Quantitative Imaging Biomarkers Research Articles

The ISMRM Open Science Initiative for Perfusion Imaging (OSIPI): Results from the OSIPI-Dynamic Contrast-Enhanced challenge.

has often been proposed as a quantitative imaging biomarker for diagnosis, prognosis, and treatment response assessment for various tumors. None of the many software tools for quantification are standardized. The ISMRM Open Science Initiative for Perfusion Imaging-Dynamic Contrast-Enhanced (OSIPI-DCE) challenge was designed to benchmark methods to better help the efforts to standardize measurement. A framework was created to evaluate values produced by DCE-MRI analysis pipelines to enable benchmarking. The perfusion MRI community was invited to apply their pipelines for quantification in glioblastoma from clinical and synthetic patients. Submissions were required to include the entrants' values, the applied software, and a standard operating procedure. These were evaluated using the proposed score defined with accuracy, repeatability, and reproducibility components. Across the 10 received submissions, the score ranged from 28% to 78% with a 59% median. The accuracy, repeatability, and reproducibility scores ranged from 0.54 to 0.92, 0.64 to 0.86, and 0.65 to 1.00, respectively (0-1 = lowest-highest). Manual arterial input function selection markedly affected the reproducibility and showed greater variability in analysis than automated methods. Furthermore, provision of a detailed standard operating procedure was critical for higher reproducibility. This study reports results from the OSIPI-DCE challenge and highlights the high inter-software variability within estimation, providing a framework for ongoing benchmarking against the scores presented. Through this challenge, the participating teams were ranked based on the performance of their software tools in the particular setting of this challenge. In a real-world clinical setting, many of these tools may perform differently with different benchmarking methodology.

Analysis of the pattern of microstructural changes in the brain after mTBI with diffusion tensor imaging and subject-specific FE models

Traumatic brain injury (TBI) is a major public health challenge. Up to 90 % of TBIs are on the mild spectrum of TBI (mTBI), where diagnosis is a major challenge. Majority of studies in this field have been conducted on human subjects, which inherently suffer from the lack of appropriate control group, selection bias, and individual differences in patients. To overcome these limitations, animal studies have been used as an alternative approach to provide deeper insights into the underlying mechanism related to the injury. Therefore our aim is to investigate various quantitative imaging biomarkers acquired from T1-W and diffusion tensor imaging (DTI) sequences to provide more information about the microstructural changes in the brain after mTBI. We then use this to generate subject-specific finite element models of the brain and examine how the changes in the brain material properties reflected in MR images affects strain distribution patterns on a subsequent head hit. Our study revealed a decrease in FA and an increase in diffusivity indices (MD, AD, RD) in the white matter tracts of the brain. This finding may represent the axonal damage, demyelination and gliosis after mild TBI, which have been shown in other animal and human studies. Moreover, our FE analysis showed that microstructural changes in the brain after mTBI might have weakened the structural integrity of the brain as the subsequent head hit led to wider and more severe brain deformations. SignificanceAnimal models have been used to investigate biomechanical and pathophysiological aspects of mild traumatic brain injuries in the past. Still, most of them used small animals such as rats and mice. These models provided valuable insight into the pathophysiology of mTBI, but their findings have limitations due to their inherent differences to human brains. We have developed a large animal model of mTBI with sheep brains by combining advanced MRI and finite element analysis as they mimic the human brain better. To the best of our knowledge, this study is the first mTBI neuroimaging study conducted on large animal brains to investigate the diffusional changes in the white matter tracts after mTBI. Our FE analysis revealed that such microstructural changes resulted in tissue softening as the extent of brain deformation increased on a subsequent head hit, indicating increased brain vulnerability after head impacts.

Evaluation of the Reproducibility of MR Elastography Measurements of the Lumbar Back Muscles.

MR elastography (MRE) may provide quantitative imaging biomarkers of lumbar back muscles (LBMs), complementing MRI in spinal diseases by assessing muscle mechanical properties. However, reproducibility analyses for MRE of LBM are lacking. To assess technical failure, within-day and inter-day reproducibility, robustness with the excitation source positioning, and inter-observer agreement of MRE of muscles. Prospective. Seventeen healthy subjects (mean age 28 ± 4 years; 11 females). 1.5 T, gradient-echo MRE, T1-weighted turbo spin echo. The pneumatic driver was centered at L3 level. Four MRE were performed during two visits, 2-4 weeks apart, each consisting of two MRE with less than 10 minutes inter-scan interval. At Visit 1, after the first MRE, the coil and driver were removed, then reinstalled. The MRE was repeated. At Visit 2, following the first MRE, only the driver was moved down 5 cm. The MRE was repeated. Two radiologists segmented the multifidus and erector spinae muscles. Paired t-test, analysis of variance, intraclass correlation coefficients (ICCs). P-values <0.05 were considered statistically significant. Mean stiffness of LBM ranged from 1.44 to 1.60 kPa. Mean technical failure rate was 2.5%. Inter-observer agreement was excellent (ICC ranging from 0.82 [0.64-0.96] to 0.99 [0.98-0.99] in the multifidus, and from 0.85 [0.69-0.92] to 0.99 [0.97-0.99] in the erector spinae muscles). Within-day reproducibility was fair in the multifidus (ICC: 0.53 [0.47-0.77]) and good in the erector spinae muscles (ICC: 0.74 [0.48-0.88]). Reproducibility after moving the driver was excellent in both multifidus (ICC: 0.85 [0.69-0.93]) and erector spinae muscles (ICC: 0.84 [0.67-0.92]). Inter-day reproducibility was excellent in the multifidus (ICC: 0.76 [0.48-0.89]) and poor in the erector spinae muscles (ICC: 0.23 [-0.61 to 0.63]). MRE of LBM provides measurements of stiffness with fair to excellent reproducibility and excellent inter-observer agreement. However, inter-day reproducibility in the multifidus muscles indicated that the herein used MRE protocol may not be optimal for this muscle. 2 TECHNICAL EFFICACY: Stage 1.

Reproducibility of diffusion-weighted magnetic resonance imaging in head and neck cancer assessed on a 1.5 T MR-Linac and comparison to parallel measurements on a 3 T diagnostic scanner

Background and purposeBefore quantitative imaging biomarkers (QIBs) acquired with magnetic resonance imaging (MRI) can be used for interventional trials in radiotherapy (RT), technical validation of these QIBs is necessary. The aim of this study was to assess the reproducibility of apparent diffusion coefficient (ADC) values, derived from diffusion-weighted (DW) MRI, in head and neck cancer using a 1.5 T MR-Linac (MRL) by comparison to a 3 T diagnostic scanner (DS). Material and methodsDW-MRIs were acquired on MRL and DS for 15 head and neck cancer patients before RT and in week 2 and rigidly registered to the planning computed tomography. Mean ADC values were calculated for submandibular (SG) and parotid (PG) glands as well as target volumes (TV, gross tumor volume and lymph nodes), which were delineated based on computed tomography. Mean absolute ADC differences as well as within-subject coefficient of variation (wCV) and intraclass correlation coefficients (ICCs) were calculated for all volumes of interest. ResultsA total of 23 datasets were analyzed. Mean ADC difference (DS-MRL) for SG, PG and TV resulted in 142, 254 and 93·10-6 mm2/s. wCVs/ICCs, comparing MRL and DS, were determined as 13.7 %/0.26, 24.4 %/0.23 and 16.1 %/0.73 for SG, PG and TV, respectively. ConclusionADC values, measured on the 1.5 T MRL, showed reasonable reproducibility with an ADC underestimation in contrast to the DS. This ADC shift must be validated in further experiments and considered for future translation of QIB candidates from DS to MRL for response adaptive RT.

A background parenchymal enhancement quantification framework of breast magnetic resonance imaging.

Background parenchymal enhancement (BPE) is defined as the enhanced proportion of normal fibroglandular tissue on enhanced magnetic resonance imaging. BPE shows promise as a quantitative imaging biomarker (QIB). However, the lack of consensus among radiologists in their semi-quantitative grading of BPE limits its clinical utility. The main objective of this study was to develop a BPE quantification model according to clinical expertise, with the BPE integral being used as a QIB to incorporate both the volume and intensity of the enhancement metrics. The model was applied to 2,786 cases to compare our quantitative results with radiologists' semi-quantitative BPE grading to evaluate the effectiveness of using the BPE integral as a QIB for analyzing BPE. Comparisons between multiple groups of nonnormally distributed BPE integrals were performed using the Kruskal-Wallis test. Our study found a considerable degree of concordance between our BPE quantitative integral and radiologists' semi-quantitative assessments. Specifically, our research results revealed significant variability in BPE integral attained through the BPE quantification framework among all semi-quantitative BPE grading groups labeled by experienced radiologists, including mild-moderate (P<0.001), mild-marked (P<0.001), and moderate-marked (P<0.001). Furthermore, there was an apparent correlation between BPE integral and BPE grades, with marked BPE displaying the highest BPE integral, followed by moderate BPE, with mild BPE exhibiting the lowest BPE integral value. The study developed and implemented a BPE quantification framework, which incorporated both the volume and intensity of enhancement and which could serve as a QIB for BPE.

Generalized Calibrated BOLD fMRI for the Relationship between Cerebral Metabolism and Hypertensive Status

AbstractBackgroundElevated blood pressure levels during mid‐life have been found to be associated with an increased susceptibility to developing dementia later in life, particularly vascular dementia. Quantitative imaging biomarkers, such as Oxygen Extraction Fraction (OEF) and Cerebral Metabolism Rate of Oxygen (CMRO2), can be indicators of impaired cerebral metabolism. Calibrated Blood Oxygen Level Dependent (BOLD) functional Magnetic Resonance Imaging (fMRI) is a multi‐parametric approach utilizing hyperoxic and hypercapnic gas challenges (Figure 1). The impact of hypertension on OEF and CMRO2 has been investigated.MethodA cohort of 84 participants was recruited from the Wake Forest Alzheimer’s Disease Research Center (ADRC) Clinical Core (Table 1). Among all the covariates, participants with prediabetes and diabetes were considered diabetic, and those who exhibited high blood pressure or were using medication to manage hypertension were classified as hypertensive. The MRI was conducted using a 3T Siemens Skyra MRI scanner with a 32‐channel head coil. Dynamic Pseudocontinuous Arterial Spin Labeling (PCASL) was used to acquire cerebral blood flow (CBF) and BOLD images during respiratory challenges with an end‐tidal forcing system. The ROI values in the frontal lobe, temporal lobe, parietal lobe, and occipital lobe regions were chosen. The correlation between hypertensive status and OEF and CMRO2 in the ROI was analyzed with a linear mixed model, with covariates of age, sex, cognitive status, APOE genotype, diabetic status, and hypertensive status.ResultA negative correlation was found between CBF, OEF, and CMRO2 with hypertensive status in frontal lobe region (Figure 2). Statistical significance was observed for OEF and CMRO2 (p = 0.049 and p = 0.028, respectively), indicating the potential influence of hypertensive status on these biomarkers.ConclusionThe study concludes that there is a significant negative association between hypertensive status and both OEF and CMRO2, particularly in the frontal lobe region. The observed reduction in OEF and CMRO2 is not solely due to decreased CBF, but rather complex and multifactorial processes involving vascular remodeling and altered oxygen metabolism. Additional research is necessary to fully understand the underlying pathophysiological mechanisms that link hypertension to decreased OEF and CMRO2, particularly in the context of regional specificity.

Tensor radiomics: paradigm for systematic incorporation of multi-flavoured radiomics features.

Radiomics features hold significant value as quantitative imaging biomarkers for diagnosis, prognosis, and treatment response assessment. To generate radiomics features and ultimately develop signatures, various factors can be manipulated, including image discretization parameters (e.g., bin number or size), convolutional filters, segmentation perturbation, or multi-modality fusion levels. Typically, only one set of parameters is employed, resulting in a single value or "flavour" for each radiomics feature. In contrast, we propose "tensor radiomics" (TR) where tensors of features calculated using multiple parameter combinations (i.e., flavours) are utilized to optimize the creation of radiomics signatures. We provide illustrative instances of TR implementation in positron emission tomography-computed tomography (PET-CT), magnetic resonance imaging (MRI), and CT by leveraging machine learning (ML) and deep learning (DL) methodologies, as well as reproducibility analyses: (I) to predict overall survival (OS) in lung cancer (CT) and head and neck cancer (PET-CT), TR was employed by varying bin sizes. This approach involved use of a hybrid deep neural network called 'TR-Net' and two ML-based techniques for combining different flavours. (II) TR was constructed by incorporating different segmentation perturbations and various bin sizes to classify the response of late-stage lung cancer to first-line immunotherapy using CT images. (III) In MRI of glioblastoma (GBM), TR was implemented to generate multi-flavour radiomics features, enabling enhanced analysis and interpretation. (IV) TR was employed via multiple PET-CT fusions in head and neck cancer. Flavours based on different fusions were created using Laplacian pyramids and wavelet transforms. Our findings demonstrated that TR outperformed conventional radiomics features in lung cancer CT and head and neck cancer PET-CT images, significantly enhancing OS prediction accuracy. TR also improved classification of lung cancer response to therapy and exhibited notable advantages in reproducibility compared to single-flavour features in MR imaging of GBM. Moreover, in head and neck cancer, TR through multiple PET-CT fusions exhibited improved performance in predicting OS. We conclude that the proposed TR paradigm has significant potential to improve performance in different medical imaging tasks. By incorporating multiple flavours of radiomics features, TR overcomes limitations associated with individual features and shows promise in enhancing prognostic capabilities in clinical settings.

Multiparametric MRI in Era of Artificial Intelligence for Bladder Cancer Therapies.

This review focuses on the principles, applications, and performance of mpMRI for bladder imaging. Quantitative imaging biomarkers (QIBs) derived from mpMRI are increasingly used in oncological applications, including tumor staging, prognosis, and assessment of treatment response. To standardize mpMRI acquisition and interpretation, an expert panel developed the Vesical Imaging-Reporting and Data System (VI-RADS). Many studies confirm the standardization and high degree of inter-reader agreement to discriminate muscle invasiveness in bladder cancer, supporting VI-RADS implementation in routine clinical practice. The standard MRI sequences for VI-RADS scoring are anatomical imaging, including T2w images, and physiological imaging with diffusion-weighted MRI (DW-MRI) and dynamic contrast-enhanced MRI (DCE-MRI). Physiological QIBs derived from analysis of DW- and DCE-MRI data and radiomic image features extracted from mpMRI images play an important role in bladder cancer. The current development of AI tools for analyzing mpMRI data and their potential impact on bladder imaging are surveyed. AI architectures are often implemented based on convolutional neural networks (CNNs), focusing on narrow/specific tasks. The application of AI can substantially impact bladder imaging clinical workflows; for example, manual tumor segmentation, which demands high time commitment and has inter-reader variability, can be replaced by an autosegmentation tool. The use of mpMRI and AI is projected to drive the field toward the personalized management of bladder cancer patients.

Abstract 351: Novel Quantitative Imaging Biomarkers From Acute Brain CT Using Artificial Intelligence for Predicting Neurological Prognosis in Cardiac Arrest Patients: A Pilot Study

Background: Effective use of information extracted from medical images has attracted attention for decision support in primary care of patients with cardiac arrest (CA). Hypothesis and Aims: This study aimed to predict the neurological prognosis of patients with cardiac arrest by quantitative imaging biomarkers (QIBs) extracted from acute CT brain images before admission using artificial intelligence. Methods: Eighty-six CA patients (good prognosis; Glasgow-Pittsburgh cerebral performance categories [CPC]=1 or 2 at 90 days after the CA : 32 patients, poor prognosis; CPC=3, 4, or 5 at 90 days after the CA : 54 patients) treated at three hospitals between 2017 and 2019 were retrospectively analyzed. None of the CA patients caused by trauma or brain damage were included in the database. The data were divided into training and test data, and the training data were used to select informative QIBs for classification between the good and poor neurological prognosis in this analysis. Results: Seven CA patients’ data were excluded from the study due to severe metal and bone artifacts broadly spreading on the brain image, and 1170 QIBs were extracted from whole brain and regional brain interest volumes on CT images of each patient. The QIBs were composed of first-order histogram and texture features and informative QIBs for the classification were selected by feature selection algorithms using the training dataset. The feature corresponding to the gray level having a maximum gradient in the image histogram from the brain region of interest on CT images showed the strongest significant difference between the good and poor neurological prognosis of the test data set with p=0.009 and AUC=0.775 (95% CI; 0.590-0.960). Conclusions: We have shown that QIB can be used to predict neurological prognosis in patients with CA. Because this is a pilot study for methodological development and there are limitations in sample size and outcomes, new studies on a larger number of patients are required.

Amide proton transfer-weighted imagingand derived radiomics in the classification of adult-type diffuse gliomas.

To evaluate the utility of amide proton transfer-weighted (APTw) MRI imaging and its derived radiomics in classifying adult-type diffuse glioma. In this prospective study, APTw imaging was performed on 129 patients with adult-type diffuse gliomas. The mean APTw-related metrics (chemical exchange saturation transfer ratio (CESTR), CESTR normalized with the reference value (CESTRnr), and relaxation-compensated inverse magnetization transfer ratio (MTRRex)) and radiomic features within 3D tumor masks were extracted. APTw-radiomics models were developed using a support vector machine (SVM) classifier. Sensitivity analysis with tumor area of interest, different histogram cutoff values, and other classifiers were conducted. CESTR, CESTRnr, and MTRRex in glioblastomas were all significantly higher (p < 0.0003) than those of oligodendrogliomas and astrocytomas, with no significant difference between oligodendrogliomas and astrocytomas. The APTw-related metrics for IDH-wildtype and high-grade gliomas were significantly higher (p < 0.001) than those for the IDH-mutant and low-grade gliomas, with area under the curve (AUCs) of 0.88 for CESTR. The CESTR-radiomics models demonstrated accuracies of 84% (AUC 0.87), 83% (AUC 0.83), 90% (AUC 0.95), and 84% (AUC 0.86) in predicting the IDH mutation status, differentiating glioblastomas from astrocytomas, distinguishing glioblastomas from oligodendrogliomas, and determining high/low grade prediction, respectively, but showed poor performance in distinguishing oligodendrogliomas from astrocytomas (accuracy 63%, AUC 0.63). The sensitivity analysis affirmed the robustness of the APTw signal and APTw-derived radiomics prediction models. APTw imaging, along with its derived radiomics, presents a promising quantitative approach for prediction IDH mutation and grading adult-type diffuse glioma. Amide proton transfer-weighted imaging, a quantitative imaging biomarker, coupled with its derived radiomics, offers a promising non-invasive approach for predicting IDH mutation status and grading adult-type diffuse gliomas, thereby informing individualized clinical diagnostics and treatment strategies. • This study evaluates the differences of different amide proton transfer-weighted metrics across three molecular subtypes and their efficacy in classifying adult-type diffuse glioma. • Chemical exchange saturation transfer ratio normalized with the reference value and relaxation-compensated inverse magnetization transfer ratio effectively predicts IDH mutation/grading, notably the first one. • Amide proton transfer-weighted imaging and its derived radiomics holds potential to be used as a diagnostic tool in routine clinical characterizing adult-type diffuse glioma.

Pre-Brachytherapy Diffusion-Weighted Magnetic Resonance Imaging (DW-MRI) Response as Predictor of Local Control in the Definitive Treatment of Cervical Cancer

Chemoradiation consisting of external beam radiotherapy (EBRT) followed by brachytherapy (BT) is the standard of care for Stage IB2-IVA cervical cancer (CC). Multi-parametric MRI (mpMRI) is a valuable tool for initial staging, treatment planning and response assessment. In this study, we aim to explore the potential of mpMRI, in particular DW-MRI, to predict overall tumor control following chemoradiation in CC. We identified 78 consecutive patients treated with chemoradiation for FIGO IB2-IVA CC between 2012-2020, who had an MRI at baseline (MRb) and post-EBRT prior to brachytherapy boost (MRpb) that included DW-MRI (b-value = 0 and 800 s/mm2). Median age was 53 years, most patients had squamous cell carcinoma (76.9%) and advanced-stage disease (56.4% stage IIIA-IVA). Median time from completion of EBRT to MRpb was 3 days. Regions of interest (ROI) in primary tumor were delineated on DW image (b = 0 s/mm2) using ITK-SNAP software. DW-MRI data were fitted to a monoexponential model to calculate apparent diffusion coefficient (ADC) values using in-house software platform (MRI-QAMPER). MRpb mean ADC values (n = 78) and relative changes (%) in mean ADC values between MRb and MRpb (n = 64) were correlated with outcomes, including local failure (LF), regional or distant failure (RDF), and failure at any site (FAS), with death without failure as a competing risk. Median follow-up time was 45 months (95% CI 38, 53). At first post treatment assessment, 72 patients (92.3%) had a complete response (CR) in the cervix and 68 patients (87.2%) had CR in all disease sites. Of patients who had CR in the cervix (n = 72), only 1 patient had local recurrence. Of patients who had CR in all disease sites, 10 later recurred (1 LF only, 1 LF&RDF, 8 RDF only). Overall, 7 patients (9%) had LF, and 19 patients (24.4%) had FAS. A higher mean ADC value in MRpb was associated with LF (HR 4.3, 95% CI 1.32, 14.6; P = 0.016), but not with RDF (P = 0.4) or FAS (P = 0.5). A higher relative change in the mean ADC value between MRb and MRpb was associated with a lower risk of LF (HR 0.94, 95% CI 0.90, 0.98; P = 0.002), but not with RDF (P = 0.8) or FAS (P = 0.4). Treatment response as measured on prebrachytherapy DW-MRI is a significant predictor of local control in patients undergoing chemoradiation for stage IB2-IVA CC. ADC values, a quantitative imaging biomarker on MRpb may be instrumental in dose intensification/de-escalation efforts in CC.

Cardiac MRI T1 and T2 Mapping as a Quantitative Imaging Biomarker in Transthyretin Amyloid Cardiomyopathy

This study aimed to assess the utility of cardiac magnetic resonance imaging (MRI) T1 and T2 mapping as quantitative imaging biomarkers in transthyretin amyloid cardiomyopathy (ATTR-CM). This study retrospectively evaluated 74 patients with confirmed wild-type ATTR-CM who underwent cardiac MRI, 99mTc-labeled pyrophosphate (99mTc-PYP) scintigraphy, and echocardiography. We assessed the quantitative disease parameters, for example, left ventricular ejection fraction (LVEF), and global longitudinal strain (GLS) by echocardiography, native T1, extracellular volume fraction (ECV), and native T2 value by cardiac MRI, heart-to-contralateral ratio (H/CL) by 99mTc-PYP, and high-sensitive cardiac troponin T. Myocardial native T2 of ≥50ms was defined as myocardial edema. Correlations between the disease's quantitative parameters were evaluated, and the ECV was compared to other parameters in ATTR-CM with/without myocardial edema. ECV in all patients with ATTR-CM revealed a strong correlation with native T1 (r=0.62), a moderate correlation with hs-TnT (r=0.59), LVEF (r=-0.48), GLS (r=0.58), and H/CL (r=0.48). Correlations between ECV and other quantitative parameters decreased in ATTR-CM with myocardial edema except for H/CL. Meanwhile, the correlations increased in ATTR-CM without myocardial edema. The presence of myocardial edema affected the interpretation of ECV assessment, although ECV can be a comprehensive imaging biomarker for ATTR-CM. ECV showed a significant correlation with various quantitative disease parameters and can be a reliable disease monitoring marker in patients with ATTR-CM when myocardial edema was excluded.

Evaluation of Non-Compressive Transvaginal Cervical Elastography as a Quantitative Imaging Biomarker in Pregnancy: A Repeatability and Reliability Analysis.

Non-compressive strain elastography has been proposed as a novel quantitative imaging biomarker for assessing the structure and function of the cervix. The current study aims to assess the repeatability, and intra- and inter-observer reliability of transvaginal non-compressive cervical strain elastography in a clinical setting. We conducted a dual-phase single-center prospective feasibility study of singleton gestations >16-weeks gestation that required a clinically-indicated transvaginal ultrasound. Each study participant, n = 43 in phase 1 and n = 13 in phase 2, had elastography performed by two trained observers that each performed multiple image acquisitions. We performed a multivariable regression to adjust for changes in clinical characteristics between study phases and calculated the repeatability coefficients, limits of agreement, and intraclass correlations for each quantitative elastography parameter. We compared quantitative elastography parameters to cervical length measurements, acquired from the same images. The repeatability coefficients and percent limits of agreement were wide for all of the quantitative elastography parameters, demonstrating poor repeatability. Intraclass correlation coefficients were poor-moderate for both intra-observer (0.31-0.77) and inter-observer reliability (0.35-0.77) in both study phases, while cervical length showed excellent reliability with intraclass correlations consistently >0.90. Non-compressive transvaginal strain cervical elastography did not demonstrate adequate repeatability or reliability. Our results highlight the importance of rigorously assessing novel quantitative imaging biomarkers before clinical application.

A Radiomic-Based Machine Learning System to Diagnose Age-Related Macular Degeneration from Ultra-Widefield Fundus Retinography.

The present study was conducted to investigate the potential of radiomics to develop an explainable AI-based system to be applied to ultra-widefield fundus retinographies (UWF-FRTs) with the objective of predicting the presence of the early signs of Age-related Macular Degeneration (AMD) and stratifying subjects with low- versus high-risk of AMD. The ultimate aim was to provide clinicians with an automatic classifier and a signature of objective quantitative image biomarkers of AMD. The use of Machine Learning (ML) and radiomics was based on intensity and texture analysis in the macular region, detected by a Deep Learning (DL)-based macular detector. Two-hundred and twenty six UWF-FRTs were retrospectively collected from two centres and manually annotated to train and test the algorithms. Notably, the combination of the ML-based radiomics model and the DL-based macular detector reported 93% sensitivity and 74% specificity when applied to the data of the centre used for external testing, capturing explainable features associated with drusen or pigmentary abnormalities. In comparison to the human operator's annotations, the system yielded a 0.79 Cohen κ, demonstrating substantial concordance. To our knowledge, these results are the first provided by a radiomic approach for AMD supporting the suitability of an explainable feature extraction method combined with ML for UWF-FRT.

Current Imaging Strategies in Patients with Abdominal Aortic Aneurysms.

An abdominal aortic aneurysm (AAA) is defined as a localized dilatation of the abdominal aorta of ≥ 3 cm. With a prevalence of 4-8 %, AAA is one of the most common vascular diseases in Western society. Radiological imaging is an elementary component in the diagnosis, monitoring, and treatment planning of AAA patients. This is a narrative review article on preoperative imaging strategies of AAA, incorporating expert opinions based on the current literature and standard-of-care practices from our own center. Examples are provided to illustrate clinical cases from our institution. Radiological imaging plays a pivotal role in the initial diagnosis and monitoring of patients with AAA. Ultrasound is the mainstay imaging modality for AAA screening and surveillance. Contrast-enhanced CT angiography is currently considered the gold standard for preoperative imaging and image-based treatment planning in AAA repair. New non-contrast MR angiography techniques are robustly applicable and allow precise determination of aortic diameters, which is of critical importance, particularly with regard to current diameter-based surgical treatment guidelines. 3D imaging with multiplanar reformation and automatic centerline positioning enables more accurate assessment of the maximum aortic diameter. Modern imaging techniques such as 4D flow MRI have the potential to further improve individualized risk stratification in patients with AAA. · Ultrasound is the mainstay imaging modality for AAA screening and monitoring. · Contrast-enhanced CT angiography is the gold standard for preoperative imaging in AAA repair. · Non-contrast MR angiography allows for accurate monitoring of aortic diameters in AAA patients. · Measurement of aortic diameters is more accurate with 3D-CT/MRI compared to ultrasound. · Research seeks new quantitative imaging biomarkers for AAA risk stratification, e. g., using 4D flow MRI.

P13.06.A RADIOMIC FEATURES FOR RISK-STRATIFICATION IN PATIENTS WITH BRAIN METASTASES

Abstract BACKGROUND Recently, non-invasive characterization of brain tumors on MRI has emerged as a promising field of research. The identification of quantitative imaging biomarkers, also known as radiomics, may complement molecular characterization and thereby improve clinical management of neuro-oncologic patients. In this study we aimed to identify imaging predictors with improved performance over clinical parameters in order to stratify patients with brain metastases into high and low risk groups for overall survival (OS). MATERIAL AND METHODS 422 patients (allocated in a 3:1 ratio to a discovery [n=317] and test [n= 105] set) with first diagnosis of brain metastases from different primary tumors from two neuro-oncologic centers were included. In each patient, eight clinical features (age, gender, KPS, systemic disease status, presence of extracranial metastases, number of cerebral metastases, primary tumor and available individual prognostic molecular status eg HER2, BRAF,⋯) were gathered and a total of 321 radiomic MRI features (including shape, first-order and higher-order features) from cerebral MRI (contrast T1-weighted and apparent diffusion coefficient maps) were extracted. Radiomic and clinical features of patients in the discovery set were subjected to different machine learning models in order to classify patients into low- and high-risk groups for OS. By performing a comprehensive grid search the best machine learning model according to the macro-evaluated score and accuracy was identified and evaluated on the testing set. In addition, a subgroup analysis, based on the primary tumor entity was done and confusion matrices were calculated in order to evaluate final predictions. RESULTS With an extra trees classifier including all clinical and 30 radiomic features, we were able to stratify patients into a high- and low-risk group for OS (test set: macro-evaluated = 0.60, accuracy = 0.67). The best performing model was a gradient boosting model only including clinical features (test set: macro-evaluated = 0.62, accuracy = 0.72), while the radiomic features alone led to the poorest results (test set: macro-evaluated = 0.52, accuracy = 0.63). Interestingly, in the subgroup of melanoma patients, the predictive power of radiomic features outperformed the clinical and the combined (radiomics and clinical) features. With a prediction solely based on radiomic features, 80% and 67% of patients were correctly classified into the low- and the high-risk group for OS, respectively. CONCLUSION In conclusion, we found that in the entire study population of patients with brain metastases radiomic features did not allow for a better prediction of the clinical outcome compared to the clinical parameters alone. However, in the subgroup of melanoma patients the predictive power of radiomic features alone was superior compared to clinical features or all features combined.

Deep learning reconstruction for brain diffusion-weighted imaging: efficacy for image quality improvement, apparent diffusion coefficient assessment, and intravoxel incoherent motion evaluation in in vitro and in vivo studies

Deep learning reconstruction (DLR) to improve imaging quality has already been introduced, but no studies have evaluated the effect of DLR on diffusion-weighted imaging (DWI) or intravoxel incoherent motion (IVIM) in in vitro or in vivo studies. The purpose of this study was to determine the effect of DLR for magnetic resonance imaging (MRI) in terms of image quality improvement, apparent diffusion coefficient (ADC) assessment, and IVIM index evaluation on DWI through in vitro and in vivo studies. For the in vitro study, a phantom recommended by the Quantitative Imaging Biomarkers Alliance was scanned and reconstructed with and without DLR, and 15 patients with brain tumors with normal-appearing gray and white matter examined using IVIM and reconstructed with and without DLR were included in the in vivo study. The ADCs of all phantoms for DWI with and without DLR, as well as the coefficient of variation percentage (CV%), and ADCs and IVIM indexes for each participant, were evaluated based on DWI with and without DLR by means of region-of-interest measurements. For the in vitro study, using the mean ADCs for all phantoms, a t-test was adopted to compare DWI with and without DLR. For the in vivo study, a Wilcoxon signed-rank test was used to compare the CV% between the two types of DWI. In addition, the Wilcoxon signed-rank test was used to compare the ADC, true diffusion coefficient (D), pseudodiffusion coefficient (D*), and percentage of water molecules in micro perfusion within 1 voxel (f) with and without DLR; the limits of agreement of each parameter were determined through a Bland-Altman analysis. The in vitro study identified no significant differences between the ADC values for DWI with and without DLR (P > 0.05), and the CV% was significantly different for DWI with and without DLR (P < 0.05) when b values ≥250 s/mm2 were used. The in vivo study revealed that D* and f with and without DLR were significantly different (P < 0.001). The limits of agreement of the ADC, D, and D* values for DWI with and without DLR were determined as 0.00 ± 0.51 × 10-3, 0.00 ± 0.06 × 10-3, and 1.13 ± 4.04 × 10-3 mm2/s, respectively. The limits of agreement of the f values for DWI with and without DLR were determined as -0.01 ± 0.07. Deep learning reconstruction for MRI has the potential to significantly improve DWI quality at higher b values. It has some effect on D* and f values in the IVIM index evaluation, but ADC and D values are less affected by DLR.

Development and benchmarking diffusion magnetic resonance imaging analysis for integration into radiation treatment planning.

The rising promise in the utility of advanced multi-parametric magnetic resonance (MR) imaging in radiotherapy (RT) treatment planning creates a necessity for testing and enhancing the accuracy of quantitative imaging analysis. Standardizing the analysis of diffusion weighted imaging (DWI) and diffusion tensor imaging (DTI) to generate meaningful and reproducible apparent diffusion coefficient (ADC) and fractional anisotropy (FA) lays the requisite needed for clinical integration. The aim of the demonstrated work is to benchmark the generation of the ADC and FA parametric map analyses using integrated tools in a commercial treatment planning system against the currently used ones. Three software packages were used for generating ADC and FA maps in this study; one tool was developed within a commercial treatment planning system, another by the Functional Magnetic Resonance Imaging of the Brain (FMRIB) Software Library FSL (Analysis Group, FMRIB, Oxford, United Kingdom), and an in-house tool developed at the M.D. Anderson Cancer Center. The ADC and FA maps generated by all three packages for 35 subjects were subtracted from one another, and the standard deviation of the images' differences was used to compare the reproducibility. The reproducibility of the ADC maps was compared with the Quantitative Imaging Biomarkers Alliance (QIBA) protocol, while that of the FA maps was compared to data in published literature. Results show that the discrepancies between the ADC maps calculated for each patient using the three different software algorithms are less than 2% which meets the 3.6% recommended QIBA requirement. Except for a small number of isolated points, the majority of differences in FA maps for each patient produced by the three methods did not exceed 0.02 which is 10 times lower than the differences seen in healthy gray and white matter. The results were also compared to the maps generated by existing MR Imaging consoles and showed that the robustness of console generated ADC and FA maps is largely dependent on the correct application of scaling factors, that only if correctly placed; the differences between the three tested methods and the console generated values were within the recommended QIBA guidelines. Cross-comparison difference maps demonstrated that quantitative reproducibility of ADC and FA metrics generated using our tested commercial treatment planning system were comparable to in-house and established tools as benchmarks. This integrated approach facilitates the clinical utility of diffusion imaging in radiation treatment planning workflow.

Emerging role of quantitative imaging (radiomics) and artificial intelligence in precision oncology.

Cancer is a fatal disease and the second most cause of death worldwide. Treatment of cancer is a complex process and requires a multi-modality-based approach. Cancer detection and treatment starts with screening/diagnosis and continues till the patient is alive. Screening/diagnosis of the disease is the beginning of cancer management and continued with the staging of the disease, planning and delivery of treatment, treatment monitoring, and ongoing monitoring and follow-up. Imaging plays an important role in all stages of cancer management. Conventional oncology practice considers that all patients are similar in a disease type, whereas biomarkers subgroup the patients in a disease type which leads to the development of precision oncology. The utilization of the radiomic process has facilitated the advancement of diverse imaging biomarkers that find application in precision oncology. The role of imaging biomarkers and artificial intelligence (AI) in oncology has been investigated by many researchers in the past. The existing literature is suggestive of the increasing role of imaging biomarkers and AI in oncology. However, the stability of radiomic features has also been questioned. The radiomic community has recognized that the instability of radiomic features poses a danger to the global generalization of radiomic-based prediction models. In order to establish radiomic-based imaging biomarkers in oncology, the robustness of radiomic features needs to be established on a priority basis. This is because radiomic models developed in one institution frequently perform poorly in other institutions, most likely due to radiomic feature instability. To generalize radiomic-based prediction models in oncology, a number of initiatives, including Quantitative Imaging Network (QIN), Quantitative Imaging Biomarkers Alliance (QIBA), and Image Biomarker Standardisation Initiative (IBSI), have been launched to stabilize the radiomic features.

Quantitative imaging biomarkers of coronary plaque morphology: insights from EVAPORATE.

Residual cardiovascular risk persists despite statin therapy. In REDUCE-IT, icosapent ethyl (IPE) reduced total events, but the mechanisms of benefit are not fully understood. EVAPORATE evaluated the effects of IPE on plaque characteristics by coronary computed tomography angiography (CCTA). Given the conclusion that the IPE-treated patients demonstrate that plaque burden decreases has already been published in the primary study analysis, we aimed to demonstrate whether the use of an analytic technique defined and validated in histological terms could extend the primary study in terms of whether such changes could be reliably seen in less time on drug, at the individual (rather than only at the cohort) level, or both, as neither of these were established by the primary study result. EVAPORATE randomized the patients to IPE 4 g/day or placebo. Plaque morphology, including lipid-rich necrotic core (LRNC), fibrous cap thickness, and intraplaque hemorrhage (IPH), was assessed using the ElucidVivo® (Elucid Bioimaging Inc.) on CCTA. The changes in plaque morphology between the treatment groups were analyzed. A neural network to predict treatment assignment was used to infer patient representation that encodes significant morphological changes. Fifty-five patients completed the 18-month visit in EVAPORATE with interpretable images at each of the three time points. The decrease of LRNC between the patients on IPE vs. placebo at 9 months (reduction of 2 mm3 vs. an increase of 41 mm3, p = 0.008), widening at 18 months (6 mm3 vs. 58 mm3 increase, p = 0.015) were observed. While not statistically significant on a univariable basis, reductions in wall thickness and increases in cap thickness motivated multivariable modeling on an individual patient basis. The per-patient response assessment was possible using a multivariable model of lipid-rich phenotype at the 9-month follow-up, p < 0.01 (sustained at 18 months), generalizing well to a validation cohort. Plaques in the IPE-treated patients acquired more characteristics of stability. Reliable assessment using histologically validated analysis of individual response is possible at 9 months, with sustained stabilization at 18 months, providing a quantitative basis to elucidate drug mechanism and assess individual patient response.