Computational Radiomics System to Decode the Radiographic Phenotype

Preoperative prediction of hepatocellular carcinoma microvascular invasion based on magnetic resonance imaging feature extraction artificial neural network

BACKGROUND

Hepatocellular carcinoma (HCC) recurrence is highly correlated with increased mortality. Microvascular invasion (MVI) is indicative of aggressive tumor biology in HCC.

AIM

To construct an artificial neural network (ANN) capable of accurately predicting MVI presence in HCC using magnetic resonance imaging.

METHODS

This study included 255 patients with HCC with tumors < 3 cm. Radiologists annotated the tumors on the T1-weighted plain MR images. Subsequently, a three-layer ANN was constructed using image features as inputs to predict MVI status in patients with HCC. Postoperative pathological examination is considered the gold standard for determining MVI. Receiver operating characteristic analysis was used to evaluate the effectiveness of the algorithm.

RESULTS

Using the bagging strategy to vote for 50 classifier classification results, a prediction model yielded an area under the curve (AUC) of 0.79. Moreover, correlation analysis revealed that alpha-fetoprotein values and tumor volume were not significantly correlated with the occurrence of MVI, whereas tumor sphericity was significantly correlated with MVI (P < 0.01).

CONCLUSION

Analysis of variable correlations regarding MVI in tumors with diameters < 3 cm should prioritize tumor sphericity. The ANN model demonstrated strong predictive MVI for patients with HCC (AUC = 0.79).

Radiomics based on diffusion tensor imaging and 3D T1-weighted MRI for essential tremor diagnosis

Background

Due to the absence of biomarkers, the misdiagnosis of essential tremor (ET) with other tremor diseases and enhanced physiologic tremor is very common in practice. Combined radiomics based on diffusion tensor imaging (DTI) and three-dimensional T1-weighted imaging (3D-T1) with machine learning (ML) give a most promising way to identify essential tremor (ET) at the individual level and further reveal the potential imaging biomarkers.

Methods

Radiomics features were extracted from 3D-T1 and DTI in 103 ET patients and 103 age-and sex-matched healthy controls (HCs). After data dimensionality reduction and feature selection, five classifiers, including the support vector machine (SVM), random forest (RF), logistic regression (LR), extreme gradient boosting (XGBoost) and multi-layer perceptron (MLP), were adopted to discriminate ET from HCs. The mean values of the area under the curve (mAUC) and accuracy were used to assess the model's performance. Furthermore, a correlation analysis was conducted between the most discriminative features and clinical tremor characteristics.

Results

All classifiers achieved good classification performance (with mAUC at 0.987, 0.984, 0.984, 0.988 and 0.981 in the test set, respectively). The most powerful discriminative features mainly located in the cerebella-thalamo-cortical (CTC) and visual pathway. Furthermore, correlation analysis revealed that some radiomics features were significantly related to the clinical tremor characteristics in ET patients.

Conclusion

These results demonstrated that combining radiomics with ML algorithms could not only achieve high classification accuracy for identifying ET but also help us to reveal the potential brain microstructure pathogenesis in ET patients.

Radiomics Nomograms Based on Multi-sequence MRI for Identifying Cognitive Impairment and Predicting Cognitive Progression in Relapsing-Remitting Multiple Sclerosis

RATIONALE AND OBJECTIVES

To build radiomics nomograms based on multi-sequence MRI to facilitate the identification of cognitive impairment (CI) and prediction of cognitive progression (CP) in patients with relapsing-remitting multiple sclerosis (RRMS).

MATERIALS AND METHODS

We retrospectively included two RRMS cohorts with multi-sequence MRI and Symbol Digit Modalities Test (SDMT) data: dataset1 (n = 149, for training and validation) and dataset2 (n = 29, for external validation). 80 patients of dataset1 had a 2-year follow-up SDMT. CI and CP were evaluated using SDMT scores at baseline and follow-up. The included DIR sequence aided in identifying cortical lesions. Lesion radiomics and structural features were extracted and selected from multi-sequence MRI, followed by the computation of radiomics and structural scores. The nomogram was developed through multivariate logistic regression, integrating clinical data, radiomics, and structural scores to identify CI in patients. Moreover, a similar method was employed to further construct a nomogram predicting CP in patients.

RESULTS

The nomogram demonstrated superior performance in identifying patients with CI, with area under the curve (AUC) values of 0.937 (95% Conf. Interval: 0.898-0.975) and 0.876 (0.810-0.943) in internal and external validation sets, compared to models solely based on clinical data, lesion radiomics, and structural features. Furthermore, another nomogram constructed in predicting CP also exhibited outstanding performance, with an AUC value of 0.969 (0.875-1.000) in the validation set.

CONCLUSION

These nomograms, integrating clinical data, multi-sequence lesions radiomics, and structural features, enable more effective identification of CI and early prediction of CP in RRMS patients, providing important support for clinical decision-making.

Non-invasive CT radiomic biomarkers predict microsatellite stability status in colorectal cancer: a multicenter validation study

BACKGROUND

Microsatellite instability (MSI) status is a strong predictor of response to immunotherapy of colorectal cancer. Radiogenomic approaches promise the ability to gain insight into the underlying tumor biology using non-invasive routine clinical images. This study investigates the association between tumor morphology and the status of MSI versus microsatellite stability (MSS), validating a novel radiomic signature on an external multicenter cohort.

METHODS

Preoperative computed tomography scans with matched MSI status were retrospectively collected for 243 colorectal cancer patients from three hospitals: Seoul National University Hospital (SNUH); Netherlands Cancer Institute (NKI); and Fondazione IRCCS Istituto Nazionale dei Tumori, Milan Italy (INT). Radiologists delineated primary tumors in each scan, from which radiomic features were extracted. Machine learning models trained on SNUH data to identify MSI tumors underwent external validation using NKI and INT images. Performances were compared in terms of area under the receiving operating curve (AUROC).

RESULTS

We identified a radiomic signature comprising seven radiomic features that were predictive of tumors with MSS or MSI (AUROC 0.69, 95% confidence interval [CI] 0.54-0.84, p = 0.018). Integrating radiomic and clinical data into an algorithm improved predictive performance to an AUROC of 0.78 (95% CI 0.60-0.91, p = 0.002) and enhanced the reliability of the predictions.

CONCLUSION

Differences in the radiomic morphological phenotype between tumors MSS or MSI could be detected using radiogenomic approaches. Future research involving large-scale multicenter prospective studies that combine various diagnostic data is necessary to refine and validate more robust, potentially tumor-agnostic MSI radiogenomic models.

RELEVANCE STATEMENT

Noninvasive radiomic signatures derived from computed tomography scans can predict MSI in colorectal cancer, potentially augmenting traditional biopsy-based methods and enhancing personalized treatment strategies.

KEY POINTS

Noninvasive CT-based radiomics predicted MSI in colorectal cancer, enhancing stratification. A seven-feature radiomic signature differentiated tumors with MSI from those with MSS in multicenter cohorts. Integrating radiomic and clinical data improved the algorithm's predictive performance.

Computed tomography-based radiomics predicts the fibroblast-related gene EZH2 expression level and survival of hepatocellular carcinoma

BACKGROUND

Hepatocellular carcinoma (HCC) is the most common subtype of liver cancer. The primary treatment strategies for HCC currently include liver transplantation and surgical resection. However, these methods often yield unsatisfactory outcomes, leading to a poor prognosis for many patients. This underscores the urgent need to identify and evaluate novel therapeutic targets that can improve the prognosis and survival rate of HCC patients.

AIM

To construct a radiomics model that can accurately predict the EZH2 expression in HCC.

METHODS

Gene expression, clinical parameters, HCC-related radiomics, and fibroblast-related genes were acquired from public databases. A gene model was developed, and its clinical efficacy was assessed statistically. Drug sensitivity analysis was conducted with identified hub genes. Radiomics features were extracted and machine learning algorithms were employed to generate a radiomics model related to the hub genes. A nomogram was used to illustrate the prognostic significance of the computed Radscore and the hub genes in the context of HCC patient outcomes.

RESULTS

EZH2 and NRAS were independent predictors for prognosis of HCC and were utilized to construct a predictive gene model. This model demonstrated robust performance in diagnosing HCC and predicted an unfavorable prognosis. A negative correlation was observed between EZH2 expression and drug sensitivity. Elevated EZH2 expression was linked to poorer prognosis, and its diagnostic value in HCC surpassed that of the risk model. A radiomics model, developed using a logistic algorithm, also showed superior efficiency in predicting EZH2 expression. The Radscore was higher in the group with high EZH2 expression. A nomogram was constructed to visually demonstrate the significant roles of the radiomics model and EZH2 expression in predicting the overall survival of HCC patients.

CONCLUSION

EZH2 plays significant roles in diagnosing HCC and therapeutic efficacy. A radiomics model, developed using a logistic algorithm, efficiently predicted EZH2 expression and exhibited strong correlation with HCC prognosis.

Delta-radiomics features of ADC maps as early predictors of treatment response in lung cancer

OBJECTIVE

Investigate the feasibility of detecting early treatment-induced tumor tissue changes in patients with advanced lung adenocarcinoma using diffusion-weighted MRI-derived radiomics features.

METHODS

This prospective observational study included 144 patients receiving either tyrosine kinase inhibitors (TKI, n = 64) or platinum-based chemotherapy (PBC, n = 80) for the treatment of pulmonary adenocarcinoma. Patients underwent diffusion-weighted MRI the day prior to therapy (baseline, all patients), as well as either + 1 (PBC) or + 7 and + 14 (TKI) days after treatment initiation. One hundred ninety-seven radiomics features were extracted from manually delineated tumor volumes. Feature changes over time were analyzed for correlation with treatment response (TR) according to CT-derived RECIST after 2 months and progression-free survival (PFS).

RESULTS

Out of 14 selected delta-radiomics features, 6 showed significant correlations with PFS or TR. Most significant correlations were found after 14 days. Features quantifying ROI heterogeneity, such as short-run emphasis (p = 0.04(pfs)/0.005(tr)), gradient short-run emphasis (p = 0.06(pfs)/0.01(tr)), and zone percentage (p = 0.02(pfs)/0.01(tr)) increased in patients with overall better TR whereas patients with worse overall response showed an increase in features quantifying ROI homogeneity, such as normalized inverse difference (p = 0.01(pfs)/0.04(tr)). Clustering of these features allows stratification of patients into groups of longer and shorter survival.

CONCLUSION

Two weeks after initiation of treatment, diffusion MRI of lung adenocarcinoma reveals quantifiable tissue-level insights that correlate well with future treatment (non-)response. Diffusion MRI-derived radiomics thus shows promise as an early, radiation-free decision-support to predict efficacy and potentially alter the treatment course early.

CRITICAL RELEVANCE STATEMENT

Delta-Radiomics texture features derived from diffusion-weighted MRI of lung adenocarcinoma, acquired as early as 2 weeks after initiation of treatment, are significantly correlated with RECIST TR and PFS as obtained through later morphological imaging.

KEY POINTS

Morphological imaging takes time to detect TR in lung cancer, diffusion-weighted MRI might identify response earlier. Several radiomics features are significantly correlated with TR and PFS. Radiomics of diffusion-weighted MRI may facilitate patient stratification and management.

Predicting gastric cancer response to anti-HER2 therapy or anti-HER2 combined immunotherapy based on multi-modal data

The sole use of single modality data often fails to capture the complex heterogeneity among patients, including the variability in resistance to anti-HER2 therapy and outcomes of combined treatment regimens, for the treatment of HER2-positive gastric cancer (GC). This modality deficit has not been fully considered in many studies. Furthermore, the application of artificial intelligence in predicting the treatment response, particularly in complex diseases such as GC, is still in its infancy. Therefore, this study aimed to use a comprehensive analytic approach to accurately predict treatment responses to anti-HER2 therapy or anti-HER2 combined immunotherapy in patients with HER2-positive GC. We collected multi-modal data, comprising radiology, pathology, and clinical information from a cohort of 429 patients: 310 treated with anti-HER2 therapy and 119 treated with a combination of anti-HER2 and anti-PD-1/PD-L1 inhibitors immunotherapy. We introduced a deep learning model, called the Multi-Modal model (MuMo), that integrates these data to make precise treatment response predictions. MuMo achieved an area under the curve score of 0.821 for anti-HER2 therapy and 0.914 for combined immunotherapy. Moreover, patients classified as low-risk by MuMo exhibited significantly prolonged progression-free survival and overall survival (log-rank test, P < 0.05). These findings not only highlight the significance of multi-modal data analysis in enhancing treatment evaluation and personalized medicine for HER2-positive gastric cancer, but also the potential and clinical value of our model.

Different radiomics annotation methods comparison in rectal cancer characterisation and prognosis prediction: a two-centre study

OBJECTIVES

To explore the performance differences of multiple annotations in radiomics analysis and provide a reference for tumour annotation in large-scale medical image analysis.

METHODS

A total of 342 patients from two centres who underwent radical resection for rectal cancer were retrospectively studied and divided into training, internal validation, and external validation cohorts. Three predictive tasks of tumour T-stage (pT), lymph node metastasis (pLNM), and disease-free survival (pDFS) were performed. Twelve radiomics models were constructed using Lasso-Logistic or Lasso-Cox to evaluate and four annotation methods, 2D detailed annotation along tumour boundaries (2D), 3D detailed annotation along tumour boundaries (3D), 2D bounding box (2DBB), and 3D bounding box (3DBB) on T2-weighted images, were compared. Radiomics models were used to establish combined models incorporating clinical risk factors. The DeLong test was performed to compare the performance of models using the receiver operating characteristic curves.

RESULTS

For radiomics models, the area under the curve values ranged from 0.627 (0.518-0.728) to 0.811 (0.705-0.917) in the internal validation cohort and from 0.619 (0.469-0.754) to 0.824 (0.689-0.918) in the external validation cohort. Most radiomics models based on four annotations did not differ significantly, except between the 3D and 3DBB models for pLNM (p = 0.0188) in the internal validation cohort. For combined models, only the 2D model significantly differed from the 2DBB (p = 0.0372) and 3D models (p = 0.0380) for pDFS.

CONCLUSION

Radiomics and combined models constructed with 2D and bounding box annotations showed comparable performances to those with 3D and detailed annotations along tumour boundaries in rectal cancer characterisation and prognosis prediction.

CRITICAL RELEVANCE STATEMENT

For quantitative analysis of radiological images, the selection of 2D maximum tumour area or bounding box annotation is as representative and easy to operate as 3D whole tumour or detailed annotations along tumour boundaries.

KEY POINTS

There is currently a lack of discussion on whether different annotation efforts in radiomics are predictively representative. No significant differences were observed in radiomics and combined models regardless of the annotations (2D, 3D, detailed, or bounding box). Prioritise selecting the more time and effort-saving 2D maximum area bounding box annotation.

A 3D Convolutional Neural Network Based on Non-enhanced Brain CT to Identify Patients with Brain Metastases

Dedicated brain imaging for cancer patients is seldom recommended in the absence of symptoms. There is increasing availability of non-enhanced CT (NE-CT) of the brain, mainly owing to a wider utilization of Positron Emission Tomography-CT (PET-CT) in cancer staging. Brain metastases (BM) are often hard to diagnose on NE-CT. This work aims to develop a 3D Convolutional Neural Network (3D-CNN) based on brain NE-CT to distinguish patients with and without BM. We retrospectively included NE-CT scans for 100 patients with single or multiple BM and 100 patients without brain imaging abnormalities. Patients whose largest lesion was < 5 mm were excluded. The largest tumor was manually segmented on a matched contrast-enhanced T1 weighted Magnetic Resonance Imaging (MRI), and shape radiomics were extracted to determine the size and volume of the lesion. The brain was automatically segmented, and masked images were normalized and resampled. The dataset was split into training (70%) and validation (30%) sets. Multiple versions of a 3D-CNN were developed, and the best model was selected based on accuracy (ACC) on the validation set. The median largest tumor Maximum-3D-Diameter was 2.29 cm, and its median volume was 2.81 cc. Solitary BM were found in 27% of the patients, while 49% had > 5 BMs. The best model consisted of 4 convolutional layers with 3D average pooling layers, dropout layers of 50%, and a sigmoid activation function. Mean validation ACC was 0.983 (SD: 0.020) and mean area under receiver-operating characteristic curve was 0.983 (SD: 0.023). Sensitivity was 0.983 (SD: 0.020). We developed an accurate 3D-CNN based on brain NE-CT to differentiate between patients with and without BM. The model merits further external validation.

Radiomic and deep learning analysis of dermoscopic images for skin lesion pattern decoding

This study aims to explore the efficacy of a hybrid deep learning and radiomics approach, supplemented with patient metadata, in the noninvasive dermoscopic imaging-based diagnosis of skin lesions. We analyzed dermoscopic images from the International Skin Imaging Collaboration (ISIC) dataset, spanning 2016-2020, encompassing a variety of skin lesions. Our approach integrates deep learning with a comprehensive radiomics analysis, utilizing a vast array of quantitative image features to precisely quantify skin lesion patterns. The dataset includes cases of three, four, and eight different skin lesion types. Our methodology was benchmarked against seven classification methods from the ISIC 2020 challenge and prior research using a binary decision framework. The proposed hybrid model demonstrated superior performance in distinguishing benign from malignant lesions, achieving area under the receiver operating characteristic curve (AUROC) scores of 99%, 95%, and 96%, and multiclass decoding AUROCs of 98.5%, 94.9%, and 96.4%, with sensitivities of 97.6%, 93.9%, and 96.0% and specificities of 98.4%, 96.7%, and 96.9% in the internal ISIC 2018 challenge, as well as in the external Jinan and Longhua datasets, respectively. Our findings suggest that the integration of radiomics and deep learning, utilizing dermoscopic images, effectively captures the heterogeneity and pattern expression of skin lesions.

Explainable machine-learning-based prediction of QCT/FEA-calculated femoral strength under stance loading configuration using radiomics features

Finite element analysis can provide precise femoral strength assessment. However, its modeling procedures were complex and time-consuming. This study aimed to develop a model to evaluate femoral strength calculated by quantitative computed tomography-based finite element analysis (QCT/FEA) under stance loading configuration, offering an effective, simple, and explainable method. One hundred participants with hip QCT images were selected from the Hong Kong part of the Osteoporotic fractures in men cohort. Radiomics features were extracted from QCT images. Filter method, Pearson correlation analysis, and least absolute shrinkage and selection operator method were employed for feature selection and dimension reduction. The remaining features were utilized as inputs, and femoral strengths were calculated as the ground truth through QCT/FEA. Support vector regression was applied to develop a femoral strength prediction model. The influence of various numbers of input features on prediction performance was compared, and the femoral strength prediction model was established. Finally, Shapley additive explanation, accumulated local effects, and partial dependency plot methods were used to explain the model. The results indicated that the model performed best when six radiomics features were selected. The coefficient of determination (R2), the root mean square error, the normalized root mean square error, and the mean squared error on the testing set were 0.820, 1016.299 N, 10.645%, and 750.827 N, respectively. Additionally, these features all positively contributed to femoral strength prediction. In conclusion, this study provided a noninvasive, effective, and explainable method of femoral strength assessment, and it may have clinical application potential.

Epigenetic profiling of prostate cancer reveals potential prognostic signatures

PURPOSE

While epigenetic profiling discovered biomarkers in several tumor entities, its application in prostate cancer is still limited. We explored DNA methylation-based deconvolution of benign and malignant prostate tissue for biomarker discovery and the potential of radiomics as a non-invasive surrogate.

METHODS

We retrospectively included 30 patients (63 [58-79] years) with prostate cancer (PCa) who had a multiparametric MRI of the prostate before radical prostatectomy between 2014 and 2019. The control group comprised four patients with benign prostate tissue adjacent to the PCa lesions and four patients with benign prostatic hyperplasia. Tissue punches of all lesions were obtained. DNA methylation analysis and reference-free in silico deconvolution were conducted to retrieve Latent Methylation Components (LCMs). LCM-based clustering was analyzed for cellular composition and correlated with clinical disease parameters. Additionally, PCa and adjacent benign lesions were analyzed using radiomics to predict the epigenetic signatures non-invasively.

RESULTS

LCMs identified two clusters with potential prognostic impact. Cluster one was associated with malignant prostate tissue (p < 0.001) and reduced immune-cell-related signatures (p = 0.004) of CD19 and CD4 cells. Cluster one comprised exclusively malignant prostate tissue enriched for significant prostate cancer and advanced tumor stages (p < 0.03 for both). No radiomics model could non-invasively predict the epigenetic clusters.

CONCLUSION

Epigenetic clusters were associated with prognostically and clinically relevant metrics in prostate cancer. Further, immune cell-related signatures differed significantly between prognostically favorable and unfavorable clusters. Further research is necessary to explore potential diagnostic and therapeutic implications.

Impact of Emerging Deep Learning-Based MR Image Reconstruction Algorithms on Abdominal MRI Radiomic Features

OBJECTIVE

This study aims to evaluate, on one MRI vendor's platform, the impact of deep learning (DL)-based reconstruction techniques on MRI radiomic features compared to conventional image reconstruction techniques.

METHODS

Under IRB approval and informed consent, we prospectively collected undersampled coronal T2-weighted MR images of the abdomen (1.5 T; Philips Healthcare) from 17 pediatric and adult subjects and reconstructed them using a conventional image reconstruction technique (compressed sensitivity encoding [C-SENSE]) and two DL-based reconstruction techniques (SmartSpeed [Philips Healthcare, US FDA cleared] and SmartSpeed with Super Resolution [SmartSpeed-SuperRes, not US FDA cleared to date]). Eight regions of interest (ROIs) across organs/tissues (liver, spleen, kidney, pancreas, fat, and muscle) were manually placed. Eighty-six MRI radiomic features were then extracted. Pearson's correlation coefficients (PCCs) and intraclass correlation coefficients (ICCs) were calculated between (A) C-SENSE versus SmartSpeed, and (B) C-SENSE versus SmartSpeed-SuperRes. To quantify the impact from the perspective of the whole MR image, cross-ROI mean PCCs and ICCs were calculated for individual radiomic features. The impact of image reconstruction on individual radiomic features in different organs/tissues was evaluated using ANOVA analyses.

RESULTS

According to cross-ROI mean PCCs, 50 out of 86 radiomic features were highly correlated (PCC, ≥0.8) between SmartSpeed and C-SENSE, whereas only 15 radiomic features were highly correlated between SmartSpeed-SuperRes and C-SENSE reconstructions. According to cross-ROI mean ICCs, 58 out of 86 radiomic features had high agreements (ICC ≥0.75) between SmartSpeed and C-SENSE, whereas only 9 radiomic features had high agreements between SmartSpeed-SuperRes and C-SENSE reconstructions. For SmartSpeed reconstruction, the psoas muscle ROI appeared to be impacted most with the lowest median (IQR) correlation of 0.57 (0.25). The circular liver ROI was impacted most by SmartSpeed-SuperRes (PCC, 0.60 [0.22]). ANOVA analyses suggest that the impact of DL reconstruction algorithms on radiomic features varies significantly among different organs/tissues (P < 0.001).

CONCLUSIONS

MRI radiomic features are significantly altered by DL-based reconstruction compared to a conventional reconstruction technique. The impact of DL reconstruction algorithms on radiomic features varies significantly between different organs/tissues.

Bridging Histopathology and Radiomics Toward Prognosis of Metastasis in Early Breast Cancer

Tumor histomorphology is crucial for the prognostication of breast cancer outcomes because it contains histological, cellular, and molecular tumor heterogeneity related to metastatic potential. To enhance breast cancer prognosis, we aimed to apply radiomics analysis-traditionally used in 3D scans-to 2D histopathology slides. This study tested radiomics analysis in a cohort of 92 breast tumor specimens for outcome prognosis, addressing -omics dimensionality by comparing models with moderate and high feature counts, using least absolute shrinkage and selection operator for feature selection and machine learning for prognostic modeling. In the test folds, models with radiomics features [area under the curves (AUCs) range 0.799-0.823] significantly outperformed the benchmark model, which only included clinicopathological (CP) parameters (AUC = 0.584). The moderate-dimensionality model with 11 CP + 93 radiomics features matched the performance of the highly dimensional models with 1,208 radiomics or 11 CP + 1,208 radiomics features, showing average AUCs of 0.823, 0.799, and 0.807 and accuracies of 79.8, 79.3, and 76.6%, respectively. In conclusion, our application of deep texture radiomics analysis to 2D histopathology showed strong prognostic performance with a moderate-dimensionality model, surpassing a benchmark based on standard CP parameters, indicating that this deep texture histomics approach could potentially become a valuable prognostic tool.

Long-term prognostic value of CT-based high-risk coronary lesion attributes and radiomic features of pericoronary adipose tissue in diabetic patients

AIMS

To investigate the long-term prognostic value of coronary computed tomography angiography (CCTA)-derived high-risk attributes and radiomic features of pericoronary adipose tissue (PCAT) in diabetic patients for predicting major adverse cardiac event (MACE).

METHODS AND RESULTS

Diabetic patients with intermediate pre-test probability of coronary artery disease were prospectively enrolled and referred for CCTA. Three models (model-1 with clinical parameters; model-2 with clinical factors + CCTA imaging parameters; model-3 with the above parameters and PCAT radiomic features) were developed in the training cohort (835 patients) and tested in the independent validation cohort (557 patients). 1392 patients were included and MACEs occurred in 108 patients (7.8%). Multivariable Cox regression analysis revealed that HbA1c, coronary calcium Agatston score, significant stenosis and high-risk plaque were independent predictors for MACE whereas none of PCAT radiomic features showed predictive value. In the training cohort, model-2 demonstrated higher predictive performance over model-1 (C-index = 0.79 vs. 0.68, p < 0.001) whereas model-3 did not show incremental value over model-2(C-index = 0.79 vs. 0.80, p = 0.408). Similar findings were found in the validation cohort.

CONCLUSIONS

The combined model (clinical and CCTA high-risk anatomical features) demonstrated high efficacy in predicting MACE in diabetes. PCAT radiomic features failed to show incremental value for risk stratification.

Deep learning-derived splenic radiomics, genomics, and coronary artery disease

Background

Despite advances in managing traditional risk factors, coronary artery disease (CAD) remains the leading cause of mortality. Circulating hematopoietic cells influence risk for CAD, but the role of a key regulating organ, spleen, is unknown. The understudied spleen is a 3-dimensional structure of the hematopoietic system optimally suited for unbiased radiologic investigations toward novel mechanistic insights.

Methods

Deep learning-based image segmentation and radiomics techniques were utilized to extract splenic radiomic features from abdominal MRIs of 42,059 UK Biobank participants. Regression analysis was used to identify splenic radiomics features associated with CAD. Genome-wide association analyses were applied to identify loci associated with these radiomics features. Overlap between loci associated with CAD and the splenic radiomics features was explored to understand the underlying genetic mechanisms of the role of the spleen in CAD.

Results

We extracted 107 splenic radiomics features from abdominal MRIs, and of these, 10 features were associated with CAD. Genome-wide association analysis of CAD-associated features identified 219 loci, including 35 previously reported CAD loci, 7 of which were not associated with conventional CAD risk factors. Notably, variants at 9p21 were associated with splenic features such as run length non-uniformity.

Conclusions

Our study, combining deep learning with genomics, presents a new framework to uncover the splenic axis of CAD. Notably, our study provides evidence for the underlying genetic connection between the spleen as a candidate causal tissue-type and CAD with insight into the mechanisms of 9p21, whose mechanism is still elusive despite its initial discovery in 2007. More broadly, our study provides a unique application of deep learning radiomics to non-invasively find associations between imaging, genetics, and clinical outcomes.

Fast prediction of personalized abdominal organ doses from CT examinations by radiomics feature-based machine learning models

The X-rays emitted during CT scans can increase solid cancer risks by damaging DNA, with the risk tied to patient-specific organ doses. This study aims to establish a new method to predict patient specific abdominal organ doses from CT examinations using minimized computational resources at a fast speed. The CT data of 247 abdominal patients were selected and exported to the auto-segmentation software named DeepViewer to generate abdominal regions of interest (ROIs). Radiomics feature were extracted based on the selected CT data and ROIs. Reference organ doses were obtained by GPU-based Monte Carlo simulations. The support vector regression (SVR) model was trained based on the radiomics features and reference organ doses to predict abdominal organ doses from CT examinations. The prediction performance of the SVR model was tested and verified by changing the abdominal patients of the train and test sets randomly. For the abdominal organs, the maximal difference between the reference and the predicted dose was less than 1 mGy. For the body and bowel, the organ doses were predicted with a percentage error of less than 5.2%, and the coefficient of determination (R2) reached up to 0.9. For the left kidney, right kidney, liver, and spinal cord, the mean absolute percentage error ranged from 5.1 to 8.9%, and the R2 values were more than 0.74. The SVR model could be trained to achieve accurate prediction of personalized abdominal organ doses in less than one second using a single CPU core.

[An MRI multi-sequence feature imputation and fusion mutual-aid model based on sequence deletion for differentiation of high-grade from low-grade glioma]

OBJECTIVE

To evaluate the performance of magnetic resonance imaging (MRI) multi-sequence feature imputation and fusion mutual model based on sequence deletion in differentiating high-grade glioma (HGG) from low-grade glioma (LGG).

METHODS

We retrospectively collected multi-sequence MR images from 305 glioma patients, including 189 HGG patients and 116 LGG patients. The region of interest (ROI) of T1-weighted images (T1WI), T2-weighted images (T2WI), T2 fluid attenuated inversion recovery (T2_FLAIR) and post-contrast enhancement T1WI (CE_T1WI) were delineated to extract the radiomics features. A mutual-aid model of MRI multi-sequence feature imputation and fusion based on sequence deletion was used for imputation and fusion of the feature matrix with missing data. The discriminative ability of the model was evaluated using 5-fold cross-validation method and by assessing the accuracy, balanced accuracy, area under the ROC curve (AUC), specificity, and sensitivity. The proposed model was quantitatively compared with other non-holonomic multimodal classification models for discriminating HGG and LGG. Class separability experiments were performed on the latent features learned by the proposed feature imputation and fusion methods to observe the classification effect of the samples in twodimensional plane. Convergence experiments were used to verify the feasibility of the model.

RESULTS

For differentiation of HGG from LGG with a missing rate of 10%, the proposed model achieved accuracy, balanced accuracy, AUC, specificity, and sensitivity of 0.777, 0.768, 0.826, 0.754 and 0.780, respectively. The fused latent features showed excellent performance in the class separability experiment, and the algorithm could be iterated to convergence with superior classification performance over other methods at the missing rates of 30% and 50%.

CONCLUSION

The proposed model has excellent performance in classification task of HGG and LGG and outperforms other non-holonomic multimodal classification models, demonstrating its potential for efficient processing of non-holonomic multimodal data.

Intratumoral habitat radiomics based on magnetic resonance imaging for preoperative prediction treatment response to neoadjuvant chemotherapy in nasopharyngeal carcinoma

PURPOSE

The aim of this study is to determine intratumoral habitat regions from multi-sequences magnetic resonance imaging (MRI) and to assess the value of those regions for prediction of patient response to neoadjuvant chemotherapy (NAC) in nasopharyngeal carcinoma (NPC).

MATERIALS AND METHODS

Two hundred and ninety seven patients with NPC were enrolled. Multi-sequences MRI data were used to outline three-dimensional volumes of interest (VOI) of the whole tumor. The original imaging data were divided into two groups, which were resampled to an isotropic resolution of 1 × 1 × 1 mm3 (group_1mm) and 3 × 3 × 3 mm3 (group_3mm). Nineteen radiomics features were computed for each voxel of three sequences in group_3mm, within the tumor region to extract local information. Then, k-means clustering was implemented to segment the whole tumor regions in two groups. After radiomics features were extracted and dimension reduction, habitat models were built using Multi-Layer Perceptron (MLP) algorithm.

RESULTS

Only T stage was included as the clinical model. The habitat3mm model, which included 10 radiomics features, achieved AUCs of 0.752 and 0.724 in the training and validation cohorts, respectively. Given the slightly better outcome of habitat3mm model, nomogram was developed in combination with habitat3mm model and T stage with the AUC of 0.749 and 0.738 in the training and validation cohorts. The decision curve analysis provides further evidence of the nomogram's clinical practicality.

CONCLUSIONS

A nomogram based on intratumoral habitat predicts the efficacy of NAC in NPC patients, offering the potential to improve both the treatment plan and patient outcomes.

Hybridization of Acoustic and Visual Features of Polish Sibilants Produced by Children for Computer Speech Diagnosis

Speech disorders are significant barriers to the balanced development of a child. Many children in Poland are affected by lisps (sigmatism)-the incorrect articulation of sibilants. Since speech therapy diagnostics is complex and multifaceted, developing computer-assisted methods is crucial. This paper presents the results of assessing the usefulness of hybrid feature vectors extracted based on multimodal (video and audio) data for the place of articulation assessment in sibilants /s/ and /ʂ/. We used acoustic features and, new in this field, visual parameters describing selected articulators' texture and shape. Analysis using statistical tests indicated the differences between various sibilant realizations in the context of the articulation pattern assessment using hybrid feature vectors. In sound /s/, 35 variables differentiated dental and interdental pronunciation, and 24 were visual (textural and shape). For sibilant /ʂ/, we found 49 statistically significant variables whose distributions differed between speaker groups (alveolar, dental, and postalveolar articulation), and the dominant feature type was noise-band acoustic. Our study suggests hybridizing the acoustic description with video processing provides richer diagnostic information.

Comparative Analysis of Repeatability in CT Radiomics and Dosiomics Features under Image Perturbation: A Study in Cervical Cancer Patients

This study aims to evaluate the repeatability of radiomics and dosiomics features via image perturbation of patients with cervical cancer. A total of 304 cervical cancer patients with planning CT images and dose maps were retrospectively included. Random translation, rotation, and contour randomization were applied to CT images and dose maps before radiomics feature extraction. The repeatability of radiomics and dosiomics features was assessed using intra-class correlation of coefficient (ICC). Pearson correlation coefficient (r) was adopted to quantify the correlation between the image characteristics and feature repeatability. In general, the repeatability of dosiomics features was lower compared with CT radiomics features, especially after small-sigma Laplacian-of-Gaussian (LoG) and wavelet filtering. More repeatable features (ICC > 0.9) were observed when extracted from the original, Large-sigma LoG filtered, and LLL-/LLH-wavelet filtered images. Positive correlations were found between image entropy and high-repeatable feature number in both CT and dose (r = 0.56, 0.68). Radiomics features showed higher repeatability compared to dosiomics features. These findings highlight the potential of radiomics features for robust quantitative imaging analysis in cervical cancer patients, while suggesting the need for further refinement of dosiomics approaches to enhance their repeatability.

Beyond hand-crafted features for pretherapeutic molecular status identification of pediatric low-grade gliomas

The use of targeted agents in the treatment of pediatric low-grade gliomas (pLGGs) relies on the determination of molecular status. It has been shown that genetic alterations in pLGG can be identified non-invasively using MRI-based radiomic features or convolutional neural networks (CNNs). We aimed to build and assess a combined radiomics and CNN non-invasive pLGG molecular status identification model. This retrospective study used the tumor regions, manually segmented from T2-FLAIR MR images, of 336 patients treated for pLGG between 1999 and 2018. We designed a CNN and Random Forest radiomics model, along with a model relying on a combination of CNN and radiomic features, to predict the genetic status of pLGG. Additionally, we investigated whether CNNs could predict radiomic feature values from MR images. The combined model (mean AUC: 0.824) outperformed the radiomics model (0.802) and CNN (0.764). The differences in model performance were statistically significant (p-values < 0.05). The CNN was able to learn predictive radiomic features such as surface-to-volume ratio (average correlation: 0.864), and difference matrix dependence non-uniformity normalized (0.924) well but was unable to learn others such as run-length matrix variance (- 0.017) and non-uniformity normalized (- 0.042). Our results show that a model relying on both CNN and radiomic-based features performs better than either approach separately in differentiating the genetic status of pLGGs, and that CNNs are unable to express all handcrafted features.

Machine learning models for differential diagnosing HER2-low breast cancer: A radiomics approach

To develop machine learning models based on preoperative dynamic enhanced magnetic resonance imaging (DCE-MRI) radiomics and to explore their potential prognostic value in the differential diagnosis of human epidermal growth factor receptor 2 (HER2)-low from HER2-positive breast cancer (BC). A total of 233 patients with pathologically confirmed invasive breast cancer admitted to our hospital between January 2018 and December 2022 were included in this retrospective analysis. Of these, 103 cases were diagnosed as HER2-positive and 130 cases were HER2 low-expression BC. The Synthetic Minority Oversampling Technique is employed to address the class imbalance problem. Patients were randomly split into a training set (163 cases) and a validation set (70 cases) in a 7:3 ratio. Radiomics features from DCE-MRI second-phase imaging were extracted. Z-score normalization was used to standardize the radiomics features, and Pearson's correlation coefficient and recursive feature elimination were used to explore the significant features. Prediction models were constructed using 6 machine learning algorithms: logistic regression, random forest, support vector machine, AdaBoost, decision tree, and auto-encoder. Receiver operating characteristic curves were constructed, and predictive models were evaluated according to the area under the curve (AUC), accuracy, sensitivity, and specificity. In the training set, the AUC, accuracy, sensitivity, and specificity of all models were 1.000. However, in the validation set, the auto-encoder model's AUC, accuracy, sensitivity, and specificity were 0.994, 0.976, 0.972, and 0.978, respectively. The remaining models' AUC, accuracy, sensitivity, and specificity were 1.000. The DeLong test showed no statistically significant differences between the machine learning models in the training and validation sets (Z = 0, P = 1). Our study investigated the feasibility of using DCE-MRI-based radiomics features to predict HER2-low BC. Certain radiomics features showed associations with HER2-low BC and may have predictive value. Machine learning prediction models developed using these radiomics features could be beneficial for distinguishing between HER2-low and HER2-positive BC. These noninvasive preoperative models have the potential to assist in clinical decision-making for HER2-low breast cancer, thereby advancing personalized clinical precision.

Voxel-based texture similarity networks reveal individual variability and correlate with biological ontologies

The human brain is organized as a complex, hierarchical network. However, the structural covariance patterns among brain regions and the underlying biological substrates of such covariance networks remain to be clarified. The present study proposed a novel individualized structural covariance network termed voxel-based texture similarity networks (vTSNs) based on 76 refined voxel-based textural features derived from structural magnetic resonance images. Validated in three independent longitudinal healthy cohorts (40, 23, and 60 healthy participants, respectively) with two common brain atlases, we found that the vTSN could robustly resolve inter-subject variability with high test-retest reliability. In contrast to the regional-based texture similarity networks (rTSNs) that calculate radiomic features based on region-of-interest information, vTSNs had higher inter- and intra-subject variability ratios and test-retest reliability in connectivity strength and network topological properties. Moreover, the Spearman correlation indicated a stronger association of the gene expression similarity network (GESN) with vTSNs than with rTSNs (vTSN: r = 0.600, rTSN: r = 0.433, z = 39.784, P < 0.001). Hierarchical clustering identified 3 vTSN subnets with differential association patterns with 13 coexpression modules, 16 neurotransmitters, 7 electrophysiology, 4 metabolism, and 2 large-scale structural and 4 functional organization maps. Moreover, these subnets had unique biological hierarchical organization from the subcortex-limbic system to the ventral neocortex and then to the dorsal neocortex. Based on 424 unrelated, qualified healthy subjects from the Human Connectome Project, we found that vTSNs could sensitively represent sex differences, especially for connections in the subcortex-limbic system and between the subcortex-limbic system and the ventral neocortex. Moreover, a multivariate variance component model revealed that vTSNs could explain a significant proportion of inter-subject behavioral variance in cognition (80.0 %) and motor functions (63.4 %). Finally, using 494 healthy adults (aged 19-80 years old) from the Southwest University Adult Lifespan Dataset, the Spearman correlation identified a significant association between aging and vTSN strength, especially within the subcortex-limbic system and between the subcortex-limbic system and the dorsal neocortex. In summary, our proposed vTSN is robust in uncovering individual variability and neurobiological brain processes, which can serve as biologically plausible measures for linking biological processes and human behavior.

Deep Learning Radiomics Features of Mediastinal Fat and Pulmonary Nodules on Lung CT Images Distinguish Benignancy and Malignancy

This study investigated the relationship between mediastinal fat and pulmonary nodule status, aiming to develop a deep learning-based radiomics model for diagnosing benign and malignant pulmonary nodules. We proposed a combined model using CT images of both pulmonary nodules and the fat around the chest (mediastinal fat). Patients from three centers were divided into training, validation, internal testing, and external testing sets. Quantitative radiomics and deep learning features from CT images served as predictive factors. A logistic regression model was used to combine data from both pulmonary nodules and mediastinal adipose regions, and personalized nomograms were created to evaluate the predictive performance. The model incorporating mediastinal fat outperformed the nodule-only model, with C-indexes of 0.917 (training), 0.903 (internal testing), 0.942 (external testing set 1), and 0.880 (external testing set 2). The inclusion of mediastinal fat significantly improved predictive performance (NRI = 0.243, p < 0.05). A decision curve analysis indicated that incorporating mediastinal fat features provided greater patient benefits. Mediastinal fat offered complementary information for distinguishing benign from malignant nodules, enhancing the diagnostic capability of this deep learning-based radiomics model. This model demonstrated strong diagnostic ability for benign and malignant pulmonary nodules, providing a more accurate and beneficial approach for patient care.

An approach leveraging radiomics and model checking for the automatic early diagnosis of adhesive capsulitis

Adhesive Capsulitis of the shoulder is a painful pathology limiting shoulder movements, commonly known as "Frozen Shoulder". Since this pathology limits movement, it is important to make an early diagnosis. Diagnosing capsulitis relies on clinical assessment, although diagnostic imaging, such as Magnetic Resonance Imaging, can provide predictive or supportive information for specific characteristic signs. However, its diagnosis is not so simple nor so immediate, indeed it remains a difficult topic for many general radiologists and expert musculoskeletal radiologists. This study aims to investigate whether it is possible to use disease signs within a medical image to automatically diagnose Adhesive Capsulitis. To this purpose, we propose an automatic Model Checking-based approach to quickly diagnose the Adhesive Capsulitis taking as input the radiomic feature values from the medical images. Furthermore, we compare the performance achieved by our method with diagnostic results obtained by professional radiologists with different levels of experience. To the best of our knowledge, this is the first method for the automatic diagnosis of Adhesive Capsulitis of the Shoulder.

Radiomic-Based Prediction of Lesion-Specific Systemic Treatment Response in Metastatic Disease

Despite sharing the same histologic classification, individual tumors in multi metastatic patients may present with different characteristics and varying sensitivities to anticancer therapies. In this study, we investigate the utility of radiomic biomarkers for prediction of lesion-specific treatment resistance in multi metastatic leiomyosarcoma patients. Using a dataset of n=202 lung metastases (LM) from n=80 patients with 1648 pre-treatment computed tomography (CT) radiomics features and LM progression determined from follow-up CT, we developed a radiomic model to predict the progression of each lesion. Repeat experiments assessed the relative predictive performance across LM volume groups. Lesion-specific radiomic models indicate up to a 4.5-fold increase in predictive capacity compared with a no-skill classifier, with an area under the precision-recall curve of 0.70 for the most precise model (FDR = 0.05). Precision varied by administered drug and LM volume. The effect of LM volume was controlled by removing radiomic features at a volume-correlation coefficient threshold of 0.20. Predicting lesion-specific responses using radiomic features represents a novel strategy by which to assess treatment response that acknowledges biological diversity within metastatic subclones, which could facilitate management strategies involving selective ablation of resistant clones in the setting of systemic therapy.

Peri- and intra-nodular radiomic features based on 18F-FDG PET/CT to distinguish lung adenocarcinomas from pulmonary granulomas

Objective

To compare the effectiveness of radiomic features based on 18F-FDG PET/CT images within (intranodular) and around (perinodular) lung nodules/masses in distinguishing between lung adenocarcinoma and pulmonary granulomas.

Methods

For this retrospective study, 18F-FDG PET/CT images were collected for 228 patients. Patients diagnosed with lung adenocarcinoma (n = 156) or granulomas (n = 72) were randomly assigned to a training (n = 159) and validation (n = 69) groups. The volume of interest (VOI) of intranodular, perinodular (1-5 voxels, termed Lesion_margin1 to Lesion_margin5) and total area (intra- plus perinodular region, termed Lesion_total1 to Lesion_total5) on PET/CT images were delineated using PETtumor and Marge tool of segmentation editor. A total of 1,037 radiomic features were extracted separately from PET and CT images, and the optimal features were selected to develop radiomic models. Model performance was evaluated using the area under the receiver operating characteristic curve (AUC).

Results

Good and acceptable performance was, respectively, observed in the training (AUC = 0.868, p < 0.001) and validation (AUC = 0.715, p = 0.004) sets for the intranodular radiomic model. Among the perinodular models, the Lesion_margin2 model demonstrated the highest AUC in both sets (0.883 and 0.616, p < 0.001 and p = 0.122). Similarly, in terms of total models, Lesion_total2 model was found to outperform others in the training (AUC = 0.879, p < 0.001) and validation (AUC = 0.742, p = 0.001) sets, slightly surpassing the intranodular model.

Conclusion

When intra- and perinodular radiomic features extracted from the immediate vicinity of the nodule/mass up to 2 voxels distance on 18F-FDG PET/CT imaging are combined, improved differential diagnostic performance in distinguishing between lung adenocarcinomas and granulomas is achieved compared to the intra- and perinodular radiomic features alone.

Utilizing radiomics and dosiomics with AI for precision prediction of radiation dermatitis in breast cancer patients

PURPOSE

This study explores integrating clinical features with radiomic and dosiomic characteristics into AI models to enhance the prediction accuracy of radiation dermatitis (RD) in breast cancer patients undergoing volumetric modulated arc therapy (VMAT).

MATERIALS AND METHODS

This study involved a retrospective analysis of 120 breast cancer patients treated with VMAT at Kaohsiung Veterans General Hospital from 2018 to 2023. Patient data included CT images, radiation doses, Dose-Volume Histogram (DVH) data, and clinical information. Using a Treatment Planning System (TPS), we segmented CT images into Regions of Interest (ROIs) to extract radiomic and dosiomic features, focusing on intensity, shape, texture, and dose distribution characteristics. Features significantly associated with the development of RD were identified using ANOVA and LASSO regression (p-value < 0.05). These features were then employed to train and evaluate Logistic Regression (LR) and Random Forest (RF) models, using tenfold cross-validation to ensure robust assessment of model efficacy.

RESULTS

In this study, 102 out of 120 VMAT-treated breast cancer patients were included in the detailed analysis. Thirty-two percent of these patients developed Grade 2+ RD. Age and BMI were identified as significant clinical predictors. Through feature selection, we narrowed down the vast pool of radiomic and dosiomic data to 689 features, distributed across 10 feature subsets for model construction. In the LR model, the J subset, comprising DVH, Radiomics, and Dosiomics features, demonstrated the highest predictive performance with an AUC of 0.82. The RF model showed that subset I, which includes clinical, radiomic, and dosiomic features, achieved the best predictive accuracy with an AUC of 0.83. These results emphasize that integrating radiomic and dosiomic features significantly enhances the prediction of Grade 2+ RD.

CONCLUSION

Integrating clinical, radiomic, and dosiomic characteristics into AI models significantly improves the prediction of Grade 2+ RD risk in breast cancer patients post-VMAT. The RF model analysis demonstrates that a comprehensive feature set maximizes predictive efficacy, marking a promising step towards utilizing AI in radiation therapy risk assessment and enhancing patient care outcomes.

Principles of artificial intelligence in radiooncology

PURPOSE

In the rapidly expanding field of artificial intelligence (AI) there is a wealth of literature detailing the myriad applications of AI, particularly in the realm of deep learning. However, a review that elucidates the technical principles of deep learning as relevant to radiation oncology in an easily understandable manner is still notably lacking. This paper aims to fill this gap by providing a comprehensive guide to the principles of deep learning that is specifically tailored toward radiation oncology.

METHODS

In light of the extensive variety of AI methodologies, this review selectively concentrates on the specific domain of deep learning. It emphasizes the principal categories of deep learning models and delineates the methodologies for training these models effectively.

RESULTS

This review initially delineates the distinctions between AI and deep learning as well as between supervised and unsupervised learning. Subsequently, it elucidates the fundamental principles of major deep learning models, encompassing multilayer perceptrons (MLPs), convolutional neural networks (CNNs), recurrent neural networks (RNNs), transformers, generative adversarial networks (GANs), diffusion-based generative models, and reinforcement learning. For each category, it presents representative networks alongside their specific applications in radiation oncology. Moreover, the review outlines critical factors essential for training deep learning models, such as data preprocessing, loss functions, optimizers, and other pivotal training parameters including learning rate and batch size.

CONCLUSION

This review provides a comprehensive overview of deep learning principles tailored toward radiation oncology. It aims to enhance the understanding of AI-based research and software applications, thereby bridging the gap between complex technological concepts and clinical practice in radiation oncology.

Visualization of Pulpal Structures by SWIR in Endodontic Access Preparation

Endodontic access preparation is one of the initial steps in root canal treatments and can be hindered by the obliteration of pulp canals and formation of tertiary dentin. Until now, methods for direct intraoperative visualization of the 3-dimensional anatomy of teeth have been missing. Here, we evaluate the use of shortwave infrared radiation (SWIR) for navigation during stepwise access preparation. Nine teeth (3 anteriors, 3 premolars, and 3 molars) were explanted en bloc with intact periodontium including alveolar bone and mucosa from the upper or lower jaw of human body donors. Analysis was performed at baseline as well as at preparation depths of 5 mm, 7 mm, and 9 mm, respectively. For reflection, SWIR was used at a wavelength of 1,550 nm from the occlusal direction, whereas for transillumination, SWIR was passed through each sample at the marginal gingiva from the buccal as well as oral side at a wavelength of 1,300 nm. Pulpal structures could be identified as darker areas approximately 2 mm before reaching the pulp chamber using SWIR transillumination, although they were indistinguishable under normal circumstances. Furcation areas in molars appeared with higher intensity than areas with canals. The location of pulpal structures was confirmed by superimposition of segmented micro-computed tomography (µCT) images. By radiomic analysis, significant differences between pulpal and parapulpal areas could be detected in image features. With hierarchical cluster analysis, both segments could be confirmed and associated with specific clusters. The local thickness of µCTs was calculated and correlated with SWIR transillumination images, by which a linear dependency of thickness and intensity could be demonstrated. Lastly, by in silico simulations of light propagation, dentin tubules were shown to be a crucial factor for understanding the visibility of the pulp. In conclusion, SWIR transillumination may allow direct clinical live navigation during endodontic access preparation.

Development and Validation of a CT-Radiomics Nomogram for the Diagnosis of Small Prevascular Mediastinal Nodules: Reducing Nontherapeutic Surgeries

RATIONALE AND OBJECTIVES

The preoperative diagnosis of small prevascular mediastinal nodules (SPMNs) presents a challenge, often leading to unnecessary surgical interventions. Our objective was to develop a nomogram based on preoperative CT-radiomics features, serving as a non-invasive diagnostic tool for SPMNs.

MATERIALS AND METHODS

Patients with surgically resected SPMNs from two medical centers between January 2018 and December 2022 were retrospectively reviewed. Radiomics features were extracted and screened from preoperative CT images. Logistic regression was employed to establish clinical, radiomics, and hybrid models for differentiating thymic epithelial tumors (TETs) from cysts. The performance of these models was validated in both internal and external test sets by area under the receiver operating characteristic curve (AUC), while also comparing their diagnostic capability with human experts.

RESULTS

The study enrolled a total of 363 patients (median age, 53 years [IQR:45-59 years]; 175 [48.2%] males) for model development and validation, including 136 TETs and 227 cysts. Lesions' enhancement status, shape, calcification, and rad-score were identified as independent factors for distinction. The hybrid model demonstrated superior diagnostic performance compared to other models and human experts, with an AUC of 0.95 (95% CI:0.92-0.98), 0.94 (95% CI:0.89-0.99), and 0.93 (95% CI:0.83-1.00) in the training set, internal test set, and external test set respectively. The calibration curve of the model demonstrated excellent fit, while decision curve analysis underscored its clinical value.

CONCLUSION

The radiomics-based nomogram effectively discriminates between the most prevalent types of SPMNs, namely TETs and cysts, thus presenting a promising tool for treatment guidance.

Bone metastasis prediction in non-small-cell lung cancer: primary CT-based radiomics signature and clinical feature

BACKGROUND

Radiomics provided opportunities to quantify the tumor phenotype non-invasively. This study extracted contrast-enhanced computed tomography (CECT) radiomic signatures and evaluated clinical features of bone metastasis in non-small-cell lung cancer (NSCLC). With the combination of the revealed radiomics and clinical features, the predictive modeling on bone metastasis in NSCLC was established.

METHODS

A total of 318 patients with NSCLC at the Tianjin Medical University Cancer Institute & Hospital was enrolled between January 2009 and December 2019, which included a feature-learning cohort (n = 223) and a validation cohort (n = 95). We trained a radiomics model in 318 CECT images from feature-learning cohort to extract the radiomics features of bone metastasis in NSCLC. The Kruskal-Wallis and the least absolute shrinkage and selection operator regression (LASSO) were used to select bone metastasis-related features and construct the CT radiomics score (Rad-score). Multivariate logistic regression was performed with the combination of the Rad-score and clinical data. A predictive nomogram was subsequently developed.

RESULTS

Radiomics models using CECT scans were significant on bone metastasis prediction in NSCLC. Model performance was enhanced with each information into the model. The radiomics nomogram achieved an AUC of 0.745 (95% confidence interval [CI]: 0.68,0.80) on predicting bone metastasis in the training set and an AUC of 0.808(95% confidence interval [CI]: 0.71,0.88) in the validation set.

CONCLUSION

The revealed invisible image features were of significance on guiding bone metastasis prediction in NSCLC. Based on the combination of the image features and clinical characteristics, the predictive nomogram was established. Such nomogram can be used for the auxiliary screening of bone metastasis in NSCLC.

A 18F-FDG PET/CT based radiomics nomogram for predicting disease-free survival in stage II/III colorectal adenocarcinoma

OBJECTIVES

This study aimed to establish a clinical nomogram model based on a radiomics signatures derived from 18F-fluorodeoxyglucose positron-emission tomography (18F-FDG PET/CT) and clinical parameters to predict disease-free survival (DFS) in patients with stage II/III colorectal adenocarcinoma. Understanding and predicting DFS in these patients is key to optimizing treatment strategies.

METHODS

A retrospective analysis included 332 cases from July 2011 to July 2021 at The Sixth Affiliated Hospital, Sun Yat-sen University, with PET/CT assessing radiomics features and clinicopathological features. Univariate Cox regression, the least absolute shrinkage and selection operator (LASSO) Cox, and multivariable Cox regression identified recurrence-related radiomics features. We used a weighted radiomics score (Rad-score) and independent risk factors to construct a nomogram. Evaluation involved time-dependent receiver operating characteristic (ROC) curves, calibration curves, and decision curve analysis (DCA).

RESULTS

The nomogram, incorporating Rad-score, pN, and pT demonstrated robust predictive ability for DFS in stage II/III colorectal adenocarcinoma. Training cohort areas under the curve (AUCs) were 0.78, 0.80, and 0.86 at 1, 2, and 3 years, respectively, and validation cohort AUCs were 0.79, 0.75, and 0.73. DCA and calibration curves affirmed the nomogram's clinical relevance.

CONCLUSION

The 18F-FDG PET/CT based radiomics nomogram, including Rad-score, pN, and pT, effectively predicted tumor recurrence in stage II/III colorectal adenocarcinoma, significantly enhancing prognostic stratification. Our findings highlight the potential of this nomogram as a guide for clinical decision making to improve patient outcomes.

Ultrasound-Based Radiomics for Predicting the WHO/ISUP Grading of Clear-Cell Renal Cell Carcinoma

OBJECTIVE

To explore the performance of ultrasound image-based radiomics in predicting World Health Organization (WHO)/International Society of Urological Pathology (ISUP) grading of clear-cell renal cell carcinoma (ccRCC).

METHODS

A retrospective study was conducted via histopathological examination on participants with ccRCC from January 2021 to August 2023. Participants were randomly allocated to a training set and a validation set in a 3:1 ratio. The maximum cross-sectional image of the lesion on the preoperative ultrasound image was obtained, with the region of interest (ROI) delineated manually. Radiomic features were computed from the ROIs and subsequently normalized using Z-scores. Wilcoxon test and least absolute shrinkage and selection operator (LASSO) regression were applied for feature reduction and model development. The performance of the model was estimated by indicators including area under the curve (AUC), sensitivity and specificity.

RESULTS

A total of 336 participants (median age, 57 y; 106 women) with ccRCC were finally included, of whom 243 had low-grade tumors (grade 1-2) and 93 had high-grade tumors (grade 3-4). A total of 1163 radiomic features were extracted from the ROIs for model construction and 117 informative radiomics features selected by Wilcoxon test were submitted to LASSO. Our ultrasound-based radiomics model included 51 features and achieved AUCs of 0.90 and 0.79 for the training and validation sets, respectively. Within the training set, the sensitivity and specificity measured 0.75 and 0.92, respectively, whereas in the validation set, the sensitivity and specificity measured 0.65 and 0.84, respectively. In the subgroup analysis, for the training and validation sets Philips AUCs were 0.91 and 0.75, Toshiba AUCs were 0.82 and 0.90, and General Electric AUCs were 0.95 and 0.82, respectively.

CONCLUSION

Ultrasound-based radiomics can effectively predict the WHO/ISUP grading of ccRCC.

Correlation Between Cognitive Impairment and Lenticulostriate Arteries: A Clinical and Radiomics Analysis

Lenticulostriate arteries (LSA) are potentially valuable for studying vascular cognitive impairment. This study aims to investigate correlations between cognitive impairment and LSA through clinical and radiomics features analysis. We retrospectively included 102 patients (mean age 62.5±10.3 years, 60 males), including 58 with mild cognitive impairment (MCI) and 44 with moderate or severe cognitive impairment (MSCI). The MRI images of these patients were subjected to z-score preprocessing, manual regions of interest (ROI) outlining, feature extraction (pyradiomics), feature selection [max-relevance and min-redundancy (mRMR), least absolute shrinkage and selection operator (LASSO), and univariate analysis], model construction (multivariate logistic regression), and evaluation [receiver operating characteristic curve (ROC), decision curve analysis (DCA), and calibration curves (CC)]. In the training dataset (71 patients, 44 MCI) and the test dataset (31 patients, 17 MCI), the area under curve (AUC) of the combined model (training 0.88 [95% CI 0.78, 0.97], test 0.76 [95% CI 0.6, 0.93]) was better than that of the clinical model and the radiomics model. The DCA results demonstrated the highest net yield of the combined model relative to the clinical and radiomics models. In addition, we found that LSA total vessel count (0.79 [95% CI 0.08, 1.59], P = 0.038) and wavelet.HLH_glcm_MCC (-1.2 [95% CI -2.2, -0.4], P = 0.008) were independent predictors of MCI. The model that combines clinical and radiomics features of LSA can predict MCI. Besides, LSA vascular parameters may serve as imaging biomarkers of cognitive impairment.

Whole-liver histogram analysis of hepatocyte-specific contrast-enhanced magnetic resonance imaging for predicting progression in patients with cirrhosis

Background

Liver cirrhosis, as the terminal phase of chronic liver disease fibrosis, is associated with high morbidity and mortality. Traditional methods for assessing liver function, such as clinical scoring systems, offer only a global evaluation and may not accurately reflect regional liver function variations. This study aimed at evaluating the diagnostic potential of whole-liver histogram analysis of gadobenate dimeglumine (Gd-BOPTA)-enhanced magnetic resonance imaging (MRI) for predicting the progression of cirrhosis.

Methods

In this retrospective study, 265 consecutive patients with cirrhosis admitted to the Department of Radiology, Shandong Provincial Hospital Affiliated to Shandong First Medical University from August 2012 to September 2019 were enrolled. After the exclusion criteria were applied, 117 patients (84 males and 33 females) were divided into Child-Pugh A cirrhosis (n=43), Child-Pugh B cirrhosis (n=49), and Child-Pugh C cirrhosis (n=25). After correction for liver signal intensity with the spleen was completed, 19 histogram features of the whole liver were extracted and modeled to evaluate liver function, with the Child-Pugh class being incorporated as a clinical parameter. Receiver operating characteristic (ROC) curves were used to assess the diagnosis capability and determine the optimal cutoffs after a mean follow-up of 42.3±19.1 (range, 8-93) months. The association between significant histogram features and the cumulative incidence of hepatic insufficiency was analyzed with the adjusted Kaplan-Meier curve model.

Results

Among 117 patients (12%), 14 developed hepatic insufficiency through a period of follow-up. Five features, including the median (P<0.01), 90th percentile (P<0.01), root mean squared (P<0.01), mean (P<0.01), and 10th percentile (P<0.05), were significantly different between the groups with and without hepatic insufficiency according to the Kruskal-Wallis test; in the ROC curve analysis, the area under the curve (AUC) of these features was 0.723 [95% confidence interval (CI): 0.653-0.793], 0.722 (95% CI: 0.652-0.792), 0.722 (95% CI: 0.652-0.792), 0.721 (95% CI: 0.651-0.791), and 0.674 (95% CI: 0.600-0.748) after correction, respectively (all P values <0.05). Median, 90th percentile, root mean squared, and mean were found to be significant factors in predicting liver insufficiency. The adjusted Kaplan-Meier curves revealed that patients with a feature level less than the cutoff, as compared to those with a level above the cutoff, showed a statistically shorter progression-free survival and higher incidences of hepatic insufficiency for significant features of median (cutoff =26.001; 21.28% versus 5.71%; P=0.02), 90th percentile (cutoff =86.263; 20.41% versus 5.88%; P<0.01), root mean squared (cutoff =1,028.477; 19.15% versus 7.14%; P=0.049), and mean (cutoff =27.484; 19.15% versus 7.14%; P=0.049). Patients with a 10th percentile less than -39.811 also showed a higher cumulative incidence of hepatic insufficiency than did those with a value higher than the cutoff (0.18% versus 7.46%; P=0.22).

Conclusions

Whole-liver histogram analysis of Gd-BOPTA-enhanced MRI may serve as a noninvasive analytical method to predict hepatic insufficiency in patients with cirrhosis.

RadShap: An Explanation Tool for Highlighting the Contributions of Multiple Regions of Interest to the Prediction of Radiomic Models

Explaining the decisions made by a radiomic model is of significant interest, as it can provide valuable insights into the information learned by complex models and foster trust in well-performing ones, thereby facilitating their clinical adoption. Promising radiomic approaches that aggregate information from multiple regions within an image currently lack suitable explanation tools that could identify the regions that most significantly influence their decisions. Here we present a model- and modality-agnostic tool (RadShap, https://github.com/ncaptier/radshap), based on Shapley values, that explains the predictions of multiregion radiomic models by highlighting the contribution of each individual region. Methods: The explanation tool leverages Shapley values to distribute the aggregative radiomic model's output among all the regions of interest of an image, highlighting their individual contribution. RadShap was validated using a retrospective cohort of 130 patients with advanced non-small cell lung cancer undergoing first-line immunotherapy. Their baseline PET scans were used to build 1,000 synthetic tasks to evaluate the degree of alignment between the tool's explanations and our data generation process. RadShap's potential was then illustrated through 2 real case studies by aggregating information from all segmented tumors: the prediction of the progression-free survival of the non-small cell lung cancer patients and the classification of the histologic tumor subtype. Results: RadShap demonstrated strong alignment with the ground truth, with a median frequency of 94% for consistently explained predictions in the synthetic tasks. In both real-case studies, the aggregative models yielded superior performance to the single-lesion models (average [±SD] time-dependent area under the receiver operating characteristic curve was 0.66 ± 0.02 for the aggregative survival model vs. 0.55 ± 0.04 for the primary tumor survival model). The tool's explanations provided relevant insights into the behavior of the aggregative models, highlighting that for the classification of the histologic subtype, the aggregative model used information beyond the biopsy site to correctly classify patients who were initially misclassified by a model focusing only on the biopsied tumor. Conclusion: RadShap aligned with ground truth explanations and provided valuable insights into radiomic models' behaviors. It is implemented as a user-friendly Python package with documentation and tutorials, facilitating its smooth integration into radiomic pipelines.

Automated MRI liver segmentation for anatomical segmentation, liver volumetry, and the extraction of radiomics

OBJECTIVES

To develop and evaluate a deep convolutional neural network (DCNN) for automated liver segmentation, volumetry, and radiomic feature extraction on contrast-enhanced portal venous phase magnetic resonance imaging (MRI).

MATERIALS AND METHODS

This retrospective study included hepatocellular carcinoma patients from an institutional database with portal venous MRI. After manual segmentation, the data was randomly split into independent training, validation, and internal testing sets. From a collaborating institution, de-identified scans were used for external testing. The public LiverHccSeg dataset was used for further external validation. A 3D DCNN was trained to automatically segment the liver. Segmentation accuracy was quantified by the Dice similarity coefficient (DSC) with respect to manual segmentation. A Mann-Whitney U test was used to compare the internal and external test sets. Agreement of volumetry and radiomic features was assessed using the intraclass correlation coefficient (ICC).

RESULTS

In total, 470 patients met the inclusion criteria (63.9±8.2 years; 376 males) and 20 patients were used for external validation (41±12 years; 13 males). DSC segmentation accuracy of the DCNN was similarly high between the internal (0.97±0.01) and external (0.96±0.03) test sets (p=0.28) and demonstrated robust segmentation performance on public testing (0.93±0.03). Agreement of liver volumetry was satisfactory in the internal (ICC, 0.99), external (ICC, 0.97), and public (ICC, 0.85) test sets. Radiomic features demonstrated excellent agreement in the internal (mean ICC, 0.98±0.04), external (mean ICC, 0.94±0.10), and public (mean ICC, 0.91±0.09) datasets.

CONCLUSION

Automated liver segmentation yields robust and generalizable segmentation performance on MRI data and can be used for volumetry and radiomic feature extraction.

CLINICAL RELEVANCE STATEMENT

Liver volumetry, anatomic localization, and extraction of quantitative imaging biomarkers require accurate segmentation, but manual segmentation is time-consuming. A deep convolutional neural network demonstrates fast and accurate segmentation performance on T1-weighted portal venous MRI.

KEY POINTS

• This deep convolutional neural network yields robust and generalizable liver segmentation performance on internal, external, and public testing data. • Automated liver volumetry demonstrated excellent agreement with manual volumetry. • Automated liver segmentations can be used for robust and reproducible radiomic feature extraction.

Adaptive Machine Learning Approach for Importance Evaluation of Multimodal Breast Cancer Radiomic Features

Breast cancer holds the highest diagnosis rate among female tumors and is the leading cause of death among women. Quantitative analysis of radiological images shows the potential to address several medical challenges, including the early detection and classification of breast tumors. In the P.I.N.K study, 66 women were enrolled. Their paired Automated Breast Volume Scanner (ABVS) and Digital Breast Tomosynthesis (DBT) images, annotated with cancerous lesions, populated the first ABVS+DBT dataset. This enabled not only a radiomic analysis for the malignant vs. benign breast cancer classification, but also the comparison of the two modalities. For this purpose, the models were trained using a leave-one-out nested cross-validation strategy combined with a proper threshold selection approach. This approach provides statistically significant results even with medium-sized data sets. Additionally it provides distributional variables of importance, thus identifying the most informative radiomic features. The analysis proved the predictive capacity of radiomic models even using a reduced number of features. Indeed, from tomography we achieved AUC-ROC 89.9 % using 19 features and 92.1 % using 7 of them; while from ABVS we attained an AUC-ROC of 72.3 % using 22 features and 85.8 % using only 3 features. Although the predictive power of DBT outperforms ABVS, when comparing the predictions at the patient level, only 8.7% of lesions are misclassified by both methods, suggesting a partial complementarity. Notably, promising results (AUC-ROC ABVS-DBT 71.8 % - 74.1 % ) were achieved using non-geometric features, thus opening the way to the integration of virtual biopsy in medical routine.

Radiomics-based prediction of nonalcoholic fatty liver disease following pancreatoduodenectomy

PURPOSE

Predicting nonalcoholic fatty liver disease (NAFLD) following pancreaticoduodenectomy (PD) is challenging, which delays therapeutic intervention and makes its prevention difficult. We conducted this study to assess the potential application of preoperative computed tomography (CT) radiomics for predicting NAFLD.

METHODS

The subjects of this retrospective study were 186 patients with PD from a single institution. We extracted the predictors of NAFLD after PD statistically from conventional clinical and radiomic features of the estimated remnant pancreas and whole liver region on preoperative nonenhanced CT images. Based on these predictors, we developed a machine-learning predictive model, which integrated clinical and radiomic features. A comparative model used only clinical features as predictors.

RESULTS

The incidence of NAFLD after PD was 43.5%. The variables of the clinicoradiomic model included one shape feature of the pancreas, two texture features of the liver, and sex; the variables of the clinical model were age, sex, and chemoradiotherapy. The accuracy%, precision%, recall%, F1 score, and area under the curve of the two models were 75.0, 72.7, 66.7, 69.6, and 0.80; and 69.6, 68.4, 54.2, 60.5, and 0.69, respectively.

CONCLUSIONS

Preoperative CT-derived radiomic features from the pancreatic and liver regions are promising for the prediction of NAFLD post-PD. Using these features enhances the predictive model, enabling earlier intervention for high-risk patients.

Machine-Learning and Radiomics-Based Preoperative Prediction of Ki-67 Expression in Glioma Using MRI Data

BACKGROUND

Gliomas are the most common primary brain tumours and constitute approximately half of all malignant glioblastomas. Unfortunately, patients diagnosed with malignant glioblastomas typically survive for less than a year. In light of this circumstance, genotyping is an effective means of categorising gliomas. The Ki67 proliferation index, a widely used marker of cellular proliferation in clinical contexts, has demonstrated potential for predicting tumour classification and prognosis. In particular, magnetic resonance imaging (MRI) plays a vital role in the diagnosis of brain tumours. Using MRI to extract glioma-related features and construct a machine learning model offers a viable avenue to classify and predict the level of Ki67 expression.

METHODS

This study retrospectively collected MRI data and postoperative immunohistochemical results from 613 glioma patients from the First Affliated Hospital of Nanjing Medical University. Subsequently, we performed registration and skull stripping on the four MRI modalities: T1-weighted (T1), T2-weighted (T2), T1-weighted with contrast enhancement (T1CE), and Fluid Attenuated Inversion Recovery (FLAIR). Each modality's segmentation yielded three distinct tumour regions. Following segmentation, a comprehensive set of features encompassing texture, first-order, and shape attributes were extracted from these delineated regions. Feature selection was conducted using the least absolute shrinkage and selection operator (LASSO) algorithm with subsequent sorting to identify the most important features. These selected features were further analysed using correlation analysis to finalise the selection for machine learning model development. Eight models: logistic regression (LR), naive bayes, decision tree, gradient boosting tree, and support vector classification (SVM), random forest (RF), XGBoost, and LightGBM were used to objectively classify Ki67 expression.

RESULTS

In total, 613 patients were enroled in the study, and 24,455 radiomic features were extracted from each patient's MRI. These features were eventually reduced to 36 after LASSO screening, RF importance ranking, and correlation analysis. Among all the tested machine learning models, LR and linear SVM exhibited superior performance. LR achieved the highest area under the curve score of 0.912 ± 0.036, while linear SVM obtained the top accuracy with a score of 0.884 ± 0.031.

CONCLUSION

This study introduced a novel approach for classifying Ki67 expression levels using MRI, which has been proven to be highly effective. With the LR model at its core, our method demonstrated its potential in signalling a promising avenue for future research. This innovative approach of predicting Ki67 expression based on MRI features not only enhances our understanding of cell activity but also represents a significant leap forward in brain glioma research. This underscores the potential of integrating machine learning with medical imaging to aid in the diagnosis and prognosis of complex diseases.

Early warning of hepatocellular carcinoma in cirrhotic patients by three-phase CT-based deep learning radiomics model: a retrospective, multicentre, cohort study

Background

The diagnosis of hepatocellular carcinoma (HCC) often experiences latency, ultimately leading to unfavorable patient outcomes due to delayed therapeutic interventions. Our study is designed to develop and validate a model that employs triple-phase computerized tomography (CT)-based deep learning radiomics and clinical variables for early warning of HCC in patients with cirrhosis.

Methods

We studied 1858 patients with cirrhosis primarily from the PreCar cohort (NCT03588442) between June 2018 and January 2020 at 11 centres, and collected triple-phase CT images and laboratory results 3-12 months prior to HCC diagnosis or non-HCC final follow-up. Using radiomics and deep learning techniques, early warning model was developed in the discovery cohort (n = 924), and then validated in an internal validation cohort (n = 231), and an external validation cohort from 10 external centres (n = 703).

Findings

We developed a hybrid model, named ALARM model, which integrates deep learning radiomics with clinical variables, enabling early warning of the majority of HCC cases. The ALARM model effectively predicted short-term HCC development in cirrhotic patients with area under the curve (AUC) of 0.929 (95% confidence interval 0.918-0.941) in the discovery cohort, 0.902 (0.818-0.987) in the internal validation cohort, and 0.918 (0.898-0.961) in the external validation cohort. By applying optimal thresholds of 0.21 and 0.65, the high-risk (n = 221, 11.9%) and medium-risk (n = 433, 23.3%) groups, which covered 94.4% (84/89) of the patients who developed HCC, had significantly higher rates of HCC occurrence compared to the low-risk group (n = 1204, 64.8%) (24.3% vs 6.4% vs 0.42%, P < 0.001). Furthermore, ALARM also demonstrated consistent performance in subgroup analysis.

Interpretation

The novel ALARM model, based on deep learning radiomics with clinical variables, provides reliable estimates of short-term HCC development for cirrhotic patients, and may have the potential to improve the precision in clinical decision-making and early initiation of HCC treatments.

Funding

This work was supported by National Key Research and Development Program of China (2022YFC2303600, 2022YFC2304800), and the National Natural Science Foundation of China (82170610), Guangdong Basic and Applied Basic Research Foundation (2023A1515011211).

Multi-sequence MRI-based radiomics: An objective method to diagnose early-stage osteonecrosis of the femoral head

OBJECTIVES

This study investigated the use of radiomics for diagnosing early-stage osteonecrosis of the femoral head (ONFH) by extracting features from multiple MRI sequences and constructing predictive models.

MATERIALS AND METHODS

We conducted a retrospective review, collected MR images of early-stage ONFH (102 from institution A and 20 from institution B) and healthy femoral heads (102 from institution A and 20 from institution B) from two institutions. We extracted radiomics features, handled batch effects using Combat, and normalized features using z-score. We employed the Least absolute shrinkage and selection operator (LASSO) algorithm, along with Max-Relevance and Min-Redundancy (mRMR), to select optimal features for constructing radiomics models based on single, double, and multi-sequence MRI data. We evaluated performance using receiver operating characteristic (ROC) and precision-recall (PR) curves, and compared area under curve of ROC (AUC-ROC) values with the DeLong test. Additionally, we studied the diagnostic performance of the multi-sequence radiomics model and radiologists, compared the diagnostic outcomes of the model and radiologists using the Fisher exact test.

RESULTS

We studied 122 early-stage ONFH and 122 normal femoral heads. The multi-sequence model exhibited the best diagnostic performance among all models (AUC-ROC, PR-AUC for training set: 0.96, 0.961; validation set: 0.96, 0.97; test set: 0.94, 0.94), and it outperformed three resident radiologists on the external testing group with an accuracy of 87.5 %, sensitivity of 85.00 %, and specificity of 90.00 % (p < 0.01), highlighting the robustness of our findings.

CONCLUSIONS

Our study underscored the novelty of the multi-sequence radiomics model in diagnosing early-stage ONFH. By leveraging features extracted from multiple imaging sequences, this approach demonstrated high efficacy, indicating its potential to advance early diagnosis for ONFH. These findings provided important guidance for enhancing early diagnosis of ONFH through radiomics methods, offering new avenues and possibilities for clinical practice and patient care.

Measurement plane of the cross-sectional area of the masseter muscle in patients with skeletal Class III malocclusion: An artificial intelligence model

INTRODUCTION

This study aimed to determine a measurement plane that could represent the maximum cross-sectional area (MCSA) of masseter muscle using an artificial intelligence model for patients with skeletal Class III malocclusion.

METHODS

The study included 197 patients, divided into subgroups according to sex, mandibular symmetry, and mandibular plane angle. The volume, MCSA, and the cross-sectional area (CSA) at different levels were calculated automatically. The vertical distance between MCSA and mandibular foramen, along with the ratio of the masseter CSA at different levels to the MCSA (R), were also calculated.

RESULTS

The MCSA and volume showed a strong correlation in the total sample and each subgroup (P <0.001). The correlation between the CSA at each level and MCSA was statistically significant (P <0.001). The peak of the r and the correlation coefficient between the CSA at different levels and MCSA were mostly present 5-10 mm above the mandibular foramen for the total sample and the subgroups. The mean of RA5 to RA10 was >0.93, whereas the corresponding correlation coefficient was >0.96, both for the entire sample and for the subgroups.

CONCLUSIONS

MCSA could be used as an indicator for masseter muscle size. For patients with skeletal Class III malocclusion, the CSA 5-10 mm above the mandibular foramen, parallel to the Frankfort plane, could be used to estimate the masseter muscle MCSA.

Developing, Purchasing, Implementing and Monitoring AI Tools in Radiology: Practical Considerations. A Multi-Society Statement From the ACR, CAR, ESR, RANZCR & RSNA

Artificial intelligence (AI) carries the potential for unprecedented disruption in radiology, with possible positive and negative consequences. The integration of AI in radiology holds the potential to revolutionize healthcare practices by advancing diagnosis, quantification, and management of multiple medical conditions. Nevertheless, the ever-growing availability of AI tools in radiology highlights an increasing need to critically evaluate claims for its utility and to differentiate safe product offerings from potentially harmful, or fundamentally unhelpful ones. This multi-society paper, presenting the views of Radiology Societies in the USA, Canada, Europe, Australia, and New Zealand, defines the potential practical problems and ethical issues surrounding the incorporation of AI into radiological practice. In addition to delineating the main points of concern that developers, regulators, and purchasers of AI tools should consider prior to their introduction into clinical practice, this statement also suggests methods to monitor their stability and safety in clinical use, and their suitability for possible autonomous function. This statement is intended to serve as a useful summary of the practical issues which should be considered by all parties involved in the development of radiology AI resources, and their implementation as clinical tools. KEY POINTS.

Radiomics for clinical decision support in radiation oncology

Radiomics is a promising tool for the development of quantitative biomarkers to support clinical decision-making. It has been shown to improve the prediction of response to treatment and outcome in different settings, particularly in the field of radiation oncology by optimising the dose delivery solutions and reducing the rate of radiation-induced side effects, leading to a fully personalised approach. Despite the promising results offered by radiomics at each of these stages, standardised methodologies, reproducibility and interpretability of results are still lacking, limiting the potential clinical impact of these tools. In this review, we briefly describe the principles of radiomics and the most relevant applications of radiomics at each stage of cancer management in the framework of radiation oncology. Furthermore, the integration of radiomics into clinical decision support systems is analysed, defining the challenges and offering possible solutions for translating radiomics into a clinically applicable tool.

Deep Learning for Automatic Gross Tumor Volumes Contouring in Esophageal Cancer Based on Contrast-Enhanced Computed Tomography Images: A Multi-Institutional Study

PURPOSE

To develop and externally validate an automatic artificial intelligence (AI) tool for delineating gross tumor volume (GTV) in patients with esophageal squamous cell carcinoma (ESCC), which can assist in neo-adjuvant or radical radiation therapy treatment planning.

METHODS AND MATERIALS

In this multi-institutional study, contrast-enhanced CT images from 580 eligible ESCC patients were retrospectively collected. The GTV contours delineated by 2 experts via consensus were used as ground truth. A 3-dimensional deep learning model was developed for GTV contouring in the training cohort and internally and externally validated in 3 validation cohorts. The AI tool was compared against 12 board-certified experts in 25 patients randomly selected from the external validation cohort to evaluate its assistance in improving contouring performance and reducing variation. Contouring performance was measured using dice similarity coefficient (DSC) and average surface distance. Additionally, our previously established radiomics model for predicting pathologic complete response was used to compare AI-generated and ground truth contours, to assess the potential of the AI contouring tool in radiomics analysis.

RESULTS

The AI tool demonstrated good GTV contouring performance in multicenter validation cohorts, with median DSC values of 0.865, 0.876, and 0.866 and median average surface distance values of 0.939, 0.789, and 0.875 mm, respectively. Furthermore, the AI tool significantly improved contouring performance for half of 12 board-certified experts (DSC values, 0.794-0.835 vs 0.856-0.881, P = .003-0.048), reduced the intra- and interobserver variations by 37.4% and 55.2%, respectively, and saved contouring time by 77.6%. In the radiomics analysis, 88.7% of radiomic features from ground truth and AI-generated contours demonstrated stable reproducibility, and similar pathologic complete response prediction performance for these contours (P = .430) was observed.

CONCLUSIONS

Our AI contouring tool can improve GTV contouring performance and facilitate radiomics analysis in ESCC patients, which indicates its potential for GTV contouring during radiation therapy treatment planning and radiomics studies.

Radiomic- and dosiomic-based clustering development for radio-induced neurotoxicity in pediatric medulloblastoma

BACKGROUND

Texture analysis extracts many quantitative image features, offering a valuable, cost-effective, and non-invasive approach for individual medicine. Furthermore, multimodal machine learning could have a large impact for precision medicine, as texture biomarkers can underlie tissue microstructure. This study aims to investigate imaging-based biomarkers of radio-induced neurotoxicity in pediatric patients with metastatic medulloblastoma, using radiomic and dosiomic analysis.

METHODS

This single-center study retrospectively enrolled children diagnosed with metastatic medulloblastoma (MB) and treated with hyperfractionated craniospinal irradiation (CSI). Histological confirmation of medulloblastoma and baseline follow-up magnetic resonance imaging (MRI) were mandatory. Treatment involved helical tomotherapy (HT) delivering a dose of 39 Gray (Gy) to brain and spinal axis and a posterior fossa boost up to 60 Gy. Clinical outcomes, such as local and distant brain control and neurotoxicity, were recorded. Radiomic and dosiomic features were extracted from tumor regions on T1, T2, FLAIR (fluid-attenuated inversion recovery) MRI-maps, and radiotherapy dose distribution. Different machine learning feature selection and reduction approaches were performed for supervised and unsupervised clustering.

RESULTS

Forty-eight metastatic medulloblastoma patients (29 males and 19 females) with a mean age of 12 ± 6 years were enrolled. For each patient, 332 features were extracted. Greater level of abstraction of input data by combining selection of most performing features and dimensionality reduction returns the best performance. The resulting one-component radiomic signature yielded an accuracy of 0.73 with sensitivity, specificity, and precision of 0.83, 0.64, and 0.68, respectively.

CONCLUSIONS

Machine learning radiomic-dosiomic approach effectively stratified pediatric medulloblastoma patients who experienced radio-induced neurotoxicity. Strategy needs further validation in external dataset for its potential clinical use in ab initio management paradigms of medulloblastoma.

CT Angiography Radiomics Combining Traditional Risk Factors to Predict Brain Arteriovenous Malformation Rupture: a Machine Learning, Multicenter Study

This study aimed to develop a machine learning model for predicting brain arteriovenous malformation (bAVM) rupture using a combination of traditional risk factors and radiomics features. This multicenter retrospective study enrolled 586 patients with unruptured bAVMs from 2010 to 2020. All patients were grouped into the hemorrhage (n = 368) and non-hemorrhage (n = 218) groups. The bAVM nidus were segmented on CT angiography images using Slicer software, and radiomic features were extracted using Pyradiomics. The dataset included a training set and an independent testing set. The machine learning model was developed on the training set and validated on the testing set by merging numerous base estimators and a final estimator based on the stacking method. The area under the receiver operating characteristic (ROC) curve, precision, and the f1 score were evaluated to determine the performance of the model. A total of 1790 radiomics features and 8 traditional risk factors were contained in the original dataset, and 241 features remained for model training after L1 regularization filtering. The base estimator of the ensemble model was Logistic Regression, whereas the final estimator was Random Forest. In the training set, the area under the ROC curve of the model was 0.982 (0.967-0.996) and 0.893 (0.826-0.960) in the testing set. This study indicated that radiomics features are a valuable addition to traditional risk factors for predicting bAVM rupture. In the meantime, ensemble learning can effectively improve the performance of a prediction model.