Radiomics: extracting more information from medical images using advanced feature analysis

Artificial intelligence-assisted precise preoperative prediction of lateral cervical lymph nodes metastasis in papillary thyroid carcinoma via a clinical-CT radiomic combined model

OBJECTIVES

This study aimed to develop an artificial intelligence-assisted model for the preoperative prediction of lateral cervical lymph node metastasis (LCLNM) in papillary thyroid carcinoma (PTC) using computed tomography (CT) radiomics, providing a new noninvasive and accurate diagnostic tool for PTC patients with LCLNM.

METHODS

This retrospective study included 389 confirmed PTC patients, randomly divided into a training set ( n = 272) and an internal validation set ( n = 117), with an additional 40 patients from another hospital as an external validation set. Patient demographics were evaluated to establish a clinical model. Radiomic features were extracted from preoperative contrast-enhanced CT images (venous phase) for each patient. Feature selection was performed using analysis of variance and the least absolute shrinkage and selection operator algorithm. We employed support vector machine, random forest (RF), logistic regression, and XGBoost algorithms to build CT radiomic models for predicting LCLNM. A radiomics score (Rad-score) was calculated using a radiomic signature-based formula. A combined clinical-radiomic model was then developed. The performance of the combined model was evaluated using the receiver operating characteristic (ROC) curve, calibration curve, and decision curve analysis (DCA).

RESULTS

A total of 1724 radiomic features were extracted from each patient's CT images, with 13 features selected based on nonzero coefficients related to LCLNM. Four clinically relevant factors (age, tumor location, thyroid capsule invasion, and central cervical lymph node metastasis) were significantly associated with LCLNM. Among the algorithms tested, the RF algorithm outperformed the others with five-fold cross-validation on the training set. After integrating the best algorithm with clinical factors, the areas under the ROC curves for the training, internal validation, and external validation sets were 0.910 (95% confidence interval [CI]: 0.729-0.851), 0.876 (95% CI: 0.747-0.911), and 0.821 (95% CI: 0.555-0.802), respectively, with DCA demonstrating the clinical utility of the combined radiomic model.

CONCLUSIONS

This study successfully established a clinical-CT radiomic combined model for predicting LCLNM, which may significantly enhance surgical decision-making for lateral cervical lymph node dissection in patients with PTC.

Evaluating the impact of the Radiomics Quality Score: a systematic review and meta-analysis

OBJECTIVES

Conduct a systematic review and meta-analysis on the application of the Radiomics Quality Score (RQS).

MATERIALS AND METHODS

A search was conducted from January 1, 2022, to December 31, 2023, for systematic reviews which implemented the RQS. Identification of articles prior to 2022 was via a previously published review. Quality scores of individual radiomics papers, their associated criteria scores, and these scores from all readers were extracted. Errors in the application of RQS criteria were noted and corrected. The RQS of radiomics papers were matched with the publication date, imaging modality, and country, where available.

RESULTS

A total of 130 systematic reviews were included, and individual quality scores 117/130 (90.0%), criteria scores 98/130 (75.4%), and multiple reader data 24/130 (18.5%) were extracted. 3258 quality scores were correlated with the radiomics study date of publication. Criteria scoring errors were discovered in 39/98 (39.8%) of articles. Overall mean RQS was 9.4 ± 6.4 (95% CI, 9.1-9.6) (26.1% ± 17.8% (25.3%-26.7%)). Quality scores were positively correlated with publication year (Pearson R = 0.32, p < 0.01) and significantly higher after publication of the RQS (year < 2018, 5.6 ± 6.1 (5.1-6.1); year ≥ 2018, 10.1 ± 6.1 (9.9-10.4); p < 0.01). Only 233/3258 (7.2%) scores were ≥ 50% of the maximum RQS. Quality scores were significantly different across imaging modalities (p < 0.01). Ten criteria were positively correlated with publication year, and one was negatively correlated.

CONCLUSION

Radiomics study adherence to the RQS is increasing with time, although a vast majority of studies are developmental and rarely provide a high level of evidence to justify the clinical translation of proposed models.

KEY POINTS

Question What level of adherence to the Radiomics Quality Score have radiomics studies achieved to date, has it increased with time, and is it sufficient? Findings A meta-analysis of 3258 quality scores extracted from 130 review articles resulted in a mean score of 9.4 ± 6.4. Quality scores were positively correlated with time. Clinical relevance Although quality scores of radiomics studies have increased with time, many studies have not demonstrated sufficient evidence for clinical translation. As new appraisal tools emerge, the current role of the Radiomics Quality Score may change.

Beyond the tumor region: Peritumoral radiomics enhances prognostic accuracy in locally advanced rectal cancer

BACKGROUND

The peritumoral region possesses attributes that promote cancer growth and progression. However, the potential prognostic biomarkers in this region remain relatively underexplored in radiomics.

AIM

To investigate the prognostic value and importance of peritumoral radiomics in locally advanced rectal cancer (LARC).

METHODS

This retrospective study included 409 patients with biopsy-confirmed LARC treated with neoadjuvant chemoradiotherapy and surgically. Patients were divided into training (n = 273) and validation (n = 136) sets. Based on intratumoral and peritumoral radiomic features extracted from pretreatment axial high-resolution small-field-of-view T2-weighted images, multivariate Cox models for progression-free survival (PFS) prediction were developed with or without clinicoradiological features and evaluated with Harrell’s concordance index (C-index), calibration curve, and decision curve analyses. Risk stratification, Kaplan-Meier analysis, and permutation feature importance analysis were performed.

RESULTS

The comprehensive integrated clinical-radiological-omics model (Model_ICRO) integrating seven peritumoral, three intratumoral, and four clinicoradiological features achieved the highest C-indices (0.836 and 0.801 in the training and validation sets, respectively). This model showed robust calibration and better clinical net benefits, effectively distinguished high-risk from low-risk patients (PFS: 97.2% vs 67.6% and 95.4% vs 64.8% in the training and validation sets, respectively; both P < 0.001). Three most influential predictors in the comprehensive Model_ICRO were, in order, a peritumoral, an intratumoral, and a clinicoradiological feature. Notably, the peritumoral model outperformed the intratumoral model (C-index: 0.754 vs 0.670; P = 0.015); peritumoral features significantly enhanced the performance of models based on clinicoradiological or intratumoral features or their combinations.

CONCLUSION

Peritumoral radiomics holds greater prognostic value than intratumoral radiomics for predicting PFS in LARC. The comprehensive model may serve as a reliable tool for better stratification and management postoperatively.

Imaging based artificial intelligence for predicting lymph node metastasis in cervical cancer patients: a systematic review and meta-analysis

Purpose

This meta-analysis was conducted to assess the diagnostic performance of artificial intelligence (AI) based on imaging for detecting lymph node metastasis (LNM) among cervical cancer patients and to compare its performance with that of radiologists.

Methods

A comprehensive literature search was conducted across PubMed, Embase, and Web of Science to identify relevant studies published up to October 2024. The search followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses of Diagnostic Test Accuracy (PRISMA-DTA) guidelines. Studies evaluating the accuracy of AI models in detecting LNM in cervical cancer through computed tomography (CT), magnetic resonance imaging (MRI), and positron emission tomography/computed tomography (PET/CT) were included. Pathology served as the reference standard for validation. A bivariate random-effects model was employed to estimate pooled sensitivity and specificity, both presented alongside 95% confidence intervals (CIs). Bias was assessed with the revised Quality Assessment of Diagnostic Accuracy Studies-2 (QUADAS-2) tool. Study heterogeneity was examined through the I2 statistic. Meta-regression was conducted when significant heterogeneity (I2 > 50%) was observed.

Results

A total of 23 studies were included in this meta-analysis. The quality and bias of the included studies were acceptable. However, substantial heterogeneity was observed among the included studies. Internal validation sets comprised 23 studies and 1,490 patients. The pooled sensitivity, specificity, and the area under the curve (AUC) for detecting LNM in cervical cancer were 0.83 (95% CI: 0.78-0.87), 0.78 (95% CI: 0.74-0.82) and 0.87 (95% CI: 0.84-0.90), respectively. External validation sets comprised six studies and 298 patients. The pooled sensitivity, specificity, and AUC for detecting LNM were 0.70 (95% CI: 0.56-0.81), 0.85 (95% CI: 0.66-0.95) and 0.76 (95% CI: 0.72-0.79), respectively. For radiologists, eight studies and 644 patients were included; the pooled sensitivity, specificity, and AUC for detecting LNM were 0.54 (95% CI: 0.42-0.66), 0.79 (95% CI: 0.59-0.91) and 0.65 (95% CI: 0.60-0.69), respectively.

Conclusions

Imaging-based AI demonstrates higher diagnostic performance than radiologists. Prospective studies with rigorous standardization as well as further research with external validation datasets, are necessary to confirm the results and assess their practical clinical applicability.

Systematic Review Registration

https://www.crd.york.ac.uk/PROSPERO, identifier CRD42024607074.

Preoperative prediction of recurrence risk factors in operable cervical cancer based on clinical-radiomics features

Objective

To investigate the value of preoperative prediction of risk factors for recurrence of operable cervical cancer based on the radiomics features of biparametric magnetic resonance imaging (bp-MRI) combined with clinical features.

Method

A retrospective collection of cervical cancer cases undergoing radical hysterectomy + pelvic and/or para-aortic lymph node dissection at the Affiliated Hospital of North Sichuan Medical College was conducted. Region of interest (ROI) was outlined using the 3D Slicer software, and radiomics after feature extraction and feature screening was performed using the least absolute shrinkage and selection operator (LASSO) algorithm. Logistic regression algorithms were used to construct a fusion clinical-radiomics model to visualize nomograms. Receiver operating characteristic (ROC), DeLong test, calibration curve (CC), and decision curve (DC) were used to evaluate the predictive performance and clinical benefit of the model.

Result

A total of 99 patients with cervical cancer were included in this study, with 79 and 20 cases in the training and test groups, respectively. Seventeen key features were selected for radiomics model construction. Three clinical features were screened to construct a clinical model. A fusion model of the radiomics model combined with the clinical model was constructed. The area under the curve (AUC) values in the training group were 0.710 (95% CI 0.602-0.819), 0.892 (95% CI 0.826-0.958), and 0.906 (95% CI 0.842-0.970), for the comparative clinical model, radiomics model, and fusion model, respectively, and the AUC values in the testing group were 0.620 (95% CI 0.366-0.874), 0.860 (95% CI 0.677-1.000), and 0.880 (95% CI 0.690-1.000), respectively. The DeLong test showed a statistically significant difference between the AUC values of the fusion model and the clinical model (p < 0.05). Decision curve analysis (DCA) showed that the fusion model had the greatest net benefit when the threshold probability was approximately 0.5.

Conclusion

The fusion model constructed based on bp-MRI radiomics features combined with clinical features provides an important reference for predicting the risk status of recurrence in operable cervical cancer. The findings of this study are preliminary exploratory results, and further large-scale, multicenter studies are needed to validate these findings.

Predicting the efficacy of neoadjuvant chemotherapy in breast cancer patients based on ultrasound longitudinal temporal depth network fusion model

OBJECTIVE

The aim of this study was to develop and validate a deep learning radiomics (DLR) model based on longitudinal ultrasound data and clinical features to predict pathologic complete response (pCR) after neoadjuvant chemotherapy (NAC) in breast cancer patients.

METHODS

Between January 2018 and June 2023, 312 patients with histologically confirmed breast cancer were enrolled and randomly assigned to a training cohort (n = 219) and a test cohort (n = 93) in a 7:3 ratio. Next, pre-NAC and post-treatment 2-cycle ultrasound images were collected, and radiomics and deep learning features were extracted from NAC pre-treatment (Pre), post-treatment 2 cycle (Post), and Delta (pre-NAC-NAC 2 cycle) images. In the training cohort, to filter features, the intraclass correlation coefficient test, the Boruta algorithm, and the least absolute shrinkage and selection operator (LASSO) logistic regression were used. Single-modality models (Pre, Post, and Delta) were constructed based on five machine-learning classifiers. Finally, based on the classifier with the optimal predictive performance, the DLR model was constructed by combining Pre, Post, and Delta ultrasound features and was subsequently combined with clinical features to develop a combined model (Integrated). The discriminative power, predictive performance, and clinical utility of the models were further evaluated in the test cohort. Furthermore, patients were assigned into three subgroups, including the HR+/HER2-, HER2+, and TNBC subgroups, according to molecular typing to validate the predictability of the model across the different subgroups.

RESULTS

After feature screening, 16, 13, and 10 features were selected to construct the Pre model, Post model, and Delta model based on the five machine learning classifiers, respectively. The three single-modality models based on the XGBoost classifier displayed optimal predictive performance. Meanwhile, the DLR model (AUC of 0.827) was superior to the single-modality model (Pre, Post, and Delta AUCs of 0.726, 0.776, and 0.710, respectively) in terms of prediction performance. Moreover, multivariate logistic regression analysis identified Her-2 status and histological grade as independent risk factors for NAC response in breast cancer. In both the training and test cohorts, the Integrated model, which included Pre, Post, and Delta ultrasound features and clinical features, exhibited the highest predictive ability, with AUC values of 0.924 and 0.875, respectively. Likewise, the Integrated model displayed the highest predictive performance across the different subgroups.

CONCLUSION

The Integrated model, which incorporated pre-NAC treatment and early treatment ultrasound data and clinical features, accurately predicted pCR after NAC in breast cancer patients and provided valuable insights for personalized treatment strategies, allowing for timely adjustment of chemotherapy regimens.

Axillary lymph node metastasis in breast cancer: from historical axillary surgery to updated advances in the preoperative diagnosis and axillary management

Axillary lymph node status, which was routinely assessed by axillary lymph node dissection (ALND) until the 1990s, is a crucial factor in determining the stage, prognosis, and therapeutic strategy used for breast cancer patients. Axillary surgery for breast cancer patients has evolved from ALND to minimally invasive approaches. Over the decades, the application of noninvasive imaging techniques, machine learning approaches and emerging clinical prediction models for the detection of axillary lymph node metastasis greatly improves clinical diagnostic efficacy and provides optimal surgical selection. In this work, we summarize the historical axillary surgery and updated perspectives of axillary management for breast cancer patients.

Towards automatical tumor segmentation in radiomics: a comparative analysis of various methods and radiologists for both region extraction and downstream diagnosis

OBJECTIVE

By discussing the difference, stability and classification ability of tumor contour extracted by artificial intelligence and doctors, can a more stable method of tumor contour extraction be obtained?

METHODS

We propose a novel framework for the automatic segmentation of lung tumor contours and the differential diagnosis of downstream tasks. This framework integrates four key modules: tumor segmentation, extraction of radiomic features, feature selection, and the development of diagnostic models for clinical applications. Using this framework, we conducted a study involving a cohort of 1,429 patients suspected of lung cancer. Four automatic segmentation methods (RNN, UNET, WFCM, and SNAKE) were evaluated against manual segmentation performed by three radiologists with varying levels of expertise. We further studied the consistency of radiomic features extracted from these methods and evaluates their diagnostic performance across three downstream tasks: benign vs. malignant classification, lung adenocarcinoma infiltration, and lung nodule density classification.

RESULTS

The Dice coefficient of RNN is the highest among the four automatic segmentation methods (0.803 > 0.751, 0.576, 0.560), and all P < 0.05. In the consistency comparison of the seven contour-extracted radiomic features, that the features extracted by RNN and S1 (the senior radiologist) showed the highest similarity which was higher than the other automatic segmentation methods and doctors with low seniority. In all three downstream tasks, the radiomic features extracted from RNN segmentation contours showed the highest diagnostic discrimination. In the classification of benign and malignant nodules, the RNN method performed slightly better than the S1 method, with an AUC of 0.840 ± 0.01 and 0.824 ± 0.015, respectively, and significantly better than the other five methods. Similarly, the RNN method had an AUC value of 0.946 in lung adenocarcinoma infiltration, and a kappa value of 0.729 in lung nodule density classification, both of which were better than the other six methods.

CONCLUSIONS

Our findings suggest that AI-driven tumor segmentation methods can enhance clinical decision-making by providing reliable and reproducible results, ultimately emphasizing the auxiliary role of automated tumor contouring in clinical practice. The findings will have important implications for the application of radiomics in clinical practice.

Predicting Microvascular Invasion and Early Recurrence in Hepatocellular Carcinoma Using DeepLab V3+ Segmentation of Multiregional MR Habitat Images

RATIONALE AND OBJECTIVES

Accurate identification of microvascular invasion (MVI) in hepatocellular carcinoma (HCC) is crucial for treatment and prognosis. Single-modality and feature fusion models using manual segmentation fail to provide insights into MVI. This study aims to develop a DeepLab V3+ model for automated segmentation of HCC magnetic resonance (MR) images and a decision fusion model to predict MVI and early recurrence (ER).

MATERIALS AND METHODS

This retrospective study included 209 HCC patients (146 in the training and 63 in the test cohorts). The performance of DeepLab V3+ for HCC MR image segmentation was evaluated using Dice Loss and F1 score. Intraclass correlation coefficients (ICCs) assessed feature extraction reliability. Spearman's correlation analyzed the relationship between tumor volumes from automated and manual segmentation, with agreement evaluated using Bland-Altman plots. Model performance was assessed using the area under the receiver operating characteristic curve (ROC AUC), calibration curves, and decision curve analysis. A nomogram predicted ER of HCC after surgery, with Kaplan-Meier analysis for 2-year recurrence-free survival (RFS).

RESULTS

The DeepLab V3+ model demonstrated high segmentation accuracy, with strong agreement in feature extraction (ICC: 0.802-0.999). The decision fusion model achieved AUCs of 0.968 and 0.878 for MVI prediction, and the nomogram for predicting ER yielded AUCs of 0.782 and 0.690 in the training and test cohorts, respectively, with significant RFS differences between the risk groups.

CONCLUSION

The DeepLab V3+ model accurately segmented HCC. The decision fusion model significantly improved MVI prediction, and the nomogram offered valuable insights into recurrence risk for clinical decision-making.

AVAILABILITY OF DATA AND MATERIALS

The datasets used and/or analyzed during the current study are available from the corresponding author upon reasonable request.

RAE-Net: a multi-modal neural network based on feature fusion and evidential deep learning algorithm in predicting breast cancer subtypes on DCE-MRI

ObjectivesAccurate identification of molecular subtypes in breast cancer is critical for personalized treatment. This study introduces a novel neural network model, RAE-Net, based on Multimodal Feature Fusion (MFF) and the Evidential Deep Learning Algorithm (EDLA) to improve breast cancer subtype prediction using dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI).MethodsA dataset of 344 patients with histologically confirmed breast cancer was divided into training (n = 200), validation (n = 60), and testing (n = 62) cohorts. RAE-Net, built on ResNet-50 with Multi-Head Attention (MHA) fusion and Multi-Layer Perceptron (MLP) mechanisms, combines radiomic and deep learning features for subtype prediction. The EDLA module adds uncertainty estimation to enhance classification reliability.ResultsThe RAE-Net model incorporating the MFF module demonstrated superior performance, achieving a mean accuracy of 0.83 and a Macro-F1 score of 0.78, surpassing traditional radiomics models (accuracy: 0.79, Macro-F1: 0.75) and standalone deep learning models (accuracy: 0.80, Macro-F1: 0.76). When an EDLA uncertainty threshold of 0.2 was applied, the performance significantly improved, with accuracy reaching 0.97 and Macro-F1 increasing to 0.92. Additionally, RAE-Net outperformed two recent deep learning networks, ResGANet and HIFUSE. Specifically, RAE-Net showed a 0.5% improvement in accuracy and a higher AUC compared to ResGANet. In comparison to HIFUSE, RAE-Net reduced both the number of parameters and computational cost by 90% while only increasing computation time by 5.7%.ConclusionsRAE-Net integrates feature fusion and uncertainty estimation to predict breast cancer subtypes from DCE-MRI. The model achieves high accuracy while maintaining computational efficiency, demonstrating its potential for clinical use as a reliable and resource-efficient diagnostic tool.

Preoperative prediction of the Lauren classification in gastric cancer using automated nnU-Net and radiomics: a multicenter study

OBJECTIVES

To develop and validate a deep learning model based on nnU-Net combined with radiomics to achieve autosegmentation of gastric cancer (GC) and preoperative prediction via the Lauren classification.

METHODS

Patients with a pathological diagnosis of GC were retrospectively enrolled in three medical centers. The nnU-Net autosegmentation model was developed using manually segmented datasets and evaluated by the Dice similarity coefficient (DSC). The CT images were processed by the nnU-Net model to obtain autosegmentation results and extract radiomic features. The least absolute shrinkage and selection operator (LASSO) method selects optimal features for calculating the Radscore and constructing a radiomic model. Clinical characteristics and the Radscore were integrated to construct a combined model. Model performance was evaluated via the receiver operating characteristic (ROC) curve.

RESULTS

A total of 433 GC patients were divided into the training set, internal validation set, external test set-1, and external test set-2. The nnU-Net model achieved a DSC of 0.79 in the test set. The areas under the curve (AUCs) of the internal validation set, external test set-1, and external test set-2 were 0.84, 0.83, and 0.81, respectively, for the radiomic model; and 0.81, 0.81, and 0.82, respectively, for the combined model. The AUCs of the radiomic and combined models showed no statistically significant difference (p > 0.05). The radiomic model was selected as the optimal model.

CONCLUSIONS

The nnU-Net model can efficiently and accurately achieve automatic segmentation of GCs. The radiomic model can preoperatively predict the Lauren classification of GC with high accuracy.

CRITICAL RELEVANCE STATEMENT

This study highlights the potential of nnU-Net combined with radiomics to noninvasively predict the Lauren classification in gastric cancer patients, enhancing personalized treatment strategies and improving patient management.

KEY POINTS

The Lauren classification influences gastric cancer treatment and prognosis. The nnU-Net model reduces doctors' manual segmentation errors and workload. Radiomics models aid in preoperative Lauren classification prediction for patients with gastric cancer.

Predictive classification of lung cancer pathological based on PET/CT radiomics

OBJECTIVES

To develop and validate a combined clinical and radiomics model for non-invasive prediction of lung cancer (LC) pathological types (lung adenocarcinoma, lung squamous cell carcinoma, and small cell lung cancer) based on patients' pre-treatment FDG PET/CT images and clinical data, as a complementary tool to aid in the diagnosis of LC pathological histological classification.

METHODS

In total, 896 patients with pathological confirmation of lung cancer were part of this retrospective study. The training and test groups included 819 patients who underwent scanning using scanner 1. The independent validation group included 77 patients who using scanner 2. The optimal features were retained by least absolute shrinkage and selection operator algorithm dimensionality reduction screening of the collected radiomics features, clinical parameters, and PET metabolic parameters. Five models were established to predict the lung cancer pathological types by the k-nearest neighbor classification (KNN) algorithm. The performance of the prediction model was assessed by calculating the area under the curve (AUC) from the receiver operator characteristic curve (ROC).

RESULTS

Of all five predictive models (the PET-only radiomics model, the CT-only radiomics model, the PET/CT radiomics model, the clinical-only model and the combined clinical and PET/CT radiomics model), the clinical combined PET/CT radiomics model exhibited best performance. The macro-AUC for the training, test and independent validation groups were 0.974, 0.931, 0.960, the micro-AUC were 0.976, 0.940, 0.970, and the accuracy were 0.963, 0.914, and 0.961, respectively.

CONCLUSIONS

Our model combined radiomics and clinical data and showed higher performance in non-invasively predicting the LC pathological types, which suggesting that PET/CT radiomics may be a promising technique for predicting LC histopathology.

Decoding breast cancer imaging trends: the role of AI and radiomics through bibliometric insights

BACKGROUND

Radiomics and AI have been widely used in breast cancer imaging, but a comprehensive systematic analysis is lacking. Therefore, this study aims to conduct a bibliometrics analysis in this field to discuss its research status and frontier hotspots and provide a reference for subsequent research.

METHODS

Publications related to AI, radiomics, and breast cancer imaging were searched in the Web of Science Core Collection. CiteSpace plotted the relevant co-occurrence network according to authors and keywords. VOSviewer and Pajek were used to draw relevant co-occurrence maps according to country and institution. In addition, R was used to conduct bibliometric analysis of relevant authors, countries/regions, journals, keywords, and annual publications and citations based on the collected information.

RESULTS

A total of 2,701 Web of Science Core Collection publications were retrieved, including 2,486 articles (92.04%) and 215 reviews (7.96%). The number of publications increased rapidly after 2018. The United States of America (n = 17,762) leads in citations, while China (n = 902) leads in the number of publications. Sun Yat-sen University (n = 75) had the largest number of publications. Bin Zheng (n = 28) was the most published author. Nico Karssemeijer (n = 72.1429) was the author with the highest average citations. "Frontiers in Oncology" was the journal with the most publications, and "Radiology" had the highest IF. The keywords with the most frequent occurrence were "breast cancer", "deep learning", and "classification". The topic trends in recent years were "explainable AI", "neoadjuvant chemotherapy", and "lymphovascular invasion".

CONCLUSION

The application of radiomics and AI in breast cancer imaging has received extensive attention. Future research hotspots may mainly focus on the progress of explainable AI in the technical field and the prediction of lymphovascular invasion and neoadjuvant chemotherapy efficacy in clinical application.

Preoperative clinical radiomics model based on deep learning in prognostic assessment of patients with gallbladder carcinoma

OBJECTIVE

We aimed to develop a preoperative clinical radiomics survival prediction model based on the radiomics features via deep learning to provide a reference basis for preoperative assessment and treatment decisions for patients with gallbladder carcinoma (GBC).

METHODS

A total of 168 GBC patients who underwent preoperative upper abdominal enhanced CT from one high-volume medical center between January 2011 to December 2020 were retrospectively analyzed. The region of interest (ROI) was manually outlined by two physicians using 3D Slicer software to establish a nnU-Net model. The DeepSurv survival prediction model was developed by combining radiomics features and preoperative clinical variables.

RESULTS

A total of 1502 radiomics features were extracted from the ROI results based on the nnU-Net model and manual segmentation, and 13 radiomics features were obtained through the 4-step dimensionality reduction methods, respectively. The C-index and AUC of 1-, 2-, and 3-year survival prediction for the nnU-Net based clinical radiomics DeepSurv model was higher than clinical and nnU-Net based radiomics DeepSurv models in the training and testing sets, and close to manual based clinical radiomics DeepSurv model. Delong-test was performed on the AUC of 1-, 2-, and 3-year survival prediction for the two preoperative clinical radiomics DeepSurv prediction models in the testing set, and the results showed that the two models had the same prediction efficiency (all P > 0.05).

CONCLUSIONS

By using the DeepSurv model via nnU-Net segmentation, postoperative survival outcomes for individual gallbladder carcinoma patients could be assessed and stratified, which can provide references for preoperative diagnosis and treatment decisions.

Predicting PD-L1 in Lung Adenocarcinoma Using 18F-FDG PET/CT Radiomic Features

Background/Objectives: This study aims to retrospectively analyze the clinical and imaging data of 101 patients with lung adenocarcinoma who underwent [18F]FDG PET/CT examination and were pathologically confirmed in the Department of Nuclear Medicine at Peking University Cancer Hospital. This study explores the predictive value and important features of [18F]FDG PET/CT radiomics for PD-L1 expression levels in lung adenocarcinoma patients, assisting in screening patients who may benefit from immunotherapy. Methods: 101 patients with histologically confirmed lung adenocarcinoma who received pre-treatment [18F] FDG PET/CT were included. Among them, 44 patients were determined to be PD-L1 positive and 57 patients were determined to be PD-L1 negative based on immunohistochemical assays. Clinical data, PET/CT radiomics parameters, conventional metabolic parameters, and observed CT characteristics were included in the modeling. Random Forest was used in feature denoising, while Forward Stepwise Regression and the Least Absolute Shrinkage and Selection Operator were used in feature selection. Models based on Tree, Discriminant, Logistic Regression, and Support Vector Machine were trained and evaluatedto explore the value of clinical data, PET/CT radiomics parameters, conventional metabolic parameters, and observed CT characteristics. Results: All models showed some predictive ability in distinguishing PD-L1 positive from PD-L1 negative samples. Among the multimodal imaging, clinical data were incorporated into the models, with clinical stage and gender selected by Forward Stepwise Regression, while clinical stage, smoking history, and gender were selected by LASSO. When incorporating clinical data and thin-section CT-derived images into the models, nodular type, spiculation, and CT Shape Flatness were selected by Forward Stepwise Regression, while nodular type and spiculation were selected by LASSO. When incorporating clinical data, PET/CT radiomics, observed CT characteristics, and conventional metabolic information. Forward Stepwise Regression selected TLGlean, MTV, nodule component, PET Shape Sphericity, while LASSO selected SULmax, MTV, nodular type, PET Shape Sphericity, and spiculation. Conclusions: The integration of clinical data, PET/CT radiomics, and conventional metabolic parameters effectively predicted PD-L1 expression, thereby assisting the selection of patients who would benefit from immunotherapy. Observed CT characteristics and conventional metabolic information play an important role in predicting PD-L1 expression levels.

Radiomics in Dermatological Optical Coherence Tomography (OCT): Feature Repeatability, Reproducibility, and Integration into Diagnostic Models in a Prospective Study

BACKGROUND/OBJECTIVES

Radiomics has seen substantial growth in medical imaging; however, its potential in optical coherence tomography (OCT) has not been widely explored. We systematically evaluate the repeatability and reproducibility of handcrafted radiomics features (HRFs) from OCT scans of benign nevi and examine the impact of bin width (BW) selection on HRF stability. The effect of using stable features on a radiomics classification model was also assessed.

METHODS

In this prospective study, 20 volunteers underwent test-retest OCT imaging of 40 benign nevi, resulting in 80 scans. The repeatability and reproducibility of HRFs extracted from manually delineated regions of interest (ROIs) were assessed using concordance correlation coefficients (CCCs) across BWs ranging from 5 to 50. A unique set of stable HRFs was identified at each BW after removing highly correlated features to eliminate redundancy. These robust features were incorporated into a multiclass radiomics classifier trained to distinguish benign nevi, basal cell carcinoma (BCC), and Bowen's disease.

RESULTS

Six stable HRFs were identified across all BWs, with a BW of 25 emerging as the optimal choice, balancing repeatability and the ability to capture meaningful textural details. Additionally, intermediate BWs (20-25) yielded 53 reproducible features. A classifier trained with six stable features achieved a 90% accuracy and AUCs of 0.96 and 0.94 for BCC and Bowen's disease, respectively, compared to a 76% accuracy and AUCs of 0.86 and 0.80 for a conventional feature selection approach.

CONCLUSIONS

This study highlights the critical role of BW selection in enhancing HRF stability and provides a methodological framework for optimizing preprocessing in OCT radiomics. By demonstrating the integration of stable HRFs into diagnostic models, we establish OCT radiomics as a promising tool to aid non-invasive diagnosis in dermatology.

Prognostic and Predictive Value of Systemic Inflammatory Markers in Epithelial Ovarian Cancer

Background and Objectives: Epithelial ovarian cancer (EOC) remains a significant global health challenge. While traditional prognostic factors are well established, emerging biomarkers continue to gain attention. Materials and Methods: This retrospective study evaluated the impact of systemic inflammatory markers on progression-free survival (PFS) and overall survival (OS) in 154 EOC patients. Pre-treatment neutrophil/lymphocyte ratio (NLR), platelet/lymphocyte ratio (PLR), and systemic inflammatory index (SII) were calculated and categorized into low and high groups. Univariate and multivariate analyses were conducted to identify independent prognostic factors, while logistic regression analysis was used to determine predictors of platinum resistance. Results: In the univariate analysis, elevated NLR and PLR were associated with poorer PFS and OS. However, these markers did not maintain statistical significance in the multivariate analysis. Although SII demonstrated a trend toward worse outcomes, it did not reach statistical significance. Histopathological type, PLR, and surgical approach were identified as independent predictors of platinum resistance. Conclusions: Our findings indicate that systemic inflammatory markers may hold prognostic value in EOC; however, further validation through larger prospective studies is necessary.

CBCT image based radiomic analysis for condylar resorption after orthognathic surgery

OBJECTIVES

To select and discover features which are highly correlated with condylar resorption (CR) after orthognathic surgery (OS) by a new method.

MATERIALS AND METHODS

The cone beam computed tomography (CBCT) images were collected from orthognathic patients taken at T1(baseline time point) and T2(within 3-36 months postoperatively). The subjects were divided into the CR group and the control group by using a 2-level scale after registering T1 and T2 images. The condyles were segmented by MVEL-Net network model and were analyzed by Pyradiomics. The D-value equaled to the absolute difference-value (D-value) between the feature values of T1 and T2. The correlation between features with statistically significant differences in T1 and D-values would be analyzed to find the specific features related to CR after OS.

RESULTS

A total of 145 subjects were included (27 males and 118 females), which had 44 subjects in the CR group, 101 subjects in the control group. For all samples, a total of 82 features were extracted (F1), which were with statistical differences at T1 time point between CR and control groups. By using the D-value, the number of features was reduced to 32 features (Fd). Among Fd, only 3 were not included in F1.

CONCLUSIONS

The D-value was proposed for selecting specific features of CR after OS and it can be observed that the D-value serves the purpose of feature specification compared to T1 values. By using the D-values, several features were found to change significantly during the process of CR after OS.

CLINICAL RELEVANCE

The features selected by D-value can be used for the establishment of a prediction model for CR after OS.

Delta Radiomics and Tumor Size: A New Predictive Radiomics Model for Chemotherapy Response in Liver Metastases from Breast and Colorectal Cancer

Background/Objectives: Radiomic features exhibit a correlation with tumor size on pretreatment images. However, on post-treatment images, this association is influenced by treatment efficacy and varies between responders and non-responders. This study introduces a novel model, called baseline-referenced Delta radiomics, which integrates the association between radiomic features and tumor size into Delta radiomics to predict chemotherapy response in liver metastases from breast cancer (BC) and colorectal cancer (CRC). Materials and Methods: A retrospective study analyzed contrast-enhanced computed tomography (CT) scans of 83 BC patients and 84 CRC patients. Among these, 57 BC patients with 106 liver lesions and 37 CRC patients with 109 lesions underwent post-treatment imaging after systemic chemotherapy. Radiomic features were extracted from up to three lesions per patient following manual segmentation. Tumor response was assessed by measuring the longest diameter and classified according to RECIST 1.1 criteria as progressive disease (PD), partial response (PR), or stable disease (SD). Classification models were developed to predict chemotherapy response using pretreatment data only, Delta radiomics, and baseline-referenced Delta radiomics. Model performance was evaluated using confusion matrix metrics. Results: Baseline-referenced Delta radiomics performed comparably or better than established radiomics models in predicting tumor response in chemotherapy-treated patients with liver metastases. The sensitivity, specificity, and balanced accuracy in predicting response ranged from 0.66 to 0.97, 0.81 to 0.97, and 80% to 90%, respectively. Conclusions: By integrating the relationship between radiomic features and tumor size into Delta radiomics, baseline-referenced Delta radiomics offers a promising approach for predicting chemotherapy response in liver metastases from breast and colorectal cancer.

CT-Based Machine Learning Radiomics Analysis to Diagnose Dysthyroid Optic Neuropathy

PURPOSE

To develop CT-based machine learning radiomics models used for the diagnosis of dysthyroid optic neuropathy (DON).

MATERIALS AND METHODS

This is a retrospective study included 57 patients (114 orbits) diagnosed with thyroid-associated ophthalmopathy (TAO) at the Beijing Tongren Hospital between December 2019 and June 2023. CT scans, medical history, examination results, and clinical data of the participants were collected. DON was diagnosed based on clinical manifestations and examinations. The DON orbits and non-DON orbits were then divided into a training set and a test set at a ratio of approximately 7:3. The 3D slicer software was used to identify the volumes of interest (VOI). Radiomics features were extracted using the Pyradiomics and selected by t-test and least absolute shrinkage and selection operator (LASSO) regression algorithm with 10-fold cross-validation. Machine-learning models, including random forest (RF) model, support vector machine (SVM) model, and logistic regression (LR) model were built and validated by receiver operating characteristic (ROC) curves, area under the curves (AUC) and confusion matrix-related data. The net benefit of the models is shown by the decision curve analysis (DCA).

RESULTS

We extracted 107 features from the imaging data, representing various image information of the optic nerve and surrounding orbital tissues. Using the LASSO method, we identified the five most informative features. The AUC ranged from 0.77 to 0.80 in the training set and the AUC of the RF, SVM and LR models based on the features were 0.86, 0.80 and 0.83 in the test set, respectively. The DeLong test showed there was no significant difference between the three models (RF model vs SVM model: p = .92; RF model vs LR model: p = .94; SVM model vs LR model: p = .98) and the models showed optimal clinical efficacy in DCA.

CONCLUSIONS

The CT-based machine learning radiomics analysis exhibited excellent ability to diagnose DON and may enhance diagnostic convenience.

Radiomics Integration of Mammography and DCE-MRI for Predicting Molecular Subtypes in Breast Cancer Patients

Background

Accurate identification of the molecular subtypes of breast cancer is essential for effective treatment selection and prognosis prediction.

Aim

This study aimed to evaluate the diagnostic performance of a radiomics model, which integrates breast mammography and dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) in predicting the molecular subtypes of breast cancer.

Methods

We retrospectively included 462 female patients with pathologically confirmed breast cancer, including 53 cases of triple-negative, 94 cases of HER2 overexpression, 95 cases of luminal A, and 215 cases of luminal B breast cancer. Radiomics analysis was performed using FAE software, wherein the radiomic features were examined about the hormone receptor status. The performance of the model was evaluated using the area under the receiver operating characteristic curve (AUC) and accuracy.

Results

In multivariate analysis, radiomic features were the only independent predictive factors for molecular subtypes. The model that incorporates multimodal fusion features from breast mammography and DCE-MRI images exhibited superior overall performance compared to using either modality independently. The AUC values (or accuracies) for six pairings were as follows: 0.648 (0.627) for luminal A vs luminal B, 0.819 (0.793) for luminal A vs HER2 overexpression, 0.725 (0.696) for luminal A vs triple-negative subtype, 0.644 (0.560) for luminal B vs HER2 overexpression, 0.625 (0.636) for luminal B vs triple-negative subtype, and 0.598 (0.500) for triple-negative subtype vs HER2 overexpression.

Conclusion

The radionics model utilizing multimodal fusion features from breast mammography combined with DCE-MRI images showed high performance in distinguishing molecular subtypes of breast cancer. It is of significance to accurately predict prognosis and determine treatment strategy of breast cancer by molecular classification.

Artificial intelligence-enhanced retinal imaging as a biomarker for systemic diseases

Retinal images provide a non-invasive and accessible means to directly visualize human blood vessels and nerve fibers. Growing studies have investigated the intricate microvascular and neural circuitry within the retina, its interactions with other systemic vascular and nervous systems, and the link between retinal biomarkers and various systemic diseases. Using the eye to study systemic health, based on these connections, has been given a term as oculomics. Advancements in artificial intelligence (AI) technologies, particularly deep learning, have further increased the potential impact of this study. Leveraging these technologies, retinal analysis has demonstrated potentials in detecting numerous diseases, including cardiovascular diseases, central nervous system diseases, chronic kidney diseases, metabolic diseases, endocrine disorders, and hepatobiliary diseases. AI-based retinal imaging, which incorporates established modalities such as digital color fundus photographs, optical coherence tomography (OCT) and OCT angiography, as well as emerging technologies like ultra-wide field imaging, shows great promises in predicting systemic diseases. This provides a valuable opportunity for systemic diseases screening, early detection, prediction, risk stratification, and personalized prognostication. As the AI and big data research field grows, with the mission of transforming healthcare, they also face numerous challenges and limitations both in data and technology. The application of natural language processing framework, large language model, and other generative AI techniques presents both opportunities and concerns that require careful consideration. In this review, we not only summarize key studies on AI-enhanced retinal imaging for predicting systemic diseases but also underscore the significance of these advancements in transforming healthcare. By highlighting the remarkable progress made thus far, we provide a comprehensive overview of state-of-the-art techniques and explore the opportunities and challenges in this rapidly evolving field. This review aims to serve as a valuable resource for researchers and clinicians, guiding future studies and fostering the integration of AI in clinical practice.

Integrating ultrasound radiomics and clinicopathological features for machine learning-based survival prediction in patients with nonmetastatic triple-negative breast cancer

OBJECTIVE

This study aimed to evaluate the predictive value of implementing machine learning models based on ultrasound radiomics and clinicopathological features in the survival analysis of triple-negative breast cancer (TNBC) patients.

METHODS AND MATERIALS

All patients, including retrospective cohort (training cohort, n = 306; internal validation cohort, n = 77) and prospective external validation cohort (n = 82), were diagnosed as locoregional TNBC and underwent pre-intervention sonographic evaluation in this multi-center study. A thorough chart review was conducted for each patient to collect clinicopathological and sonographic features, and ultrasound radiomics features were obtained by PyRadiomics. Deep learning algorithms were utilized to delineate ROIs on ultrasound images. Radiomics analysis pipeline modules were developed for analyzing features. Radiomic scores, clinical scores, and combined nomograms were analyzed to predict 2-year, 3-year, and 5-year overall survival (OS) and disease-free survival (DFS). Receiver operating characteristic (ROC) curves, calibration curves, and decision curves were used to evaluate the prediction performance.

FINDINGS

Both clinical and radiomic scores showed good performance for overall survival and disease-free survival prediction in internal (median AUC of 0.82 and 0.72 respectively) and external validation (median AUC of 0.70 and 0.74 respectively). The combined nomograms had AUCs of 0.80-0.93 and 0.73-0.89 in the internal and external validation, which had best predictive performance in all tasks (p < 0.05), especially for 5-year OS (p < 0.01). For the overall evaluation of six tasks, combined models obtained better performance than clinical and radiomic scores [AUCs of 0.83 (0.73,0.93), 0.81 (0.72,0.93), and 0.70 (0.61,0.85) respectively].

INTERPRETATION

The combined nomograms based on pre-intervention ultrasound radiomics and clinicopathological features demonstrated exemplary performance in survival analysis. The new models may allow us to non-invasively classify TNBC patients with various disease outcome.

MRI in Oral Tongue Squamous Cell Carcinoma: A Radiomic Approach in the Local Recurrence Evaluation

(1) Background: Oral tongue squamous cell carcinoma (OTSCC) is a prevalent malignancy with high loco-regional recurrence. Advanced imaging biomarkers are critical for stratifying patients at a high risk of recurrence. This study aimed to develop MRI-based radiomic models to predict loco-regional recurrence in OTSCC patients undergoing surgery. (2) Methods: We retrospectively selected 92 patients with OTSCC who underwent MRI, followed by surgery and cervical lymphadenectomy. A total of 31 patients suffered from a loco-regional recurrence. Radiomic features were extracted from preoperative post-contrast high-resolution MRI and integrated with clinical and pathological data to develop predictive models, including radiomic-only and combined radiomic-clinical approaches, trained and validated with stratified data splitting. (3) Results: Textural features, such as those derived from the Gray-Level Size-Zone Matrix, Gray-Level Dependence Matrix, and Gray-Level Run-Length Matrix, showed significant associations with recurrence. The radiomic-only model achieved an accuracy of 0.79 (95% confidence interval: 0.69, 0.87) and 0.74 (95% CI: 0.54, 0.89) in the training and validation set, respectively. Combined radiomic and clinical models, incorporating features like the pathological depth of invasion and lymph node status, provided comparable diagnostic performances. (4) Conclusions: MRI-based radiomic models demonstrated the potential for predicting loco-regional recurrence, highlighting their increasingly important role in advancing precision oncology for OTSCC.

Tumour surface regularity predicts survival and benefit from gross total resection in IDH-wildtype glioblastoma patients

OBJECTIVES

To evaluate the ability of sphericity in glioblastomas (GBMs) for predicting overall survival (OS) and the survival benefit from gross tumour resection (GTR).

METHODS

Preoperative MRI scans were retrospectively analysed in IDH-wildtype GBM patients from two datasets. After MRI preprocessing and tumour segmentation, tumour sphericity was calculated based on the tumour core region. The prognostic value of tumour surface regularity was evaluated via Kaplan-Meier (K-M) plots, univariate and multivariate Cox proportional hazards analyses. In different surface regularity subgroups, the OS benefit from GTR was evaluated via K-M plots and the restricted mean survival time (RMST).

RESULTS

This study included 367 patients (median age, 62.0 years [IQR, 54.5-70.5 years]) in the discovery cohort and 475 patients (median age, 63.6 years [IQR, 56.2-71.3 years]) in the validation cohort. Sphericity was an independent predictor of OS in the discovery (p = 0.022, hazard ratio (HR) = 1.45, 95% confidence interval (CI) 1.06-1.99) and validation groups (p = 0.007, HR = 1.38, 95% CI: 1.09-1.74) according to multivariate analysis. Age, extent of resection, and surface regularity composed a prognostic model that separated patients into subgroups with distinct prognoses. Patients in the surface-irregular subgroup benefited from GTR, but patients in the surface-regular subgroup did not in the discovery (p < 0.001 vs. p = 0.056) and validation datasets (p < 0.001 vs. p = 0.11).

CONCLUSIONS

The high surface regularity of IDH-wildtype GBM is significantly correlated with better OS and does not benefit substantially from GTR.

CRITICAL RELEVANCE STATEMENT

The proposed imaging marker has the potential to increase the survival prediction efficacy for IDH-wildtype glioblastomas (GBMs), offering a valuable indicator for clinical decision-making.

KEY POINTS

Sphericity is an independent prognostic factor in IDH-wildtype glioblastomas (GBMs). High sphericity in IDH-wildtype GBM is significantly correlated with better survival. GBM patients with low sphericity could receive survival benefits from gross tumour resection.

Prediction of right recurrent laryngeal nerve lymph node metastasis in esophageal cancer based on computed tomography imaging histology

Purpose

This study aimed to identify risk factors for right recurrent laryngeal nerve lymph node (RRLNLN) metastasis using computed tomography (CT) imaging histology and clinical data from patients with esophageal squamous cell carcinoma (ESCC), ultimately developing a clinical prediction model.

Methods

Data were collected from 370 patients who underwent surgical resection at the Department of Thoracic Surgery, Zhongshan Hospital of Xiamen University, from December 2014 to December 2020. Subsequently, the venous-stage chest-enhanced CT images of the patients were imported into 3DSlicer 4.11 software, allowing for the extraction of imaging histological features. Additionally, by combining the clinical data of the patients, single- and multifactor analyses were conducted to screen the risk factors and build a predictive model in the form of a nomogram. The area under the curve (AUC) was used as a discriminant for model accuracy, while differentiation and calibration methods were applied to further evaluate the model's accuracy. Finally, the Bootstrap resampling method was employed to repeat sampling 2,000 times to draw calibration curves, while the K-fold crossvalidation method was used for the internal validation of the prediction model.

Results

The RRLNLN lymph node metastasis rate was 17.3%. Four significant factors-Maximum2DDiameterSlice, Mean, Imc1, and Dependence Entropy-were identified. Alignment diagrams were subsequently constructed, yielding an AUC of 0.938 and a C-index of 0.904 during internal validation.

Conclusion

The model demonstrates high predictive accuracy, making it a valuable tool for guiding the development of preoperative protocols.

Value of a combined magnetic resonance-enhanced and diffusion-weighted imaging dual-sequence radiomics model in predicting the efficacy of high-intensity focused ultrasound ablation for uterine fibroids

OBJECTIVE

To establish a joint radiomics model based on T1 contrast-enhanced (T1C) imaging and diffusion-weighted imaging (DWI), and investigate its value in predicting the efficacy of high-intensity focused ultrasound (HIFU) in ablating uterine fibroids.

METHODS

This multicenter retrospective study included 195 patients with uterine fibroids. Their data were divided into training (n = 120), internal test (n = 30), and external test (n = 45) sets. The radiomic features were extracted from T1C and DWI sequences. Logistic regression was used to develop the T1C, DWI, integration, and joint models, and receiver operating characteristic curves were used to assess model performance. The Delong test was used to compare the predictive efficacies of different models, and the best model was used for external validation and development of the nomogram.

RESULTS

Eight T1C features, six DWI features, and three imaging features were retained for the modeling. The areas under the curve were 0.852 and 0.769 for the integrated model on the training and internal test sets, respectively; 0.857 and 0.824 for the joint model on the training and internal test sets, respectively, which were higher than those of the single-sequence model; and 0.857 for the joint model on the external test set.

CONCLUSIONS

A joint radiomics model based on T1C and DWI data can effectively predict the efficacy of HIFU for ablating uterine fibroids and guide the development of individualized clinical treatment plans.

Multicenter study on predicting postoperative upper limb muscle strength improvement in cervical spinal cord injury patients using radiomics and deep learning

Cervical spinal cord injury is often catastrophic, frequently leading to irreversible impairment. MRI has become the gold standard for evaluating spinal cord injuries (SCI). Our study aimed to assess the accuracy of a radiomics approach, based on machine learning and utilizing conventional MRI, in predicting the prognosis of patients with SCI. In a retrospective analysis of 82 SCI patients from three hospitals, we categorized them into good (n = 49) and poor (n = 33) prognosis groups. Preoperative T2-weighted MRI images were segmented using 3D-Region of Interest (ROI) techniques, and both radiomic and deep transfer learning features were extracted. These features were normalized using Z-score and harmonized via ComBat. Feature selection was performed using a greedy algorithm and Least absolute shrinkage and selection operator (LASSO), and others, followed by the calculation of radiomics scores through linear regression. Machine learning was then used to identify the most predictive radiomic features. Model performance was evaluated by analyzing the area under the curve (AUC) and other indicators.Univariate analysis indicated that the demographic characteristics of cervical spinal cord injury were not statistically significant. In the test dataset, the random forest (RF) combined with radiomics and ResNet34 demonstrated better performance, with an accuracy of 0.800 and an AUC of 0.893.Using MRI, deep learning-based radiomics signals show promise in evaluating and predicting the postoperative prognosis of these patients.

A novel MSN-II feature extracted from T1-weighted MRI for discriminating between BD patients and MDD patients

BACKGROUND

Differentiating between patients with bipolar disorder (BD) and major depressive disorder (MDD) is clinically challenging. This study aimed to explore the potential of radiomic textural features for discriminating BD and MDD.

METHODS

A total 253 subjects (114 patients with BD, 139 patients with MDD) with T1-weighted MRI data were recruited. Radiomics features and gray matter volume (GMV) features were extracted from each brain region. A novel high-level MSN_II feature method based on radiomic features was proposed. And a total of 21 MSN features (5 MSN_I and 16 MSN_II) based on different combinations of the 5 types of radiomic textural feature were calculated. Classification models were constructed using various combinations of MSNs or GMV, and their performance and stability was evaluated through 2000 repeated experiments.

RESULTS

The model built with combined features (GMV and GMV + MSN_II_GLCM_GLSZM_NGTDM) showed the best classification performance (AUC = 0.896±0.058, ACC = 0.831±0.064) in the validation cohort. After MANOVA analysis and FDR correlation, the MSN_II_GLCM_GLSZM_NGTDM values in 4 regions (right rectus gyrus, right temporal pole: middle temporal gyrus, Vermis3 and Vermis10) showed significant difference between BD and MDD.

LIMITATION

The main limitation of this study is that the data is derived from a single center without an external independent test set.

CONCLUSIONS

Incorporating the high-level MSN_II based on radiomics features can improve the classification performance compared to models solely relying on GMV features alone. This result implied the potential application of the proposed high level MSN method and radiomics textural features on the MDD and BD clinical studies.

Application of artificial intelligence in forecasting survival in high-grade glioma: systematic review and meta-analysis involving 79,638 participants

High-grade glioma (HGG) is an aggressive brain tumor with poor survival rates. Predicting survival outcomes is critical for personalized treatment planning. In recent years, artificial intelligence (AI), particularly machine learning (ML) and deep learning (DL) models, has emerged as a promising approach for enhancing prognostic accuracy in HGG but this study especially focused on the potential of AI in the recurrence of HGG. A systematic review and meta-analysis were conducted to assess the performance of AI-based models in predicting survival outcomes for HGG patients. Relevant studies were retrieved from PubMed, Embase, Scopus, and Web of Science until 2 Dec 2024, using predefined keywords ("High-Grade Glioma", "Survival" and "Machine Learning") without date or language restrictions. Data extraction and quality assessment were performed in accordance with PRISMA and PROBAST guidelines. In this study were included. The pooled diagnostic metric, the area under the curve (AUC), was analyzed using random-effects models. A total of 39 studies with 29 various algorithms and 79,638 patients were included, with 15 studies contributing to the meta-analysis. The most commonly used algorithms were random forest (RF) and logistic regression (LR), which demonstrated robust predictive accuracy. The pooled AUCs for one-year, two-year, three-year and overall survival predictions were 0.816, 0.854, 0.871 and 0.789 respectively. Subgroup analysis revealed that RSF achieved the highest predictive accuracy with an AUC of 0.91 (95% CI: 0.84-0.98), while LR followed with an AUC of 0.89 (95% CI: 0.82-0.96). Models integrating clinical, radiomics, and genetic features consistently outperformed single-data-type models. MRI was the most frequently utilized imaging modality. AI-based models, particularly ML and DL algorithms, show significant potential for improving survival prediction in HGG patients. By integrating multimodal data, these models offer valuable tools for personalized treatment planning, although further validation in prospective, multicenter studies is needed to ensure clinical applicability.

An interpretable ensemble model combining handcrafted radiomics and deep learning for predicting the overall survival of hepatocellular carcinoma patients after stereotactic body radiation therapy

PURPOSE

Hepatocellular carcinoma (HCC) remains a global health concern, marked by increasing incidence rates and poor outcomes. This study seeks to develop a robust predictive model by integrating radiomics and deep learning features with clinical data to predict 2-year survival in HCC patients treated with stereotactic body radiation therapy (SBRT).

METHODS

This study analyzed a cohort of 186 HCC patients who underwent SBRT. Radiomics features were extracted from CT scans, complemented by collection of clinical data. Training and validation of machine learning models were conducted using nested cross-validation techniques. Deep learning models, leveraging various convolutional neural networks (CNNs), were employed to effectively integrate both image and clinical data. Post-hoc explainability techniques were applied to elucidate the contribution of imaging data to predictive outcomes.

RESULTS

Handcrafted radiomics features demonstrated moderate predictive performance, with area under the receiver operating characteristic curve (AUC) values ranging from 0.59 to 0.72. Deep learning models, harnessing the fusion of image and clinical data, exhibited improved predictive accuracy, with AUC values ranging from 0.71 to 0.81. Notably, the ensemble model, amalgamating handcrafted radiomics and deep learning features with clinical data, demonstrated the most robust predictive capability, achieving an AUC of 0.86 (95% CI: 0.80-0.93).

CONCLUSION

The ensemble model represents a significant advancement, providing a comprehensive tool for predicting survival outcomes in HCC patients undergoing SBRT. The inclusion of interpretability methods such as Grad-CAM enhances transparency and understanding of these complex predictive models.

Support Vector Machine for Stratification of Cognitive Impairment Using 3D T1WI in Patients with Type 2 Diabetes Mellitus

Purpose

To explore the potential of MRI-based radiomics in predicting cognitive dysfunction in patients with diagnosed type 2 diabetes mellitus (T2DM).

Patients and Methods

In this study, data on 158 patients with T2DM were retrospectively collected between September 2019 and December 2020. The participants were categorized into a normal cognitive function (N) group (n=30), a mild cognitive impairment (MCI) group (n=90), and a dementia (DM) group (n=38) according to the Chinese version of the Montréal Cognitive Assessment Scale-B (MoCA-B). Radiomics features were extracted from the brain tissue except ventricles and sulci in the 3D T1WI images, support vector machine (SVM) model was then established to identify the CI and N groups, and the MCI and DM groups, respectively. The models were evaluated based on their area under the receiver operating characteristic curve (AUC), Precision (P), Recall rate (Recall, R), F1-score, and Support. Finally, ROC curves were plotted for each model.

Results

The study consisted of 68 cases in the N and CI group, with 54 cases in the training set and 14 in the verification set, and 128 cases were included in the MCI and DM groups, with 90 training sets and 38 verification sets. The consistency for inter-group and intra-group of radiomics features in two physicians were 0.86 and 0.90, respectively. After features selection, there were 11 optimal features to distinguish N and CI and 12 optimal features to MCI and DM. In the test set, the AUC for the SVM classifier was 0.857 and the accuracy was 0.830 in distinguishing CI and N, while AUC was 0.821 and the accuracy was 0.830 in distinguishing MCI and DM.

Conclusion

The SVM model based on MRI radiomics exhibits high efficacy in the diagnosis of cognitive dysfunction and evaluation of its severity among patients with T2DM.

Impact of MRI Texture Analysis on Complication Rate in MRI-Guided Liver Biopsies

Magnetic resonance imaging (MRI)-derived texture features are quantitative imaging parameters that may have valuable associations with clinical aspects. Their prognostic ability in patients undergoing percutaneous MRI-guided liver biopsy to identify associations with post-interventional bleeding complications and biopsy success rate has not been sufficiently investigated. The patient sample consisted 79 patients (32 females, 40.5%) with a mean age of 58.7 ± 12.4 years. Clinical parameters evaluated included comorbidities, pre-existing liver disease, known cancer diagnosis, and hemostaseological parameters. Several puncture-related parameters such as biopsy angle, distance of needle entry to capsule, and lesion were analyzed. MRI texture features of the target lesion were extracted from the planning sequence of the MRI-guided liver biopsy. Mann-Whitney U test and Fisher's exact test were used for group comparison; multivariate regression model was used for outcome prediction. Overall, the diagnostic outcome of biopsy was malignant in 38 cases (48.1%) and benign in 32 cases (40.5%). A total of 11 patients (13.9%) had post-interventional bleeding, while nine patients (11.4%) had a negative biopsy result. Several texture features were statistically significantly different between patients with and without hemorrhage. The texture feature GrVariance (1.37 ± 0.78 vs. 0.80 ± 0.35, p = 0.007) reached the highest statistical significance. Regarding unsuccessful biopsy results, S(1,1)DifEntrp (0.80 ± 0.10 vs. 0.89 ± 0.12, p = 0.022) and S(0,4)DifEntrp (1.14 ± 0.10 vs. 1.22 ± 0.11, p = 0.021) reached statistical significance between groups. Several MRI texture features of the target lesion were associated with bleeding complications or negative biopsy after MRI-guided percutaneous liver biopsy. This could be used to identify at-risk patients at the beginning of the procedure and should be further analyzed.

Combined morphology and radiomics of intravoxel incoherent movement as a predictive model for the pathologic complete response before neoadjuvant chemotherapy in patients with breast cancer

Background

To develop a predictive model using baseline imaging of morphology and radiomics derived from intravoxel incoherent motion diffusion-weighted imaging (IVIM-DWI) to determine the pathologic complete response (pCR) to neoadjuvant chemotherapy (NACT) in breast cancer patients.

Methods

A total of 265 patients who underwent 3.0 T MRI scans before NACT were examined. Among them, 113 female patients with stage II-III breast cancer were included. The training data set consisted of 79 patients (31/48=pCR/Non-PCR, npCR), while the remaining 34 cases formed the validation cohort (13/21=pCR/npCR). Radiomics and conventional magnetic resonance imaging features analysis were performed. To build a nomogram model that integrates the radiomics signature and conventional imaging, a logistic regression method was employed. The performance evaluation of the nomogram involved the area under the receiver operating characteristic curve (AUC), a decision curve analysis, and the calibration slope.

Results

In an assessment for predicting pCR, the radiomics model displayed an AUC of 0.778 and 0.703 for the training and testing cohorts, respectively. Conversely, the morphology model exhibited an AUC of 0.721 and 0.795 for the training and testing cohorts, respectively. The nomogram displayed superior predictive discrimination with an AUC of 0.862 for the training cohort and 0.861 for the testing cohort. Decision curve analyses indicated that the nomogram provided the highest clinical net benefit.

Conclusion

Performing a nomogram consisting of integrated morphology and radiomics assessment using IVIM-DWI before NACT enables effective prediction of pCR in breast cancer. This predictive model therefore can facilitate medical professionals in making individualized treatment decisions.

Multimodal Imaging of Osteosarcoma: From First Diagnosis to Radiomics

Osteosarcoma is a primary malignant bone tumor characterized by the production of an osteoid matrix. Although histology remains the definitive diagnostic standard, imaging plays a crucial role in diagnosis, therapeutic planning, and follow-up. Conventional radiography serves as the initial checkpoint for detecting this pathology, which often presents diagnostic challenges due to vague and nonspecific symptoms, especially in its early stages. Today, the integration of different imaging techniques enables an increasingly personalized diagnosis and management, with each contributing unique and complementary information. Conventional radiography typically initiates the imaging assessment, and the Bone Reporting and Data System (Bone-RADS) of the Society of Skeletal Radiology (SSR) is a valuable tool for stratifying the risk of suspicious bone lesions. CT is the preferred modality for evaluating the bone matrix, while bone scans and PET/CT are effective for detecting distant metastases. MRI reveals the extent of the lesion in adjacent soft tissues, the medullary canal, and joints, as well as its relationship to neurovascular structures and the presence of skip lesions. Advanced techniques such as dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI), diffusion-weighted imaging (DWI), and perfusion MRI help characterize the tumor environment and assess treatment response. Osteosarcoma comprises a range of subtypes with differing clinical and imaging characteristics, some of which are particularly distinctive, such as in the case of telangiectatic osteosarcoma. Knowledge of these variants can guide radiologists in the differential diagnosis, which includes both central and surface forms, ranging from highly aggressive to more indolent types. In this review, we present a wide range of representative cases from our hospital case series to illustrate both typical and atypical imaging presentations. Finally, we discuss recent advancements and challenges in applying artificial intelligence approaches to the imaging of osteosarcoma.

Comparative performance of multiple ensemble learning models for preoperative prediction of tumor deposits in rectal cancer based on MR imaging

Ensemble learning can effectively mitigate the risk of model overfitting during training. This study aims to evaluate the performance of ensemble learning models in predicting tumor deposits in rectal cancer (RC) and identify the optimal model for preoperative clinical decision-making. A total of 199 RC patients were analyzed, with radiomic features extracted from T2-weighted and apparent diffusion coefficient images and selected through advanced statistical methods. After that, the bagging-ensemble learning model (random forest), boosting-ensemble learning model (XGBoost, AdaBoost, LightGBM, and CatBoost), and voting-ensemble learning model (integrating 5 classifiers) were applied and optimized using grid search with tenfold cross-validation. The area under the receiver operator characteristic curve, calibration curve, t-distributed stochastic neighbor embedding (t-SNE), and decision curve analysis were adopted to evaluate the performance of each model. The voting-ensemble learning model (VELM) performs best in the testing cohort, with an AUC of 0.875 and an accuracy of 0.800. Notably, Calibration plots confirmed VELM's stability and t-SNE visualization illustrated clear clustering of radiomic features. Decision curve analysis further validated the VELM's superior net benefit across a range of clinical thresholds, underscoring its potential as a reliable tool for clinical decision-making in RC.

A Hybrid Machine Learning CT-Based Radiomics Nomogram for Predicting Cancer-Specific Survival in Curatively Resected Colorectal Cancer

RATIONALE AND OBJECTIVES

To develop and validate a computed tomography-based radiomics nomogram for cancer-specific survival (CSS) prediction in curatively resected colorectal cancer (CRC), and its performance was compared with the American Joint Committee on Cancer (AJCC) staging and clinical-pathological models.

MATERIALS AND METHODS

A total of 794 patients with curatively resected CRC from a prospective cancer registry program were included and randomly divided into the training (n = 556) and validation (n = 238) cohorts. A radiomics signature (RS) predicting CSS was constructed with a hybrid automatic machine learning strategy, and the prognostic value was assessed with Kaplan-Meier (KM) survival analysis. The performance of the established models was assessed by the discrimination, calibration, and clinical utility.

RESULTS

A 10-feature-based RS with independent prognostic value was developed. KM survival curves showed that high-risk patients defined by RS had a worse CSS than the low-risk patients (log-rank P<0.001). The radiomics nomogram integrating the RS and clinical-pathological factors had the optimal performance in predicting CSS in terms of Harrell's concordance index (0.803 [95% confidence interval: 0.761-0.845] for the primary cohort, 0.772 [95% confidence interval: 0.702-0.841] for the validation cohort), time-dependent receiver operating curves (time-ROC) (the area under the time-ROC curves [AUC] at three years were 84.06±2.86 and at five years were 86.35±2.19 in the primary cohort, the AUC at three years were 77.6±4.76, and at five years were 84±3.66 in the validation cohort), calibration curves and decision curve analysis, in comparison with the AJCC staging model, clinical-pathological model, and the RS alone.

CONCLUSION

The radiomics nomogram integrating the RS and clinical-pathological factors could be a valuable individualized predictor of the CSS for curatively resected CRC patients.

Research progress of MRI-based radiomics in hepatocellular carcinoma

Background

Primary liver cancer (PLC), notably hepatocellular carcinoma (HCC), stands as a formidable global health challenge, ranking as the sixth most prevalent malignant tumor and the third leading cause of cancer-related deaths. HCC presents a daunting clinical landscape characterized by nonspecific early symptoms and late-stage detection, contributing to its poor prognosis. Moreover, the limited efficacy of existing treatments and high recurrence rates post-surgery compound the challenges in managing this disease. While histopathologic examination remains the cornerstone for HCC diagnosis, its utility in guiding preoperative decisions is constrained. Radiomics, an emerging field, harnesses high-throughput imaging data, encompassing shape, texture, and intensity features, alongside clinical parameters, to elucidate disease characteristics through advanced computational techniques such as machine learning and statistical modeling. MRI radiomics specifically holds significant importance in the diagnosis and treatment of hepatocellular carcinoma (HCC).

Objective

This study aims to evaluate the methodology of radiomics and delineate the clinical advancements facilitated by MRI-based radiomics in the realm of hepatocellular carcinoma diagnosis and treatment.

Methods

A systematic review of the literature was conducted, encompassing peer-reviewed articles published between July 2018 and Jan 2025, sourced from PubMed and Google Scholar. Key search terms included Hepatocellular carcinoma, HCC, Liver cancer, Magnetic resonance imaging, MRI, radiomics, deep learning, machine learning, and artificial intelligence.

Results

A comprehensive analysis of 93 articles underscores the efficacy of MRI radiomics, a noninvasive imaging analysis modality, across various facets of HCC management. These encompass tumor differentiation, subtype classification, histopathological grading, prediction of microvascular invasion (MVI), assessment of treatment response, early recurrence prognostication, and metastasis prediction.

Conclusion

MRI radiomics emerges as a promising adjunctive tool for early HCC detection and personalized preoperative decision-making, with the overarching goal of optimizing patient outcomes. Nevertheless, the current lack of interpretability within the field underscores the imperative for continued research and validation efforts.

Simplicity within biological complexity

Motivation

Heterogeneous, interconnected, systems-level, molecular (multi-omic) data have become increasingly available and key in precision medicine. We need to utilize them to better stratify patients into risk groups, discover new biomarkers and targets, repurpose known and discover new drugs to personalize medical treatment. Existing methodologies are limited and a paradigm shift is needed to achieve quantitative and qualitative breakthroughs.

Results

In this perspective paper, we survey the literature and argue for the development of a comprehensive, general framework for embedding of multi-scale molecular network data that would enable their explainable exploitation in precision medicine in linear time. Network embedding methods (also called graph representation learning) map nodes to points in low-dimensional space, so that proximity in the learned space reflects the network's topology-function relationships. They have recently achieved unprecedented performance on hard problems of utilizing few omic data in various biomedical applications. However, research thus far has been limited to special variants of the problems and data, with the performance depending on the underlying topology-function network biology hypotheses, the biomedical applications, and evaluation metrics. The availability of multi-omic data, modern graph embedding paradigms and compute power call for a creation and training of efficient, explainable and controllable models, having no potentially dangerous, unexpected behaviour, that make a qualitative breakthrough. We propose to develop a general, comprehensive embedding framework for multi-omic network data, from models to efficient and scalable software implementation, and to apply it to biomedical informatics, focusing on precision medicine and personalized drug discovery. It will lead to a paradigm shift in the computational and biomedical understanding of data and diseases that will open up ways to solve some of the major bottlenecks in precision medicine and other domains.

Dual-Time-Point Radiomics for Prognosis Prediction in Colorectal Liver Metastasis Treated with Neoadjuvant Therapy Before Radical Resection: A Two-Center Study

BACKGROUND

Optimal prognostic stratification for colorectal liver metastases (CRLM) patients undergoing surgery with neoadjuvant therapy (NAT) remains elusive. This study aimed to develop and validate dual-time-point radiomic models for CRLM prognosis prediction using pre- and post-NAT imaging features.

METHODS

Radiomic features were extracted from four MRI sequences in 100 cases of CRLM patients who underwent NAT and radical resection. RAD scores were generated, and clinical/pathologic variables were incorporated into uni- and multivariate Cox regression analyses to construct prognosis models. Time-ROC, time-C index, decision curve analysis (DCA), and calibration curves assessed the predictive performance of Fong score and pre- and post-NAT models for overall survival (OS) and disease-free survival (DFS) in a testing set.

RESULTS

The final models included four variables for OS and three variables for DFS. The post-NAT models outperformed the pre-NAT models in time-ROC, time-C index, calibration, and DCA analysis, except for the 1-year DFS area under the curve (AUC). The Fong score models underperformed. The post-NAT OS RAD score effectively stratified patients into prognostic subgroups.

CONCLUSIONS

The radiomic models incorporating pre- and post-NAT MRI features and clinical/pathologic variables effectively stratified CRLM patients prognositically. The post-NAT models demonstrated superior performance.

Magnetic resonance imaging -based radiomics of the pituitary gland is highly predictive of precocious puberty in girls: a pilot study

Background

The aim of the study was to explore a radiomic model that could assist physicians in the diagnosis of central precocious puberty (CPP). A predictive model based on radiomic features (RFs), extracted form magnetic resonance imaging (MRI) of the pituitary gland, was thus developed to distinguish between CPP and control subjects.

Methods

45 girls with confirmed diagnosis of CPP (CA:8.4 ± 0.9 yr) according to the current criteria and 47 age-matched pre-pubertal control subjects (CA:8.7 ± 1.2 yr) were retrospectively enrolled. Two readers (R1, R2) blindly segmented the pituitary gland on MRI studies for RFs and performed a manual estimation of the pituitary volume. Radiomics was compared against pituitary volume in terms of predictive performances (metrics: ROC-AUC, accuracy, sensitivity and specificity) and reliability (metric: intraclass correlation coefficient, ICC). Pearson correlation between RFs and auxological, biochemical, and ultrasound data was also computed.

Results

Two different radiomic parameters, Shape Surface Volume Ratio and Glrlm Gray Level Non-Uniformity, predicted CPP with a high diagnostic accuracy (ROC-AUC 0.81 ± 0.08) through the application of our ML algorithm. Anthropometric variables were not confounding factors of these RFs suggesting that premature thelarche and/or pubarche would not be potentially misclassified. The selected RFs correlated with baseline and peak LH (p < 0.05) after GnRH stimulation. The diagnostic sensitivity was improved compared to pituitary volume only (0.76 versus 0.68, p<0.001) and demonstrated higher inter-reader reliability (ICC>0.57 versus ICC=0.46).

Discussion

Radiomics is a promising tool to diagnose CPP as it reflects also functional aspects. Further studies are warranted to validate these preliminary data.

Age-stratified deep learning model for thyroid tumor classification: a multicenter diagnostic study

OBJECTIVES

Thyroid cancer, the only cancer that uses age as a specific predictor of survival, is increasing in incidence, yet it has a low mortality rate, which can lead to overdiagnosis and overtreatment. We developed an age-stratified deep learning (DL) model (hereafter, ASMCNet) for classifying thyroid nodules and aimed to investigate the effect of age stratification on the accuracy of a DL model, exploring how ASMCNet can help radiologists improve diagnostic performance and avoid unnecessary biopsies.

METHODS

In this retrospective study, we used ultrasound images from three hospitals, a total of 10,391 images of 5934 patients were used for training, validation, and testing. The performance of ASMCNet was compared with that of model-trained non-age-stratified radiologists with different experience levels on the test data set with the DeLong method.

RESULTS

The area under the receiver operating characteristic curve (AUROC), sensitivity, and specificity of ASMCNet were 0.906, 86.1%, and 85.1%, respectively, which exceeded those of model-trained non-age-stratified (0.867, 83.2%, and 75.5%, respectively; p < 0.001) and higher than all of the radiologists (p < 0.001). Reader studies show that radiologists' performances are improved when assisted by the explaining heatmaps (p < 0.001).

CONCLUSIONS

Our study demonstrates that age stratification based on DL can further improve the performance of thyroid tumor classification models, which also suggests that age is an important factor in the diagnosis of thyroid tumors. The ASMCNet model shows promising clinical applicability and can assist radiologists in improving diagnostic accuracy.

KEY POINTS

Question Age is crucial for differentiated thyroid carcinoma (DTC) prognosis, yet its diagnostic impact lacks research. Findings Adding age stratification to DL models can further improve the accuracy of thyroid nodule diagnosis. Clinical relevance Age-stratified multimodal classification network is a reliable tool used to help radiologists diagnose thyroid nodules, and integrating it into clinical practice can improve diagnostic accuracy and reduce unnecessary biopsies or treatments.

Radiomic features of dynamic contrast-enhanced MRI can predict Ki-67 status in head and neck squamous cell carcinoma

PURPOSE

This study aimed to investigate the potential of radiomic features derived from dynamic contrast-enhanced MRI (DCE-MRI) in predicting Ki-67 and p16 status in head and neck squamous cell carcinoma (HNSCC).

MATERIALS AND METHODS

A cohort of 124 HNSCC patients who underwent pre-surgery DCE-MRI were included and divided into training and test set (7:3), further subgroup analysis was performed for 104 cases with oral squamous cell carcinoma (OSCC). Radiomics features were extracted from DCE images. The least absolute shrinkage and selection operator (LASSO) was used for radiomics features selection, and receiver operating characteristics analysis for predictive performance assessment. The nomogram's performance was evaluated using decision curve analysis (DCA).

RESULTS

Ten DCE-MRI features were identified to build the predictive model of HNSCC, demonstrating excellent predictive value for Ki-67 status in both the training set (AUC of 0.943) and test set (AUC of 0.801). The nomograms based on the predictive model showed good fit in the calibration curves (p > 0.05), and DCA indicated its high clinical usefulness. In subgroup analysis of OSCC, fourteen features were selected to build the predictive model for Ki-67 status with an AUC of 0.960 in training set and 0.817 in test set. No features could be included to establish a model to predict p16 status.

CONCLUSION

The radiomics model utilizing DCE-MRI features could effectively predict Ki-67 status in HNSCC patients, offering potential for noninvasive preoperative prediction of Ki-67 status.

Artificial Intelligence and Machine Learning Predicting Transarterial Chemoembolization Outcomes: A Systematic Review

BACKGROUND

Major society guidelines recommend transarterial chemoembolization (TACE) as the standard of care for intermediate-stage hepatocellular carcinoma (HCC) patients. However, predicting treatment response remains challenging.

AIMS

As artificial intelligence (AI) may predict therapeutic responses, this systematic review aims to assess the performance and effectiveness of radiomics and AI-based models in predicting TACE outcomes in patients with HCC.

METHODS

A systemic search was conducted on Medline and Embase databases from inception to 7th April 2024. Included studies generated a predictive model for TACE response and evaluated its performance by area under the curve (AUC), specificity, or sensitivity analysis. Systematic reviews, meta-analyses, case series and reports, pediatric, and animal studies were excluded. Secondary search of references of included articles ensured comprehensiveness.

RESULTS

64 articles, with 13,412 TACE-treated patients, were included. AI in pre-treatment CT scans provided value in predicting the efficacy of TACE in HCC treatment. A positive association was observed for AI in pre-treatment MRI scans. Models incorporating radiomics had numerically better performance than those incorporating manual measured radiological variables. 39 studies demonstrated that combined predictive models had numerically better performance than models based solely on imaging or non-imaging features.

CONCLUSION

A combined predictive model incorporating clinical features, laboratory investigations, and radiological characteristics may effectively predict response to TACE treatment for HCC.

Radiomics analysis of thoracic vertebral bone marrow microenvironment changes before bone metastasis of breast cancer based on chest CT

Bone metastasis from breast cancer significantly elevates patient morbidity and mortality, making early detection crucial for improving outcomes. This study utilizes radiomics to analyze changes in the thoracic vertebral bone marrow microenvironment from chest computerized tomography (CT) images prior to bone metastasis in breast cancer, and constructs a model to predict metastasis.

METHODS

This study retrospectively gathered data from breast cancer patients who were diagnosed and continuously monitored for five years from January 2013 to September 2023. Radiomic features were extracted from the bone marrow of thoracic vertebrae on non-contrast chest CT scans. Multiple machine learning algorithms were utilized to construct various radiomics models for predicting the risk of bone metastasis, and the model with optimal performance was integrated with clinical features to develop a nomogram. The effectiveness of this combined model was assessed through receiver operating characteristic (ROC) analysis as well as decision curve analysis (DCA).

RESULTS

The study included a total of 106 patients diagnosed with breast cancer, among whom 37 developed bone metastases within five years. The radiomics model's area under the curve (AUC) for the test set, calculated using logistic regression, is 0.929, demonstrating superior predictive performance compared to alternative machine learning models. Furthermore, DCA demonstrated the potential of radiomics models in clinical application, with a greater clinical benefit in predicting bone metastasis than clinical model and nomogram.

CONCLUSION

CT-based radiomics can capture subtle changes in the thoracic vertebral bone marrow before breast cancer bone metastasis, offering a predictive tool for early detection of bone metastasis in breast cancer.

Predicting progression in triple-negative breast cancer patients undergoing neoadjuvant chemotherapy: Insights from peritumoral radiomics

OBJECTIVE

To investigate whether radiomic features obtained from the intratumoral and peritumoral regions of pretreatment magnetic resonance imaging (MRI) can predict progression in patients with triple-negative breast cancer (TNBC) undergoing neoadjuvant chemotherapy (NAC) in comparison with the previously determined clinical score.

METHODS

This single-center retrospective study evaluated 224 women with TNBC who underwent NAC between 2010 and 2019. Women were randomly allocated to the training set (n = 169) for model development and the test set (n = 55) for model validation. The clinical score consisted of the histologic type, Ki-67 index, and degree of edema on T2-weighted imaging. Intratumoral and peritumoral radiomic features were extracted from T2-weighted images and the first- and last-phase images of dynamic contrast-enhanced MRI. The radiomics model was built using only radiomic features, whereas the combined model incorporated the clinical score along with radiomic features. The area under the receiver operating characteristic curve (AUC) was used to assess performance.

RESULTS

Progression occurred in 18 and five patients in the training and test sets, respectively. The radiomics model selected three radiomic features (two peritumoral and one intratumoral), while the combined model selected the clinical score and five radiomic features (four peritumoral and one intratumoral). Among the total radiomic features, Inverse Difference Normalized of the peritumoral region of the T2-weighted images, reflective of peritumoral heterogeneity, demonstrated the highest level of association with tumor progression. In the test set, the AUC values of the radiomics-only model, the combined model, and the clinical score were 0.592, 0.764, and 0.720, respectively. Compared to the clinical score, the radiomics-only model (0.720 vs. 0.592, p = 0.468) and the combined model (0.720 vs. 0.764, p = 0.553) did not show superior performance.

CONCLUSION

The radiomics features were not superior in predicting the progression of TNBC compared to the clinical score, although the peritumoral heterogeneity on T2-weighted images showed a potential.

Radiomics based on multiple machine learning methods for diagnosing early bone metastases not visible on CT images

OBJECTIVES

This study utilizes [99mTc]-methylene diphosphate (MDP) single photon emission computed tomography (SPECT) images as a reference standard to evaluate whether the integration of radiomics features from computed tomography (CT) and machine learning algorithms can identify microscopic early bone metastases. Additionally, we also determine the optimal machine learning approach.

MATERIALS AND METHODS

We retrospectively studied 63 patients with early bone metastasis from July 2020 to March 2023. The ITK-SNAP software was used to delineate early bone metastases and normal bone tissue in SPECT images of each patient, which were then registered onto CT images to outline the volume of interest (VOI). The VOI includes 63 early bone metastasis volumes and 63 normal bone tissue volumes. 126 VOIs were randomly distributed in a 7:3 ratio between the training and testing groups, and 944 radiomics features were extracted from every VOI. We established 20 machine learning models using 5 feature selection algorithms and 4 classification methods. Evaluate the performance of the model using the area under the receiver operating characteristic curve (AUC).

RESULTS

Most machine learning models demonstrated outstanding discriminative capacity, with AUCs higher than 0.70. Notably, the K-Nearest Neighbors (KNN) classifier exhibited significant performance improvement compared to the other four classifiers. Specifically, the model constructed utilizing eXtreme Gradient Boosting (XGBoost) feature selection method integrated with KNN classifier achieved the maximum AUC, which is 0.989 in the training set and 0.975 in the testing set.

CONCLUSIONS

Radiomics features integrated with machine learning methods can identify early bone metastases that are not visible on CT images. In our analysis, KNN is considered the optimal classification method.