Segmental quantification of hepatic lipid content based on volumetric MRI data in patients with suspected iron overload.

PURPOSE: To investigate the segmental distribution of hepatic fat fraction, determined with MRI (MR proton density fat fraction, short MR-PDFF) in patients suspected of having liver iron overload. METHODS: The liver of 44 patients examined with MRI using a 3D multi-echo gradient-echo sequence was segmented semiautomatically and subdivided into nine segments (segment 4 divided in 4a and 4b). Segmental fat content was determined on MR-PDFF maps. Whole-liver steatosis grades were compared to those found in individual segments. Segmental MR-PDFF differences were tested for statistical significance. RESULTS: The most common diseases were thalassemia, various forms of anemia, and hereditary hemochromatosis. No patients suffered from fat metabolism disease. Iron overload was present in 37/44 (84 %) patients. For the whole liver, 22 patients showed a steatosis grade of 0, 21 patients were graded S1, and one patient had a steatosis grade of 2. The grade of steatosis was underestimated in 5 of 21 patients (24 %) in segment 8 and in 8 of 21 patients (38 %) in segment 7. Highly significant segmental MR-PDFF differences were detected with p < 0.00 001, e. g., comparing segment 2 to 5. Segments 1 to 3 had the highest fat content, segments 7 and 8 had the lowest. CONCLUSION: Our results suggest that the storage of fat in the liver is inhomogeneous, so that segment-wise differing fat concentrations were found. Fat distribution in patients with suspected hepatic iron overload was similar to living liver donors. However, it showed significant differences compared with the values published for NAFLD patients, which were less pronounced in the group with high average hepatic MR-PDFF values than in the group with normal lipid content. In patients suspected of having iron overload, segment 8, which is mainly targeted for biopsy, and segment 7 may underestimate steatosis grade. KEY POINTS: · A volumetric analysis of 3D MRI data of patients with suspected hepatic iron overload yielded a markedly elevated MR proton density fat fraction (MR-PDFF) in hepatic segments 1 to 3.. · This hepatic fat distribution, observed for the whole patient cohort, is similar to healthy living liver donors.. · The subgroup of patients with a high average MR-PDFF ≥ 6.5 % shows this effect with lower segmental deviations.. · In patients without fat metabolic disorders, the steatosis grade may be underestimated when taking biopsies in segment 8 or 7..

Self-supervised pre-training with contrastive and masked autoencoder methods for dealing with small datasets in deep learning for medical imaging.

Deep learning in medical imaging has the potential to minimize the risk of diagnostic errors, reduce radiologist workload, and accelerate diagnosis. Training such deep learning models requires large and accurate datasets, with annotations for all training samples. However, in the medical imaging domain, annotated datasets for specific tasks are often small due to the high complexity of annotations, limited access, or the rarity of diseases. To address this challenge, deep learning models can be pre-trained on large image datasets without annotations using methods from the field of self-supervised learning. After pre-training, small annotated datasets are sufficient to fine-tune the models for a specific task. The most popular self-supervised pre-training approaches in medical imaging are based on contrastive learning. However, recent studies in natural image processing indicate a strong potential for masked autoencoder approaches. Our work compares state-of-the-art contrastive learning methods with the recently introduced masked autoencoder approach "SparK" for convolutional neural networks (CNNs) on medical images. Therefore, we pre-train on a large unannotated CT image dataset and fine-tune on several CT classification tasks. Due to the challenge of obtaining sufficient annotated training data in medical imaging, it is of particular interest to evaluate how the self-supervised pre-training methods perform when fine-tuning on small datasets. By experimenting with gradually reducing the training dataset size for fine-tuning, we find that the reduction has different effects depending on the type of pre-training chosen. The SparK pre-training method is more robust to the training dataset size than the contrastive methods. Based on our results, we propose the SparK pre-training for medical imaging tasks with only small annotated datasets.

Medical volume segmentation by overfitting sparsely annotated data.

Purpose: Semantic segmentation is one of the most significant tasks in medical image computing, whereby deep neural networks have shown great success. Unfortunately, supervised approaches are very data-intensive, and obtaining reliable annotations is time-consuming and expensive. Sparsely labeled approaches, such as bounding boxes, have shown some success in reducing the annotation time. However, in 3D volume data, each slice must still be manually labeled. Approach: We evaluate approaches that reduce the annotation effort by reducing the number of slices that need to be labeled in a 3D volume. In a two-step process, a similarity metric is used to select slices that should be annotated by a trained radiologist. In the second step, a predictor is used to predict the segmentation mask for the rest of the slices. We evaluate different combinations of selectors and predictors on medical CT and MRI volumes. Thus we can determine that combination works best, and how far slice annotations can be reduced. Results: Our results show that for instance for the Medical Segmentation Decathlon-heart dataset, some selector, and predictor combinations allow for a Dice score 0.969 when only annotating 20% of slices per volume. Experiments on other datasets show a similarly positive trend. Conclusions: We evaluate a method that supports experts during the labeling of 3D medical volumes. Our approach makes it possible to drastically reduce the number of slices that need to be manually labeled. We present a recommendation in which selector predictor combination to use for different tasks and goals.

Artificial intelligence in radiology - beyond the black box. / Künstliche Intelligenz in der Radiologie ­ jenseits der Black-Box.

BACKGROUND: Artificial intelligence is playing an increasingly important role in radiology. However, more and more often it is no longer possible to reconstruct decisions, especially in the case of new and powerful methods from the field of deep learning. The resulting models fulfill their function without the users being able to understand the internal processes and are used as so-called black boxes. Especially in sensitive areas such as medicine, the explainability of decisions is of paramount importance in order to verify their correctness and to be able to evaluate alternatives. For this reason, there is active research going on to elucidate these black boxes. METHOD: This review paper presents different approaches for explainable artificial intelligence with their advantages and disadvantages. Examples are used to illustrate the introduced methods. This study is intended to enable the reader to better assess the limitations of the corresponding explanations when meeting them in practice and strengthen the integration of such solutions in new research projects. RESULTS AND CONCLUSION: Besides methods to analyze black-box models for explainability, interpretable models offer an interesting alternative. Here, explainability is part of the process and the learned model knowledge can be verified with expert knowledge. KEY POINTS: · The use of artificial intelligence in radiology offers many possibilities to provide safer and more efficient medical care. This includes, but is not limited to support during image acquisition and processing or for diagnosis.. · Complex models can achieve high accuracy, but make it difficult to understand data processing.. · If the explainability is already taken into account during the planning of the model, methods can be developed that are powerful and interpretable at the same time.. CITATION FORMAT: · Gallée L, Kniesel H, Ropinski T et al. Artificial intelligence in radiology - beyond the black box. Fortschr Röntgenstr 2023; 195: 797 - 803.

Noninvasive liver iron quantification by MRI using refocused gradient-echo (bSSFP): preliminary results.

PURPOSE: To evaluate the feasibility of using a balanced steady-state free precession sequence (bSSFP) to determine liver iron content (LIC). METHOD: Thirty-five consecutive patients with liver iron overload were examined with bSSFP. Signal intensity ratios of liver parenchyma to paraspinal muscles were retrospectively correlated with LIC values obtained by FerriScan, which was used as the reference method. Combinations of bSSFP protocols were also evaluated. The best combination was utilized to calculate LIC from bSSFP data. The sensitivity and specificity for the therapeutically relevant LIC threshold of 80 µmol/g (4.5 mg/g) were determined. RESULTS: LIC values ranged from 24 to 756 µmol/g. The best SIR-to-LIC correlation of a single protocol was obtained with a 3.5-ms repetition time (TR) and 17° excitation flip angle (FA). A combination of protocols with TRs of 3.5, 5, and 6.5 ms, each at 17° FA, yielded a superior correlation. LIC values calculated using this combination resulted in a sensitivity/specificity of 0.91/0.85. CONCLUSION: bSSFP is basically suitable to determine LIC. Its advantages are high SNR efficiency and the ability to acquire the entire liver in a breath hold without acceleration techniques. KEY POINTS: · The bSSFP sequence is suited to quantify liver iron overload.. · bSSFP has a high scanning efficiency and potential for LIC screening.. · Despite susceptibility artifacts, the LIC determined from bSSFP data showed high accuracy.. CITATION FORMAT: · Wunderlich AP, Cario H, Götz M et al. Noninvasive liver iron quantification by MRI using refocused gradient-echo (bSSFP): preliminary results. Fortschr Röntgenstr 2023; 195: 804 - 808.

Prediction of Bone Marrow Biopsy Results From MRI in Multiple Myeloma Patients Using Deep Learning and Radiomics.

OBJECTIVES: In multiple myeloma and its precursor stages, plasma cell infiltration (PCI) and cytogenetic aberrations are important for staging, risk stratification, and response assessment. However, invasive bone marrow (BM) biopsies cannot be performed frequently and multifocally to assess the spatially heterogenous tumor tissue. Therefore, the goal of this study was to establish an automated framework to predict local BM biopsy results from magnetic resonance imaging (MRI). MATERIALS AND METHODS: This retrospective multicentric study used data from center 1 for algorithm training and internal testing, and data from center 2 to 8 for external testing. An nnU-Net was trained for automated segmentation of pelvic BM from T1-weighted whole-body MRI. Radiomics features were extracted from these segmentations, and random forest models were trained to predict PCI and the presence or absence of cytogenetic aberrations. Pearson correlation coefficient and the area under the receiver operating characteristic were used to evaluate the prediction performance for PCI and cytogenetic aberrations, respectively. RESULTS: A total of 672 MRIs from 512 patients (median age, 61 years; interquartile range, 53-67 years; 307 men) from 8 centers and 370 corresponding BM biopsies were included. The predicted PCI from the best model was significantly correlated ( P ≤ 0.01) to the actual PCI from biopsy in all internal and external test sets (internal test set: r = 0.71 [0.51, 0.83]; center 2, high-quality test set: r = 0.45 [0.12, 0.69]; center 2, other test set: r = 0.30 [0.07, 0.49]; multicenter test set: r = 0.57 [0.30, 0.76]). The areas under the receiver operating characteristic of the prediction models for the different cytogenetic aberrations ranged from 0.57 to 0.76 for the internal test set, but no model generalized well to all 3 external test sets. CONCLUSIONS: The automated image analysis framework established in this study allows for noninvasive prediction of a surrogate parameter for PCI, which is significantly correlated to the actual PCI from BM biopsy.

Artificial intelligence in coronary computed tomography angiography: Demands and solutions from a clinical perspective.

Coronary computed tomography angiography (CCTA) is increasingly the cornerstone in the management of patients with chronic coronary syndromes. This fact is reflected by current guidelines, which show a fundamental shift towards non-invasive imaging - especially CCTA. The guidelines for acute and stable coronary artery disease (CAD) of the European Society of Cardiology from 2019 and 2020 emphasize this shift. However, to fulfill this new role, a broader availability in adjunct with increased robustness of data acquisition and speed of data reporting of CCTA is needed. Artificial intelligence (AI) has made enormous progress for all imaging methodologies concerning (semi)-automatic tools for data acquisition and data post-processing, with outreach toward decision support systems. Besides onco- and neuroimaging, cardiac imaging is one of the main areas of application. Most current AI developments in the scenario of cardiac imaging are related to data postprocessing. However, AI applications (including radiomics) for CCTA also should enclose data acquisition (especially the fact of dose reduction) and data interpretation (presence and extent of CAD). The main effort will be to integrate these AI-driven processes into the clinical workflow, and to combine imaging data/results with further clinical data, thus - beyond the diagnosis of CAD- enabling prediction and forecast of morbidity and mortality. Furthermore, data fusing for therapy planning (e.g., invasive angiography/TAVI planning) will be warranted. The aim of this review is to present a holistic overview of AI applications in CCTA (including radiomics) under the umbrella of clinical workflows and clinical decision-making. The review first summarizes and analyzes applications for the main role of CCTA, i.e., to non-invasively rule out stable coronary artery disease. In the second step, AI applications for additional diagnostic purposes, i.e., to improve diagnostic power (CAC = coronary artery classifications), improve differential diagnosis (CT-FFR and CT perfusion), and finally improve prognosis (again CAC plus epi- and pericardial fat analysis) are reviewed.

A CRR2-Dependent sRNA Sequence Supports Papillomavirus Vaccine Expression in Tobacco Chloroplasts.

Human papillomavirus (HPV) infection is the leading cause of cervical cancer, and vaccination with HPV L1 capsid proteins has been successful in controlling it. However, vaccination coverage is not universal, particularly in developing countries, where 80% of all cervical cancer cases occur. Cost-effective vaccination could be achieved by expressing the L1 protein in plants. Various efforts have been made to produce the L1 protein in plants, including attempts to express it in chloroplasts for high-yield performance. However, manipulating chloroplast gene expression requires complex and difficult-to-control expression elements. In recent years, a family of nuclear-encoded, chloroplast-targeted RNA-binding proteins, the pentatricopeptide repeat (PPR) proteins, were described as key regulators of chloroplast gene expression. For example, PPR proteins are used by plants to stabilize and translate chloroplast mRNAs. The objective is to demonstrate that a PPR target site can be used to drive HPV L1 expression in chloroplasts. To test our hypothesis, we used biolistic chloroplast transformation to establish tobacco lines that express two variants of the HPV L1 protein under the control of the target site of the PPR protein CHLORORESPIRATORY REDUCTION2 (CRR2). The transgenes were inserted into a dicistronic operon driven by the plastid rRNA promoter. To determine the effectiveness of the PPR target site for the expression of the HPV L1 protein in the chloroplasts, we analyzed the accumulation of the transgenic mRNA and its processing, as well as the accumulation of the L1 protein in the transgenic lines. We established homoplastomic lines carrying either the HPV18 L1 protein or an HPV16B Enterotoxin::L1 fusion protein. The latter line showed severe growth retardation and pigment loss, suggesting that the fusion protein is toxic to the chloroplasts. Despite the presence of dicistronic mRNAs, we observed very little accumulation of monocistronic transgenic mRNA and no significant increase in CRR2-associated small RNAs. Although both lines expressed the L1 protein, quantification using an external standard suggested that the amounts were low. Our results suggest that PPR binding sites can be used to drive vaccine expression in plant chloroplasts; however, the factors that modulate the effectiveness of target gene expression remain unclear. The identification of dozens of PPR binding sites through small RNA sequencing expands the set of expression elements available for high-value protein production in chloroplasts.

Volumetric Evaluation of 3D Multi-Gradient-Echo MRI Data to Assess Whole Liver Iron Distribution by Segmental R2* Analysis: First Experience.

PURPOSE: MR transverse relaxation rate R2* has been shown to be useful for monitoring liver iron overload. A sequence enabling acquisition of the whole liver in a single breath hold is now available, thus allowing volumetric hepatic R2* distribution studies. We evaluated the feasibility of computer-assisted whole liver segmentation of 3 D multi-gradient-echo MRI data, and compared whole liver R2* determination to analyzing only a single slice. Also, segmental R2* differences were studied. MATERIALS AND METHODS: The liver of 44 patients, investigated by multi-gradient echo MRI at 1.5 T, was segmented and divided into nine segments. Segmental R2* values were examined for all patients together and with respect to two criteria: average R2* values, and reason for iron overload. Correlation of single-slice and volumetric data was tested with Spearman's rank test, segmental and group differences were evaluated by analysis of variance. RESULTS: Whole-liver R2* values correlated excellent to single slice data (p < 0.001). The lowest R2* occurred in segment 1 (S1), differences of S1 with regard to other segments were significant in five cases and highly significant in two cases. Patients with high average R2* showed significant differences between S1 and segments 2, 6, and 7. Disease-related differences with respect to S1 were significant in segments 3 to 5 and 7. CONCLUSION: Our results suggest inhomogeneous hepatic iron distribution. Low R2* in S1 may be explained by its special vascularization. KEY POINTS: · Hepatic R2* distribution is not as homogeneous as previously thought.. · Liver segments might have a functional relevance.. · Segmental and total liver R2* values coincide best in segment 8.. CITATION FORMAT: · Wunderlich AP, Cario H, Kannengießer S et al. Volumetric Evaluation of 3D Multi-Gradient-Echo MRI Data to Assess Whole Liver Iron Distribution by Segmental R2* Analysis: First Experience. Fortschr Röntgenstr 2023; 195: 224 - 233.

In Vivo Repeatability and Multiscanner Reproducibility of MRI Radiomics Features in Patients With Monoclonal Plasma Cell Disorders: A Prospective Bi-institutional Study.

OBJECTIVES: Despite the extensive number of publications in the field of radiomics, radiomics algorithms barely enter large-scale clinical application. Supposedly, the low external generalizability of radiomics models is one of the main reasons, which hinders the translation from research to clinical application. The objectives of this study were to investigate reproducibility of radiomics features (RFs) in vivo under variation of patient positioning, magnetic resonance imaging (MRI) sequence, and MRI scanners, and to identify a subgroup of RFs that shows acceptable reproducibility across all different acquisition scenarios. MATERIALS AND METHODS: Between November 30, 2020 and February 16, 2021, 55 patients with monoclonal plasma cell disorders were included in this prospective, bi-institutional, single-vendor study. Participants underwent one reference scan at a 1.5 T MRI scanner and several retest scans: once after simple repositioning, once with a second MRI protocol, once at another 1.5 T scanner, and once at a 3 T scanner. Radiomics feature from the bone marrow of the left hip bone were extracted, both from original scans and after different image normalizations. Intraclass correlation coefficient (ICC) was used to assess RF repeatability and reproducibility. RESULTS: Fifty-five participants (mean age, 59 ± 7 years; 36 men) were enrolled. For T1-weighted images after muscle normalization, in the simple test-retest experiment, 110 (37%) of 295 RFs showed an ICC ≥0.8: 54 (61%) of 89 first-order features (FOFs), 35 (95%) of 37 volume and shape features, and 21 (12%) of 169 texture features (TFs). When the retest was performed with different technical settings, even after muscle normalization, the number of FOF/TF with an ICC ≥0.8 declined to 58/13 for the second protocol, 29/7 for the second 1.5 T scanner, and 49/7 for the 3 T scanner, respectively. Twenty-five (28%) of the 89 FOFs and 6 (4%) of the 169 TFs from muscle-normalized T1-weighted images showed an ICC ≥0.8 throughout all repeatability and reproducibility experiments. CONCLUSIONS: In vivo, only few RFs are reproducible with different MRI sequences or different MRI scanners, even after application of a simple image normalization. Radiomics features selected by a repeatability experiment only are not necessarily suited to build radiomics models for multicenter clinical application. This study isolated a subset of RFs, which are robust to variations in MRI acquisition observed in scanners from 1 vendor, and therefore are candidates to build reproducible radiomics models for monoclonal plasma cell disorders for multicentric applications, at least when centers are equipped with scanners from this vendor.

Combining Deep Learning and Radiomics for Automated, Objective, Comprehensive Bone Marrow Characterization From Whole-Body MRI: A Multicentric Feasibility Study.

OBJECTIVES: Disseminated bone marrow (BM) involvement is frequent in multiple myeloma (MM). Whole-body magnetic resonance imaging (wb-MRI) enables to evaluate the whole BM. Reading of such whole-body scans is time-consuming, and yet radiologists can transfer only a small fraction of the information of the imaging data set to the report. This limits the influence that imaging can have on clinical decision-making and in research toward precision oncology. The objective of this feasibility study was to implement a concept for automatic, comprehensive characterization of the BM from wb-MRI, by automatic BM segmentation and subsequent radiomics analysis of 30 different BM spaces (BMS). MATERIALS AND METHODS: This retrospective multicentric pilot study used a total of 106 wb-MRI from 102 patients with (smoldering) MM from 8 centers. Fifty wb-MRI from center 1 were used for training of segmentation algorithms (nnU-Nets) and radiomics algorithms. Fifty-six wb-MRI from 8 centers, acquired with a variety of different MRI scanners and protocols, were used for independent testing. Manual segmentations of 2700 BMS from 90 wb-MRI were performed for training and testing of the segmentation algorithms. For each BMS, 296 radiomics features were calculated individually. Dice score was used to assess similarity between automatic segmentations and manual reference segmentations. RESULTS: The "multilabel nnU-Net" segmentation algorithm, which performs segmentation of 30 BMS and labels them individually, reached mean dice scores of 0.88 ± 0.06/0.87 ± 0.06/0.83 ± 0.11 in independent test sets from center 1/center 2/center 3-8 (interrater variability between radiologists, 0.88 ± 0.01). The subset from the multicenter, multivendor test set (center 3-8) that was of high imaging quality was segmented with high precision (mean dice score, 0.87), comparable to the internal test data from center 1. The radiomic BM phenotype consisting of 8880 descriptive parameters per patient, which result from calculation of 296 radiomics features for each of the 30 BMS, was calculated for all patients. Exemplary cases demonstrated connections between typical BM patterns in MM and radiomic signatures of the respective BMS. In plausibility tests, predicted size and weight based on radiomics models of the radiomic BM phenotype significantly correlated with patients' actual size and weight ( P = 0.002 and P = 0.003, respectively). CONCLUSIONS: This pilot study demonstrates the feasibility of automatic, objective, comprehensive BM characterization from wb-MRI in multicentric data sets. This concept allows the extraction of high-dimensional phenotypes to capture the complexity of disseminated BM disorders from imaging. Further studies need to assess the clinical potential of this method for automatic staging, therapy response assessment, or prediction of biopsy results.

Longitudinal CT Imaging to Explore the Predictive Power of 3D Radiomic Tumour Heterogeneity in Precise Imaging of Mantle Cell Lymphoma (MCL).

The study's primary aim is to evaluate the predictive performance of CT-derived 3D radiomics for MCL risk stratification. The secondary objective is to search for radiomic features associated with sustained remission. Included were 70 patients: 31 MCL patients and 39 control subjects with normal axillary lymph nodes followed over five years. Radiomic analysis of all targets (n = 745) was performed and features selected using the Mann Whitney U test; the discriminative power of identifying "high-risk MCL" was evaluated by receiver operating characteristics (ROC). The four radiomic features, "Uniformity", "Entropy", "Skewness" and "Difference Entropy" showed predictive significance for relapse (p < 0.05)-in contrast to the routine size measurements, which showed no relevant difference. The best prognostication for relapse achieved the feature "Uniformity" (AUC-ROC-curve 0.87; optimal cut-off ≤0.0159 to predict relapse with 87% sensitivity, 65% specificity, 69% accuracy). Several radiomic features, including the parameter "Short Axis," were associated with sustained remission. CT-derived 3D radiomics improves the predictive estimation of MCL patients; in combination with the ability to identify potential radiomic features that are characteristic for sustained remission, it may assist physicians in the clinical management of MCL.

Pre-examinations Improve Automated Metastases Detection on Cranial MRI.

MATERIALS AND METHODS: Our local ethics committee approved this retrospective monocenter study.First, a dual-time approach was assessed, for which the CNN was provided sequences of the MRI that initially depicted new MM (diagnosis MRI) as well as of a prediagnosis MRI: inclusion of only contrast-enhanced T1-weighted images (CNNdual_ce) was compared with inclusion of also the native T1-weighted images, T2-weighted images, and FLAIR sequences of both time points (CNNdual_all).Second, results were compared with the corresponding single time approaches, in which the CNN was provided exclusively the respective sequences of the diagnosis MRI.Casewise diagnostic performance parameters were calculated from 5-fold cross-validation. RESULTS: In total, 94 cases with 494 MMs were included. Overall, the highest diagnostic performance was achieved by inclusion of only the contrast-enhanced T1-weighted images of the diagnosis and of a prediagnosis MRI (CNNdual_ce, sensitivity = 73%, PPV = 25%, F1-score = 36%). Using exclusively contrast-enhanced T1-weighted images as input resulted in significantly less false-positives (FPs) compared with inclusion of further sequences beyond contrast-enhanced T1-weighted images (FPs = 5/7 for CNNdual_ce/CNNdual_all, P < 1e-5). Comparison of contrast-enhanced dual and mono time approaches revealed that exclusion of prediagnosis MRI significantly increased FPs (FPs = 5/10 for CNNdual_ce/CNNce, P < 1e-9).Approaches with only native sequences were clearly inferior to CNNs that were provided contrast-enhanced sequences. CONCLUSIONS: Automated MM detection on contrast-enhanced T1-weighted images performed with high sensitivity. Frequent FPs due to artifacts and vessels were significantly reduced by additional inclusion of prediagnosis MRI, but not by inclusion of further sequences beyond contrast-enhanced T1-weighted images. Future studies might investigate different change detection architectures for computer-aided detection.

The Image Biomarker Standardization Initiative: Standardized Quantitative Radiomics for High-Throughput Image-based Phenotyping.

Background Radiomic features may quantify characteristics present in medical imaging. However, the lack of standardized definitions and validated reference values have hampered clinical use. Purpose To standardize a set of 174 radiomic features. Materials and Methods Radiomic features were assessed in three phases. In phase I, 487 features were derived from the basic set of 174 features. Twenty-five research teams with unique radiomics software implementations computed feature values directly from a digital phantom, without any additional image processing. In phase II, 15 teams computed values for 1347 derived features using a CT image of a patient with lung cancer and predefined image processing configurations. In both phases, consensus among the teams on the validity of tentative reference values was measured through the frequency of the modal value and classified as follows: less than three matches, weak; three to five matches, moderate; six to nine matches, strong; 10 or more matches, very strong. In the final phase (phase III), a public data set of multimodality images (CT, fluorine 18 fluorodeoxyglucose PET, and T1-weighted MRI) from 51 patients with soft-tissue sarcoma was used to prospectively assess reproducibility of standardized features. Results Consensus on reference values was initially weak for 232 of 302 features (76.8%) at phase I and 703 of 1075 features (65.4%) at phase II. At the final iteration, weak consensus remained for only two of 487 features (0.4%) at phase I and 19 of 1347 features (1.4%) at phase II. Strong or better consensus was achieved for 463 of 487 features (95.1%) at phase I and 1220 of 1347 features (90.6%) at phase II. Overall, 169 of 174 features were standardized in the first two phases. In the final validation phase (phase III), most of the 169 standardized features could be excellently reproduced (166 with CT; 164 with PET; and 164 with MRI). Conclusion A set of 169 radiomics features was standardized, which enabled verification and calibration of different radiomics software. © RSNA, 2020 Online supplemental material is available for this article. See also the editorial by Kuhl and Truhn in this issue.

Optimal Statistical Incorporation of Independent Feature Stability Information into Radiomics Studies.

Conducting side experiments termed robustness experiments, to identify features that are stable with respect to rescans, annotation, or other confounding effects is an important element in radiomics research. However, the matter of how to include the finding of these experiments into the model building process still needs to be explored. Three different methods for incorporating prior knowledge into a radiomics modelling process were evaluated: the naïve approach (ignoring feature quality), the most common approach consisting of removing unstable features, and a novel approach using data augmentation for information transfer (DAFIT). Multiple experiments were conducted using both synthetic and publicly available real lung imaging patient data. Ignoring additional information from side experiments resulted in significantly overestimated model performances meaning the estimated mean area under the curve achieved with a model was increased. Removing unstable features improved the performance estimation, while slightly decreasing the model performance, i.e. decreasing the area under curve achieved with the model. The proposed approach was superior both in terms of the estimation of the model performance and the actual model performance. Our experiments show that data augmentation can prevent biases in performance estimation and has several advantages over the plain omission of the unstable feature. The actual gain that can be obtained depends on the quality and applicability of the prior information on the features in the given domain. This will be an important topic of future research.

MITK Phenotyping: An open-source toolchain for image-based personalized medicine with radiomics.

Radiomics - The extraction of quantitative features from radiologic images - shows increasing potential in contributing to modern personalized medicine approaches. MITK Phenotyping is an openly distributed radiomics framework implementing an exhaustive set of features, adhering to most recent international standards, and supporting a variety of different user interfaces and programming languages.

Correlation between genomic index lesions and mpMRI and 68Ga-PSMA-PET/CT imaging features in primary prostate cancer.

Magnetic resonance imaging (MRI) and prostate specific membrane antigen (PSMA)- positron emission tomography (PET)/computed tomography (CT)-imaging of prostate cancer (PCa) are emerging techniques to assess the presence of significant disease and tumor progression. It is not known, however, whether and to what extent lesions detected by these imaging techniques correlate with genomic features of PCa. The aim of this study was therefore to define a genomic index lesion based on chromosomal copy number alterations (CNAs) as marker for tumor aggressiveness in prostate biopsies in direct correlation to multiparametric (mp) MRI and 68Ga-PSMA-PET/CT imaging features. CNA profiles of 46 biopsies from five consecutive patients with clinically high-risk PCa were obtained from radiologically suspicious and unsuspicious areas. All patients underwent mpMRI, MRI/TRUS-fusion biopsy, 68Ga-PSMA-PET/CT and a radical prostatectomy. CNAs were directly correlated to imaging features and radiogenomic analyses were performed. Highly significant CNAs (≥10 Mbp) were found in 22 of 46 biopsies. Chromosome 8p, 13q and 5q losses were the most common findings. There was an strong correspondence between the radiologic and the genomic index lesions. The radiogenomic analyses suggest the feasibility of developing radiologic signatures that can distinguish between genomically more or less aggressive lesions. In conclusion, imaging features of mpMRI and 68Ga-PSMA-PET/CT can guide to the genomically most aggressive lesion of a PCa. Radiogenomics may help to better differentiate between indolent and aggressive PCa in the future.

Radiomic Machine Learning for Characterization of Prostate Lesions with MRI: Comparison to ADC Values.

Purpose To compare biparametric contrast-free radiomic machine learning (RML), mean apparent diffusion coefficient (ADC), and radiologist assessment for characterization of prostate lesions detected during prospective MRI interpretation. Materials and Methods This single-institution study included 316 men (mean age ± standard deviation, 64.0 years ± 7.8) with an indication for MRI-transrectal US fusion biopsy between May 2015 and September 2016 (training cohort, 183 patients; test cohort, 133 patients). Lesions identified by prospective clinical readings were manually segmented for mean ADC and radiomics analysis. Global and zone-specific random forest RML and mean ADC models for classification of clinically significant prostate cancer (Gleason grade group ≥ 2) were developed on the training set and the fixed models tested on an independent test set. Clinical readings, mean ADC, and radiomics were compared by using the McNemar test and receiver operating characteristic (ROC) analysis. Results In the test set, radiologist interpretation had a per-lesion sensitivity of 88% (53 of 60) and specificity of 50% (79 of 159). Quantitative measurement of the mean ADC (cut-off 732 mm2/sec) significantly reduced false-positive (FP) lesions from 80 to 60 (specificity 62% [99 of 159]) and false-negative (FN) lesions from seven to six (sensitivity 90% [54 of 60]) (P = .048). Radiologist interpretation had a per-patient sensitivity of 89% (40 of 45) and specificity of 43% (38 of 88). Quantitative measurement of the mean ADC reduced the number of patients with FP lesions from 50 to 43 (specificity 51% [45 of 88]) and the number of patients with FN lesions from five to three (sensitivity 93% [42 of 45]) (P = .496). Comparison of the area under the ROC curve (AUC) for the mean ADC (AUCglobal = 0.84; AUCzone-specific ≤ 0.87) vs the RML (AUCglobal = 0.88, P = .176; AUCzone-specific ≤ 0.89, P ≥ .493) showed no significantly different performance. Conclusion Quantitative measurement of the mean apparent diffusion coefficient (ADC) improved differentiation of benign versus malignant prostate lesions, compared with clinical assessment. Radiomic machine learning had comparable but not better performance than mean ADC assessment. © RSNA, 2018 Online supplemental material is available for this article.

Early postoperative delineation of residual tumor after low-grade glioma resection by probabilistic quantification of diffusion-weighted imaging.

OBJECTIVEIn WHO grade II low-grade gliomas (LGGs), early postoperative MRI (epMRI) may overestimate residual tumor on FLAIR sequences. Consequently, MRI at 3-6 months follow-up (fuMRI) is used for delineation of residual tumor. This study sought to evaluate if integration of apparent diffusion coefficient (ADC) maps permits an accurate estimation of residual tumor early on epMRI.METHODSFrom a consecutive cohort, 43 cases with an initial surgery for an LGG, and complete epMRI (< 72 hours after resection) and fuMRI including ADC maps, were retrospectively identified. Residual FLAIR hyperintense tumor was manually segmented on epMRI and corresponding ADC maps were coregistered. Using an expectation maximization algorithm, residual tumor segments were probabilistically clustered into areas of residual tumor, ischemia, or normal white matter (NWM) by fitting a mixture model of superimposed Gaussian curves to the ADC histogram. Tumor volumes from epMRI, clustering, and fuMRI were statistically compared and agreement analysis was performed.RESULTSMean FLAIR hyperintensity suggesting residual tumor was significantly larger on epMRI compared to fuMRI (19.4 ± 16.5 ml vs 8.4 ± 10.2 ml, p < 0.0001). Probabilistic clustering of corresponding ADC histograms on epMRI identified subsegments that were interpreted as mean residual tumor (7.6 ± 10.2 ml), ischemia (8.1 ± 5.9 ml), and NWM (3.7 ± 4.9 ml). Therefore, mean tumor quantification error between epMRI and fuMRI was significantly reduced (11.0 ± 10.6 ml vs -0.8 ± 3.7 ml, p < 0.0001). Mean clustered tumor volumes on epMRI were no longer significantly different from the fuMRI reference (7.6 ± 10.2 ml vs 8.4 ± 10.2 ml, p = 0.16). Correlation (Pearson r = 0.96, p < 0.0001), concordance correlation coefficient (0.89, 95% confidence interval 0.83), and Bland-Altman analysis suggested strong agreement between both measures after clustering.CONCLUSIONSProbabilistic segmentation of ADC maps facilitates accurate assessment of residual tumor within 72 hours after LGG resection. Multiparametric image analysis detected FLAIR signal alterations attributable to surgical trauma, which led to overestimation of residual LGG on epMRI compared to fuMRI. The prognostic value and clinical impact of this method has to be evaluated in larger case series in the future.