Introduction to Radiomics

Prediction of clinical outcome for high-intensity focused ultrasound ablation of adenomyosis based on non-enhanced MRI radiomics

OBJECTIVES

The study aimed to develop a non-enhanced MRI-based radiomics model for the preoperative prediction of the efficacy of adenomyosis after high-intensity focused ultrasound (HIFU) treatment.

METHODS

The data of 130 patients with adenomyosis who underwent HIFU treatment were reviewed. Based on a non-perfused volume ratio (NPVR) of 50%, the patients were assigned to high ablation rate and low ablation rate groups. A radiomics model was constructed from the screened radiomics features and its output probability was calculated as the radiomics score (Radscore). The clinical-imaging model was constructed from the independent predictors of clinical-imaging characteristics. The combined model was constructed by integrating Radscore and clinical-imaging independent predictors. Receiver operating characteristic (ROC) curves, the Delong test, and decision curve analysis (DCA) were used to evaluate the models.

RESULTS

The combined model had the best overall performance among the three models. The AUC (95% CI), specificity, sensitivity, accuracy, and precision of the combined model were 0.860 (0.786-0.935), 0.780, 0.756, 0.769, 0.738 in the training set, and 0.878 (0.774-0.983), 0.859, 0.667, 0.769, 0.800 in the test set, respectively. The Delong test showed that the performance of both the radiomics and combined models differed significantly from the clinical-imaging model. But the performance of the combined and the radiomics model was statistically equivalent. The DCA indicated that the combined model had better clinical net benefit.

CONCLUSION

The combined model based on non-enhanced MRI radiomics was effective in predicting the outcome of HIFU ablation of adenomyosis before surgery.

Personalized Medicine for Systemic Sclerosis-Associated Interstitial Lung Disease

Purpose of the review

Systemic sclerosis (SSc) is a rare immune-mediated connective tissue disease with high morbidity and mortality. Interstitial lung disease (ILD) is now the leading cause of death for patients with SSc. While several therapeutic agents have been approved for SSc-ILD, opportunities remain for a personalized medicine approach to improve patient outcomes. The purpose of this narrative review is to summarize the current state of personalized medicine for SSc-ILD and future directions to facilitate earlier diagnosis, disease stratification, prognostication, and determination of treatment response. We also review opportunities for personalized medicine approaches within clinical trial design for SSc-ILD.

Recent findings

The management of SSc-ILD remains challenging due to its variable clinical course and current deficits in predicting which individuals will develop progressive pulmonary fibrosis. There have additionally been many challenges in clinical trial design due to limitations in enrichment strategies. Emerging data suggest that serum, radiologic, and other novel biomarkers could be utilized to assess disease activity and treatment response on an individual level.

Summary

Personalized medicine is emerging as a way to address unmet challenges in SSc-ILD and has applicability for identifying stratifying, prognostic, and therapeutic markers for routine clinical care and clinical trial design.

A computed tomography angiography-based radiomics model for prognostic prediction of endovascular abdominal aortic repair

OBJECTIVE

This study aims to develop a radiomics machine learning (ML) model that uses preoperative computed tomography angiography (CTA) data to predict the prognosis of endovascular aneurysm repair (EVAR) for abdominal aortic aneurysm (AAA) patients.

METHODS

In this retrospective study, 164 AAA patients underwent EVAR and were categorized into shrinkage (good prognosis) or stable (poor prognosis) groups based on post-EVAR sac regression. From preoperative AAA and perivascular adipose tissue (PVAT) image, radiomics features (RFs) were extracted for model creation. Patients were split into 80 % training and 20 % test sets. A support vector machine model was constructed for prediction. Accuracy is evaluated via the area under the receiver operating characteristic curve (AUC).

RESULTS

Demographics and comorbidities showed no significant differences between shrinkage and stable groups. The model containing 5 AAA RFs (which are original_firstorder_InterquartileRange, log-sigma-3-0-mm-3D_glrlm_GrayLevelNonUniformityNormalized, log-sigma-3-0-mm-3D_glrlm_RunPercentage, log-sigma-4-0-mm-3D_glrlm_ShortRunLowGrayLevelEmphasis, wavelet-LLH_glcm_SumEntropy) had AUCs of 0.86 (training) and 0.77 (test). The model containing 7 PVAT RFs (which are log-sigma-3-0-mm-3D_firstorder_InterquartileRange, log-sigma-3-0-mm-3D_glcm_Correlation, wavelet-LHL_firstorder_Energy, wavelet-LHL_firstorder_TotalEnergy, wavelet-LHH_firstorder_Mean, wavelet-LHH_glcm_Idmn, wavelet-LHH_glszm_GrayLevelNonUniformityNormalized) had AUCs of 0.76 (training) and 0.78 (test). Combining AAA and PVAT RFs yielded the highest accuracy: AUCs of 0.93 (training) and 0.87 (test).

CONCLUSIONS

Radiomics-based CTA model predicts aneurysm sac regression post-EVAR in AAA patients. PVAT RFs from preoperative CTA images were closely related to AAA prognosis after EVAR, enhancing accuracy when combined with AAA RFs. This preliminary study explores a predictive model designed to assist clinicians in optimizing therapeutic strategies during clinical decision-making processes.

MRI Radiomics Analysis in the Diagnostic Differentiation of Malignant Soft Tissue Myxoid Sarcomas From Benign Soft Tissue Musculoskeletal Myxomas

BACKGROUND

Differentiation of benign myxomas and malignant myxoid sarcomas can be difficult with an overlapping spectrum of morphologic MR findings.

PURPOSE

To assess the diagnostic utility of MRI radiomics in the differentiation of musculoskeletal myxomas and myxoid sarcomas.

STUDY TYPE

Retrospective.

POPULATION

A total of 523 patients were included; histologically proven myxomas (N = 201) and myxoid sarcomas (N = 322), randomly divided (70:30) into training:test subsets.

SEQUENCE/FIELD STRENGTH

T1-weighted (T1W), T2-weighted fat-suppressed (fluid-sensitive), and T1-weighted post-contrast (T1W + C) sequences at 1.0 T, 1.5 T, or 3.0 T.

ASSESSMENT

Seven semantic (qualitative) tumor features were assessed in each case. Manual 3D tumor segmentations performed with radiomics features extracted from T1W, fluid-sensitive, and T1W + C acquisitions. Models were constructed based on radiomic features from individual sequences and from their combination, both with and without the addition of qualitative tumor features.

STATISTICAL TESTS

Intraclass correlation evaluated in 60 cases segmented by three readers. Features with intraclass correlation <0.7 excluded from further analysis. Boruta feature selection and Random Forest modeling performed using the training-dataset, with resultant models used to assess class discrimination (myxoma vs. myxoid sarcoma) in the test dataset. Radiomics score defined as probability class = myxoma. Logistic regression modeling employed to estimate performance of the radiomics score. Area under the receiver operating characteristic curve (AUC) was used to assess diagnostic performance, and DeLong's test to assess performance between constructed models. A P-value <0.05 was considered significant.

RESULTS

Four qualitative semantic features showed significant predictive power in class discrimination. Radiomic models demonstrated excellent differentiation of myxomas from myxoid sarcomas: AUC of 0.9271 (T1W), 0.9049 (fluid-sensitive), and 0.9179 (T1W + C). Incorporation of multiparametric data or semantic features did not significantly improve model performance (P ≥ 0.08) compared to radiomic models derived from any individual MRI sequence alone.

DATA CONCLUSION

MRI radiomics appears to be accurate in the differentiation of myxomas from myxoid sarcomas. Classification performance did not improve when incorporating qualitative features or multiparametric imaging data.

PLAIN LANGUAGE SUMMARY

Accurately distinguishing between benign soft tissue myxomas and malignant myxoid sarcomas is essential for guiding appropriate management but remains challenging with conventional MRI interpretation. This study utilized radiomics, a method that extracts quantitative mathematically derived features from images, to develop predictive models based on routine MRI examination. Analyzing over 500 cases, MRI radiomics demonstrated excellent diagnostic accuracy in differentiating between benign myxomas and malignant myxoid sarcomas, highlighting the potential of the technique, as a powerful non-invasive tool that could complement current diagnostic approaches, and enhance clinical decision-making in patients with soft tissue myxoid tumors of the musculoskeletal system.

LEVEL OF EVIDENCE

3 TECHNICAL EFFICACY: Stage 2.

Radiomics analysis of 18F-FDG PET/CT for visceral pleural invasion in non-small cell lung cancer with pleural attachment

AIM

This study aimed to establish and validate a preoperative model that integrates clinical factors and radiomic features from 2-[18F]-fluoro-2-deoxy-D-glucose (18F-FDG) positron emission tomography (PET)/computed tomography (CT) for predicting visceral pleural invasion (VPI) in non-small cell lung cancer (NSCLC) with radiological pleural attachment.

MATERIALS AND METHODS

A total of 974 NSCLC patients (408 with VPI-present and 566 with VPI-absent) were retrospectively included from two medical centres. Clinical data and PET/CT radiomic features were collected. The optimal predictors from these radiomic features were selected to create the radiomics score (Rad-score) for the PET/CT radiomics model. Significant clinical factors and Rad-scores were incorporated into a combined PET/CT radiomics-clinical model. The predictive performance of the models was assessed using receiver operating characteristic (ROC) analysis.

RESULTS

The combined PET/CT radiomics-clinical model predicted VPI status with areas under the ROC curve (AUCs) of 0.869, 0.858, and 0.863 in the training set (n=569), internal validation set (n=245), and external validation set (n=160), respectively. These were significantly higher than the AUCs of the PET/CT radiomics model, which were 0.828, 0.782, and 0.704 (all P<0.001). In patients with a maximum tumour diameter (Dmax) ≤ 3 cm (n=537) and in patients with adenocarcinoma (n=659), the AUCs of the combined model were 0.876 and 0.877, respectively. A nomogram based on the combined model was developed, with well-fitted calibration curves.

CONCLUSION

The combined PET/CT radiomics-clinical model provides an advantage in predicting VPI status in NSCLC with pleural attachment.

Enhancing radiomics features via a large language model for classifying benign and malignant breast tumors in mammography

BACKGROUND AND OBJECTIVES

Radiomics is widely used to assist in clinical decision-making, disease diagnosis, and treatment planning for various target organs, including the breast. Recent advances in large language models (LLMs) have helped enhance radiomics analysis.

MATERIALS AND METHODS

Herein, we sought to improve radiomics analysis by incorporating LLM-learned clinical knowledge, to classify benign and malignant tumors in breast mammography. We extracted radiomics features from the mammograms based on the region of interest and retained the features related to the target task. Using prompt engineering, we devised an input sequence that reflected the selected features and the target task. The input sequence was fed to the chosen LLM (LLaMA variant), which was fine-tuned using low-rank adaptation to enhance radiomics features. This was then evaluated on two mammogram datasets (VinDr-Mammo and INbreast) against conventional baselines.

RESULTS

The enhanced radiomics-based method performed better than baselines using conventional radiomics features tested on two mammogram datasets, achieving accuracies of 0.671 for the VinDr-Mammo dataset and 0.839 for the INbreast dataset. Conventional radiomics models require retraining from scratch for an unseen dataset using a new set of features. In contrast, the model developed in this study effectively reused the common features between the training and unseen datasets by explicitly linking feature names with feature values, leading to extensible learning across datasets. Our method performed better than the baseline method in this retraining setting using an unseen dataset.

CONCLUSIONS

Our method, one of the first to incorporate LLM into radiomics, has the potential to improve radiomics analysis.

Radiomics across modalities: a comprehensive review of neurodegenerative diseases

Radiomics allows extraction from medical images of quantitative features that are able to reveal tissue patterns that are generally invisible to human observers. Despite the challenges in visually interpreting radiomic features and the computational resources required to generate them, they hold significant value in downstream automated processing. For instance, in statistical or machine learning frameworks, radiomic features enhance sensitivity and specificity, making them indispensable for tasks such as diagnosis, prognosis, prediction, monitoring, image-guided interventions, and evaluating therapeutic responses. This review explores the application of radiomics in neurodegenerative diseases, with a focus on Alzheimer's disease, Parkinson's disease, Huntington's disease, and multiple sclerosis. While radiomics literature often focuses on magnetic resonance imaging (MRI) and computed tomography (CT), this review also covers its broader application in nuclear medicine, with use cases of positron emission tomography (PET) and single-photon emission computed tomography (SPECT) radiomics. Additionally, we review integrated radiomics, where features from multiple imaging modalities are fused to improve model performance. This review also highlights the growing integration of radiomics with artificial intelligence and the need for feature standardisation and reproducibility to facilitate its translation into clinical practice.

Radiomics of lung ventilation/perfusion tomographic imaging in pulmonary embolism diagnosis

PURPOSE

The aim of this study was to develop a machine learning model (named V/P-mics) to identify pulmonary embolism based on lung ventilation/perfusion single-photon emission tomography (V/P-SPECT) images.

METHODS

We retrospectively collected the data of 260 patients from one hospital who underwent V/P-SPECT. Patients were randomly assigned to training and testing groups in a 7:3 ratio. We created an internal further validation group using data of an additional 35 patients from the same hospital, and an external further validation group using data of 30 patients from another hospital. We constructed 35 models and selected one for further optimization. The generalizability of V/P-mics was proven by comparing the area under the curve (AUC) of the testing group, internal and external further validation groups. The diagnostic accuracy and efficiency of V/P-mics was compared with that of nuclear physicians.

RESULTS

V/P-mics showed excellent generalizability, with no statistical difference in AUC among the testing, internal further validation, and external further validation groups (0.938 vs. 0.923 vs. 0.990, all P values > 0.05). The AUC of V/P-mics was close to that of the senior physician (0.923 vs. 0.975, P = 0.332), but significantly higher than the junior physician (0.923 vs. 0.725, P = 0.050). Furthermore, V/P-mics significantly shortened the diagnosis time as compared to the junior physician (100 ± 16 s vs. 240 ± 37 s, P = 0.001).

CONCLUSION

The V/P-mics had good discrimination and generalizability and significantly shortened the diagnosis time for patients with pulmonary embolism. Of note, the model showed excellent interpretability.

Predicting hepatocellular carcinoma response to TACE: A machine learning study based on 2.5D CT imaging and deep features analysis

OBJECTIVES

Prior to the commencement of treatment, it is essential to establish an objective method for accurately predicting the prognosis of patients with hepatocellular carcinoma (HCC) undergoing transarterial chemoembolization (TACE). In this study, we aimed to develop a machine learning (ML) model to predict the response of HCC patients to TACE based on CT images analysis.

MATERIALS AND METHODS

Public dataset from The Cancer Imaging Archive (TCIA), uploaded in August 2022, comprised a total of 105 cases, including 68 males and 37 females. The external testing dataset was collected from March 1, 2019 to July 1, 2022, consisting of total of 26 patients who underwent TACE treatment at our institution and were followed up for at least 3 months after TACE, including 22 males and 4 females. The public dataset was utilized for ResNet50 transfer learning and ML model construction, while the external testing dataset was used for model performance evaluation. All CT images with the largest lesions in axial, sagittal, and coronal orientations were selected to construct 2.5D images. Pre-trained ResNet50 weights were adapted through transfer learning to serve as a feature extractor to derive deep features for building ML models. Model performance was assessed using area under the curve (AUC), accuracy, F1-Score, confusion matrix analysis, decision curves, and calibration curves.

RESULTS

The AUC values for the external testing dataset were 0.90, 0.90, 0.91, and 0.89 for random forest classifier (RFC), support vector classifier (SVC), logistic regression (LR), and extreme gradient boosting (XGB), respectively. The accuracy values for the external testing dataset were 0.79, 0.81, 0.80, and 0.80 for RFC, SVC, LR, and XGB, respectively. The F1-score values for the external testing dataset were 0.75, 0.77, 0.78, and 0.79 for RFC, SVC, LR, and XGB, respectively.

CONCLUSION

The ML model constructed using deep features from 2.5D images has the potential to be applied in predicting the prognosis of HCC patients following TACE treatment.

Comparisons among radiologist, MR findings and radiomics-clinical models in predicting placenta accreta spectrum disorders: a multicenter study

OBJECTIVE

To assess and compare the diagnostic accuracy of radiologist, MR findings, and radiomics-clinical models in the diagnosis of placental implantation disorders.

METHODS

Retrospective collection of MR images from patients suspected of having placenta accreta spectrum (PAS) was conducted across three institutions: Institution I (n = 505), Institution II (n = 67), and Institution III (n = 58). Data from Institution I were utilized to form a training set, while data from Institutions II and III served as an external test set. Radiologist diagnosis was performed by radiologists of varying levels of experience. The interpretation of MR findings was conducted by two radiologists with 10-15 years of experience in pelvic MR diagnosis, following the guidelines for diagnosis. Radiomics analysis extracted features from sagittal T2-weighted images and combined them with prenatal clinical features to construct predictive models. These models were then evaluated for discrimination and calibration to assess their performance.

RESULTS

As measured by the area under the receiver operating characteristic curve (AUC), the diagnostic efficacy was 0.587 (0.542-0.630) for junior radiologists from Institution I, 0.568 (0.441-0.689) from Institution II, and 0.507 (0.373-0.641) from Institution III. The AUC was 0.623 (0.580-0.666) for senior radiologists from Institution I, 0.635 (0.508-0.749) from Institution II, and 0.632 (0.495-0.755) from Institution III. The diagnostic efficacy of MR findings was 0.648 (0.601-0.695) for Institution I, 0.569 (0.429-0.709) for Institution II, and 0.588 (0.442-0.735) for Institution III. The diagnostic efficacy of the radiomics-clinical model was significantly higher, with an AUC of 0.794 (0.754-0.833) for Institution I, 0.783 (0.664-0.903) for Institution II, and 0.816 (0.704-0.927) for Institution III. The diagnostic efficacy of the Fusion model was significantly higher, with an AUC of 0.867 (0.836-0.899) for Institution I, 0.849 (0.753-0.944) for Institution II, and 0.823 (0.708-0.939) for Institution III.

CONCLUSION

The fusion models demonstrated superior diagnostic efficacy compared to radiologists, MR findings, and the radiomics-clinical models. Furthermore, the diagnostic accuracy of PAS was notably higher when utilizing the radiomics-clinical models than when relying solely on radiologist diagnosis or MR findings.

ADVANCES IN KNOWLEDGE

Radiomics analysis substantially augments the diagnostic precision in PAS, providing a significant enhancement over conventional radiologist and MRI findings. The diagnostic efficacy of the fusion model is notably superior to that of individual diagnostic modalities.

Image normalization techniques and their effect on the robustness and predictive power of breast MRI radiomics

BACKGROUND AND PURPOSE

Radiomics analysis has emerged as a promising approach to aid in cancer diagnosis and treatment. However, radiomics research currently lacks standardization, and radiomics features can be highly dependent on acquisition and pre-processing techniques used. In this study, we aim to investigate the effect of various image normalization techniques on robustness of radiomics features extracted from breast cancer patient MRI scans.

MATERIALS AND METHODS

MRI scans from the publicly available MAMA-MIA dataset and an internal breast MRI test set depicting triple negative breast cancer (TNBC) were used. We compared the effect of commonly used image normalization techniques on radiomics feature robustnessusing Concordance-Correlation-Coefficient (CCC) between multiple combinations of normalization approaches. We also trained machine learning-based prediction models of pathologic complete response (pCR) on radiomics after different normalization techniques were used and compared their areas under the receiver operating characteristic curve (ROC-AUC).

RESULTS

For predicting complete pathological response from pre-treatment breast cancer MRI radiomics, the highest overall ROC-AUC was achieved by using a combination of three different normalization techniques indicating their potentially powerful role when working with heterogeneous imaging data. The effect of normalization was more pronounced with smaller training data and normalization may be less important with increasing abundance of training data. Additionally, we observed considerable differences between MRI data sets and their feature robustness towards normalization.

CONCLUSION

Overall, we were able to demonstrate the importance of selecting and standardizing normalization methods for accurate and reliable radiomics analysis in breast MRI scans especially with small training data sets.

Prediction of lymph node metastasis in papillary thyroid carcinoma using non-contrast CT-based radiomics and deep learning with thyroid lobe segmentation: A dual-center study

Objectives

This study aimed to develop a predictive model for lymph node metastasis (LNM) in papillary thyroid carcinoma (PTC) patients by deep learning radiomic (DLRad) and clinical features.

Methods

This study included 271 thyroid lobes from 228 PTC patients who underwent preoperative neck non-contrast CT at Center 1 (May 2021-April 2024). LNM status was confirmed via postoperative pathology, with each thyroid lobe labeled accordingly. The cohort was divided into training (n = 189) and validation (n = 82) cohorts, with additional temporal (n = 59 lobes, Center 1, May-August 2024) and external (n = 66 lobes, Center 2) test cohorts. Thyroid lobes were manually segmented from the isthmus midline, ensuring interobserver consistency (ICC ≥ 0.8). Deep learning and radiomics features were selected using LASSO algorithms to compute DLRad scores. Logistic regression identified independent predictors, forming DLRad, clinical, and combined models. Model performance was evaluated using AUC, calibration, decision curves, and the DeLong test, compared against radiologists' assessments.

Results

Independent predictors of LNM included age, gender, multiple nodules, tumor size group, and DLRad. The combined model demonstrated superior diagnostic performance with AUCs of 0.830 (training), 0.799 (validation), 0.819 (temporal test), and 0.756 (external test), outperforming the DLRad model (AUCs: 0.786, 0.730, 0.753, 0.642), clinical model (AUCs: 0.723, 0.745, 0.671, 0.660), and radiologist evaluations (AUCs: 0.529, 0.606, 0.620, 0.503). It also achieved the lowest Brier scores (0.167, 0.184, 0.175, 0.201) and the highest net benefit in decision-curve analysis at threshold probabilities > 20 %.

Conclusions

The combined model integrating DLRad and clinical features exhibits good performance in predicting LNM in PTC patients.

Performance analysis of 2D and 3D image features for computer-assisted speech diagnosis of dental sibilants in Polish children

BACKGROUND AND OBJECTIVE

Sigmatism is a speech disorder concerning sibilants, and its diagnosis affects many Polish children of preschool age. The success of therapy often depends on early and accurate diagnosis. This paper presents research findings on using 2D and 3D (time-related) visual features to analyze the place of articulation, sibilance (the character of a gap between teeth that allows the articulation of sibilant sounds), and tongue positioning in four of twelve Polish sibilants:/s/,/z/,/ʦ/, and/dz/.

METHODS

A dedicated data acquisition system captured the stereovision stream during the speech therapy examination (201 speakers aged 4-8). The material contains 23 words and four logatomes. This study introduces 3D texture and shape features extracted for the mouth, lips, and tongue. The third dimension is the time of articulation, and the volumes reflect the movements of speech organs. The research compares the usability of 3D mode to a 2D approach (mouth texture features; mouth, lips, and tongue shape parameters) described in previous works. The statistical analysis includes Mann-Whitney U test to indicate the significant differences between selected articulation patterns for each sibilant and pronunciation aspect (considering p<0.05).

RESULTS

Overall outcomes suggest the dominance of 3D time-related statistically significant features, especially describing the shape of a tongue. Analysis considering features with at least medium effect size showed that 3D features differentiate dental and interdental articulation in case of/s/,/z/, and/ʦ/, while in case of/dz/ significant parameters were 2D. The 3D mode prevails also in terms of sibilance: analysis of sounds/z/ and/ʦ/ results in 3D features only, but for/s/ and/dz/ outcomes include both 3D and 2D parameters. Analysis of the tongue positioning during articulation in terms of at least moderate effect size suggests a presence of features only in the case of affricates:/ʦ/ (3D features) and/dz/ (2D features). All parameters with at least medium effect size describe the shape of the tongue.

CONCLUSIONS

This research proves the potential of visual data in building computer-aided speech diagnosis systems using non-contact recording tools. It highlights the usability of a 3D approach introduced in this paper. Results also emphasize the importance of tongue movement analysis.

Progress in the application of fludeoxyglucose positron emission tomography computed tomography in biliary tract cancer

Biliary tract cancer (BTC) is a group of heterogeneous sporadic diseases, including intrahepatic, hilar, and distal cholangiocarcinoma, as well as gallbladder cancer. BTC is characterized by high invasiveness and extremely poor prognosis, with a global increased incidence due to intrahepatic cholangiocarcinoma (ICC). The 18F-fludeoxyglucose positron emission tomography (PET) computed tomography (18F-FDG PET/CT) combines glucose metabolic information (reflecting the glycolytic activity of tumor cells) with anatomical structure to assess tumor metabolic heterogeneity, systemic metastasis, and molecular characteristics noninvasively, overcoming the limitations of traditional imaging in the detection of micrometastases and recurrent lesions. 18F-FDG PET/CT offers critical insights in clinical staging, therapeutic evaluation, and prognostic prediction of BTC. This article reviews research progress in this field over the past decade, with a particular focus on the advances made in the last 3 years, which have not been adequately summarized and recognized. The research paradigm in this field is shifting from qualitative to quantitative studies, and there have been significant breakthroughs in using 18F-FDG PET/CT metabolic information to predict gene expression in ICC. Radiomics and deep learning techniques have been applied to ICC for prognostic prediction and differential diagnosis. Additionally, PET/magnetic resonance imaging is increasingly demonstrating its value in this field.

Brain Fractal Dimension and Machine Learning can predict first-episode psychosis and risk for transition to psychosis

Although there are notable structural abnormalities in the brain associated with psychotic diseases, it is still unclear how these abnormalities relate to clinical presentation. However, the fractal dimension (FD), which offers details on the complexity and irregularity of brain microstructures, may be a promising feature, as demonstrated by neuropsychiatric disorders such as Parkinson's and Alzheimer's. It may offer a possible biomarker for the detection and prognosis of psychosis when paired with machine learning. The purpose of this study is to investigate FD as a structural magnetic resonance imaging (sMRI) feature from individuals with a high clinical risk of psychosis who did not transit to psychosis (CHR_NT), clinical high risk who transit to psychosis (CHR_T), patients with first-episode psychosis (FEP) and healthy controls (HC). Using a machine learning approach that ultimately classifies sMRI images, the goals are (a) to evaluate FD as a potential biomarker and (b) to investigate its ability to predict a subsequent transition to psychosis from the high-risk clinical condition. We obtained sMRI images from 194 subjects, including 44 HCs, 77 FEPs, 16 CHR_Ts, and 57 CHR_NTs. We extracted the FD features and analyzed them using machine learning methods under five classification schemas (a) FEP vs. HC, (b) FEP vs. CHR_NT, (c) FEP vs. CHR_T, (d) CHR_NT vs. CHR_T, (d) CHR_NT vs. HC and (e) CHR_T vs. HC. In addition, the CHR_T group was used as external validation in (a), (b) and (d) comparisons to examine whether the progression of the disorder followed the FEP or CHR_NT patterns. The proposed algorithm resulted in a balanced accuracy greater than 0.77. This study has shown that FD can function as a predictive neuroimaging marker, providing fresh information on the microstructural alterations triggered throughout the course of psychosis. The effectiveness of FD in the detection of psychosis and transition to psychosis should be established by further research using larger datasets.

Multimodal fusion model for prognostic prediction and radiotherapy response assessment in head and neck squamous cell carcinoma

Accurate prediction of prognosis and postoperative radiotherapy response is critical for personalized treatment in head and neck squamous cell carcinoma (HNSCC). We developed a multimodal deep learning model (MDLM) integrating computed tomography, whole-slide images, and clinical features from 1087 HNSCC patients across multiple centers. The MDLM exhibited good performance in predicting overall survival (OS) and disease-free survival in external test cohorts. Additionally, the MDLM outperformed unimodal models. Patients with a high-risk score who underwent postoperative radiotherapy exhibited prolonged OS compared to those who did not (P = 0.016), whereas no significant improvement in OS was observed among patients with a low-risk score (P = 0.898). Biological exploration indicated that the model may be related to changes in the cytochrome P450 metabolic pathway, tumor microenvironment, and myeloid-derived cell subpopulations. Overall, the MDLM effectively predicts prognosis and postoperative radiotherapy response, offering a promising tool for personalized HNSCC therapy.

Development and validation of a radiomics-based prediction model for variceal bleeding in patients with Budd-Chiari syndrome-related gastroesophageal varices

BACKGROUND

Budd-Chiari syndrome (BCS) is caused by obstruction of the hepatic veins or suprahepatic inferior vena cava, leading to portal hypertension and the development of gastroesophageal varices (GEVs), which are associated with an increased risk of bleeding. Existing risk models for variceal bleeding in cirrhotic patients have limited applicability to BCS due to differences in pathophysiology. Radiomics, as a noninvasive technique, holds promise as a tool for more accurate prediction of bleeding risk in BCS-related GEVs.

AIM

To develop and validate a personalized risk model for predicting variceal bleeding in BCS patients with GEVs.

METHODS

We retrospectively analyzed clinical data from 444 BCS patients with GEVs in two centers. Radiomic features were extracted from portal venous phase computed tomography (CT) scans. A training cohort of 334 patients was used to develop the model, with 110 patients serving as an external validation cohort. LASSO Cox regression was used to select radiomic features for constructing a radiomics score (Radscore). Univariate and multivariate Cox regression identified independent clinical predictors. A combined radiomics + clinical (R + C) model was developed using stepwise regression. Model performance was assessed using the area under the receiver operating characteristic curve (AUC), calibration plots, and decision curve analysis (DCA), with external validation to evaluate generalizability.

RESULTS

The Radscore comprised four hepatic and six splenic CT features, which predicted the risk of variceal bleeding. Multivariate analysis identified invasive treatment to relieve hepatic venous outflow obstruction, anticoagulant therapy, and hemoglobin levels as independent clinical predictors. The R + C model achieved C-indices of 0.906 (training) and 0.859 (validation), outperforming the radiomics and clinical models alone (AUC: training 0.936 vs 0.845 vs 0.823; validation 0.876 vs 0.712 vs 0.713). DCA showed higher clinical net benefit across the thresholds. The model stratified patients into low-, medium- and high-risk groups with significant differences in bleeding rates (P < 0.001). An online tool is available at https://bcsvh.shinyapps.io/BCS_Variceal_Bleeding_Risk_Tool/.

CONCLUSION

We developed and validated a novel radiomics-based model that noninvasively and conveniently predicted risk of variceal bleeding in BCS patients with GEVs, aiding early identification and management of high-risk patients.

Advanced feature fusion of radiomics and deep learning for accurate detection of wrist fractures on X-ray images

OBJECTIVE

The aim of this study was to develop a hybrid diagnostic framework integrating radiomic and deep features for accurate and reproducible detection and classification of wrist fractures using X-ray images.

MATERIALS AND METHODS

A total of 3,537 X-ray images, including 1,871 fracture and 1,666 non-fracture cases, were collected from three healthcare centers. Radiomic features were extracted using the PyRadiomics library, and deep features were derived from the bottleneck layer of an autoencoder. Both feature modalities underwent reliability assessment via Intraclass Correlation Coefficient (ICC) and cosine similarity. Feature selection methods, including ANOVA, Mutual Information (MI), Principal Component Analysis (PCA), and Recursive Feature Elimination (RFE), were applied to optimize the feature set. Classifiers such as XGBoost, CatBoost, Random Forest, and a Voting Classifier were used to evaluate diagnostic performance. The dataset was divided into training (70%) and testing (30%) sets, and metrics such as accuracy, sensitivity, and AUC-ROC were used for evaluation.

RESULTS

The combined radiomic and deep feature approach consistently outperformed standalone methods. The Voting Classifier paired with MI achieved the highest performance, with a test accuracy of 95%, sensitivity of 94%, and AUC-ROC of 96%. The end-to-end model achieved competitive results with an accuracy of 93% and AUC-ROC of 94%. SHAP analysis and t-SNE visualizations confirmed the interpretability and robustness of the selected features.

CONCLUSIONS

This hybrid framework demonstrates the potential for integrating radiomic and deep features to enhance diagnostic performance for wrist and forearm fractures, providing a reliable and interpretable solution suitable for clinical applications.

Association of [18F]-FDG PET/CT-Derived Radiomic Features with Clinical Outcomes and Genomic Profiles in Patients with Chronic Lymphocytic Leukemia

Background: The role of PET/CT imaging in chronic lymphoproliferative syndromes (CLL) is debated. This study examines the potential of PET/CT radiomics in predicting outcomes and genetic profiles in CLL patients. Methods: A retrospective analysis was conducted on 50 CLL patients treated at Policlinico Tor Vergata, Rome, and screened, at diagnosis, with [18F]-FDG PET/CT. Potentially pathological lymph nodes were semi-automatically segmented. Genetic mutations in TP53, NOTCH1, and IGVH were assessed. Eight hundred and sixty-five radiomic features were extracted, with the cohort split into training (70%) and validation (30%) sets. Four machine learning models, each with Random Forest, Stochastic Gradient Descent, and Support Vector Machine learners, were trained. Results: Progression occurred in 10 patients. The selected radiomic features from CT and PET datasets were correlated with four models of progression and mutations (TP53, NOTCH1, IGVH). The Random Forest models outperformed others in predicting progression (AUC = 0.94/0.88, CA = 0.87/0.75, TP = 80.00%/87.50%, TN = 72.70%/87.50%) and the occurrence of TP53 (AUC = 0.94/0.96, CA = 0.87/0.80, TP = 87.50%/90.21%, TN = 85.70%/90.90%), and NOTCH1 (AUC = 0.94/0.85, CA = 0.87/0.67, TP = 80.00%/88.90%, TN = 80.00%/83.30%)mutations. The IGVH models showed poorer performance. Conclusions: ML models based on PET/CT radiomic features effectively predict outcomes and genetic profiles in CLL patients.

Development and validation of ultrasound-based radiomics deep learning model to identify bone erosion in rheumatoid arthritis

OBJECTIVE

To develop and validate a deep learning radiomics fusion model (DLR) based on ultrasound (US) images to identify bone erosion in rheumatoid arthritis (RA) patients.

METHODS

A total of 432 patients with RA at two institutions were collected. Three hundred twelve patients from center 1 were randomly divided into a training set (N = 218) and an internal test set (N = 94) in a 7:3 ratio; meanwhile, 124 patients from center 2 were as an external test set. Radiomics (Rad) and deep learning (DL) features were extracted based on hand-crafted radiomics and deep transfer learning networks. The least absolute shrinkage and selection operator regression was employed to establish DLR fusion feature from the Rad and DL features. Subsequently, 10 machine learning algorithms were used to construct models and the final optimal model was selected. The performance of models was evaluated using receiver operating characteristic (ROC) and decision curve analysis (DCA). The diagnostic efficacy of sonographers was compared with and without the assistance of the optimal model.

RESULTS

LR was chosen as the optimal algorithm for model construction account for superior performance (Rad/DL/DLR: area under the curve [AUC] = 0.906/0.974/0.979) in the training set. In the internal test set, DLR_LR as the final model had the highest AUC (AUC = 0.966), which was also validated in the external test set (AUC = 0.932). With the aid of DLR_LR model, the overall performance of both junior and senior sonographers improved significantly (P < 0.05), and there was no significant difference between the junior sonographer with DLR_LR model assistance and the senior sonographer without assistance (P > 0.05).

CONCLUSION

DLR model based on US images is the best performer and is expected to become an important tool for identifying bone erosion in RA patients. Key Points • DLR model based on US images is the best performer in identifying BE in RA patients. • DLR model may assist the sonographers to improve the accuracy of BE evaluations.

MRI Radiomics to Predict Early Treatment Response to TACE Combined with Lenvatinib Plus a PD-1 Inhibitor for Hepatocellular Carcinoma with Portal Vein Tumor Thrombus

Purpose

To develop and validate a predictor for early treatment response in hepatocellular carcinoma (HCC) patients accompanied by portal vein tumor thrombus (PVTT) undergoing transarterial chemoembolization (TACE), lenvatinib and a programmed cell death protein 1 (PD-1) inhibitor (TLP) therapy.

Patients and Methods

In this retrospective study, patients with HCC and PVTT from two institutions receiving triple TLP therapy were enrolled. Radiomics features derived from pretreatment contrast-enhanced MRI were curated using intraclass correlation coefficient (ICC), Student's t-test, least absolute shrinkage and selection operator (LASSO), and recursive feature elimination (RFE) to ensure robust selection. Various machine learning (ML) algorithms were then used to construct the models. The meaningful clinical indicators were obtained via logistic regression analysis and ultimately integrated with radiomics features to develop a combined model. In addition, we used Shapley Additive exPlanation (SHAP) to clarify the model's operational dynamics.

Results

Our study ultimately included 115 patients (7:3 randomization, 80 and 35 in the training and test cohorts, respectively) in total. No patients achieved complete remission, 47 achieved partial remission, 29 achieved stable disease, and 39 experienced disease progression. Among objective response rates (ORRs) and disease control rates (DCRs), 40.9% and 66.1% were reported. One of the four ML classifiers with optimal performance, namely random forest, was adopted as the radiomics model after testing. Regarding the performance assessment, the radiomics model's area under the curve (AUC) values reached 0.92 (95% CI: 0.86-0.97) and 0.79 (95% CI: 0.61-0.95), inferior to the combined model's AUCs of 0.95 (95% CI: 0.68-0.98) and 0.84 (95% CI: 0.91-0.99). Moreover, the SHAP plots illustrate the importance of global variables and the prediction process for individual samples.

Conclusion

The model based on machine learning and radiomics showed favorable performance, and the operating mode was visualized through SHAP.

Deep learning MRI-based radiomic models for predicting recurrence in locally advanced nasopharyngeal carcinoma after neoadjuvant chemoradiotherapy: a multi-center study

Local recurrence and distant metastasis were a common manifestation of locoregionally advanced nasopharyngeal carcinoma (LA-NPC) after neoadjuvant chemoradiotherapy (NACT). To validate the clinical value of MRI radiomic models based on deep learning for predicting the recurrence of LA-NPC patients. A total of 328 NPC patients from four hospitals were retrospectively included and divided into the training(n = 229) and validation (n = 99) cohorts randomly. Extracting 975 traditional radiomic features and 1000 deep radiomic features from contrast enhanced T1-weighted (T1WI + C) and T2-weighted (T2WI) sequences, respectively. Least absolute shrinkage and selection operator (LASSO) was applied for feature selection. Five machine learning classifiers were conducted to develop three models for LA-NPC prediction in training cohort, namely Model I: traditional radiomic features, Model II: combined the deep radiomic features with Model I, and Model III: combined Model II with clinical features. The predictive performance of these models were evaluated by receive operating characteristic (ROC) curve analysis, area under the curve (AUC), accuracy, sensitivity and specificity in both cohorts. The clinical characteristics in two cohorts showed no significant differences. Choosing 15 radiomic features and 6 deep radiomic features from T1WI + C. Choosing 9 radiomic features and 6 deep radiomic features from T2WI. In T2WI, the Model II based on Random forest (RF) (AUC = 0.87) performed best compared with other models in validation cohort. Traditional radiomic model combined with deep radiomic features shows excellent predictive performance. It could be used assist clinical doctors to predict curative effect for LA-NPC patients after NACT.

Artificial Intelligence and Uterine Fibroids: A Useful Combination for Diagnosis and Treatment

This manuscript examines the role of artificial intelligence (AI) in the diagnosis and treatment of uterine fibroids and uterine sarcomas, offering a comprehensive assessment of AI-supported diagnostic and therapeutic techniques. Through the use of radiomics, machine learning, and deep neural network models, AI shows promise in identifying benign and malignant uterine lesions, directing therapeutic decisions, and improving diagnostic accuracy. It also demonstrates significant capabilities in the timely detection of fibroids. Additionally, AI improves surgical precision, real-time structure detection, and patient outcomes by transforming surgical techniques such as myomectomy, robot-assisted laparoscopic surgery, and High-Intensity Focused Ultrasound (HIFU) ablation. By helping to forecast treatment outcomes and monitor progress during procedures like uterine fibroid embolization, AI also offers a fresh and fascinating perspective for improving the clinical management of these conditions. This review critically assesses the current literature, identifies the advantages and limitations of various AI approaches, and provides future directions for research and clinical implementation.

Comparison of clinical, radiomics, deep learning, and fusion models for predicting early recurrence in locally advanced rectal cancer based on multiparametric MRI: a multicenter study

OBJECTIVE

Predicting early recurrence (ER) in locally advanced rectal cancer (LARC) is critical for clinical decision-making. This study aimed at comparing clinical, deep learning (DL), radiomics, and two fusion models for ER prediction based on multiparametric MRI.

METHODS

This retrospective study involved 337 LARC patients from four centers between January 2016 and September 2021. Radiomics and DL features were extracted from preoperative multiparametric MRI, including T2WI, DWI, T1WI, and contrast-enhanced T1WI (CET1WI). The extreme gradient boosting (XGBoost) classifier was applied to establish the clinical model, radiomics model, DL model, and two fusion models (the feature-based early fusion model and the decision-based late fusion model). The area under the curve (AUC), DeLong test, calibration curve, and decision curve analysis (DCA) were used to assess models. Kaplan-Meier analysis was conducted to determine the prognostic value of the models by evaluating the differences in recurrence-free survival (RFS) between the high- and low-risk patients of ER.

RESULTS

The late fusion model demonstrated the best performance compared with the early fusion model, clinical, radiomics and DL models, with the highest AUC (0.863-0.880) across all cohorts. In addition, the late fusion model exhibited the highest clinical net benefit, and good calibration. Kaplan-Meier survival curves showed that high-risk patients of ER defined by the late fusion model had a worse RFS than low-risk ones of ER (log-rank p < 0.001).

CONCLUSIONS

The late fusion model can accurately predict ER in LARC and may serve as a clinically useful, non-invasive tool for optimizing treatment strategies and monitoring disease progression.

Machine learning-based ultrasound radiomics for predicting risk of recurrence in breast cancer

Purpose

To develop a radiomics model based on ultrasound images for predicting risk of recurrence in breast cancer patients.

Methods

In this retrospective study, 420 patients with pathologically confirmed breast cancer were included, randomly divided into training (70%) and test (30%) sets, with an independent external validation cohort of 90 patients. According to St. Gallen recurrence risk criteria, patients were categorized into two groups, low-medium-risk and high-risk. Radiomics features were extracted from a radiomics analysis set using Pyradiomics. The informative radiomics features were screened using the minimum redundancy maximum relevance (mRMR) and the least absolute shrinkage and selection operator (LASSO) algorithms. Subsequently, radiomics models were constructed with eight machine learning algorithms. Three distinct nomogram models were created using the features selected through multivariate logistic regression, including the Clinic-Ultrasound (Clin-US), Clinic-Radiomics (Clin-Rad), and Clinic-Ultrasound-Radiomics (Clin-US-Rad) models. The receiver operating characteristic (ROC), calibration, and decision curve analysis (DCA) curves were used to evaluate the model's clinical applicability and predictive performance.

Results

A total of 12 ultrasound radiomics features were screened, of which wavelet.LHL first order Mean features weighed more and tended to have a high risk of recurrence. The higher the risk of recurrence, the higher the radiomics score (Rad-score) in all three sets (training, test, and external validation set, all p < 0.05). Rad-score is equally applicable in four different subtypes of breast cancer. In the test set and external validation set, the Clin-US-Rad model achieved the highest AUC values (AUC = 0.817 and 0.851, respectively). The calibration and DCA curves also demonstrated the good clinical utility of the combined model.

Conclusion

The machine learning-based ultrasound radiomics model were useful for predicting the risk of recurrence in breast cancer. The nomograms show promising potential in assessing the recurrence risk of breast cancer. This non-invasive approach offers crucial guidance for the diagnosis and treatment of the condition.

Preoperative radiomics models using CT and MRI for microsatellite instability in colorectal cancer: a systematic review and meta-analysis

OBJECTIVE

Microsatellite instability (MSI) is a novel predictive biomarker for chemotherapy and immunotherapy response, as well as prognostic indicator in colorectal cancer (CRC). The current standard for MSI identification is polymerase chain reaction (PCR) testing or the immunohistochemical analysis of tumor biopsy samples. However, tumor heterogeneity and procedure complications pose challenges to these techniques. CT and MRI-based radiomics models offer a promising non-invasive approach for this purpose.

MATERIALS AND METHODS

A systematic search of PubMed, Embase, Cochrane Library and Scopus was conducted to identify studies evaluating the diagnostic performance of CT and MRI-based radiomics models for detecting MSI status in CRC. Pooled area under the curve (AUC), sensitivity, and specificity were calculated in RStudio using a random-effects model. Forest plots and a summary ROC curve were generated. Heterogeneity was assessed using I² statistics and explored through sensitivity analyses, threshold effect assessment, subgroup analyses and meta-regression.

RESULTS

17 studies with a total of 6,045 subjects were included in the analysis. All studies extracted radiomic features from CT or MRI images of CRC patients with confirmed MSI status to train machine learning models. The pooled AUC was 0.815 (95% CI: 0.784-0.840) for CT-based studies and 0.900 (95% CI: 0.819-0.943) for MRI-based studies. Significant heterogeneity was identified and addressed through extensive analysis.

CONCLUSION

Radiomics models represent a novel and promising tool for predicting MSI status in CRC patients. These findings may serve as a foundation for future studies aimed at developing and validating improved models, ultimately enhancing the diagnosis, treatment, and prognosis of colorectal cancer.

CT-based quantification of intratumoral heterogeneity for predicting distant metastasis in retroperitoneal sarcoma

OBJECTIVES

Retroperitoneal sarcoma (RPS) is highly heterogeneous, leading to different risks of distant metastasis (DM) among patients with the same clinical stage. This study aims to develop a quantitative method for assessing intratumoral heterogeneity (ITH) using preoperative contrast-enhanced CT (CECT) scans and evaluate its ability to predict DM risk.

METHODS

We conducted a retrospective analysis of 274 PRS patients who underwent complete surgical resection and were monitored for ≥ 36 months at two centers. Conventional radiomics (C-radiomics), ITH radiomics, and deep-learning (DL) features were extracted from the preoperative CECT scans and developed single-modality models. Clinical indicators and high-throughput CECT features were integrated to develop a combined model for predicting DM. The performance of the models was evaluated by measuring the receiver operating characteristic curve and Harrell's concordance index (C-index). Distant metastasis-free survival (DMFS) was also predicted to further assess survival benefits.

RESULTS

The ITH model demonstrated satisfactory predictive capability for DM in internal and external validation cohorts (AUC: 0.735, 0.765; C-index: 0.691, 0.729). The combined model that combined clinicoradiological variables, ITH-score, and DL-score achieved the best predictive performance in internal and external validation cohorts (AUC: 0.864, 0.801; C-index: 0.770, 0.752), successfully stratified patients into high- and low-risk groups for DM (p < 0.05).

CONCLUSIONS

The combined model demonstrated promising potential for accurately predicting the DM risk and stratifying the DMFS risk in RPS patients undergoing complete surgical resection, providing a valuable tool for guiding treatment decisions and follow-up strategies.

CRITICAL RELEVANCE STATEMENT

The intratumoral heterogeneity analysis facilitates the identification of high-risk retroperitoneal sarcoma patients prone to distant metastasis and poor prognoses, enabling the selection of candidates for more aggressive surgical and post-surgical interventions.

KEY POINTS

Preoperative identification of retroperitoneal sarcoma (RPS) with a high potential for distant metastasis (DM) is crucial for targeted interventional strategies. Quantitative assessment of intratumoral heterogeneity achieved reasonable performance for predicting DM. The integrated model combining clinicoradiological variables, ITH radiomics, and deep-learning features effectively predicted distant metastasis-free survival.

Using baseline MRI radiomic features to predict the efficacy of repetitive transcranial magnetic stimulation in Alzheimer's patients

The efficacy of repetitive transcranial magnetic stimulation (rTMS) as a treatment for Alzheimer's disease (AD) is uncertain at baseline. Herein, we aimed to investigate whether radiomic features from the pre-treatment MRI data could predict rTMS efficacy for AD treatment. Out of 110 participants with AD in the active (n = 75) and sham (n = 35) rTMS treatment groups having T1-weighted brain MRI data, we had two groups of responders (active = 55 and sham = 24) and non-responders (active = 20 and sham = 11). We extracted histogram-based radiomic features from MRI data using 3D Slicer software; the most important features were selected utilizing a combination of a two-sample t-test, correlation test, least absolute shrinkage, and selection operator. The support vector machine classified rTMS responders and non-responders with a cross-validated mean accuracy/AUC of 81.9%/90.0% in the active group and 87.4%/95.8% in the sham group. Further, the radiomic features of the active group significantly correlated with participants' AD assessment scale-cognitive subscale (ADAS-Cog) change after treatment (false discovery rate corrected p < 0.05). Given that baseline radiomic features were able to accurately predict AD patients' responses to rTMS treatment, these radiomic features warrant further investigation for personalizing AD therapeutic strategies.

Ultrasound-based deep learning radiomics for enhanced axillary lymph node metastasis assessment: a multicenter study

BACKGROUND

Accurate preoperative assessment of axillary lymph node metastasis (ALNM) in breast cancer is crucial for guiding treatment decisions. This study aimed to develop a deep-learning radiomics model for assessing ALNM and to evaluate its impact on radiologists' diagnostic accuracy.

METHODS

This multicenter study included 866 breast cancer patients from 6 hospitals. The data were categorized into training, internal test, external test, and prospective test sets. Deep learning and handcrafted radiomics features were extracted from ultrasound images of primary tumors and lymph nodes. The tumor score and LN score were calculated following feature selection, and a clinical-radiomics model was constructed based on these scores along with clinical-ultrasonic risk factors. The model's performance was validated across the 3 test sets. Additionally, the diagnostic performance of radiologists, with and without model assistance, was evaluated.

RESULTS

The clinical-radiomics model demonstrated robust discrimination with AUCs of 0.94, 0.92, 0.91, and 0.95 in the training, internal test, external test, and prospective test sets, respectively. It surpassed the clinical model and single score in all sets (P < .05). Decision curve analysis and clinical impact curves validated the clinical utility of the clinical-radiomics model. Moreover, the model significantly improved radiologists' diagnostic accuracy, with AUCs increasing from 0.71 to 0.82 for the junior radiologist and from 0.75 to 0.85 for the senior radiologist.

CONCLUSIONS

The clinical-radiomics model effectively predicts ALNM in breast cancer patients using noninvasive ultrasound features. Additionally, it enhances radiologists' diagnostic accuracy, potentially optimizing resource allocation in breast cancer management.

Practical use of radiomic features as a metric for image quality discrimination in [18F] FDG-PET: a pilot study

PURPOSE

Radiomics features have been utilised as group metrics of image quality in many areas of diagnostic radiology. In this pilot study, the relationship between technical metrics used in image quality assurance and visual grading scores provided by a radiologist were evaluated. Image dataset harmonisation allowed comparison between the two and allowed trends to be extracted. We propose a reproducible technique to identify the metrics.

METHODS

A retrospective chart review of 30 [18F] FDG-PET/CT performed in a nuclear medicine referral centre was performed. Image datasets were reprocessed to correspond to a bed duration of 180, 120, 60, 30 s per bed position and were analysed according to a pre-set bank of semi-quantitative features by a radiology resident. The extraction of radiomic features in PET images was performed using SLICER-RADIOMICS Module version 5.2.2. To facilitate the comparison of radiomic features and radiologist scoring data, normalisation was performed on both data sets. Fréchet distance analysis, Mean Square Error and Mean Absolute Error display the level of agreement between features and radiologist following the rescale of the data.

RESULTS

Of the 120 reprocessed image datasets, 115 were included in the study. We focused on overall image quality score rather than individual radiomic metrics as this identified the most robust trend. A significant difference in the 30 s image dataset with respect to each group individually and combined for the radiologist overall score was observed.

CONCLUSION

Our results show that a large percentage change in certain features can indicate a significant change in quality in clinically processed images.

An overview of the applications of AI for detecting anatomical configurations in endodontics

BACKGROUND

Artificial intelligence (AI), which uses algorithms to replicate human intellect, allows robots to learn from data and complete complex tasks on their own.

STUDY DESIGN

Mini-narrative review.

OBJECTIVE

This mini-narrative review evaluates AI's potential to improve the detection of anatomical features of root and canal systems, focusing on its benefits, challenges, and future applications.

METHODS

A comprehensive literature search was conducted using PubMed, Scopus, and Google Scholar to identify studies on AI applications in detecting anatomical features of root and root canal systems in endodontics. Inclusion criteria encompassed all relevant literature focused on anatomical feature detection, with no restrictions on time or language. Studies were excluded if they were unrelated to the topic or focused on pathological rather than anatomical feature detection.

CONCLUSION

AI has significantly improved the detection of root and canal anatomy, including minor constrictions, working length, second mesio-buccal canals, and complex systems like C-shaped canals, with diagnostic accuracy comparable to or surpassing experienced practitioners. While challenges remain in technology, ethics, and regulation, AI enhances precision, efficiency, and patient outcomes. Addressing these hurdles will further advance its integration into endodontic practice and shape its future positively.

Decoding the Rotation Effect: A Retrospective Analysis of Lesion Orientation and Its Impact on Wavelet-Based Radiomics Feature Extraction and Lung Cancer Classification

Wavelet decomposition (WD), widely used in radiomics, redistributes information among derived wavelet components when the input is rotated. This redistribution may alter predictions for the same lesion when scanned at different angles. Despite its potential significance, this vulnerability has frequently been overlooked in radiomic studies while its impact remains poorly understood. Therefore, this study aims to investigate how variations in lesion orientation affect both WD and non-WD radiomic feature values, and subsequently, model performance. We analyzed CT radiomics of primary non-small-cell lung cancer (NSCLC). Prior to feature extraction, we introduced random rotations ranging from 5° to 80° to the tumors. Their effects were quantified by evaluating the percentage difference ( % Δ ) between the rotated and unrotated feature values, and validated using Spearman's rank test. Additionally, radiomics models were trained to discriminate between three histological subtypes of NSCLC using the original features, and then tested on rotated inputs. The correlation between the model accuracies and the degree of rotation was again evaluated using Spearman's rank test. Four-hundred nineteen NSCLC patients (mean age: 68.1 ± 10.1, 289 men) were evaluated. Significant correlations between feature values and rotations (Spearman's correlation [CC] magnitude ≥ 0.1, p < .05) were found in 23.7% (176/744) of the WD and 0.5% (5/930) of the non-WD texture features. Significant association between performance and rotation was observed in WD-based models built to discriminate between NSCLC histological subtypes (CC =  - 0.44, p < .001) but not in non-WD-based models (CC = 0.03, p = 0.07). Input lesion orientation affects radiomic feature values and model reproducibility. WD features exhibited significantly greater instability to orientation variations compared to non-WD features.

Preoperative prediction of WHO/ISUP grade of ccRCC using intratumoral and peritumoral habitat imaging: multicenter study

OBJECTIVES

The World Health Organization/International Society of Urological Pathology (WHO/ISUP) grading of clear cell renal cell carcinoma (ccRCC) is crucial for prognosis and treatment planning. This study aims to predict the grade using intratumoral and peritumoral subregional CT radiomics analysis for better clinical interventions.

METHODS

Data from two hospitals included 513 ccRCC patients, who were divided into training (70%), validation (30%), and an external validation set (testing) of 67 patients. Using ITK-SNAP, two radiologists annotated tumor regions of interest (ROI) and extended surrounding areas by 1 mm, 3 mm, and 5 mm. The K-means clustering algorithm divided the tumor region into three sub-regions, and the Least Absolute Shrinkage and Selection Operator (LASSO) regression identified the most predictive features. Various machine learning models were established, including radiomics models, peritumoral radiomics models, models based on intratumoral heterogeneity (ITH) score, clinical models, and comprehensive models. Predictive ability was evaluated using receiver operating characteristic (ROC) curves, area under the curve (AUC) values, DeLong tests, calibration curves, and decision curves.

RESULTS

The combined model showed strong predictive power with an AUC of 0.852 (95% CI: 0.725-0.979) on the test data, outperforming individual models. The ITH score model was highly precise, with AUCs of 0.891 (95% CI: 0.854-0.927) in training, 0.877 (95% CI: 0.814-0.941) in validation, and 0.847 (95% CI: 0.725-0.969) in testing, proving its superior predictive ability across datasets.

CONCLUSION

A comprehensive model combining Habitat, Peri1mm, and salient clinical features was significantly more accurate in predicting ccRCC pathologic grading.

KEY POINTS

Question: Characterize tumor heterogeneity to non-invasively predict WHO/ISUP pathological grading preoperatively.

FINDINGS

An integrated model combining subregion characterization, peritumoral characteristics, and clinical features can predict ccRCC grade preoperatively.

CLINICAL RELEVANCE

Subregion tumor characterization outperforms the single-entity approach. The integrated model, compared with the radiomics model, boosts grading and prognostic accuracy for more targeted clinical actions.

Echocardiographic Tissue Characterization Using Radiomics in Patients With Transthyretin-Related Cardiac Amyloidosis

BACKGROUND

Transthyretin-related cardiac amyloidosis (ATTR-CA) is often diagnosed at an advanced stage. Emerging evidence suggests that radiomics applied to echocardiographic images (ie, ultrasonomics) can detect early myocardial texture changes in ATTR-CA.

OBJECTIVES

This study aimed to develop a radiomic model for characterizing ATTR-infiltrated myocardium via echocardiography.

METHODS

Echocardiographic images in parasternal long-axis and apical 4-chamber views from ATTR-CA and control patients were collected across 4 Italian centers. A region of interest (ROI) within the interventricular septum was delineated. Ninety-four radiomic features were extracted and classified into 2 categories for analysis, based on whether they were ROI-dependent or independent. Five logistic regression models analyzed data from 3 centers (229 ATTR-CA, 224 controls) to assess diagnostic accuracy and area under the curve (AUC) of different sets of radiomic features, with external validation conducted on patients from a fourth center (32 ATTR-CA, 32 controls).

RESULTS

Models analyzing the entire ROI using both ROI-dependent and ROI-independent features demonstrated high cross-validated accuracies (93%-95%) and AUC values (0.97-0.99). Using a fixed-size 0.5 × 0.5 cm ROI, these values decreased to 85% and 0.91, respectively, highlighting previous models' dependence on ROI size. The fifth model used 73 ROI-independent features on the entire ROI and demonstrated significantly better accuracy and AUC (92% and 0.97, respectively, P < 0.001), confirmed in the external validation cohort (87% and 0.95, respectively). Removing the least informative features slightly improved the model, achieving 90% accuracy and 0.95 precision.

CONCLUSIONS

This study showcases ultrasonomics potential to differentiate ATTR-CA and control patients by capturing disease-specific textural features independent of ROI dimensions.

Machine learning models for enhanced diagnosis and risk assessment of prostate cancer with 68Ga-PSMA-617 PET/CT

OBJECTIVE

Prostate cancer (PCa) is highly heterogeneous, making early detection of adverse pathological features crucial for improving patient outcomes. This study aims to predict PCa aggressiveness and identify radiomic and protein biomarkers associated with poor pathology, ultimately developing a multi-omics marker model for better clinical risk stratification.

METHODS

In this retrospective study, 191 patients with PCa or benign prostatic hyperplasia confirmed via 68Ga-PSMA-617 PET/CT scans were analyzed. Radiomic features were extracted from scan contours, and six machine learning algorithms were used to predict malignancy and adverse pathological features like Gleason score, ISUP group, tumor stage, lymph node infiltration, and perineural invasion. Feature selection and dimensionality reduction were performed using minimum redundancy maximum relevance and least absolute shrinkage and selection operator methods. Proteomics analysis on 39 patients identified protein biomarkers, followed by correlation analysis between radiomic features and identified proteins.

RESULTS

The radiomics model showed an AUC of 0.938 for predicting malignant prostate lesions and 0.916 for adverse pathological features in the test set, with validation set AUCs of 0.918 and 0.855, respectively. Three quantitative radiomic features and ten protein molecules associated with adverse pathology were identified, with significant correlations observed between radiomic features and protein biomarkers. Radioproteomic analysis revealed that molecular changes in protein molecules could influence imaging biomarkers.

CONCLUSIONS

The machine learning models based on 68 Ga-PSMA-617 PET/CT radiomic features performed well in stratifying patients, supporting clinical risk stratification and highlighting connections between radiomic characteristics and protein biomarkers.

Novel MRI-based Hyper-Fused Radiomics for Predicting Pathologic Complete Response to Neoadjuvant Therapy in Breast Cancer

RATIONALE AND OBJECTIVES

To propose a novel MRI-based hyper-fused radiomic approach to predict pathologic complete response (pCR) to neoadjuvant therapy (NAT) in breast cancer (BC).

MATERIALS AND METHODS

Pretreatment dynamic contrast-enhanced (DCE) MRI and ultra-multi-b-value (UMB) diffusion-weighted imaging (DWI) data were acquired in BC patients who received NAT followed by surgery at two centers. Hyper-fused radiomic features (RFs) and conventional RFs were extracted from DCE-MRI or UMB-DWI. After feature selection, the following models were built using logistic regression and the retained RFs: hyper-fused model, conventional model, and compound model that integrates the hyper-fused and conventional RFs. The output probability of each model was used to generate a radiomic signature. The model's performance was quantified by the area under the receiver-operating characteristic curve (AUC). Multivariable logistic regression was used to identify variables (clinicopathological variables and the generated radiomic signatures) associated with pCR.

RESULTS

The training/external test set (center 1/2) included 547/295 women. The hyper-fused models (AUCs=0.81-0.85) outperformed (p<0.05) the conventional models (AUCs=0.74-0.80) in predicting pCR. The compound models (AUCs=0.88-0.93) outperformed (p<0.05) the hyper-fused models and conventional models for pCR prediction. The hyper-fused radiomic signatures (odds ratios=5.70-12.98; p<0.05) and compound radiomic signatures (odds ratios=1.57-7.71; p<0.05) were independently associated with pCR. These are true for the training and external test sets.

CONCLUSION

The hyper-fused radiomic approach had significantly better performance for predicting pCR to NAT than the conventional radiomic approach, and the hyper-fused RFs provided incremental discrimination of pCR beyond the conventional RFs. The generated hyper-fused radiomic signatures were independent predictors of pCR.

CT-based radiomics model for predicting perineural invasion status in gastric cancer

PURPOSE

Perineural invasion (PNI) is an independent risk factor for poor prognosis in gastric cancer (GC) patients. This study aimed to develop and validate predictive models based on CT imaging and clinical features to predict PNI status in GC patients.

METHODS

This retrospective study included 291 GC patients (229 in the training cohort and 62 in the validation cohort) who underwent gastrectomy between January 2020 and August 2022. The clinical data and preoperative abdominal contrast-enhanced computed tomography (CECT) images were collected. Radiomics features were extracted from the venous phase of CECT images. The intraclass correlation coefficient (ICC), Pearson correlation coefficient, and t-test were applied for radiomics feature selection. The random forest algorithm was used to construct a radiomics signature and calculate the radiomics feature score (Rad-score). A hybrid model was built by aggregating the Rad-score and clinical predictors. The area under the receiver operating characteristic curve (ROC) and decision curve analysis (DCA) were used to evaluate the prediction performance of the radiomics, clinical, and hybrid models.

RESULTS

A total of 994 radiomics features were extracted from the venous phase images of each patient. Finally, 5 radiomics features were selected and used to construct a radiomics signature. The hybrid model demonstrated strong predictive ability for PNI, with AUCs of 0.833 (95% CI: 0.779-0.887) and 0.806 (95% CI: 0.628-0.983) in the training and validation cohorts, respectively. The DCA showed that the hybrid model had good clinical utility.

CONCLUSION

We established three models, and the hybrid model that combined the Rad-score and clinical predictors had a high potential for predicting PNI in GC patients.

Radiomics from dual-energy CT-derived iodine maps for predicting lymph node metastases in patients with resectable rectal cancer

BackgroundLymph node metastasis (LNM) is a poor prognostic predictor and is highly correlated with local recurrence in rectal cancer patients.ObjectiveTo investigate the value of radiomics from dual-energy CT-derived iodine maps for the preoperative prediction of LNM in rectal cancer patients.MethodsA total of 176 patients were enrolled in this study (training group, n = 123; validation group, n = 53). A radiomic signature was constructed via support vector machine (SVM) modeling. Seven models, including a clinical feature model (Model 1), an arterial model (Model 2), a venous model (Model 3), an arterial-venous model (Model 4), an arterial-clinical model (Model 5), a venous-clinical model (Model 6) and an arterial-venous-clinical model (Model 7), were established via logistic regression modeling. Diagnostic performance was assessed via receiver operating characteristic (ROC) curves.ResultsTumor location and carcinoembryonic antigen levels were used to construct Model 1 (training group, AUC [area under the ROC curve] = 0.721, 95% CI [confidence intervals], 0.630-0.813; validation group, AUC = 0.729, 95% CI, 0.593-0.865). Model 6 and Model 7 further improved the discriminatory performance in the training (AUC = 0.850 and 0.869, 95% CI, 0.782-0.919 and 0.807-0.932, respectively; p = 0.250) and validation groups (AUC = 0.780 and 0.716, 95% CI, 0.653-0.906 and 0.576-0.856, respectively; p = 0.115). Moreover, decision curve analysis revealed a greater net benefit with Model 6.ConclusionsThe combination of radiomic features based on dual-energy CT-derived iodine maps and clinical features provides better diagnostic performance for predicting LNM in rectal cancer patients.

The Value of Whole-Volume Radiomics Machine Learning Model Based on Multiparametric MRI in Predicting Triple-Negative Breast Cancer

OBJECTIVE

This study aimed to investigate the value of radiomics analysis in the precise diagnosis of triple-negative breast cancer (TNBC) based on breast dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) and apparent diffusion coefficient (ADC) maps.

METHODS

This retrospective study included 326 patients with pathologically proven breast cancer (TNBC: 129, non-TNBC: 197). The lesions were segmented using the ITK-SNAP software, and whole-volume radiomics features were extracted using a radiomics platform. Radiomics features were obtained from DCE-MRI and ADC maps. The least absolute shrinkage and selection operator regression method was employed for feature selection. Three prediction models were constructed using a support vector machine classifier: Model A (based on the selected features of the ADC maps), Model B (based on the selected features of DCE-MRI), and Model C (based on the selected features of both combined). Receiver operating characteristic curves were used to evaluate the diagnostic performance of the conventional MR image model and the 3 radiomics models in predicting TNBC.

RESULTS

In the training dataset, the AUCs for the conventional MR image model and the 3 radiomics models were 0.749, 0.801, 0.847, and 0.896. The AUCs for the conventional MR image model and 3 radiomics models in the validation dataset were 0.693, 0.742, 0.793, and 0.876, respectively.

CONCLUSIONS

Radiomics based on the combination of whole volume DCE-MRI and ADC maps is a promising tool for distinguishing between TNBC and non-TNBC.

Infection and Inflammation in Nuclear Medicine Imaging: The Role of Artificial Intelligence

Infectious and inflammatory diseases represent a global challenge. Delayed diagnosis and treatment lead to death, disabilities and impairment of the quality of life. The detection of low-grade inflammation and occult infections remains challenging. Nuclear medicine techniques are well established in the assessment of the severity and extent of the disease. However, high-level expertise is required to process and interpret the images. Additionally, the workflows are frequently time consuming. Artificial intelligence (AI)-based techniques can be efficiently applied in this setting. We reviewed the literature to assess the state of the application of AI in nuclear medicine imaging in infectious and inflammatory diseases. We included 22 studies, which applied AI-based methods for any of the steps of their workflow. In this review we report and critically discuss the state-of-the-art knowledge on the application of AI models in Infection and Inflammation nuclear medicine imaging.

Development and validation of a multivariable risk model based on clinicopathological characteristics, mammography, and MRI imaging features for predicting axillary lymph node metastasis in patients with upgraded ductal carcinoma in situ

Background

Axillary surgical staging is required for patients with upgraded ductal carcinoma in situ (DCIS) (DCIS is diagnosed on core biopsy with invasive cancer found on pathology after complete surgical excision), which may lead to complications in axillary surgery. At present, there is no reliable and accurate method for predicting axillary lymph node metastasis (ALNM) in patients with upgraded DCIS; however, such a method could prevent unnecessary axillary surgical interventions from being performed. In this study, we aimed to construct a non-invasive model for predicting ALNM in DCIS patients based on clinicopathological characteristics, mammography (MG) features, and magnetic resonance imaging (MRI) features.

Methods

Between February 2018 and June 2020, 326 patients with upgraded DCIS were enrolled in this retrospective analysis. These patients were randomly divided into the training cohort (80%) and validation cohort (20%). Univariate and multivariable regression analyses were conducted to identify the candidate pathological features, which then used to develop a clinicopathological model. The features of the 2-mm, 4-mm, and 6-mm intratumoral and peritumoral regions (T-PTR) were extracted to develop the MRI radiomics model, and two deep learning classification models were developed based on the medial-lateral oblique (MLO) and craniocaudal (CC) views of the MG. A fusion model was then established that combined these sub-models. The receiver operating characteristic (ROC) curve, area under the curve (AUC), and other indicators were used to evaluate the performance of these models.

Results

The clinicopathological characteristics of the two cohorts were basically balanced. The AUC values of the clinicopathological model were 0.675 and 0.690 in the training and validation cohorts, respectively. The model based on the T-PTR of MRI showed promising predictive ability. Among the three MRI models, the T-PTR (4 mm) model showed the best predictivity both in the training (AUC =0.885) and validation cohorts (AUC =0.843). The AUC values for the deep learning models of the MG CC and MLO positions all exceeded 0.7, indicating reliable predictive performance. The fusion model that combined the three methods significantly improved the accuracy and robustness of ALNM prediction. In both the training (AUC =0.975) and validation (AUC =0.877) cohorts, the fusion model showed excellent performance.

Conclusions

We developed a fusion model that combined clinicopathological characteristics, MRI T-PTR (4 mm) radiomics, and MG-based deep learning. Our combined model showed promising performance in predicting ALNM in patients with upgraded DCIS.

Machine learning algorithms for predicting malignancy grades of lung adenocarcinoma and guiding treatments: CT radiomics-based comparisons

Background

Lung adenocarcinoma (LUAD) is the most frequently diagnosed subtype of non-small cell lung cancer (NSCLC). Notably, prognosis can vary significantly among LUAD patients with different tumor subtypes. The advent of radiomics and machine learning (ML) technologies enables the development of non-invasive pathology predictive models. We attempted to develop computed tomography (CT) radiomics-based diagnostic models, enhanced by ML, to predict LUAD malignancy grade and guide surgical strategies.

Methods

In this retrospective analysis, a total of 168 surgical patients with histology-confirmed LUAD were divided into low-risk group (n=93) and intermediate-to-high-risk group (n=75) based on postoperative pathology. The region of interest (ROI) was delineated on the preoperative CT images for all patients, followed by the extraction of radiomic features. Patients were randomly allocated to a training set (n=117) and a testing set (n=51) in a 7:3 ratio. Within the training set, clinical-radiological model (CM) and radiomics model (RM) were developed utilizing patients' clinical characteristics, radiological semantic features, and radiomic features, along with the calculation of Rad scores. After the Rad scores were combined with independent risk factors among clinical-radiological features, logistic regression (LR), decision tree (DT), random forest (RF), extreme gradient boosting (XGBoost), support vector machine (SVM), K-nearest neighbors (KNN), and naïve Bayes model (NBM) were employed to create different comprehensive models (COMs). The optimal model was identified based on the receiver operating characteristic (ROC) curves and the DeLong test. Finally, Shapley additive explanations (SHAP) were utilized to visualize the predictive processes of the models.

Results

Among the 168 patients enrolled, there were 50 males (29.76%) aged 56 (49.25, 67.00) years and 118 females (70.24%) aged 56.5 (42.00, 64.00) years; Diameter (P<0.001), and consolidation-to-tumor ratio (CTR) ≥0.5 (P=0.002) were identified as independent risk factors for the malignancy degree of LUAD during CM creation. The CM had an area under the ROC curve (AUC) of 0.909 [95% confidence interval (CI): 0.856-0.962] in the training set and 0.920 (95% CI: 0.846-0.994) in the testing set. The RM, comprising seven radiomic features, achieved an AUC of 0.961 (95% CI: 0.926-0.996) in the training set and 0.957 (95% CI: 0.905-1.000) in the testing set. Among models created using various ML algorithms, the XGBoost model was identified as the optimal model. SHAP visualization revealed the model prediction processes and the values of different features.

Conclusions

We constructed and validated a robust, integrative model leveraging ML and CT radiomics, which amalgamates radiomics, clinical, and radiological attributes to precisely identify LUADs with elevated postoperative pathological grades. This enables doctors to formulate different surgical plans according to the pathology of the patients' tumors before the operation.

Development of a radiomic model to predict CEACAM1 expression and prognosis in ovarian cancer

We aimed to investigate the prognostic role of CEACAM1 and to construct a radiomic model to predict CEACAM1 expression and prognosis in ovary cancer (OC). Sequencing data and CT scans in OC were sourced from TCGA and TCIA databases. CEACAM1 expression was assessed by Cox regression analyses, Kaplan-Meier curves and GSVA enrichment analysis. Furthermore, radiomic features were extracted from CT scans and selected by LASSO and ROC. The selected radiomic features were used to construct a radiomic model to predict CEACAM1 expression. In addition, the radiomic score (RS) and its relationship with OC survival were investigated by Kaplan-Meier and ROC curves. At last, RS and clinical features were included into LASSO, using nomogram to predict OC prognosis. Cox regression analyses showed that CEACAM1 expression was an independent prognostic factor and associated with immune cell infiltration in OC. By LASSO and ROC, six radiomic features were selected and used to construct a radiomic model. The PR, calibration, DCA and ROC curves revealed the good performance and clinical utility of the radiomic model to predict CEACAM1 expression. In addition, RS based on radiomic features was found to be associated with OC survival. At last, a nomogram based on RS, age, chemotherapy and tumor residual disease was constructed and was found to have high accuracy in predicting OC prognosis. For the first time, our study constructed a radiomic model to predict CEACAM1 expression and prognosis of OC patients. Those findings may guide novel diagnosis and treatment for OC patients.

Integrated clinical-radiomic model for predicting treatment response of concurrent chemo-radiotherapy and radiotherapy alone in controversial subgroup of AJCC/UICC ninth edition stage I nasopharyngeal cancer

Objective

Radiotherapy (RT) is the definitive treatment for stage II nasopharyngeal carcinoma (NPC), which is classified as stages IA and IB in the latest ninth edition of American Joint Committee on Cancer (AJCC)/Union for International Cancer Control (UICC). A crucial question is whether concurrent chemo-radiotherapy (CCRT) could derive additional benefits to this recent "down-staging" subgroup of NPC patients. This study aimed to interrogate clinical and radiomic features for predicting 5-year progression-free survival (PFS) of stage II NPC treated with RT alone or CCRT.

Methods

Imaging and clinical data of 166 stage II NPC (eighth edition AJCC/UICC) patients were collected. Data were allocated into training, internal testing, and external testing sets. For each case, 851 radiomic features were extracted and 10 clinical features were collected. Radiomic and clinical features most associated with the 5-year PFS were selected separately. A combined model was developed using multivariate logistic regression by integrating selected features and treatment option to predict 5-year PFS. Model performances were evaluated by area under the receiver operating curve (AUC), prediction accuracy, and decision curve analysis. Survival analyses including Kaplan-Meier analysis and Cox regression model were performed for further analysis.

Results

Thirteen radiomic features, three clinical features, and treatment option were considered for model development. The combined model showed higher prognostic performance than using either. For the merged testing set (internal and external testing sets), AUC is 0.76 (combined) vs. 0.56-0.80 (clinical or radiomic alone) and accuracy is 0.75 (combined) vs. 0.62-0.73 (clinical or radiomic alone). Kaplan-Meier analysis using the combined model showed significant discrimination in PFS of the predicted low-risk and high-risk groups in the training and internal testing cohorts (P<0.05).

Conclusions

Integrating with clinical and radiomic features could provide prognostic information on 5-year PFS under either treatment regimen, guiding individualized decisions of chemotherapy based on the predicted treatment outcome.

Unsupervised machine learning analysis of optical coherence tomography radiomics features for predicting treatment outcomes in diabetic macular edema

This study aimed to identify distinct clusters of diabetic macular edema (DME) patients with differential anti-vascular endothelial growth factor (VEGF) treatment outcomes using an unsupervised machine learning (ML) approach based on radiomic features extracted from pre-treatment optical coherence tomography (OCT) images. Retrospective data from 234 eyes with DME treated with three anti-VEGF therapies between January 2020 and March 2024 were collected from two clinical centers. Radiomic analysis was conducted on pre-treatment OCT images. Following principal component analysis (PCA) for dimensionality reduction, two unsupervised clustering methods (K-means and hierarchical clustering) were applied. Baseline characteristics and treatment outcomes were compared across clusters to assess clustering efficacy. Feature selection employed a three-stage pipeline: exclusion of collinear features (Pearson's r > 0.8); sequential filtering through ANOVA (P < 0.05) and Boruta algorithm (500 iterations); multivariate stepwise regression (entry criteria: univariate P < 0.1) to identify outcome-associated predictors. From 1165 extracted radiomic features, four distinct DME clusters were identified. Cluster 4 exhibited a significantly lower incidence of residual/recurrent DME (RDME) (34.29%) compared to Clusters 1-3 (P = 0.003, P = 0.005 and P = 0.002, respectively). This cluster also demonstrated the highest proportion of eyes (71.43%) with best-corrected visual acuity (BCVA) exceeding 20/63 (P = 0.003, P = 0.005 and P = 0.002, respectively). Multivariate analysis identified logarithm_gldm_DependenceVariance as an independent risk factor for RDME (OR 1.75, 95% CI 1.28-2.40; P < 0.001), while Wavelet-LH_Firstorder_Mean correlated with worse visual outcomes (OR 8.76, 95% CI 1.22-62.84; P = 0.031). Unsupervised ML leveraging pre-treatment OCT radiomics successfully stratifies DME eyes into clinically distinct subgroups with divergent therapeutic responses. These quantitative features may serve as non-invasive biomarkers for personalized outcome prediction and retinal pathology assessment.

Personalized predictions of neoadjuvant chemotherapy response in breast cancer using machine learning and full-field digital mammography radiomics

Objective

This study aimed to develop a comprehensive nomogram model by integrating clinical pathological and full-field digital mammography (FFDM) radiomic features to predict the efficacy of neoadjuvant chemotherapy (NAC) in breast cancer patients, thereby providing personalized treatment recommendations.

Methods

A retrospective analysis was conducted on the clinical and imaging data of 227 breast cancer patients from 2016 to 2024 at the Second Affiliated Hospital of Harbin Medical University. The patients were divided into a training set (n = 159) and a test set (n = 68) with a 7:3 ratio. The region of interest (ROI) was manually segmented on FFDM images, and features were extracted and gradually selected. The rad-score was calculated for each patient. Five machine learning classifiers were used to build radiomics models, and the optimal model was selected. Univariate and multivariate regression analyses were performed to identify independent risk factors for predicting the efficacy of NAC in breast cancer patients. A nomogram prediction model was further developed by combining the independent risk factors and rad-score, and probability-based stratification was applied. An independent cohort was collected from an external hospital to evaluate the performance of the model.

Results

The radiomics model based on support vector machine (SVM) demonstrated the best predictive performance. FFDM tumor density and HER-2 status were identified as independent risk factors for achieving pathologic complete response (PCR) after NAC (p < 0.05). The nomogram prediction model, developed by combining the independent risk factors and rad-score, outperformed other models, with areas under the curve (AUC) of 0.91 and 0.85 for the training and test sets, respectively. Based on the optimal cutoff points of 103.42 from the nomogram model, patients were classified into high-probability and low-probability groups. When the nomogram model was applied to an independent cohort of 47 patients, only four patients had incorrect diagnoses. The nomogram model demonstrated stable and accurate predictive performance.

Conclusion

The nomogram prediction model, developed by integrating clinical pathological and radiomic features, demonstrated significant performance in predicting the efficacy of NAC in breast cancer, providing valuable reference for clinical personalized prediction planning.

Radiomics analysis of dual-layer detector spectral CT-derived iodine maps for predicting Ki-67 PI in pancreatic ductal adenocarcinoma

OBJECTIVE

To evaluate the feasibility of radiomics analysis using dual-layer detector spectral CT (DLCT)-derived iodine maps for the preoperative prediction of the Ki-67 proliferation index (PI) in pancreatic ductal adenocarcinoma (PDAC).

MATERIALS AND METHODS

A total of 168 PDAC patients who underwent DLCT examination were included and randomly allocated to the training (n = 118) and validation (n = 50) sets. A clinical model was constructed using independent clinicoradiological features identified through multivariate logistic regression analysis in the training set. The radiomics signature was generated based on the coefficients of selected features from iodine maps in the arterial and portal venous phases of DLCT. Finally, a radiomics-clinical model was developed by integrating the radiomics signature and significant clinicoradiological features. The predictive performance of three models was evaluated using the Receiver Operating Characteristic (ROC) curve and Decision Curve Analysis. The optimal model was then used to develop a nomogram, with goodness-of-fit evaluated through the calibration curve.

RESULTS

The radiomics-clinical model integrating radiomics signature, CA19-9, and CT-reported regional lymph node status demonstrated excellent performance in predicting Ki-67 PI in PDAC, which showed an area under the ROC curve of 0.979 and 0.956 in the training and validation sets, respectively. The radiomics-clinical nomogram demonstrated the improved net benefit and exhibited satisfactory consistency.

CONCLUSIONS

This exploratory study demonstrated the feasibility of using DLCT-derived iodine map-based radiomics to predict Ki-67 PI preoperatively in PDAC patients. While preliminary, our findings highlight the potential of functional imaging combined with radiomics for personalized treatment planning.

Dual-radiomics based on SHapley additive explanations for predicting hematologic toxicity in concurrent chemoradiotherapy patients

BACKGROUND

This study investigates the application of a machine learning model that integrates radiomic features and dosiomic features to predict hematologic toxicity (HT) in patients with advanced cervical cancer undergoing concurrent chemoradiotherapy (CCRT). Two integration methods based on SHapley Additive exPlanations (SHAP) values for dual-radiomic features were compared.

METHODS

Clinical information, planning CT images, and dose distribution files from 205 patients with advanced cervical cancer treated with CCRT were retrospectively collected. Patients were categorized by HT severity, with 80% of the data used for training and 20% for testing. Radiomic features and dosiomic features were extracted from the same regions of interest, and SHAP-based feature selection was employed. Extreme gradient boosting models were developed using two feature selection schemes: single-step and multi-step. Sensitivity, specificity, and area under the curve (AUC) values on the test set were used to evaluate model performance.

RESULTS

For the single-step feature selection scheme, the best hybrid model achieved an AUC of 0.79, sensitivity of 0.67, and specificity of 0.72. For the multi-step feature selection scheme, the best hybrid model achieved an AUC of 0.81, sensitivity of 0.75, and specificity of 0.83. Furthermore, the hybrid models outperformed those using radiomic or dosiomic features alone.

CONCLUSIONS

Combining radiomic and dosiomic features improves classification performance in predicting HT in patients with advanced cervical cancer undergoing CCRT, with the multi-step SHAP-based feature selection scheme offering additional advantages. These models hold promise for optimizing patient treatment strategies.

Understandings 18 FDG PET radiomics and its application to lymphoma

The early identification of lymphoma patients who fail front-line treatment is crucial for optimizing disease management. Positron emission tomography, a well-established tool for staging and response evaluation in lymphoma, is typically assessed visually or semiquantitatively, leaving much of its latent information unexploited. Radiomic analysis, which employs mathematical descriptors, can enable the extraction of quantitative features from baseline images that correlate with the disease's biological characteristics. Emerging radiomic features such as metabolic tumour volume, total lesion glycolysis and markers of disease dissemination and metabolic heterogeneity are proving to be powerful prognostic biomarkers in lymphoma. Texture analysis, the most advanced area of radiomics, offers highly complex features that require further standardization and validation before being adopted as reliable biomarkers. Combining radiomic features with clinical risk factors and genomic data holds promising potential for improving clinical risk prediction. This review explores the current state of radiomic analysis, progress towards its standardization and its incorporation into clinical practice and trial designs. The integration of radiomic markers with circulating tumour DNA may provide a comprehensive approach to developing baseline and dynamic risk scores, facilitating the testing of novel treatments and advancing personalized treatment of aggressive lymphomas.

Reproducibility of MRI-derived radiomic features in prostate cancer detection: a methodological approach

Purpose

We aim to evaluate the reproducibility of these features and apply machine learning algorithms to predict cancer diagnosis.

Material and methods

We analyzed magnetic resonance (MR) images from a cohort of 82 individuals, split between 41 prostate cancer patients and 41 healthy controls. A total of 215 radiomic features were extracted from T2-weighted and ADC images using the Software Environment for Radiomic Analysis (SERA). Intraclass correlation coefficient (ICC) analysis was used to assess the reproducibility of features, and Pearson's correlation was applied to remove redundant features. After feature selection, seven dimensionality reduction techniques, including principal component analysis (PCA), kernel PCA, linear discriminant analysis, and locally linear embedding, were applied to preprocess the radiomic features. Ten machine learning algorithms, including support vector machines (SVM), random forests, neural networks, logistic regression, and ensemble methods such as CatBoost and AdaBoost, were utilized to classify cancerous versus non-cancerous tissues. Model performance was evaluated using accuracy and AUC-ROC metrics.

Results

The results showed that features with high reproducibility (ICC > 0.75) contributed significantly to the performance of machine learning models. SVM, neural networks, and logistic regression achieved the highest accuracy (0.88-0.9) and AUC (up to 0.93) when using features from the good and excellent reproducibility categories. PCA emerged as the most effective dimensionality reduction method, preserving the discriminative power of reproducible features across all models.

Conclusion

The results indicate that radiomic feature extraction from MR images, combined with dimensionality reduction and machine learning algorithms, provides a robust approach for prostate cancer diagnosis.