Effects of Interobserver Variability on 2D and 3D CT- and MRI-Based Texture Feature Reproducibility of Cartilaginous Bone Tumors

Effects of Interobserver Segmentation Variability and Intensity Discretization on MRI-Based Radiomic Feature Reproducibility of Lipoma and Atypical Lipomatous Tumor

Segmentation and image intensity discretization impact on radiomics workflow. The aim of this study is to investigate the influence of interobserver segmentation variability and intensity discretization methods on the reproducibility of MRI-based radiomic features in lipoma and atypical lipomatous tumor (ALT). Thirty patients with lipoma or ALT were retrospectively included. Three readers independently performed manual contour-focused segmentation on T1-weighted and T2-weighted sequences, including the whole tumor volume. Additionally, a marginal erosion was applied to segmentations to evaluate its influence on feature reproducibility. After image pre-processing, with included intensity discretization employing both fixed bin number and width approaches, 1106 radiomic features were extracted from each sequence. Intraclass correlation coefficient (ICC) 95% confidence interval lower bound ≥ 0.75 defined feature stability. In contour-focused vs. margin shrinkage segmentation, the rates of stable features extracted from T1-weighted and T2-weighted images ranged from 92.68 to 95.21% vs. 90.69 to 95.66% after fixed bin number discretization and from 95.75 to 97.65% vs. 95.39 to 96.47% after fixed bin width discretization, respectively, with no difference between the two segmentation approaches (p ≥ 0.175). Higher stable feature rates and higher feature ICC values were found when implementing discretization with fixed bin width compared to fixed bin number, regardless of the segmentation approach (p < 0.001). In conclusion, MRI radiomic features of lipoma and ALT are reproducible regardless of the segmentation approach and intensity discretization method, although a certain degree of interobserver variability highlights the need for a preliminary reliability analysis in future studies.

Influence of image preprocessing on the segmentation-based reproducibility of radiomic features: in vivo experiments on discretization and resampling parameters

PURPOSE

To systematically investigate the impact of image preprocessing parameters on the segmentation-based reproducibility of magnetic resonance imaging (MRI) radiomic features.

METHODS

The MRI scans of 50 patients were included from the multi-institutional Brain Tumor Segmentation 2021 public glioma dataset. Whole tumor volumes were manually segmented by two independent readers, with the participation of eight readers. Radiomic features were extracted from two sequences: T2-weighted (T2) and contrast-enhanced T1-weighted (T1ce). Two methods were considered for discretization: bin count (i.e., relative discretization) and bin width (i.e., absolute discretization). Ten discretization (five for each method) and five resampling parameters were varied while other parameters were fixed. The intraclass correlation coefficient (ICC) was used for reliability analysis based on two commonly used cut-off values (0.75 and 0.90).

RESULTS

Image preprocessing parameters had a significant impact on the segmentation-based reproducibility of radiomic features. The bin width method yielded more reproducible features than the bin count method. In discretization experiments using the bin width on both sequences, according to the ICC cut-off values of 0.75 and 0.90, the rate of reproducible features ranged from 70% to 84% and from 35% to 57%, respectively, with an increasing percentage trend as parameter values decreased (from 84 to 5 for T2; 100 to 6 for T1ce). In the resampling experiments, these ranged from 53% to 74% and from 10% to 20%, respectively, with an increasing percentage trend from lower to higher parameter values (physical voxel size; from 1 x 1 x 1 to 2 x 2 x 2 mm3).

CONCLUSION

The segmentation-based reproducibility of radiomic features appears to be substantially influenced by discretization and resampling parameters. Our findings indicate that the bin width method should be used for discretization and lower bin width and higher resampling values should be used to allow more reproducible features.

Bone Reporting and Data System (Bone-RADS) and Other Proposed Practice Guidelines for Reporting Bone Tumors

BACKGROUND

 The purpose of this article is to review the different bone tumor radiology reporting systems [Bone Reporting and Data System (Bone-RADS), Osseous Tumor Reporting and Data System (OT-RADS), Solitary Bone Tumor Imaging Reporting and Data System (BTI-RADS), and Radiological Evaluation Score for Bone Tumors (REST)] and summarize their advantages and disadvantages.

METHODS

 A selective search of PubMed was performed for literature regarding the definition and discussion of bone tumor reporting systems. No time frame was selected, but the search was particularly focused on current literature on musculoskeletal radiology lexicon.

RESULTS

 To date, four major reporting systems has been proposed to standardize and systematize the reporting of imaging studies of bone tumors: Bone-RADS, OT-RADS, BTI-RADS, and REST. Both Bone-RADS and OT-RADS aid in the characterization and management of bone lesions on CT and MRI. OT-RADS and REST can be applied to MRI and radiography, respectively.

CONCLUSION

 Radiologists play a central role in the detection and characterization of asymptomatic (or incidentally detected) and symptomatic bone tumors. There are several existing bone tumor reporting systems with various advantages and disadvantages including emphasis on lesion characterization as well as management of incidentally detected bone lesions.

KEY POINTS

  · Four bone tumor reporting systems have been proposed thus far.. · Bone-RADS guides management of incidental bone lesions on CT and MRI.. · OT-RADS guides management of bone lesions on MRI with high accuracy.. · BTI-RADS classifies bone tumors on CT and MRI..

X-rays radiomics-based machine learning classification of atypical cartilaginous tumour and high-grade chondrosarcoma of long bones

BACKGROUND

Atypical cartilaginous tumour (ACT) and high-grade chondrosarcoma (CS) of long bones are respectively managed with active surveillance or curettage and wide resection. Our aim was to determine diagnostic performance of X-rays radiomics-based machine learning for classification of ACT and high-grade CS of long bones.

METHODS

This retrospective, IRB-approved study included 150 patients with surgically treated and histology-proven lesions at two tertiary bone sarcoma centres. At centre 1, the dataset was split into training (n = 71 ACT, n = 24 high-grade CS) and internal test (n = 19 ACT, n = 6 high-grade CS) cohorts, respectively, based on the date of surgery. At centre 2, the dataset constituted the external test cohort (n = 12 ACT, n = 18 high-grade CS). Manual segmentation was performed on frontal view X-rays, using MRI or CT for preliminary identification of lesion margins. After image pre-processing, radiomic features were extracted. Dimensionality reduction included stability, coefficient of variation, and mutual information analyses. In the training cohort, after class balancing, a machine learning classifier (Support Vector Machine) was automatically tuned using nested 10-fold cross-validation. Then, it was tested on both the test cohorts and compared to two musculoskeletal radiologists' performance using McNemar's test.

FINDINGS

Five radiomic features (3 morphology, 2 texture) passed dimensionality reduction. After tuning on the training cohort (AUC = 0.75), the classifier had 80%, 83%, 79% and 80%, 89%, 67% accuracy, sensitivity, and specificity in the internal (temporally independent) and external (geographically independent) test cohorts, respectively, with no difference compared to the radiologists (p ≥ 0.617).

INTERPRETATION

X-rays radiomics-based machine learning accurately differentiates between ACT and high-grade CS of long bones.

FUNDING

AIRC Investigator Grant.

CT and MRI radiomics of bone and soft-tissue sarcomas: an updated systematic review of reproducibility and validation strategies

OBJECTIVE

To systematically review radiomic feature reproducibility and model validation strategies in recent studies dealing with CT and MRI radiomics of bone and soft-tissue sarcomas, thus updating a previous version of this review which included studies published up to 2020.

METHODS

A literature search was conducted on EMBASE and PubMed databases for papers published between January 2021 and March 2023. Data regarding radiomic feature reproducibility and model validation strategies were extracted and analyzed.

RESULTS

Out of 201 identified papers, 55 were included. They dealt with radiomics of bone (n = 23) or soft-tissue (n = 32) tumors. Thirty-two (out of 54 employing manual or semiautomatic segmentation, 59%) studies included a feature reproducibility analysis. Reproducibility was assessed based on intra/interobserver segmentation variability in 30 (55%) and geometrical transformations of the region of interest in 2 (4%) studies. At least one machine learning validation technique was used for model development in 34 (62%) papers, and K-fold cross-validation was employed most frequently. A clinical validation of the model was reported in 38 (69%) papers. It was performed using a separate dataset from the primary institution (internal test) in 22 (40%), an independent dataset from another institution (external test) in 14 (25%) and both in 2 (4%) studies.

CONCLUSIONS

Compared to papers published up to 2020, a clear improvement was noted with almost double publications reporting methodological aspects related to reproducibility and validation. Larger multicenter investigations including external clinical validation and the publication of databases in open-access repositories could further improve methodology and bring radiomics from a research area to the clinical stage.

CRITICAL RELEVANCE STATEMENT

An improvement in feature reproducibility and model validation strategies has been shown in this updated systematic review on radiomics of bone and soft-tissue sarcomas, highlighting efforts to enhance methodology and bring radiomics from a research area to the clinical stage.

KEY POINTS

• 2021-2023 radiomic studies on CT and MRI of musculoskeletal sarcomas were reviewed. • Feature reproducibility was assessed in more than half (59%) of the studies. • Model clinical validation was performed in 69% of the studies. • Internal (44%) and/or external (29%) test datasets were employed for clinical validation.

AI applications in musculoskeletal imaging: a narrative review

This narrative review focuses on clinical applications of artificial intelligence (AI) in musculoskeletal imaging. A range of musculoskeletal disorders are discussed using a clinical-based approach, including trauma, bone age estimation, osteoarthritis, bone and soft-tissue tumors, and orthopedic implant-related pathology. Several AI algorithms have been applied to fracture detection and classification, which are potentially helpful tools for radiologists and clinicians. In bone age assessment, AI methods have been applied to assist radiologists by automatizing workflow, thus reducing workload and inter-observer variability. AI may potentially aid radiologists in identifying and grading abnormal findings of osteoarthritis as well as predicting the onset or progression of this disease. Either alone or combined with radiomics, AI algorithms may potentially improve diagnosis and outcome prediction of bone and soft-tissue tumors. Finally, information regarding appropriate positioning of orthopedic implants and related complications may be obtained using AI algorithms. In conclusion, rather than replacing radiologists, the use of AI should instead help them to optimize workflow, augment diagnostic performance, and keep up with ever-increasing workload.Relevance statement This narrative review provides an overview of AI applications in musculoskeletal imaging. As the number of AI technologies continues to increase, it will be crucial for radiologists to play a role in their selection and application as well as to fully understand their potential value in clinical practice. Key points • AI may potentially assist musculoskeletal radiologists in several interpretative tasks.• AI applications to trauma, age estimation, osteoarthritis, tumors, and orthopedic implants are discussed.• AI should help radiologists to optimize workflow and augment diagnostic performance.

Radiomics in Musculoskeletal Tumors

Sarcomas are heterogeneous rare tumors predominantly affecting the musculoskeletal (MSK) system. Due to significant variations in their natural history and variable response to conventional treatments, the discovery of novel diagnostic and prognostic biomarkers to guide therapeutic decision-making is an active and ongoing field of research. As new cellular, molecular, and metabolic biomarkers continue to be discovered, quantitative radiologic imaging is becoming increasingly important in sarcoma management. Radiomics offers the potential for discovering novel imaging diagnostic and predictive biomarkers using standard-of-care medical imaging. In this review, we detail the core concepts of radiomics and the application of radiomics to date in MSK sarcoma research. Also described are specific challenges related to radiomic studies, as well as viewpoints on clinical adoption and future perspectives in the field.

Segmentation methods applied to MRI-derived radiomic analysis for the prediction of placenta accreta spectrum in patients with placenta previa

PURPOSE

To retrospectively evaluate the performance of different manual segmentation methods of placenta MR images for predicting Placenta Accreta Spectrum (PAS) disorders in patients with placenta previa (PP) using a Machine Learning (ML) Radiomics analysis.

METHODS

64 patients (n=41 with PAS and n= 23 without PAS) with PP who underwent MRI examination for suspicion of PAS were retrospectively selected. All MRI examinations were acquired on a 1.5 T using T2-weighted (T2w) sequences on axial, sagittal and coronal planes. Ten different manual segmentation methods were performed on sagittal placental T2-weighted images obtaining five sets of 2D regions of interest (ROIs) and five sets of 3D volumes of interest (VOIs) from each patient. In detail, ROIs and VOIs were positioned on the following areas: placental tissue, retroplacental myometrium, cervix, placenta with underneath myometrium, placenta with underneath myometrium and cervix. For feature stability testing, the same process was repeated on 30 randomly selected placental MRI examinations by two additional radiologists, working independently and blinded to the original segmentation. Radiomic features were extracted from all available ROIs and VOIs. 100 iterations of 5-fold cross-validation with nested feature selection, based on recursive feature elimination, were subsequently run on each ROI/VOI to identify the best-performing method to classify instances correctly.

RESULTS

Among the segmentation methods, the best performance in predicting PAS was obtained by the VOIs covering the retroplacental myometrium (Mean validation score: 0.761, standard deviation: 0.116).

CONCLUSION

Our preliminary results show that the VOI including the retroplacental myometrium using T2w images seems to be the best method when segmenting images for the development of ML radiomics predictive models to identify PAS in patients with PP.

Prediction of 2-[18F]FDG PET-CT SUVmax for Adrenal Mass Characterization: A CT Radiomics Feasibility Study

BACKGROUND

Indeterminate adrenal masses (AM) pose a diagnostic challenge, and 2-[¹⁸F]FDG PET-CT serves as a problem-solving tool. Aim of this study was to investigate whether CT radiomics features could be used to predict the 2-[¹⁸F]FDG SUVmax of AM.

METHODS

Patients with AM on 2-[¹⁸F]FDG PET-CT scan were grouped based on iodine contrast injection as CT contrast-enhanced (CE) or CT unenhanced (NCE). Two-dimensional segmentations of AM were manually obtained by multiple operators on CT images. Image resampling and discretization (bin number = 16) were performed. 919 features were calculated using PyRadiomics. After scaling, unstable, redundant, and low variance features were discarded. Using linear regression and the Uniform Manifold Approximation and Projection technique, a CT radiomics synthetic value (RadSV) was obtained. The correlation between CT RadSV and 2-[¹⁸F]FDG SUVmax was assessed with Pearson test.

RESULTS

A total of 725 patients underwent PET-CT from April 2020 to April 2021. In 150 (21%) patients, a total of 179 AM (29 bilateral) were detected. Group CE consisted of 84 patients with 108 AM (size = 18.1 ± 4.9 mm) and Group NCE of 66 patients with 71 AM (size = 18.5 ± 3.8 mm). In both groups, 39 features were selected. No statisticallyf significant correlation between CT RadSV and 2-[¹⁸F]FDG SUVmax was found (Group CE, r = 0.18 and p = 0.058; Group NCE, r = 0.13 and p = 0.27).

CONCLUSIONS

It might not be feasible to predict 2-[¹⁸F]FDG SUVmax of AM using CT RadSV. Its role as a problem-solving tool for indeterminate AM remains fundamental.

The diagnostic value of magnetic resonance imaging-based texture analysis in differentiating enchondroma and chondrosarcoma

OBJECTIVE

To assess the diagnostic performance of MRI-based texture analysis for differentiating enchondromas and chondrosarcomas, especially on fat-suppressed proton density (FS-PD) images.

MATERIALS AND METHODS

The whole tumor volumes of 23 chondrosarcomas and 24 enchondromas were manually segmented on both FS-PD and T1-weighted images. A total of 861 radiomic features were extracted. SelectKBest was used to select the features. The data were randomly split into training (n = 36) and test (n = 10) for T1-weighted and training (n = 37) and test (n = 10) for FS-PD datasets. Fivefold cross-validation was performed. Fifteen machine learning models were created using the training set. The best models for T1-weighted, FS-PD, and T1-weighted + FS-PD images were selected in terms of accuracy and area under the curve (AUC).

RESULTS

There were 7 men and 16 women in the chondrosarcoma group (mean ± standard deviation age, 45.65 ± 11.24) and 7 men and 17 women in the enchondroma group (mean ± standard deviation age, 46.17 ± 11.79). Naive Bayes was the best model for accuracy and AUC for T1-weighted images (AUC = 0.76, accuracy = 80%, recall = 80%, precision = 80%, F1 score = 80%). The best model for FS-PD images was the K neighbors classifier for accuracy and AUC (AUC = 1.00, accuracy = 80%, recall = 80%, precision = 100%, F1 score = 89%). The best model for T1-weighted + FS-PD images was logistic regression for accuracy and AUC (AUC = 0.84, accuracy = 80%, recall = 60%, precision = 100%, F1 score = 75%).

CONCLUSION

MRI-based machine learning models have promising results in the discrimination of enchondroma and chondrosarcoma based on radiomic features obtained from both FS-PD and T1-weighted images.

T2-weighted MRI radiomics in high-grade intramedullary osteosarcoma: predictive accuracy in assessing histologic response to chemotherapy, overall survival, and disease-free survival

OBJECTIVE

To analyze radiomic features obtained from pre-treatment T2-weighted MRI acquisitions in patients with histologically proven intramedullary high-grade osteosarcomas and assess the accuracy of radiomic modelling as predictive biomarker of tumor necrosis following neoadjuvant chemotherapy (NAC), overall survival (OS), and disease-free survival (DFS).

MATERIALS AND METHODS

Pre-treatment MRI exams in 105 consecutive patients who underwent NAC and resection of high-grade intramedullary osteosarcoma were evaluated. Histologic necrosis following NAC, and clinical outcome-survival data was collected for each case. Radiomic features were extracted from segmentations performed by two readers, with poorly reproducible features excluded from further analysis. Cox proportional hazard model and Spearman correlation with multivariable modelling were used for assessing relationships of radiomics features with OS, DFS, and histologic tumor necrosis.

RESULTS

Study included 74 males, 31 females (mean 32.5yrs, range 15-77 years). Histologic assessment of tumor necrosis following NAC was available in 104 cases, with good response (≥ 90% necrosis) in 41, and poor response in 63. Fifty-three of 105 patients were alive at follow-up (median 40 months, range: 2-213 months). Median OS was 89 months. Excluding 14 patients with metastases at presentation, median DFS was 19 months. Eleven radiomics features were employed in final radiomics model predicting histologic tumor necrosis (mean AUC 0.708 ± 0.046). Thirteen radiomic features were used in model predicting OS (mean concordance index 0.741 ± 0.011), and 12 features retained in predicting DFS (mean concordance index 0.745 ± 0.010).

CONCLUSIONS

T2-weighted MRI radiomic models demonstrate promising results as potential prognostic biomarkers of prospective tumor response to neoadjuvant chemotherapy and prediction of clinical outcomes in conventional osteosarcoma.

Radiomics Applications in Head and Neck Tumor Imaging: A Narrative Review

Recent advances in machine learning and artificial intelligence technology have ensured automated evaluation of medical images. As a result, quantifiable diagnostic and prognostic biomarkers have been created. We discuss radiomics applications for the head and neck region in this paper. Molecular characterization, categorization, prognosis and therapy recommendation are given special consideration. In a narrative manner, we outline the fundamental technological principles, the overall idea and usual workflow of radiomic analysis and what seem to be the present and potential challenges in normal clinical practice. Clinical oncology intends for all of this to ensure informed decision support for personalized and useful cancer treatment. Head and neck cancers present a unique set of diagnostic and therapeutic challenges. These challenges are brought on by the complicated anatomy and heterogeneity of the area under investigation. Radiomics has the potential to address these barriers. Future research must be interdisciplinary and focus on the study of certain oncologic functions and outcomes, with external validation and multi-institutional cooperation in order to achieve this.

Robustness of Radiomic Features: Two-Dimensional versus Three-Dimensional MRI-Based Feature Reproducibility in Lipomatous Soft-Tissue Tumors

This retrospective study aimed to compare the intra- and inter-observer manual-segmentation variability in the feature reproducibility between two-dimensional (2D) and three-dimensional (3D) magnetic-resonance imaging (MRI)-based radiomic features. The study included patients with lipomatous soft-tissue tumors that were diagnosed with histopathology and underwent MRI scans. Tumor segmentation based on the 2D and 3D MRI images was performed by two observers to assess the intra- and inter-observer variability. In both the 2D and the 3D segmentations, the radiomic features were extracted from the normalized images. Regarding the stability of the features, the intraclass correlation coefficient (ICC) was used to evaluate the intra- and inter-observer segmentation variability. Features with ICC > 0.75 were considered reproducible. The degree of feature robustness was classified as low, moderate, or high. Additionally, we compared the efficacy of 2D and 3D contour-focused segmentation in terms of the effects of the stable feature rate, sensitivity, specificity, and diagnostic accuracy of machine learning on the reproducible features. In total, 93 and 107 features were extracted from the 2D and 3D images, respectively. Only 35 features from the 2D images and 63 features from the 3D images were reproducible. The stable feature rate for the 3D segmentation was more significant than for the 2D segmentation (58.9% vs. 37.6%, p = 0.002). The majority of the features for the 3D segmentation had moderate-to-high robustness, while 40.9% of the features for the 2D segmentation had low robustness. The diagnostic accuracy of the machine-learning model for the 2D segmentation was close to that for the 3D segmentation (88% vs. 90%). In both the 2D and the 3D segmentation, the specificity values were equal to 100%. However, the sensitivity for the 2D segmentation was lower than for the 3D segmentation (75% vs. 83%). For the 2D + 3D radiomic features, the model achieved a diagnostic accuracy of 87% (sensitivity, 100%, and specificity, 80%). Both 2D and 3D MRI-based radiomic features of lipomatous soft-tissue tumors are reproducible. With a higher stable feature rate, 3D contour-focused segmentation should be selected for the feature-extraction process.

Ovarian imaging radiomics quality score assessment: an EuSoMII radiomics auditing group initiative

OBJECTIVE

To evaluate the methodological rigor of radiomics-based studies using noninvasive imaging in ovarian setting.

METHODS

Multiple medical literature archives (PubMed, Web of Science, and Scopus) were searched to retrieve original studies focused on computed tomography (CT), magnetic resonance imaging (MRI), ultrasound (US), or positron emission tomography (PET) radiomics for ovarian disorders' assessment. Two researchers in consensus evaluated each investigation using the radiomics quality score (RQS). Subgroup analyses were performed to assess whether the total RQS varied according to first author category, study aim and topic, imaging modality, and journal quartile.

RESULTS

From a total of 531 items, 63 investigations were finally included in the analysis. The studies were greatly focused (94%) on the field of oncology, with CT representing the most used imaging technique (41%). Overall, the papers achieved a median total RQS 6 (IQR, -0.5 to 11), corresponding to a percentage of 16.7% of the maximum score (IQR, 0-30.6%). The scoring was low especially due to the lack of prospective design and formal validation of the results. At subgroup analysis, the 4 studies not focused on oncological topic showed significantly lower quality scores than the others.

CONCLUSIONS

The overall methodological rigor of radiomics studies in the ovarian field is still not ideal, limiting the reproducibility of results and potential translation to clinical setting. More efforts towards a standardized methodology in the workflow are needed to allow radiomics to become a viable tool for clinical decision-making.

KEY POINTS

• The 63 included studies using noninvasive imaging for ovarian applications were mostly focused on oncologic topic (94%). • The included investigations achieved a median total RQS 6 (IQR, -0.5 to 11), indicating poor methodological rigor. • The RQS was low especially due to the lack of prospective design and formal validation of the results.

Prediction of response to preoperative neoadjuvant chemotherapy in extremity high-grade osteosarcoma using X-ray and multiparametric MRI radiomics

PURPOSE

This study aims to evaluate the value of applying X-ray and magnetic resonance imaging (MRI) models based on radiomics feature to predict response of extremity high-grade osteosarcoma to neoadjuvant chemotherapy (NAC).

MATERIALS AND METHODS

A retrospective dataset was assembled involving 102 consecutive patients (training dataset, n = 72; validation dataset, n = 30) diagnosed with extremity high-grade osteosarcoma. The clinical features of age, gender, pathological type, lesion location, bone destruction type, size, alkaline phosphatase (ALP), and lactate dehydrogenase (LDH) were evaluated. Imaging features were extracted from X-ray and multi-parametric MRI (T1-weighted, T2-weighted, and contrast-enhanced T1-weighted) data. Features were selected using a two-stage process comprising minimal-redundancy-maximum-relevance (mRMR) and least absolute shrinkage and selection operator (LASSO) regression. Logistic regression (LR) modelling was then applied to establish models based on clinical, X-ray, and multi-parametric MRI data, as well as combinations of these datasets. Each model was evaluated using sensitivity, specificity, and area under the receiver operating characteristic curve (AUC) with a 95% confidence interval (CI).

RESULTS

AUCs of 5 models using clinical, X-ray radiomics, MRI radiomics, X-ray plus MRI radiomics, and combination of all were 0.760 (95% CI: 0.583-0.937), 0.706 (95% CI: 0.506-0.905), 0.751 (95% CI: 0.572-0.930), 0.796 (95% CI: 0.629-0.963), 0.828 (95% CI: 0.676-0.980), respectively. The DeLong test showed no significant difference between any pair of models (p > 0.05). The combined model yielded higher performance than the clinical and radiomics models as demonstrated by net reclassification improvement (NRI) and integrated difference improvement (IDI) values, respectively. This combined model was also found to be clinically useful in the decision curve analysis (DCA).

CONCLUSION

Modelling based on combination of clinical and radiomics data improves the ability to predict pathological responses to NAC in extremity high-grade osteosarcoma compared to the models based on either clinical or radiomics data.

A modified method for CT radiomics region-of-interest segmentation in adrenal lipid-poor adenomas: a two-institution comparative study

Objective

This study aimed to investigate the application of modified region-of-interest (ROI) segmentation method in unenhanced computed tomography in the radiomics model of adrenal lipid-poor adenoma, and to evaluate the diagnostic performance using an external medical institution data set and select the best ROI segmentation method.

Methods

The imaging data of 135 lipid-poor adenomas and 102 non-adenomas in medical institution A and 30 lipid-poor adenomas and 43 non-adenomas in medical institution B were retrospectively analyzed, and all cases were pathologically or clinically confirmed. The data of Institution A builds the model, and the data of Institution B verifies the diagnostic performance of the model. Semi-automated ROI segmentation of tumors was performed using uAI software, using maximum area single-slice method (MAX) and full-volume method (ALL), as well as modified single-slice method (MAX_E) and full-volume method (ALL_E) to segment tumors, respectively. The inter-rater correlation coefficients (ICC) was performed to assess the stability of the radiomics features of the four ROI segmentation methods. The area under the curve (AUC) and at least 95% specificity pAUC (Partial AUC) were used as measures of the diagnostic performance of the model.

Results

A total of 104 unfiltered radiomics features were extracted using each of the four segmentation methods. In the ROC analysis of the radiomics model, the AUC value of the model constructed by MAX was 0.925, 0.919, and 0.898 on the training set, the internal validation set, and the external validation set, respectively, and the AUC value of MAX_E was 0.937, 0.931, and 0.906, respectively. The AUC value of ALL was 0.929, 0.929, and 0.918, and the AUC value of ALL_E was 0.942, 0.926, and 0.927, respectively. In all samples, the pAUCs of MAX, MAX_E, ALL, and ALL_E were 0.021, 0.025, 0.018, and 0.028, respectively.

Conclusion

The diagnostic performance of the radiomics model constructed based on the full-volume method was better than that of the model based on the single-slice method. The model constructed using the ALL_E method had a stronger generalization ability and the highest AUC and pAUC value.

3D vs. 2D MRI radiomics in skeletal Ewing sarcoma: Feature reproducibility and preliminary machine learning analysis on neoadjuvant chemotherapy response prediction

OBJECTIVE

The extent of response to neoadjuvant chemotherapy predicts survival in Ewing sarcoma. This study focuses on MRI radiomics of skeletal Ewing sarcoma and aims to investigate feature reproducibility and machine learning prediction of response to neoadjuvant chemotherapy.

MATERIALS AND METHODS

This retrospective study included thirty patients with biopsy-proven skeletal Ewing sarcoma, who were treated with neoadjuvant chemotherapy before surgery at two tertiary sarcoma centres. 7 patients were poor responders and 23 were good responders based on pathological assessment of the surgical specimen. On pre-treatment T1-weighted and T2-weighted MRI, 2D and 3D tumour segmentations were manually performed. Features were extracted from original and wavelet-transformed images. Feature reproducibility was assessed through small geometrical transformations of the regions of interest mimicking multiple manual delineations, and intraclass correlation coefficient >0.75 defined feature reproducibility. Feature selection also consisted of collinearity and significance analysis. After class balancing in the training cohort, three machine learning classifiers were trained and tested on unseen data using hold-out cross-validation.

RESULTS

1303 (77%) 3D and 620 (65%) 2D radiomic features were reproducible. 4 3D and 4 2D features passed feature selection. Logistic regression built upon 3D features achieved the best performance with 85% accuracy (AUC=0.9) in predicting response to neoadjuvant chemotherapy.

CONCLUSION

Compared to 2D approach, 3D MRI radiomics of Ewing sarcoma had superior reproducibility and higher accuracy in predicting response to neoadjuvant chemotherapy, particularly when using logistic regression classifier.

Radiomics in Abdominopelvic Solid-Organ Oncologic Imaging: Current Status

Radiomics is the process of extraction of high-throughput quantitative imaging features from medical images. These features represent noninvasive quantitative biomarkers that go beyond the traditional imaging features visible to the human eye. This article first reviews the steps of the radiomics pipeline, including image acquisition, ROI selection and image segmentation, image preprocessing, feature extraction, feature selection, and model development and application. Current evidence for the application of radiomics in abdominopelvic solid-organ cancers is then reviewed. Applications including diagnosis, subtype determination, treatment response assessment, and outcome prediction are explored within the context of hepatobiliary and pancreatic cancer, renal cell carcinoma, prostate cancer, gynecologic cancer, and adrenal masses. This literature review focuses on the strongest available evidence, including systematic reviews, meta-analyses, and large multicenter studies. Limitations of the available literature are highlighted, including marked heterogeneity in radiomics methodology, frequent use of small sample sizes with high risk of overfitting, and lack of prospective design, external validation, and standardized radiomics workflow. Thus, although studies have laid a foundation that supports continued investigation into radiomics models, stronger evidence is needed before clinical adoption.

The effect of preprocessing filters on predictive performance in radiomics

Background

Radiomics is a noninvasive method using machine learning to support personalised medicine. Preprocessing filters such as wavelet and Laplacian-of-Gaussian filters are commonly used being thought to increase predictive performance. However, the use of preprocessing filters increases the number of features by up to an order of magnitude and can produce many correlated features. Both substantially increase the dataset complexity, which in turn makes modeling with machine learning techniques more challenging, possibly leading to poorer performance. We investigated the impact of these filters on predictive performance.

Methods

Using seven publicly available radiomic datasets, we measured the impact of adding features preprocessed with eight different preprocessing filters to the unprocessed features on the predictive performance of radiomic models. Modeling was performed using five feature selection methods and five classifiers, while predictive performance was measured using area-under-the-curve at receiver operating characteristics analysis (AUC-ROC) with nested, stratified 10-fold cross-validation.

Results

Significant improvements of up to 0.08 in AUC-ROC were observed when all image preprocessing filters were applied compared to using only the original features (up to p = 0.024). Decreases of -0.04 and -0.10 were observed on some data sets, but these were not statistically significant (p > 0.179). Tuning of the image preprocessing filters did not result in decreases in AUC-ROC but further improved results by up to 0.1; however, these improvements were not statistically significant (p > 0.086) except for one data set (p = 0.023).

Conclusions

Preprocessing filters can have a significant impact on the predictive performance and should be used in radiomic studies.

Supplementary Information

The online version contains supplementary material available at 10.1186/s41747-022-00294-w.

Using pre-operative radiomics to predict microvascular invasion of hepatocellular carcinoma based on Gd-EOB-DTPA enhanced MRI

Objectives

We aimed to investigate the value of performing gadolinium-ethoxybenzyl-diethylenetriamine pentaacetic acid (Gd-EOB-DTPA) enhanced magnetic resonance imaging (MRI) radiomics for preoperative prediction of microvascular invasion (MVI) of hepatocellular carcinoma (HCC) based on multiple sequences.

Methods

We randomly allocated 165 patients with HCC who underwent partial hepatectomy to training and validation sets. Stepwise regression and the least absolute shrinkage and selection operator algorithm were used to select significant variables. A clinicoradiological model, radiomics model, and combined model were constructed using multivariate logistic regression. The performance of the models was evaluated, and a nomogram risk-prediction model was built based on the combined model. A concordance index and calibration curve were used to evaluate the discrimination and calibration of the nomogram model.

Results

The tumour margin, peritumoural hypointensity, and seven radiomics features were selected to build the combined model. The combined model outperformed the radiomics model and the clinicoradiological model and had the highest sensitivity (90.89%) in the validation set. The areas under the receiver operating characteristic curve were 0.826, 0.755, and 0.708 for the combined, radiomics, and clinicoradiological models, respectively. The nomogram model based on the combined model exhibited good discrimination (concordance index = 0.79) and calibration.

Conclusions

The combined model based on radiomics features of Gd-EOB-DTPA enhanced MRI, tumour margin, and peritumoural hypointensity was valuable for predicting HCC microvascular invasion. The nomogram based on the combined model can intuitively show the probabilities of MVI.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12880-022-00855-w.

The stability of oncologic MRI radiomic features and the potential role of deep learning: a review

The use of MRI radiomic models for the diagnosis, prognosis and treatment response prediction of tumors has been increasingly reported in literature. However, its widespread adoption in clinics is hampered by issues related to features stability. In the MRI radiomic workflow, the main factors that affect radiomic features computation can be found in the image acquisition and reconstruction phase, in the image pre-processing steps, and in the segmentation of the region of interest on which radiomic indices are extracted. Deep Neural Networks (DNNs), having shown their potentiality in the medical image processing and analysis field, can be seen as an attractive strategy to partially overcome the issues related to radiomic stability and mitigate their impact. In fact, DNN approaches can be prospectively integrated in the MRI radiomic workflow to improve image quality, obtain accurate and reproducible segmentations and generate standardized images. In this review, DNN methods that can be included in the image processing steps of the radiomic workflow are described and discussed, in the light of a detailed analysis of the literature in the context of MRI radiomic reliability.

Can Radiomics Provide Additional Diagnostic Value for Identifying Adrenal Lipid-Poor Adenomas From Non-Adenomas on Unenhanced CT?

Background

It is difficult for radiologists to differentiate adrenal lipid-poor adenomas from non-adenomas; nevertheless, this differentiation is important as the clinical interventions required are different for adrenal lipid-poor adenomas and non-adenomas.

Purpose

To develop an unenhanced computed tomography (CT)-based radiomics model for identifying adrenal lipid-poor adenomas to assist in clinical decision-making.

Materials and methods

Patients with adrenal lesions who underwent CT between January 2015 and August 2021 were retrospectively recruited from two independent institutions. Patients from institution 1 were randomly divided into training and test sets, while those from institution 2 were used as the external validation set. The unenhanced attenuation and tumor diameter were measured to build a conventional model. Radiomics features were extracted from unenhanced CT images, and selected features were used to build a radiomics model. A nomogram model combining the conventional and radiomic features was also constructed. All the models were developed in the training set and validated in the test and external validation sets. The diagnostic performance of the models for identifying adrenal lipid-poor adenomas was compared.

Results

A total of 292 patients with 141 adrenal lipid-poor adenomas and 151 non-adenomas were analyzed. Patients with adrenal lipid-poor adenomas tend to have lower unenhanced attenuation and smoother image textures. In the training set, the areas under the curve of the conventional, radiomic, and nomogram models were 0.94, 0.93, and 0.96, respectively. There was no difference in diagnostic performance between the conventional and nomogram models in all datasets (all p < 0.05).

Conclusions

Our unenhanced CT-based nomogram model could effectively distinguish adrenal lipid-poor adenomas. The diagnostic power of conventional unenhanced CT imaging features may be underestimated, and further exploration is worthy.

Diffusion-weighted MRI radiomics of spine bone tumors: feature stability and machine learning-based classification performance

PURPOSE

To evaluate stability and machine learning-based classification performance of radiomic features of spine bone tumors using diffusion- and T2-weighted magnetic resonance imaging (MRI).

MATERIAL AND METHODS

This retrospective study included 101 patients with histology-proven spine bone tumor (22 benign; 38 primary malignant; 41 metastatic). All tumor volumes were manually segmented on morphologic T2-weighted sequences. The same region of interest (ROI) was used to perform radiomic analysis on ADC map. A total of 1702 radiomic features was considered. Feature stability was assessed through small geometrical transformations of the ROIs mimicking multiple manual delineations. Intraclass correlation coefficient (ICC) quantified feature stability. Feature selection consisted of stability-based (ICC > 0.75) and significance-based selections (ranking features by decreasing Mann-Whitney p-value). Class balancing was performed to oversample the minority (i.e., benign) class. Selected features were used to train and test a support vector machine (SVM) to discriminate benign from malignant spine tumors using tenfold cross-validation.

RESULTS

A total of 76.4% radiomic features were stable. The quality metrics for the SVM were evaluated as a function of the number of selected features. The radiomic model with the best performance and the lowest number of features for classifying tumor types included 8 features. The metrics were 78% sensitivity, 68% specificity, 76% accuracy and AUC 0.78.

CONCLUSION

SVM classifiers based on radiomic features extracted from T2- and diffusion-weighted imaging with ADC map are promising for classification of spine bone tumors. Radiomic features of spine bone tumors show good reproducibility rates.

Radiomics of Musculoskeletal Sarcomas: A Narrative Review

Bone and soft-tissue primary malignant tumors or sarcomas are a large, diverse group of mesenchymal-derived malignancies. They represent a model for intra- and intertumoral heterogeneities, making them particularly suitable for radiomics analyses. Radiomic features offer information on cancer phenotype as well as the tumor microenvironment which, combined with other pertinent data such as genomics and proteomics and correlated with outcomes data, can produce accurate, robust, evidence-based, clinical-decision support systems. Our purpose in this narrative review is to offer an overview of radiomics studies dealing with Magnetic Resonance Imaging (MRI)-based radiomics models of bone and soft-tissue sarcomas that could help distinguish different histotypes, low-grade from high-grade sarcomas, predict response to multimodality therapy, and thus better tailor patients’ treatments and finally improve their survivals. Although showing promising results, interobserver segmentation variability, feature reproducibility, and model validation are three main challenges of radiomics that need to be addressed in order to translate radiomics studies to clinical applications. These efforts, together with a better knowledge and application of the “Radiomics Quality Score” and Image Biomarker Standardization Initiative reporting guidelines, could improve the quality of sarcoma radiomics studies and facilitate radiomics towards clinical translation.

MRI radiomics-based machine learning classification of atypical cartilaginous tumour and grade II chondrosarcoma of long bones

Background

Atypical cartilaginous tumour (ACT) and grade II chondrosarcoma (CS2) of long bones are respectively managed with watchful waiting or curettage and wide resection. Preoperatively, imaging diagnosis can be challenging due to interobserver variability and biopsy suffers from sample errors. The aim of this study is to determine diagnostic performance of MRI radiomics-based machine learning in differentiating ACT from CS2 of long bones.

Methods

One-hundred-fifty-eight patients with surgically treated and histology-proven cartilaginous bone tumours were retrospectively included at two tertiary bone tumour centres. The training cohort consisted of 93 MRI scans from centre 1 (n=74 ACT; n=19 CS2). The external test cohort consisted of 65 MRI scans from centre 2 (n=45 ACT; n=20 CS2). Bidimensional segmentation was performed on T1-weighted MRI. Radiomic features were extracted. After dimensionality reduction and class balancing in centre 1, a machine-learning classifier (Extra Trees Classifier) was tuned on the training cohort using 10-fold cross-validation and tested on the external test cohort. In centre 2, its performance was compared with an experienced musculoskeletal oncology radiologist using McNemar's test.

Findings

After tuning on the training cohort (AUC=0.88), the machine-learning classifier had 92% accuracy (60/65, AUC=0.94) in identifying the lesions in the external test cohort. Its accuracies in correctly classifying ACT and CS2 were 98% (44/45) and 80% (16/20), respectively. The radiologist had 98% accuracy (64/65) with no difference compared to the classifier (p=0.134).

Interpretation

Machine learning showed high accuracy in classifying ACT and CS2 of long bones based on MRI radiomic features.

Funding

ESSR Young Researchers Grant.