Interpretation of radiomics features-A pictorial review

Multiparametric MRI dataset for susceptibility-based radiomic feature extraction and analysis

Multiple sclerosis (MS) is a progressive demyelinating disease impacting the central nervous system. Conventional Magnetic Resonance Imaging (MRI) techniques (e.g., T2w images) help diagnose MS, although they sometimes reveal non-specific lesions. Quantitative MRI techniques are capable of quantifying imaging biomarkers in vivo, offering the potential to identify specific signs related to pre-clinical inflammation. Among those techniques, Quantitative Susceptibility Mapping (QSM) is particularly useful for studying processes that influence the magnetic properties of brain tissue, such as alterations in myelin concentration. Because of its intrinsic quantitative nature, it is particularly well-suited to be analyzed through radiomics, including techniques that extract a high number of complex and multi-dimensional features from radiological images. The dataset presented in this work provides information about normal-appearing white matter (NAWM) in a cohort of MS patients and healthy controls. It includes QSM-based radiomic features from NAWM and its tracts, and MR sequences necessary to implement the pipeline: T1w, T2w, QSM, DWI. The workflow is outlined in this article, along with an application showing feature reliability assessment.

Radiomics Features in Predicting Human Papillomavirus Status in Oropharyngeal Squamous Cell Carcinoma: A Systematic Review, Quality Appraisal, and Meta-Analysis

We sought to determine the diagnostic accuracy of radiomics features in predicting HPV status in oropharyngeal squamous cell carcinoma (SCC) compared to routine paraclinical measures used in clinical practice. Twenty-six articles were included in the systematic review, and thirteen were used for the meta-analysis. The overall sensitivity of the included studies was 0.78, the overall specificity was 0.76, and the overall area under the ROC curve was 0.84. The diagnostic odds ratio (DOR) equaled 12 (8, 17). Subgroup analysis showed no significant difference between radiomics features extracted from CT or MR images. Overall, the studies were of low quality in regard to radiomics quality score, although most had a low risk of bias based on the QUADAS-2 tool. Radiomics features showed good overall sensitivity and specificity in determining HPV status in OPSCC, though the low quality of the included studies poses problems for generalizability.

A Radiomic "Warning Sign" of Progression on Brain MRI in Individuals with MS

BACKGROUND AND PURPOSE

MS is a chronic progressive, idiopathic, demyelinating disorder whose diagnosis is contingent on the interpretation of MR imaging. New MR imaging lesions are an early biomarker of disease progression. We aimed to evaluate a machine learning model based on radiomics features in predicting progression on MR imaging of the brain in individuals with MS.

MATERIALS AND METHODS

This retrospective cohort study with external validation on open-access data obtained full ethics approval. Longitudinal MR imaging data for patients with MS were collected and processed for machine learning. Radiomics features were extracted at the future location of a new lesion in the patients' prior MR imaging ("prelesion"). Additionally, "control" samples were obtained from the normal-appearing white matter for each participant. Machine learning models for binary classification were trained and tested and then evaluated the external data of the model.

RESULTS

The total number of participants was 167. Of the 147 in the training/test set, 102 were women and 45 were men. The average age was 42 (range, 21-74 years). The best-performing radiomics-based model was XGBoost, with accuracy, precision, recall, and F1-score of 0.91, 0.91, 0.91, and 0.91 on the test set, and 0.74, 0.74, 0.74, and 0.70 on the external validation set. The 5 most important radiomics features to the XGBoost model were associated with the overall heterogeneity and low gray-level emphasis of the segmented regions. Probability maps were produced to illustrate potential future clinical applications.

CONCLUSIONS

Our machine learning model based on radiomics features successfully differentiated prelesions from normal-appearing white matter. This outcome suggests that radiomics features from normal-appearing white matter could serve as an imaging biomarker for progression of MS on MR imaging.

AdaRes: A deep learning-based model for ultrasound image denoising: Results of image quality metrics, radiomics, artificial intelligence, and clinical studies

PURPOSE

The quality of ultrasound images is degraded by speckle and Gaussian noises. This study aims to develop a deep-learning (DL)-based filter for ultrasound image denoising.

METHODS

A novel DL-based filter using adaptive residual (AdaRes) learning was proposed. Five image quality metrics (IQMs) and 27 radiomics features were used to evaluate denoising results. The effect of our proposed filter, AdaRes, on four pre-trained convolutional neural network (CNN) classification models and three radiologists was assessed.

RESULTS

AdaRes filter was tested on both natural and ultrasound image databases. IQMs results indicate that AdaRes could remove noises in three different noise levels with the highest performances. In addition, a radiomics study proved that AdaRes did not distort tissue textures and it could preserve most radiomics features. AdaRes could also improve the performance classification using CNNs in different settings. Finally, AdaRes also improved the mean overall performance (AUC) of three radiologists from 0.494 to 0.702 in the classification of benign and malignant lesions.

CONCLUSIONS

AdaRes filtered out noises on ultrasound images more effectively and can be used as an auxiliary preprocessing step in computer-aided diagnosis systems. Radiologists may use it to remove unwanted noises and improve the ultrasound image quality before the interpretation.

Machine Learning-Based Breast Tumor Ultrasound Radiomics for Pre-operative Prediction of Axillary Sentinel Lymph Node Metastasis Burden in Early-Stage Invasive Breast Cancer

OBJECTIVE

The aim of the work described here was to assess the application of ultrasound (US) radiomics with machine learning (ML) classifiers to the prediction of axillary sentinel lymph node metastasis (SLNM) burden in early-stage invasive breast cancer (IBC).

METHODS

In this study, 278 early-stage IBC patients with at least one SLNM (195 in the training set and 83 in the test set) were studied at our institution. Pathologic SLNM burden was used as the reference standard. The US radiomics features of breast tumors were extracted by using 3D-Slicer and PyRadiomics software. Four ML classifiers-linear discriminant analysis (LDA), support vector machine (SVM), random forest (RF) and decision tree (DT)-were used to construct radiomics models for the prediction of SLNM burden. The combined clinicopathologic-radiomics models were also assessed with respect to sensitivity, specificity, accuracy and areas under the curve (AUCs).

RESULTS

Among the US radiomics models, the SVM classifier achieved better predictive performance with an AUC of 0.920 compared with RF (AUC = 0.874), LDA (AUC = 0.835) and DT (AUC = 0.800) in the test set. The clinicopathologic model had low efficacy, with AUCs of 0.678 and 0.710 in the training and test sets, respectively. The combined clinicopathologic (C) factors and SVM classifier (C + SVM) model improved the predictive ability with an AUC of 0.934, sensitivity of 86.7%, specificity of 89.9% and accuracy of 91.0% in the test set.

CONCLUSION

ML-based US radiomics analysis, as a novel and promising predictive tool, is conducive to a precise clinical treatment strategy.

Application of machine learning-based multi-sequence MRI radiomics in diagnosing anterior cruciate ligament tears

OBJECTIVE

To compare the diagnostic power among various machine learning algorithms utilizing multi-sequence magnetic resonance imaging (MRI) radiomics in detecting anterior cruciate ligament (ACL) tears. Additionally, this research aimed to create and validate the optimal diagnostic model.

METHODS

In this retrospective analysis, 526 patients were included, comprising 178 individuals with ACL tears and 348 with a normal ACL. Radiomics features were derived from multi-sequence MRI scans, encompassing T1-weighted imaging and proton density (PD)-weighted imaging. The process of selecting the most reliable radiomics features involved using interclass correlation coefficient (ICC) testing, t tests, and the least absolute shrinkage and selection operator (LASSO) technique. After the feature selection process, five machine learning classifiers were created. These classifiers comprised logistic regression (LR), support vector machine (SVM), K-nearest neighbors (KNN), light gradient boosting machine (LightGBM), and multilayer perceptron (MLP). A thorough performance evaluation was carried out, utilizing diverse metrics like the area under the receiver operating characteristic curve (ROC), specificity, accuracy, sensitivity positive predictive value, and negative predictive value. The classifier exhibiting the best performance was chosen. Subsequently, three models were developed: the PD model, the T1 model, and the combined model, all based on the optimal classifier. The diagnostic performance of these models was assessed by employing AUC values, calibration curves, and decision curve analysis.

RESULTS

Out of 2032 features, 48 features were selected. The SVM-based multi-sequence radiomics outperformed all others, achieving AUC values of 0.973 and 0.927, sensitivities of 0.933 and 0.857, and specificities of 0.930 and 0.829, in the training and validation cohorts, respectively.

CONCLUSION

The multi-sequence MRI radiomics model, which is based on machine learning, exhibits exceptional performance in diagnosing ACL tears. It provides valuable insights crucial for the diagnosis and treatment of knee joint injuries, serving as an accurate and objective supplementary diagnostic tool for clinical practitioners.

Research on automatic recognition radiomics algorithm for early sacroiliac arthritis based on sacroiliac MRI imaging

OBJECTIVE

To create an automated machine learning model using sacroiliac joint MRI imaging for early sacroiliac arthritis detection, aiming to enhance diagnostic accuracy.

METHODS

We conducted a retrospective analysis involving 71 patients with early sacroiliac arthritis and 85 patients with normal sacroiliac joint MRI scans. Transverse T1WI and T2WI sequences were collected and subjected to radiomics analysis by two physicians. Patients were randomly divided into training and test groups at a 7:3 ratio. Initially, we extracted the region of interest on the sacroiliac joint surface using ITK-SNAP 3.6.0 software and extracted radiomic features. We retained features with an Intraclass Correlation Coefficient > 0.80, followed by filtering using max-relevance and min-redundancy (mRMR) and LASSO algorithms to establish an automatic identification model for sacroiliac joint surface injury. Receiver operating characteristic (ROC) curves were plotted, and the area under the ROC curve (AUC) was calculated. Model performance was assessed by accuracy, sensitivity, and specificity.

RESULTS

We evaluated model performance, achieving an AUC of 0.943 for the SVM-T1WI training group, with accuracy, sensitivity, and specificity values of 0.878, 0.836, and 0.943, respectively. The SVM-T1WI test group exhibited an AUC of 0.875, with corresponding accuracy, sensitivity, and specificity values of 0.909, 0.929, and 0.875, respectively. For the SVM-T2WI training group, the AUC was 0.975, with accuracy, sensitivity, and specificity values of 0.933, 0.889, and 0.750. The SVM-T2WI test group produced an AUC of 0.902, with accuracy, sensitivity, and specificity values of 0.864, 0.889, and 0.800. In the SVM-bimodal training group, we achieved an AUC of 0.974, with accuracy, sensitivity, and specificity values of 0.921, 0.889, and 0.971, respectively. The SVM-bimodal test group exhibited an AUC of 0.964, with accuracy, sensitivity, and specificity values of 0.955, 1.000, and 0.875, respectively.

CONCLUSION

The radiomics-based detection model demonstrates excellent automatic identification performance for early sacroiliitis.

Radiomics analysis of multiparametric MRI for preoperative prediction of microsatellite instability status in endometrial cancer: a dual-center study

Objective

To develop and validate a multiparametric MRI-based radiomics model for prediction of microsatellite instability (MSI) status in patients with endometrial cancer (EC).

Methods

A total of 225 patients from Center I including 158 in the training cohort and 67 in the internal testing cohort, and 132 patients from Center II were included as an external validation cohort. All the patients were pathologically confirmed EC who underwent pelvic MRI before treatment. The MSI status was confirmed by immunohistochemistry (IHC) staining. A total of 4245 features were extracted from T2-weighted imaging (T2WI), contrast enhanced T1-weighted imaging (CE-T1WI) and apparent diffusion coefficient (ADC) maps for each patient. Four feature selection steps were used, and then five machine learning models, including Logistic Regression (LR), k-Nearest Neighbors (KNN), Naive Bayes (NB), Support Vector Machine (SVM), and Random Forest (RF), were built for MSI status prediction in the training cohort. Receiver operating characteristics (ROC) curve and decision curve analysis (DCA) were used to evaluate the performance of these models.

Results

The SVM model showed the best performance with an AUC of 0.905 (95%CI, 0.848-0.961) in the training cohort, and was subsequently validated in the internal testing cohort and external validation cohort, with the corresponding AUCs of 0.875 (95%CI, 0.762-0.988) and 0.862 (95%CI, 0.781-0.942), respectively. The DCA curve demonstrated favorable clinical utility.

Conclusion

We developed and validated a multiparametric MRI-based radiomics model with gratifying performance in predicting MSI status, and could potentially be used to facilitate the decision-making on clinical treatment options in patients with EC.

A nomogram model combining ultrasound-based radiomics features and clinicopathological factors to identify germline BRCA1/2 mutation in invasive breast cancer patients

Objective

BRCA1/2 status is a key to personalized therapy for invasive breast cancer patients. This study aimed to explore the association between ultrasound radiomics features and germline BRCA1/2 mutation in patients with invasive breast cancer.

Materials and methods

In this retrospective study, 100 lesions in 92 BRCA1/2-mutated patients and 390 lesions in 357 non-BRCA1/2-mutated patients were included and randomly assigned as training and validation datasets in a ratio of 7:3. Gray-scale ultrasound images of the largest plane of the lesions were used for feature extraction. Maximum relevance minimum redundancy (mRMR) algorithm and multivariate logistic least absolute shrinkage and selection operator (LASSO) regression were used to select features. The multivariate logistic regression method was used to construct predictive models based on clinicopathological factors, radiomics features, or a combination of them.

Results

In the clinical model, age at first diagnosis, family history of BRCA1/2-related malignancies, HER2 status, and Ki-67 level were found to be independent predictors for BRCA1/2 mutation. In the radiomics model, 10 significant features were selected from the 1032 radiomics features extracted from US images. The AUCs of the radiomics model were not inferior to those of the clinical model in both training dataset [0.712 (95% CI, 0.647-0.776) vs 0.768 (95% CI, 0.704-0.835); p = 0.429] and validation dataset [0.705 (95% CI, 0.597-0.808) vs 0.723 (95% CI, 0.625-0.828); p = 0.820]. The AUCs of the nomogram model combining clinical and radiomics features were 0.804 (95% CI, 0.748-0.861) in the training dataset and 0.811 (95% CI, 0.724-0.894) in the validation dataset, which were proved significantly higher than those of the clinical model alone by DeLong's test (p = 0.041; p = 0.007). To be noted, the negative predictive values (NPVs) of the nomogram model reached a favorable 0.93 in both datasets.

Conclusion

This machine nomogram model combining ultrasound-based radiomics and clinical features exhibited a promising performance in identifying germline BRCA1/2 mutation in patients with invasive breast cancer and may help avoid unnecessary gene tests in clinical practice.

AI-powered interpretable imaging phenotypes noninvasively characterize tumor microenvironment associated with diverse molecular signatures and survival in breast cancer

BACKGROUND AND OBJECTIVES

Tumor microenvironment (TME) is a determining factor in decision-making and personalized treatment for breast cancer, which is highly intra-tumor heterogeneous (ITH). However, the noninvasive imaging phenotypes of TME are poorly understood, even invasive genotypes have been largely known in breast cancer.

METHODS

Here, we develop an artificial intelligence (AI)-driven approach for noninvasively characterizing TME by integrating the predictive power of deep learning with the explainability of human-interpretable imaging phenotypes (IMPs) derived from 4D dynamic imaging (DCE-MRI) of 342 breast tumors linked to genomic and clinical data, which connect cancer phenotypes to genotypes. An unsupervised dual-attention deep graph clustering model (DGCLM) is developed to divide bulk tumor into multiple spatially segregated and phenotypically consistent subclusters. The IMPs ranging from spatial heterogeneity to kinetic heterogeneity are leveraged to capture architecture, interaction, and proximity between intratumoral subclusters.

RESULTS

We demonstrate that our IMPs correlate with well-known markers of TME and also can predict distinct molecular signatures, including expression of hormone receptor, epithelial growth factor receptor and immune checkpoint proteins, with the performance of accuracy, reliability and transparency superior to recent state-of-the-art radiomics and 'black-box' deep learning methods. Moreover, prognostic value is confirmed by survival analysis accounting for IMPs.

CONCLUSIONS

Our approach provides an interpretable, quantitative, and comprehensive perspective to characterize TME in a noninvasive and clinically relevant manner.

CT-based radiomics nomogram for overall survival prediction in patients with cervical cancer treated with concurrent chemoradiotherapy

Background and purpose

To establish and validate a hybrid radiomics model to predict overall survival in cervical cancer patients receiving concurrent chemoradiotherapy (CCRT).

Methods

We retrospectively collected 367 cervical cancer patients receiving chemoradiotherapy from the First Affiliated Hospital of Soochow University in China and divided them into a training set and a test set in a ratio of 7:3. Handcrafted and deep learning (DL)-based radiomics features were extracted from the contrast-enhanced computed tomography (CT), and the two types of radiomics signatures were calculated based on the features selected using the least absolute shrinkage and selection operator (LASSO) Cox regression. A hybrid radiomics nomogram was constructed by integrating independent clinical risk factors, handcrafted radiomics signature, and DL-based radiomics signature in the training set and was validated in the test set.

Results

The hybrid radiomics nomogram exhibited favorable performance in predicting overall survival, with areas under the receiver operating characteristic curve (AUCs) for 1, 3, and 5 years in the training set of 0.833, 0.777, and 0.871, respectively, and in the test set of 0.811, 0.713, and 0.730, respectively. Furthermore, the hybrid radiomics nomogram outperformed the single clinical model, handcrafted radiomics signature, and DL-based radiomics signature in both the training (C-index: 0.793) and test sets (C-index: 0.721). The calibration curves and decision curve analysis (DCA) indicated that our hybrid nomogram had good calibration and clinical benefits. Finally, our hybrid nomogram demonstrated value in stratifying patients into high- and low-risk groups (cutoff value: 5.6).

Conclusion

A high-performance hybrid radiomics model based on pre-radiotherapy CT was established, presenting strengths in risk stratification.

A radiomics based approach using adrenal gland and periadrenal fat CT images to allocate COVID-19 health care resources fairly

BACKGROUND

The value of radiomics features from the adrenal gland and periadrenal fat CT images for predicting disease progression in patients with COVID-19 has not been studied extensively. We assess the value of radiomics features from the adrenal gland and periadrenal fat CT images in predicting COVID-19 disease exacerbation.

METHODS

A total of 1,245 patients (685 moderate and 560 severe patients) were enrolled in a retrospective study. We proposed a 3D V-net to segment adrenal glands in onset CT images automatically, and periadrenal fat was obtained using inflation operation around the adrenal gland. Next, we built a clinical model (CM), three radiomics models (adrenal gland model [AM], periadrenal fat model [PM], and fusion of adrenal gland and periadrenal fat model [FM]), and radiomics nomogram (RN) after radiomics features extracted.

RESULTS

The auto-segmentation framework yielded a dice value 0.79 in the training set. CM, AM, PM, FM, and RN obtained AUCs of 0.717, 0.716, 0.736, 0.760, and 0.833 in the validation set. FM and RN had better predictive efficacy than CM (P < 0.0001) in the training set. RN showed that there was no significant difference in the validation set (mean absolute error [MAE] = 0.04) and test set (MAE = 0.075) between predictive and actual results. Decision curve analysis showed that if the threshold probability was between 0.4 and 0.8 in the validation set or between 0.3 and 0.7 in the test set, it could gain more net benefits using RN than FM and CM.

CONCLUSIONS

Radiomics features extracted from the adrenal gland and periadrenal fat CT images are related to disease exacerbation in patients with COVID-19.

Association Between MRI Radiomics and Intratumoral Tertiary Lymphoid Structures in Intrahepatic Cholangiocarcinoma and Its Prognostic Significance

BACKGROUND

Tertiary lymphoid structures (TLSs) have prognostic value in intrahepatic cholangiocarcinoma (ICC) patients. Noninvasive tool to preoperatively evaluate TLSs is still lacking.

PURPOSE

To explore the association between TLSs status of ICC and preoperative MRI radiomics analysis.

STUDY TYPE

Retrospective.

SUBJECTS

One hundred and ninety-two patients with ICC, divided into training (T = 105), internal validation groups (V1 = 46), and external validation group (V2 = 41).

SEQUENCE

Coronal and axial single-shot fast spin-echo T2-weighted, diffusion-weighted imaging, T1-weighted, and T1WI fat-suppressed spoiled gradient-recall echo LAVA sequence at 3.0 T.

ASSESSMENT

The VOIs were drawn manually within the visible borders of the tumors using ITK-SNAP version 3.8.0 software in the axial T2WI, DWI, and portal vein phase sequences. Radiomics features were subjected to least absolute shrinkage and selection operator regression to select the associated features of TLSs and construct the radiomics model. Univariate and multivariate analyses were used to identify the clinical radiological variables associated with TLSs. The performances were evaluated by the area under the receiver operator characteristic curve (AUC).

STATISTICAL TESTS

Logistic regression analysis, ROC and AUC, Hosmer-Lemeshow test, Kaplan-Meier method with the log-rank test, calibration curves, and decision curve analysis. P < 0.05 was considered statistically significant.

RESULTS

The AUCs of arterial phase diffuse hyperenhancement were 0.59 (95% confidence interval [CI], 0.50-0.67), 0.52 (95% CI, 0.43-0.61), and 0.66 (95% CI, 0.52-0.80) in the T, V1, and V2 cohorts. The AUCs of Rad-score were 0.85 (95% CI, 0.77-0.92), 0.81 (95% CI, 0.67-0.94), and 0.84 (95% CI, 0.71-0.96) in the T, V1, and V2 cohorts, respectively. In cohort T, low-risk group showed significantly better median recurrence-free survival (RFS) than that of the high-risk group, which was also confirmed in cohort V1 and V2.

DATA CONCLUSION

A preoperative MRI radiomics signature is associated with the intratumoral TLSs status of ICC patients and correlate significantly with RFS.

LEVEL OF EVIDENCE

3 TECHNICAL EFFICACY: Stage 2.

Axial Skeleton Radiomics of 18F-FDG PET/CT: Impact on Event-Free Survival Prediction in High-Risk Pediatric Neuroblastoma

OBJECTIVES

To construct and validate a combined model based on axial skeleton radiomics of 18F-FDG PET/CT for predicting event-free survival in high-risk pediatric neuroblastoma patients.

MATERIALS AND METHODS

Eighty-seven high-risk neuroblastoma patients were retrospectively enrolled in this study and randomized in a 7:3 ratio to the training and validation cohorts. The radiomics model was constructed using radiomics features that were extracted from the axial skeleton. A univariate Cox regression analysis was then performed to screen clinical risk factors associated with event-free survival for building clinical model. Radiomics features and clinical risk factors were incorporated to construct the combined model for predicting the event-free survival in high-risk neuroblastoma patients. The performance of the models was evaluated by the C-index.

RESULTS

Eighteen radiomics features were selected to build the radiomics model. The radiomics model achieved better event-free survival prediction than the clinical model in the training cohort (C-index: 0.846 vs. 0.612) and validation cohort (C-index: 0.754 vs. 0.579). The combined model achieved the best prognostic prediction performance with a C-index of 0.863 and 0.799 in the training and validation cohorts, respectively.

CONCLUSION

The combined model integrating radiomics features and clinical risk factors showed more accurate predictive performance for event-free survival in high-risk pediatric neuroblastoma patients, which helps to design individualized treatment strategies and regular follow-ups.

The value of radiomics model based on ultrasound image features in the differentiation between minimal breast cancer and small benign breast masses

BACKGROUND

Female breast cancer has surpassed lung cancer as the most common cancer, and is also the main cause of cancer death for women worldwide. Breast cancer <1 cm showed excellent survival rate. However, the diagnosis of minimal breast cancer (MBC) is challenging.

OBJECTIVE

The purpose of our research is to develop and validate an radiomics model based on ultrasound images for early recognition of MBC.

METHODS

302 breast masses with a diameter of <10 mm were retrospectively studied, including 159 benign and 143 malignant breast masses. The radiomics features were extracted from the gray-scale ultrasound image of the largest face of each breast mass. The maximum relevance minimum reduncancy and recursive feature elimination methods were used to screen. Finally, 10 features with the most discriminating value were selected for modeling. The random forest was used to establish the prediction model, and the rad-score of each mass was calculated. In order to evaluate the effectiveness of the model, we calculated and compared the area under the curve (AUC) value, sensitivity, specificity, accuracy, positive predictive value, and negative predictive value of the model and three groups with different experience in predicting small breast masses, and drew calibration curves and decision curves to test the stability and consistency of the model.

RESULTS

When we selected 10 radiomics features to calculate the rad-score, the prediction efficiency was the best, the AUC values for the training set and testing set were 0.840 and 0.793, which was significantly better than the insufficient experience group (AUC = 0.673), slightly better than the moderate experience group (AUC = 0.768), and was inferior to the experienced group (AUC = 0.877). The calibration curve and decision curve also showed that the radiomics model had satisfied stability and clinical application value.

CONCLUSION

The radiomics model based on ultrasound image features has a satisfied predictive ability for small breast masses, and is expected to become a potential tool for the diagnosis of MBC, and it is a zero cost (in terms of patient participation and imaging time).

Radiomics-guided prognostic assessment of early-stage hepatocellular carcinoma recurrence post-radical resection

PURPOSE

The prognosis of early-stage hepatocellular carcinoma (HCC) patients after radical resection has received widespread attention, but reliable prediction methods are lacking. Radiomics derived from enhanced computed tomography (CT) imaging offers a potential avenue for practical prognostication in HCC patients.

METHODS

We recruited early-stage HCC patients undergoing radical resection. Statistical analyses were performed to identify clinicopathological and radiomic features linked to recurrence. Clinical, radiomic, and combined models (incorporating clinicopathological and radiomic features) were built using four algorithms. The performance of these models was scrutinized via fivefold cross-validation, with evaluation metrics including the area under the curve (AUC), accuracy (ACC), sensitivity (SEN), and specificity (SPE) being calculated and compared. Ultimately, an integrated nomogram was devised by combining independent clinicopathological predictors with the Radscore.

RESULTS

From January 2016 through December 2020, HCC recurrence was observed in 167 cases (64.5%), with a median time to recurrence of 26.7 months following initial resection. Combined models outperformed those solely relying on clinicopathological or radiomic features. Notably, among the combined models, those employing support vector machine (SVM) algorithms exhibited the most promising predictive outcomes (AUC: 0.840 (95% Confidence interval (CI): [0.696, 0.984]), ACC: 0.805, SEN: 0.849, SPE: 0.733). Hepatitis B infection, tumour size > 5 cm, and alpha-fetoprotein (AFP) > 400 ng/mL were identified as independent recurrence predictors and were subsequently amalgamated with the Radscore to create a visually intuitive nomogram, delivering robust and reliable predictive performance.

CONCLUSION

Machine learning models amalgamating clinicopathological and radiomic features provide a valuable tool for clinicians to predict postoperative HCC recurrence, thereby informing early preventative strategies.

Radioproteomics modeling of metformin-enhanced radiosensitivity: an animal study

PURPOSE

Metformin is considered as radiation modulator in both tumors and healthy tissues. Radiomics has the potential to decode biological mechanisms of radiotherapy response. The aim of this study was to apply radiomics analysis in metformin-induced radiosensitivity and finding radioproteomics associations of computed tomography (CT) imaging features and proteins involved in metformin radiosensitivity signaling pathways.

MATERIALS AND METHODS

A total of 32 female BALB/c mice were used in this study and were subjected to injection of breast cancer cells. When tumors reached a mean volume of 150 mm3, mice were randomly divided into the four groups including Control, Metformin, Radiation, and Radiation + Metformin. Western blot analysis was performed after treatment to measure expression of proteins including AMPK-alpha, phospho-AMPK-alpha (Thr172), mTOR, phospho-mTOR (Ser2448), phospho-4EBP1 (Thr37/46), phospho-ACC (Ser79), and β-actin. CT imaging was performed before treatment and at the end of treatment in all groups. Radiomics features extracted from segmented tumors were selected using Elastic-net regression and were assessed in terms of correlation with expression of the proteins.

RESULTS

It was observed that proteins including phospho-mTOR, phospho-4EBP1, and mTOR had positive correlations with changes in tumor volumes in days 28, 24, 20, 16, and 12, while tumor volume changes at these days had negative correlations with AMPK-alpha, phospho-AMPK-alpha, and phospho-ACC proteins. Furthermore, median feature had a positive correlation with AMPK-alpha, phospho-ACC, and phospho-AMPK-alpha proteins. Also, Cluster shade feature had positive correlations with mTOR and p-mTOR. On the other hand, LGLZE feature had negative correlations with AMPK-alpha and phospho-AMPK-alpha.

CONCLUSION

Radiomics features can decode proteins that involved in response to metformin and radiation, although further studies are warranted to investigate the optimal way to integrate radiomics into biological experiments.

Assessing Machine Learning Models for Predicting Age with Intracranial Vessel Tortuosity and Thickness Information

This study aimed to develop and validate machine learning (ML) models that predict age using intracranial vessels' tortuosity and diameter features derived from magnetic resonance angiography (MRA) data. A total of 171 subjects' three-dimensional (3D) time-of-flight MRA image data were considered for analysis. After annotations of two endpoints in each arterial segment, tortuosity features such as the sum of the angle metrics, triangular index, relative length, and product of the angle distance, as well as the vessels' diameter features, were extracted and used to train and validate the ML models for age prediction. Features extracted from the right and left internal carotid arteries (ICA) and basilar arteries were considered as the inputs to train and validate six ML regression models with a four-fold cross validation. The random forest regression model resulted in the lowest root mean square error of 14.9 years and the highest average coefficient of determination of 0.186. The linear regression model showed the lowest average mean absolute percentage error (MAPE) and the highest average Pearson correlation coefficient (0.532). The mean diameter of the right ICA vessel segment was the most important feature contributing to prediction of age in two out of the four regression models considered. An ML of tortuosity descriptors and diameter features extracted from MRA data showed a modest correlation between real age and ML-predicted age. Further studies are warranted for the assessment of the model's age predictions in patients with intracranial vessel diseases.

Assessing robustness of quantitative susceptibility-based MRI radiomic features in patients with multiple sclerosis

Multiple Sclerosis (MS) is an autoimmune demyelinating disease characterised by changes in iron and myelin content. These biomarkers are detectable by Quantitative Susceptibility Mapping (QSM), an advanced Magnetic Resonance Imaging technique detecting magnetic properties. When analysed with radiomic techniques that exploit its intrinsic quantitative nature, QSM may furnish biomarkers to facilitate early diagnosis of MS and timely assessment of progression. In this work, we explore the robustness of QSM radiomic features by varying the number of grey levels (GLs) and echo times (TEs), in a sample of healthy controls and patients with MS. We analysed the white matter in total and within six clinically relevant tracts, including the cortico-spinal tract and the optic radiation. After optimising the number of GLs (n = 64), at least 65% of features were robust for each Volume of Interest (VOI), with no difference (p > .05) between left and right hemispheres. Different outcomes in feature robustness among the VOIs depend on their characteristics, such as volume and variance of susceptibility values. This study validated the processing pipeline for robustness analysis and established the reliability of QSM-based radiomics features against GLs and TEs. Our results provide important insights for future radiomics studies using QSM in clinical applications.

Performance Analysis of Six Semi-Automated Tumour Delineation Methods on [18F] Fluorodeoxyglucose Positron Emission Tomography/Computed Tomography (FDG PET/CT) in Patients with Head and Neck Cancer

Background. Head and neck cancer (HNC) is the seventh most common neoplastic disorder at the global level. Contouring HNC lesions on [18F] Fluorodeoxyglucose positron emission tomography/computed tomography (FDG PET/CT) scans plays a fundamental role for diagnosis, risk assessment, radiotherapy planning and post-treatment evaluation. However, manual contouring is a lengthy and tedious procedure which requires significant effort from the clinician. Methods. We evaluated the performance of six hand-crafted, training-free methods (four threshold-based, two algorithm-based) for the semi-automated delineation of HNC lesions on FDG PET/CT. This study was carried out on a single-centre population of n=103 subjects, and the standard of reference was manual segmentation generated by nuclear medicine specialists. Figures of merit were the Sørensen-Dice coefficient (DSC) and relative volume difference (RVD). Results. Median DSC ranged between 0.595 and 0.792, median RVD between -22.0% and 87.4%. Click and draw and Nestle's methods achieved the best segmentation accuracy (median DSC, respectively, 0.792 ± 0.178 and 0.762 ± 0.107; median RVD, respectively, -21.6% ± 1270.8% and -32.7% ± 40.0%) and outperformed the other methods by a significant margin. Nestle's method also resulted in a lower dispersion of the data, hence showing stronger inter-patient stability. The accuracy of the two best methods was in agreement with the most recent state-of-the art results. Conclusions. Semi-automated PET delineation methods show potential to assist clinicians in the segmentation of HNC lesions on FDG PET/CT images, although manual refinement may sometimes be needed to obtain clinically acceptable ROIs.

Cine MRI-Derived Radiomics Features of the Cardiac Blood Pool: Periodicity, Specificity, and Reproducibility

BACKGROUND

Although radiomics features of the left ventricular wall have been used to assess cardiac diseases, radiomics features of the cardiac blood pool have been relatively ignored.

PURPOSE

To test the hypothesis that cine MRI-derived radiomics features of the cardiac blood pool are associated with cardiac function and motion.

STUDY TYPE

Retrospective.

POPULATION

A total of 26 healthy volunteers (51.2 ± 15.6 years; 17 males).

FIELD STRENGTH/SEQUENCE

A 1.5 T/balanced steady-state free precession (bSSFP).

ASSESSMENT

The radiomics features (107 features in seven classes) of the blood pool of the left/right ventricle/atrium (LV/RV/LA/RA) were extracted on four-chamber cine images (25 phases). Conventional cardiac function parameters (volumes, ejection fraction [EF] and longitudinal strain) were assessed in each cardiac chamber. Intraobserver- and interobserver agreements of radiomics features of all chambers acquired at all phases were assessed, as well as scan-rescan agreement in a subset of 13 volunteers.

STATISTICAL TESTS

Pearson correlation coefficients (r) were used to assess the associations between peak values of radiomics features and end-diastolic (or maximal) volume, end-systolic (or minimal) volume, EF, and longitudinal strain of corresponding chambers. Good intraobserver, interobserver, and scan-rescan agreements for radiomics features acquired were defined as intraclass correlation coefficient (ICC) > 0.7 or coefficient of variation (CoV) < 20%.

RESULTS

Most radiomics features of the blood pool varied periodically throughout the cardiac cycle. Peak values of chamber-specific blood pool radiomics features were correlated with traditional cardiac function and motion indices of corresponding chambers (r: 0.4-0.87). Ninety-three (87%), 86 (80%), and 73 (68%) radiomics features demonstrated good intraobserver, interobserver, and scan-rescan reproducibility, respectively.

CONCLUSION

Cine MRI-derived radiomics features within LV/RV/LA/RA are associated with traditional cardiac function and motion indices of corresponding chambers and may have the potential to become novel quantitative imaging biomarkers in cardiovascular medicine.

EVIDENCE LEVEL

3.

TECHNICAL EFFICACY

1.

Radiomics optimizing the evaluation of endometrial receptivity for women with unexplained recurrent pregnancy loss

Background

The optimization of endometrial receptivity (ER) through individualized therapies has been shown to enhance the likelihood of successful gestation. However, current practice lacks comprehensive methods for evaluating the ER of patients with recurrent pregnancy loss (RPL). Radiomics, an emerging AI-based technique that enables the extraction of mineable information from medical images, holds potential to offer a more objective and quantitative approach to ER assessment. This innovative tool may facilitate a deeper understanding of the endometrial environment and enable clinicians to optimize ER evaluation in RPL patients.

Objective

This study aimed to identify ultrasound radiomics features associated with ER, with the purpose of predicting successful ongoing pregnancies in RPL patients, and to assess the predictive accuracy of these features against regular ER parameters.

Methods

This retrospective, controlled study involved 262 patients with unexplained RPL and 273 controls with a history of uncomplicated full-term pregnancies. Radiomics features were extracted from ultrasound endometrial segmentation images to derive a radiomics score (rad-score) for each participant. Associations between rad-scores, baseline clinical variables, and sonographic data were evaluated using univariate and multivariate logistic regression analyses to identify potential indicators of RPL. Receiver operating characteristic (ROC) curve analysis was performed to evaluate the predictive accuracy of the rad-score and other identified indicators in discriminating RPL cases. Furthermore, the relationships between age and these identified indicators were assessed via Pearson correlation analysis.

Results

From the 1312 extracted radiomics features, five non-zero coefficient radiomics signatures were identified as significantly associated with RPL, forming the basis of the rad-score. Following multivariate logistic regression analysis, age, spiral artery pulsatility index (SA-PI), vascularisation index (VI), and rad-score emerged as independent correlates of RPL (all P<0.05). ROC curve analyses revealed the superior discriminative capability of the rad-score (AUC=0.882) over age (AUC=0.778), SA-PI (AUC=0.771), and VI (AUC=0.595). There were notable correlations between age and rad-score (r=0.275), VI (r=-0.224), and SA-PI (r=0.211), indicating age-related variations in RPL predictors.

Conclusion

This study revealed a significant association between unexplained RPL and elevated endometrial rad-scores during the WOI. Furthermore, it demonstrated the potential of rad-scores as a promising predictive tool for successful ongoing pregnancies in RPL patients.

Predictive role of radiomics features extracted from preoperative cross-sectional imaging of pancreatic ductal adenocarcinoma in detecting lymph node metastasis: a systemic review and meta-analysis

Lymph node metastases are associated with poor clinical outcomes in pancreatic ductal adenocarcinoma (PDAC). In preoperative imaging, conventional diagnostic modalities do not provide the desired accuracy in diagnosing lymph node metastasis. The current review aims to determine the pooled diagnostic profile of studies examining the role of radiomics features in detecting lymph node metastasis in PDAC. PubMed, Google Scholar, and Embase databases were searched for relevant articles. The quality of the studies was examined using the Radiomics Quality Score and Quality Assessment of Diagnostic Accuracy Studies 2 (QUADAS-2) tools. Pooled results for sensitivity, specificity, likelihood, and odds ratios with the corresponding 95% confidence intervals (CIs) were calculated using a random-effect model (DerSimonian-Liard method). No significant publication bias was detected among the studies included in this meta-analysis. The pooled sensitivity of the validation datasets included in the study was 77.4% (72.7%, 81.5%) and pooled specificity was 72.4% (63.8, 79.6%). The diagnostic odds ratio of the validation datasets was 9.6 (6.0, 15.2). No statistically significant heterogeneity was detected for sensitivity and odds ratio (P values of 0.3 and 0.08, respectively). However, there was significant heterogeneity concerning specificity (P = 0.003). The pretest probability of having lymph node metastasis in the pooled databases was 52% and a positive post-test probability was 76% after the radiomics features were used, showing a net benefit of 24%. Classifiers trained on radiomics features extracted from preoperative images can improve the sensitivity and specificity of conventional cross-sectional imaging in detecting lymph node metastasis in PDAC.

Breast Cancer Surrogate Subtype Classification Using Pretreatment Multi-Phase Dynamic Contrast-Enhanced Magnetic Resonance Imaging Radiomics: A Retrospective Single-Center Study

This study aimed to explore the potential of multi-phase dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) radiomics for classifying breast cancer surrogate subtypes. This retrospective study analyzed 360 breast cancers from 319 patients who underwent pretreatment DCE-MRI between January 2015 and January 2019. The cohort consisted of 33 triple-negative, 26 human epidermal growth factor receptor 2 (HER2)-positive, 109 luminal A-like, 144 luminal B-like HER2-negative, and 48 luminal B-like HER2-positive lesions. A total of 1781 radiomic features were extracted from manually segmented breast cancers in each DCE-MRI sequence. The model was internally validated and selected using ten times repeated five-fold cross-validation on the primary cohort, with further evaluation using a validation cohort. The most successful models were logistic regression models applied to the third post-contrast subtraction images. These models exhibited the highest area under the curve (AUC) for discriminating between luminal A like vs. others (AUC: 0.78), luminal B-like HER2 negative vs. others (AUC: 0.57), luminal B-like HER2 positive vs. others (AUC: 0.60), HER2 positive vs. others (AUC: 0.81), and triple negative vs. others (AUC: 0.83). In conclusion, the radiomic features extracted from multi-phase DCE-MRI are promising for discriminating between breast cancer subtypes. The best-performing models relied on tissue changes observed during the mid-stage of the imaging process.

Multi-Omic Biomarkers Improve Indeterminate Pulmonary Nodule Malignancy Risk Assessment

The clinical management of patients with indeterminate pulmonary nodules is associated with unintended harm to patients and better methods are required to more precisely quantify lung cancer risk in this group. Here, we combine multiple noninvasive approaches to more accurately identify lung cancer in indeterminate pulmonary nodules. We analyzed 94 quantitative radiomic imaging features and 41 qualitative semantic imaging variables with molecular biomarkers from blood derived from an antibody-based microarray platform that determines protein, cancer-specific glycan, and autoantibody-antigen complex content with high sensitivity. From these datasets, we created a PSR (plasma, semantic, radiomic) risk prediction model comprising nine blood-based and imaging biomarkers with an area under the receiver operating curve (AUROC) of 0.964 that when tested in a second, independent cohort yielded an AUROC of 0.846. Incorporating known clinical risk factors (age, gender, and smoking pack years) for lung cancer into the PSR model improved the AUROC to 0.897 in the second cohort and was more accurate than a well-characterized clinical risk prediction model (AUROC = 0.802). Our findings support the use of a multi-omics approach to guide the clinical management of indeterminate pulmonary nodules.

Radiogenomics Analysis Linking Multiparametric MRI and Transcriptomics in Prostate Cancer

Prostate cancer (PCa) is a highly prevalent cancer type with a heterogeneous prognosis. An accurate assessment of tumor aggressiveness can pave the way for tailored treatment strategies, potentially leading to better outcomes. While tumor aggressiveness is typically assessed based on invasive methods (e.g., biopsy), radiogenomics, combining diagnostic imaging with genomic information can help uncover aggressive (imaging) phenotypes, which in turn can provide non-invasive advice on individualized treatment regimens. In this study, we carried out a parallel analysis on both imaging and transcriptomics data in order to identify features associated with clinically significant PCa (defined as an ISUP grade ≥ 3), subsequently evaluating the correlation between them. Textural imaging features were extracted from multi-parametric MRI sequences (T2W, DWI, and DCE) and combined with DCE-derived parametric pharmacokinetic maps obtained using magnetic resonance dispersion imaging (MRDI). A transcriptomic analysis was performed to derive functional features on transcription factors (TFs), and pathway activity from RNA sequencing data, here referred to as transcriptomic features. For both the imaging and transcriptomic features, different machine learning models were separately trained and optimized to classify tumors in either clinically insignificant or significant PCa. These models were validated in an independent cohort and model performance was used to isolate a subset of relevant imaging and transcriptomic features to be further investigated. A final set of 31 imaging features was correlated to 33 transcriptomic features obtained on the same tumors. Five significant correlations (p < 0.05) were found, of which, three had moderate strength (|r| ≥ 0.5). The strongest significant correlations were seen between a perfusion-based imaging feature-MRDI A median-and the activities of the TFs STAT6 (-0.64) and TFAP2A (-0.50). A higher-order T2W textural feature was also significantly correlated to the activity of the TF STAT6 (-0.58). STAT6 plays an important role in controlling cell proliferation and migration. Loss of the AP2alpha protein expression, quantified by TFAP2A, has been strongly associated with aggressiveness and progression in PCa. According to our findings, a combination of texture features extracted from T2W and DCE, as well as perfusion-based pharmacokinetic features, can be considered for the prediction of clinically significant PCa, with the pharmacokinetic MRDI A feature being the most correlated with the underlying transcriptomic information. These results highlight a link between quantitative imaging features and the underlying transcriptomic landscape of prostate tumors.

The Additive Value of Radiomics Features Extracted from Baseline MR Images to the Barcelona Clinic Liver Cancer (BCLC) Staging System in Predicting Transplant-Free Survival in Patients with Hepatocellular Carcinoma: A Single-Center Retrospective Analysis

BACKGROUND

To study the additive value of radiomics features to the BCLC staging system in clustering HCC patients.

METHODS

A total of 266 patients with HCC were included in this retrospective study. All patients had undergone baseline MR imaging, and 95 radiomics features were extracted from 3D segmentations representative of lesions on the venous phase and apparent diffusion coefficient maps. A random forest algorithm was utilized to extract the most relevant features to transplant-free survival. The selected features were used alongside BCLC staging to construct Kaplan-Meier curves.

RESULTS

Out of 95 extracted features, the three most relevant features were incorporated into random forest classifiers. The Integrated Brier score of the prediction error curve was 0.135, 0.072, and 0.048 for the BCLC, radiomics, and combined models, respectively. The mean area under the receiver operating curve (ROC curve) over time for the three models was 81.1%, 77.3%, and 56.2% for the combined radiomics and BCLC models, respectively.

CONCLUSIONS

Radiomics features outperformed the BCLC staging system in determining prognosis in HCC patients. The addition of a radiomics classifier increased the classification capability of the BCLC model. Texture analysis features could be considered as possible biomarkers in predicting transplant-free survival in HCC patients.

Multisequence magnetic resonance imaging-based radiomics models for the prediction of microsatellite instability in endometrial cancer

PURPOSE

To evaluate the performance of multisequence magnetic resonance imaging (MRI)-based radiomics models in the assessment of microsatellite instability (MSI) status in endometrial cancer (EC).

MATERIALS AND METHODS

This retrospective multicentre study included 338 EC patients with available MSI status and preoperative MRI scans, divided into training (37 MSI, 123 microsatellite stability [MSS]), internal validation (15 MSI, 52 MSS), and external validation cohorts (30 MSI, 81 MSS). Radiomics features were extracted from T2-weighted images, diffusion-weighted images, and contrast-enhanced T1-weighted images. The ComBat harmonisation method was applied to remove intrascanner variability. The Boruta wrapper algorithm was used for key feature selection. Three classification algorithms, logistic regression (LR), random forest (RF), and support vector machine (SVM), were applied to build the radiomics models. The area under the receiver operating characteristic curve (AUC) was calculated to compare the diagnostic performance of the models. Decision curve analysis (DCA) was conducted to determine the clinical usefulness of the models.

RESULTS

Among the 1980 features, Boruta finally selected nine radiomics features. A higher MSI prediction performance was achieved after running the ComBat harmonisation method. The SVM algorithm had the best performance, with AUCs of 0.921, 0.903, and 0.937 in the training, internal validation, and external validation cohorts, respectively. The DCA results showed that the SVM algorithm achieved higher net benefits than the other classifiers over a threshold range of 0.581-0.783.

CONCLUSION

The multisequence MRI-based radiomics models showed promise in preoperatively predicting the MSI status in EC in this multicentre setting.

Asymmetry of radiomics features in the white matter of patients with primary progressive aphasia

Introduction

Primary Progressive Aphasia (PPA) is a neurological disease characterized by linguistic deficits. Semantic (svPPA) and non-fluent/agrammatic (nfvPPA) variants are the two main clinical subtypes. We applied a novel analytical framework, based on radiomic analysis, to investigate White Matter (WM) asymmetry and to examine whether asymmetry is associated with verbal fluency performance.

Methods

Analyses were performed on T1-weighted images including 56 patients with PPA (31 svPPA and 25 nfvPPA) and 53 age- and sex-matched controls. Asymmetry Index (AI) was computed for 86 radiomics features in 34 white matter regions. The relationships between AI, verbal fluency performance (semantic and phonemic) and Boston Naming Test score (BNT) were explored through Spearman correlation analysis.

Results

Relative to controls, WM asymmetry in svPPA patients involved regions adjacent to middle temporal cortex as part of the inferior longitudinal (ILF), fronto-occipital (IFOF) and superior longitudinal fasciculi. Conversely, nfvPPA patients showed an asymmetry of WM in lateral occipital regions (ILF/IFOF). A higher lateralization involving IFOF, cingulum and forceps minor was found in nfvPPA compared to svPPA patients. In nfvPPA patients, semantic fluency was positively correlated to asymmetry in ILF/IFOF tracts. Performances at BNT were associated with AI values of the middle temporal (ILF/SLF) and parahippocampal (ILF/IFOF) gyri in svPPA patients.

Discussion

Radiomics features depicted distinct pathways of asymmetry in svPPA and nfvPPA involving damage of principal fiber tracts associated with speech and language. Assessing asymmetry of radiomics in PPA allows achieving a deeper insight into the neuroanatomical damage and may represent a candidate severity marker for language impairments in PPA patients.

Radiomics in clinical trials: perspectives on standardization

The term biomarker is used to describe a biological measure of the disease behavior. The existing imaging biomarkers are associated with the known tissue biological characteristics and follow a well-established roadmap to be implemented in routine clinical practice. Recently, a new quantitative imaging analysis approach named radiomics has emerged. It refers to the extraction of a large number of advanced imaging features with high-throughput computing. Extensive research has demonstrated its value in predicting disease behavior, progression, and response to therapeutic options. However, there are numerous challenges to establishing it as a clinically viable solution, including lack of reproducibility and transparency. The data-driven nature also does not offer insights into the underpinning biology of the observed relationships. As such, additional effort is needed to establish it as a qualified biomarker to inform clinical decisions. Here we review the technical difficulties encountered in the clinical applications of radiomics and current effort in addressing some of these challenges in clinical trial designs. By addressing these challenges, the true potential of radiomics can be unleashed.

A combined predictive model based on radiomics features and clinical factors for disease progression in early-stage non-small cell lung cancer treated with stereotactic ablative radiotherapy

Purpose

To accurately assess disease progression after Stereotactic Ablative Radiotherapy (SABR) of early-stage Non-Small Cell Lung Cancer (NSCLC), a combined predictive model based on pre-treatment CT radiomics features and clinical factors was established.

Methods

This study retrospectively analyzed the data of 96 patients with early-stage NSCLC treated with SABR. Clinical factors included general information (e.g. gender, age, KPS, Charlson score, lung function, smoking status), pre-treatment lesion status (e.g. diameter, location, pathological type, T stage), radiation parameters (biological effective dose, BED), the type of peritumoral radiation-induced lung injury (RILI). Independent risk factors were screened by logistic regression analysis. Radiomics features were extracted from pre-treatment CT. The minimum Redundancy Maximum Relevance (mRMR) and the Least Absolute Shrinkage and Selection Operator (LASSO) were adopted for the dimensionality reduction and feature selection. According to the weight coefficient of the features, the Radscore was calculated, and the radiomics model was constructed. Multiple logistic regression analysis was applied to establish the combined model based on radiomics features and clinical factors. Receiver Operating Characteristic (ROC) curve, DeLong test, Hosmer-Lemeshow test, and Decision Curve Analysis (DCA) were used to evaluate the model’s diagnostic efficiency and clinical practicability.

Results

With the median follow-up of 59.1 months, 29 patients developed progression and 67 remained good controlled within two years. Among the clinical factors, the type of peritumoral RILI was the only independent risk factor for progression (P< 0.05). Eleven features were selected from 1781 features to construct a radiomics model. For predicting disease progression after SABR, the Area Under the Curve (AUC) of training and validation cohorts in the radiomics model was 0.88 (95%CI 0.80-0.96) and 0.80 (95%CI 0.62-0.98), and AUC of training and validation cohorts in the combined model were 0.88 (95%CI 0.81-0.96) and 0.81 (95%CI 0.62-0.99). Both the radiomics and the combined models have good prediction efficiency in the training and validation cohorts. Still, DeLong test shows that there is no difference between them.

Conclusions

Compared with the clinical model, the radiomics model and the combined model can better predict the disease progression of early-stage NSCLC after SABR, which might contribute to individualized follow-up plans and treatment strategies.