlit.io/whuh-ml/chondrosarcoma/Predict/app.py.

Conclusions

Time-to-event prediction models based on deep learning algorithms are successful in predicting chondrosarcoma prognosis, with DeepSurv producing the best discriminative performance and calibration.

An automated approach for predicting glioma grade and survival of LGG patients using CNN and radiomics

Objectives

To develop and validate an efficient and automatically computational approach for stratifying glioma grades and predicting survival of lower-grade glioma (LGG) patients using an integration of state-of-the-art convolutional neural network (CNN) and radiomics.

Method

This retrospective study reviewed 470 preoperative MR images of glioma from BraTs public dataset (n=269) and Jinling hospital (n=201). A fully automated pipeline incorporating tumor segmentation and grading was developed, which can avoid variability and subjectivity of manual segmentations. First, an integrated approach by fusing CNN features and radiomics features was employed to stratify glioma grades. Then, a deep-radiomics signature based on the integrated approach for predicting survival of LGG patients was developed and subsequently validated in an independent cohort.

Results

The performance of tumor segmentation achieved a Dice coefficient of 0.81. The intraclass correlation coefficients (ICCs) of the radiomics features between the segmentation network and physicians were all over 0.75. The performance of glioma grading based on integrated approach achieved the area under the curve (AUC) of 0.958, showing the effectiveness of the integrated approach. The multivariable Cox regression results demonstrated that the deep-radiomics signature remained an independent prognostic factor and the integrated nomogram showed significantly better performance than the clinical nomogram in predicting overall survival of LGG patients (C-index: 0.865 vs. 0.796, P=0.005).

Conclusion

The proposed integrated approach can be noninvasively and efficiently applied in prediction of gliomas grade and survival. Moreover, our fully automated pipeline successfully achieved computerized segmentation instead of manual segmentation, which shows the potential to be a reproducible approach in clinical practice.

Evaluation of Epidermal Growth Factor Receptor 2 Status in Gastric Cancer by CT-Based Deep Learning Radiomics Nomogram

Purpose

To explore the role of computed tomography (CT)-based deep learning and radiomics in preoperative evaluation of epidermal growth factor receptor 2 (HER2) status in gastric cancer.

Materials and methods

The clinical data on gastric cancer patients were evaluated retrospectively, and 357 patients were chosen for this study (training cohort: 249; test cohort: 108). The preprocessed enhanced CT arterial phase images were selected for lesion segmentation, radiomics and deep learning feature extraction. We integrated deep learning features and radiomic features (Inte). Four methods were used for feature selection. We constructed models with support vector machine (SVM) or random forest (RF), respectively. The area under the receiver operating characteristics curve (AUC) was used to assess the performance of these models. We also constructed a nomogram including Inte-feature scores and clinical factors.

Results

The radiomics-SVM model showed good classification performance (AUC, training cohort: 0.8069; test cohort: 0.7869). The AUC of the ResNet50-SVM model and the Inte-SVM model in the test cohort were 0.8955 and 0.9055. The nomogram also showed excellent discrimination achieving greater AUC (training cohort, 0.9207; test cohort, 0.9224).

Conclusion

CT-based deep learning radiomics nomogram can accurately and effectively assess the HER2 status in patients with gastric cancer before surgery and it is expected to assist physicians in clinical decision-making and facilitates individualized treatment planning.

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.

Evaluation of cancer outcome assessment using MRI: A review of deep-learning methods

Accurate evaluation of tumor response to treatment is critical to allow personalized treatment regimens according to the predicted response and to support clinical trials investigating new therapeutic agents by providing them with an accurate response indicator. Recent advances in medical imaging, computer hardware, and machine-learning algorithms have resulted in the increased use of these tools in the field of medicine as a whole and specifically in cancer imaging for detection and characterization of malignant lesions, prognosis, and assessment of treatment response. Among the currently available imaging techniques, magnetic resonance imaging (MRI) plays an important role in the evaluation of treatment assessment of many cancers, given its superior soft-tissue contrast and its ability to allow multiplanar imaging and functional evaluation. In recent years, deep learning (DL) has become an active area of research, paving the way for computer-assisted clinical and radiological decision support. DL can uncover associations between imaging features that cannot be visually identified by the naked eye and pertinent clinical outcomes. The aim of this review is to highlight the use of DL in the evaluation of tumor response assessed on MRI. In this review, we will first provide an overview of common DL architectures used in medical imaging research in general. Then, we will review the studies to date that have applied DL to magnetic resonance imaging for the task of treatment response assessment. Finally, we will discuss the challenges and opportunities of using DL within the clinical workflow.

Multimodal MRI Image Decision Fusion-Based Network for Glioma Classification

Purpose

Glioma is the most common primary brain tumor, with varying degrees of aggressiveness and prognosis. Accurate glioma classification is very important for treatment planning and prognosis prediction. The main purpose of this study is to design a novel effective algorithm for further improving the performance of glioma subtype classification using multimodal MRI images.

Method

MRI images of four modalities for 221 glioma patients were collected from Computational Precision Medicine: Radiology-Pathology 2020 challenge, including T1, T2, T1ce, and fluid-attenuated inversion recovery (FLAIR) MRI images, to classify astrocytoma, oligodendroglioma, and glioblastoma. We proposed a multimodal MRI image decision fusion-based network for improving the glioma classification accuracy. First, the MRI images of each modality were input into a pre-trained tumor segmentation model to delineate the regions of tumor lesions. Then, the whole tumor regions were centrally clipped from original MRI images followed by max-min normalization. Subsequently, a deep learning-based network was designed based on a unified DenseNet structure, which extracts features through a series of dense blocks. After that, two fully connected layers were used to map the features into three glioma subtypes. During the training stage, we used the images of each modality after tumor segmentation to train the network to obtain its best accuracy on our testing set. During the inferring stage, a linear weighted module based on a decision fusion strategy was applied to assemble the predicted probabilities of the pre-trained models obtained in the training stage. Finally, the performance of our method was evaluated in terms of accuracy, area under the curve (AUC), sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), etc.

Results

The proposed method achieved an accuracy of 0.878, an AUC of 0.902, a sensitivity of 0.772, a specificity of 0.930, a PPV of 0.862, an NPV of 0.949, and a Cohen's Kappa of 0.773, which showed a significantly higher performance than existing state-of-the-art methods.

Conclusion

Compared with current studies, this study demonstrated the effectiveness and superiority in the overall performance of our proposed multimodal MRI image decision fusion-based network method for glioma subtype classification, which would be of enormous potential value in clinical practice.

Preoperative prediction of lymph node metastasis using deep learning-based features

Lymph node involvement increases the risk of breast cancer recurrence. An accurate non-invasive assessment of nodal involvement is valuable in cancer staging, surgical risk, and cost savings. Radiomics has been proposed to pre-operatively predict sentinel lymph node (SLN) status; however, radiomic models are known to be sensitive to acquisition parameters. The purpose of this study was to develop a prediction model for preoperative prediction of SLN metastasis using deep learning-based (DLB) features and compare its predictive performance to state-of-the-art radiomics. Specifically, this study aimed to compare the generalizability of radiomics vs DLB features in an independent test set with dissimilar resolution. Dynamic contrast-enhancement images from 198 patients (67 positive SLNs) were used in this study. Of these subjects, 163 had an in-plane resolution of 0.7 × 0.7 mm², which were randomly divided into a training set (approximately 67%) and a validation set (approximately 33%). The remaining 35 subjects with a different in-plane resolution (0.78 × 0.78 mm²) were treated as independent testing set for generalizability. Two methods were employed: (1) conventional radiomics (CR), and (2) DLB features which replaced hand-curated features with pre-trained VGG-16 features. The threshold determined using the training set was applied to the independent validation and testing dataset. Same feature reduction, feature selection, model creation procedures were used for both approaches. In the validation set (same resolution as training), the DLB model outperformed the CR model (accuracy 83% vs 80%). Furthermore, in the independent testing set of the dissimilar resolution, the DLB model performed markedly better than the CR model (accuracy 77% vs 71%). The predictive performance of the DLB model outperformed the CR model for this task. More interestingly, these improvements were seen particularly in the independent testing set of dissimilar resolution. This could indicate that DLB features can ultimately result in a more generalizable model.

Treatment Response and Prognosis Evaluation in High-Grade Glioma: An Imaging Review Based on MRI

In recent years, the development of advanced magnetic resonance imaging (MRI) technology and machine learning (ML) have created new tools for evaluating treatment response and prognosis of patients with high-grade gliomas (HGG); however, patient prognosis has not improved significantly. This is mainly due to the heterogeneity between and within HGG tumors, resulting in standard treatment methods not benefitting all patients. Moreover, the survival of patients with HGG is not only related to tumor cells, but also to noncancer cells in the tumor microenvironment (TME). Therefore, during preoperative diagnosis and follow-up treatment of patients with HGG, noninvasive imaging markers are needed to characterize intratumoral heterogeneity, and then to evaluate treatment response and predict prognosis, timeously adjust treatment strategies, and achieve individualized diagnosis and treatment. In this review, we summarize the research progress of conventional MRI, advanced MRI technology, and ML in evaluation of treatment response and prognosis of patients with HGG. We further discuss the significance of the TME in the prognosis of HGG patients, associate imaging features with the TME, indirectly reflecting the heterogeneity within the tumor, and shifting treatment strategies from tumor cells alone to systemic therapy of the TME, which may be a major development direction in the future. LEVEL OF EVIDENCE: 5 TECHNICAL EFFICACY STAGE: 4.

Developing and validating a deep learning and radiomic model for glioma grading using multiplanar reconstructed magnetic resonance contrast-enhanced T1-weighted imaging: a robust, multi-institutional study

Background

Although surgical pathology or biopsy are considered the gold standard for glioma grading, these procedures have limitations. This study set out to evaluate and validate the predictive performance of a deep learning radiomics model based on contrast-enhanced T1-weighted multiplanar reconstruction images for grading gliomas.

Methods

Patients from three institutions who diagnosed with gliomas by surgical specimen and multiplanar reconstructed (MPR) images were enrolled in this study. The training cohort included 101 patients from institution 1, including 43 high-grade glioma (HGG) patients and 58 low-grade glioma (LGG) patients, while the test cohorts consisted of 50 patients from institutions 2 and 3 (25 HGG patients, 25 LGG patients). We then extracted radiomics features and deep learning features using six pretrained models from the MPR images. The Spearman correlation test and the recursive elimination feature selection method were used to reduce the redundancy and select most predictive features. Subsequently, three classifiers were used to construct classification models. The performance of the grading models was evaluated using the area under the receiver operating curve, sensitivity, specificity, accuracy, precision, and negative predictive value. Finally, the prediction performances of the test cohort were compared to determine the optimal classification model.

Results

For the training cohort, 62% (13 out of 21) of the classification models constructed with MPR images from multiple planes outperformed those constructed with single-plane MPR images, and 61% (11 out of 18) of classification models constructed with both radiomics features and deep learning features had higher area under the curve (AUC) values than those constructed with only radiomics or deep learning features. The optimal model was a random forest model that combined radiomic features and VGG16 deep learning features derived from MPR images, which achieved AUC of 0.847 in the training cohort and 0.898 in the test cohort. In the test cohort, the sensitivity, specificity, and accuracy of the optimal model were 0.840, 0.760, and 0.800, respectively.

Conclusions

Multiplanar CE-T1W MPR imaging features are more effective than features from single planes when differentiating HGG and LGG. The combination of deep learning features and radiomics features can effectively grade glioma and assist clinical decision-making.

Prognostic aspects of lymphovascular invasion in localized gastric cancer: new insights into the radiomics and deep transfer learning from contrast-enhanced CT imaging

OBJECTIVES

Lymphovascular invasion (LVI) is a factor significantly impacting treatment and outcome of patients with gastric cancer (GC). We aimed to investigate prognostic aspects of a preoperative LVI prediction in GC using radiomics and deep transfer learning (DTL) from contrast-enhanced CT (CECT) imaging.

METHODS

A total of 1062 GC patients (728 training and 334 testing) between Jan 2014 and Dec 2018 undergoing gastrectomy were retrospectively included. Based on CECT imaging, we built two gastric imaging (GI) markers, GI-marker-1 from radiomics and GI-marker-2 from DTL features, to decode LVI status. We then integrated demographics, clinical data, GI markers, radiologic interpretation, and biopsies into a Gastric Cancer Risk (GRISK) model for predicting LVI. The performance of GRISK model was tested and applied to predict survival outcomes in GC patients. Furthermore, the prognosis between LVI (+) and LVI (-) patients was compared in chemotherapy and non-chemotherapy cohorts, respectively.

RESULTS

GI-marker-1 and GI-marker-2 yield similar performance in predicting LVI in training and testing dataset. The GRISK model yields the diagnostic performance with AUC of 0.755 (95% CI 0.719-0.790) and 0.725 (95% CI 0.669-0.781) in training and testing dataset. Patients with LVI (+) trend toward lower progression-free survival (PFS) and overall survival (OS). The difference of prognosis between LVI (+) and LVI (-) was more noticeable in non-chemotherapy than that in chemotherapy group.

CONCLUSION

Radiomics and deep transfer learning features on CECT demonstrate potential power for predicting LVI in GC patients. Prospective use of a GRISK model can help to optimize individualized treatment decisions and predict survival outcomes.

Preoperative Contrast-Enhanced MRI in Differentiating Glioblastoma From Low-Grade Gliomas in The Cancer Imaging Archive Database: A Proof-of-Concept Study

Purpose

This study aimed to develop a repeatable MRI-based machine learning model to differentiate between low-grade gliomas (LGGs) and glioblastoma (GBM) and provide more clinical information to improve treatment decision-making.

Methods

Preoperative MRIs of gliomas from The Cancer Imaging Archive (TCIA)–GBM/LGG database were selected. The tumor on contrast-enhanced MRI was segmented. Quantitative image features were extracted from the segmentations. A random forest classification algorithm was used to establish a model in the training set. In the test phase, a random forest model was tested using an external test set. Three radiologists reviewed the images for the external test set. The area under the receiver operating characteristic curve (AUC) was calculated. The AUCs of the radiomics model and radiologists were compared.

Results

The random forest model was fitted using a training set consisting of 142 patients [mean age, 52 years ± 16 (standard deviation); 78 men] comprising 88 cases of GBM. The external test set included 25 patients (14 with GBM). Random forest analysis yielded an AUC of 1.00 [95% confidence interval (CI): 0.86–1.00]. The AUCs for the three readers were 0.92 (95% CI 0.74–0.99), 0.70 (95% CI 0.49–0.87), and 0.59 (95% CI 0.38–0.78). Statistical differences were only found between AUC and Reader 1 (1.00 vs. 0.92, respectively; p = 0.16).

Conclusion

An MRI radiomics-based random forest model was proven useful in differentiating GBM from LGG and showed better diagnostic performance than that of two inexperienced radiologists.

Practice of the new supervised machine learning predictive analytics for glioma patient survival after tumor resection: Experiences in a high-volume Chinese center

OBJECTIVE

This study aims to assess the effectiveness of the Gradient Boosting (GB) algorithm on glioma prognosis prediction and to explore new predictive models for glioma patient survival after tumor resection.

METHODS

A cohort of 776 glioma cases (WHO grades II-IV) between 2010 and 2017 was obtained. Clinical characteristics and biomarker information were reviewed. Subsequently, we constructed the conventional Cox survival model and three different supervised machine learning models, including support vector machine (SVM), random survival forest (RSF), Tree GB, and Component GB. Then, the model performance was compared with each other. At last, we also assessed the feature importance of models.

RESULTS

The concordance indexes of the conventional survival model, SVM, RSF, Tree GB, and Component GB were 0.755, 0.787, 0.830, 0.837, and 0.840, respectively. All areas under the cumulative receiver operating characteristic curve of both GB models were above 0.800 at different survival times. Their calibration curves showed good calibration of survival prediction. Meanwhile, the analysis of feature importance revealed Karnofsky performance status, age, tumor subtype, extent of resection, and so on as crucial predictive factors.

CONCLUSION

Gradient Boosting models performed better in predicting glioma patient survival after tumor resection than other models.

Deep Learning for Outcome Prediction in Neurosurgery: A Systematic Review of Design, Reporting, and Reproducibility

Deep learning (DL) is a powerful machine learning technique that has increasingly been used to predict surgical outcomes. However, the large quantity of data required and lack of model interpretability represent substantial barriers to the validity and reproducibility of DL models. The objective of this study was to systematically review the characteristics of DL studies involving neurosurgical outcome prediction and to assess their bias and reporting quality. Literature search using the PubMed, Scopus, and Embase databases identified 1949 records of which 35 studies were included. Of these, 32 (91%) developed and validated a DL model while 3 (9%) validated a pre-existing model. The most commonly represented subspecialty areas were oncology (16 of 35, 46%), spine (8 of 35, 23%), and vascular (6 of 35, 17%). Risk of bias was low in 18 studies (51%), unclear in 5 (14%), and high in 12 (34%), most commonly because of data quality deficiencies. Adherence to transparent reporting of a multivariable prediction model for individual prognosis or diagnosis reporting standards was low, with a median of 12 transparent reporting of a multivariable prediction model for individual prognosis or diagnosis items (39%) per study not reported. Model transparency was severely limited because code was provided in only 3 studies (9%) and final models in 2 (6%). With the exception of public databases, no study data sets were readily available. No studies described DL models as ready for clinical use. The use of DL for neurosurgical outcome prediction remains nascent. Lack of appropriate data sets poses a major concern for bias. Although studies have demonstrated promising results, greater transparency in model development and reporting is needed to facilitate reproducibility and validation.

Artificial intelligence, machine learning, and deep learning for clinical outcome prediction

AI is a broad concept, grouping initiatives that use a computer to perform tasks that would usually require a human to complete. AI methods are well suited to predict clinical outcomes. In practice, AI methods can be thought of as functions that learn the outcomes accompanying standardized input data to produce accurate outcome predictions when trialed with new data. Current methods for cleaning, creating, accessing, extracting, augmenting, and representing data for training AI clinical prediction models are well defined. The use of AI to predict clinical outcomes is a dynamic and rapidly evolving arena, with new methods and applications emerging. Extraction or accession of electronic health care records and combining these with patient genetic data is an area of present attention, with tremendous potential for future growth. Machine learning approaches, including decision tree methods of Random Forest and XGBoost, and deep learning techniques including deep multi-layer and recurrent neural networks, afford unique capabilities to accurately create predictions from high dimensional, multimodal data. Furthermore, AI methods are increasing our ability to accurately predict clinical outcomes that previously were difficult to model, including time-dependent and multi-class outcomes. Barriers to robust AI-based clinical outcome model deployment include changing AI product development interfaces, the specificity of regulation requirements, and limitations in ensuring model interpretability, generalizability, and adaptability over time.

Differentiation of Brain Abscess From Cystic Glioma Using Conventional MRI Based on Deep Transfer Learning Features and Hand-Crafted Radiomics Features

Objectives: To develop and validate the model for distinguishing brain abscess from cystic glioma by combining deep transfer learning (DTL) features and hand-crafted radiomics (HCR) features in conventional T1-weighted imaging (T1WI) and T2-weighted imaging (T2WI).

Methods: This single-center retrospective analysis involved 188 patients with pathologically proven brain abscess (102) or cystic glioma (86). One thousand DTL and 105 HCR features were extracted from the T1WI and T2WI of the patients. Three feature selection methods and four classifiers, such as k-nearest neighbors (KNN), random forest classifier (RFC), logistic regression (LR), and support vector machine (SVM), for distinguishing brain abscess from cystic glioma were compared. The best feature combination and classifier were chosen according to the quantitative metrics including area under the curve (AUC), Youden Index, and accuracy.

Results: In most cases, deep learning-based radiomics (DLR) features, i.e., DTL features combined with HCR features, contributed to a higher accuracy than HCR and DTL features alone for distinguishing brain abscesses from cystic gliomas. The AUC values of the model established, based on the DLR features in T2WI, were 0.86 (95% CI: 0.81, 0.91) in the training cohort and 0.85 (95% CI: 0.75, 0.95) in the test cohort, respectively.

Conclusions: The model established with the DLR features can distinguish brain abscess from cystic glioma efficiently, providing a useful, inexpensive, convenient, and non-invasive method for differential diagnosis. This is the first time that conventional MRI radiomics is applied to identify these diseases. Also, the combination of HCR and DTL features can lead to get impressive performance.

CT-Guided Survival Prediction of Esophageal Cancer

Survival prediction of esophageal cancer is an essential task for doctors to make personalized cancer treatment plans. However, handcrafted features from medical images need prior medical knowledge, which is usually limited and not complete, yielding unsatisfying survival predictions. To address these challenges, we propose a novel and efficient deep learning-based survival prediction framework for evaluating clinical outcomes before concurrent chemoradiotherapy. The proposed model consists of two key components: a 3D Coordinate Attention Convolutional Autoencoder (CACA) and an uncertainty-based jointly Optimizing Cox Model (UOCM). The CACA is built upon an autoencoder structure with 3D coordinate attention layers, capturing latent representations and encoding 3D spatial characteristics with precise positional information. Additionally, we designed an Uncertainty-based jointly Optimizing Cox Model, which jointly optimizes the CACA and survival prediction task. The survival prediction task models the interactions between a patient's feature signatures and clinical outcome to predict a reliable hazard ratio of patients. To verify the effectiveness of our model, we conducted extensive experiments on a dataset including computed tomography of 285 patients with esophageal cancer. Experimental results demonstrated that the proposed method achieved a C-index of 0.72, outperforming the state-of-the-art method.

Augmenting Osteoporosis Imaging with Machine Learning

PURPOSE OF REVIEW

In this paper, we discuss how recent advancements in image processing and machine learning (ML) are shaping a new and exciting era for the osteoporosis imaging field. With this paper, we want to give the reader a basic exposure to the ML concepts that are necessary to build effective solutions for image processing and interpretation, while presenting an overview of the state of the art in the application of machine learning techniques for the assessment of bone structure, osteoporosis diagnosis, fracture detection, and risk prediction.

RECENT FINDINGS

ML effort in the osteoporosis imaging field is largely characterized by "low-cost" bone quality estimation and osteoporosis diagnosis, fracture detection, and risk prediction, but also automatized and standardized large-scale data analysis and data-driven imaging biomarker discovery. Our effort is not intended to be a systematic review, but an opportunity to review key studies in the recent osteoporosis imaging research landscape with the ultimate goal of discussing specific design choices, giving the reader pointers to possible solutions of regression, segmentation, and classification tasks as well as discussing common mistakes.

Clinical-Deep Neural Network and Clinical-Radiomics Nomograms for Predicting the Intraoperative Massive Blood Loss of Pelvic and Sacral Tumors

Background

Patients with pelvic and sacral tumors are prone to massive blood loss (MBL) during surgery, which may endanger their lives.

Purposes

This study aimed to determine the feasibility of using deep neural network (DNN) and radiomics nomogram (RN) based on 3D computed tomography (CT) features and clinical characteristics to predict the intraoperative MBL of pelvic and sacral tumors.

Materials and Methods

This single-center retrospective analysis included 810 patients with pelvic and sacral tumors. 1316 CT and CT enhanced radiomics features were extracted. RN1 and RN2 were constructed by random grouping and time node grouping, respectively. The DNN models were constructed for comparison with RN. Clinical factors associated with the MBL were also evaluated. The area under the receiver operating characteristic curve (AUC) and accuracy (ACC) were used to evaluate different models.

Results

Radscore, tumor type, tumor location, and sex were significant predictors of the MBL of pelvic and sacral tumors (P < 0.05), of which radscore (OR, ranging from 2.109 to 4.706, P < 0.001) was the most important. The clinical-DNN and clinical-RN performed better than DNN and RN. The best-performing clinical-DNN model based on CT features exhibited an AUC of 0.92 and an ACC of 0.97 in the training set, and an AUC of 0.92 and an ACC of 0.75 in the validation set.

Conclusions

The clinical-DNN and clinical-RN had good performance in predicting the MBL of pelvic and sacral tumors, which could be used for clinical decision-making.