Differentiation of benign from malignant solid renal lesions using CT-based radiomics and machine learning: comparison with radiologist interpretation.

PURPOSE: To assess the performance of a machine learning model trained with contrast-enhanced CT-based radiomics features in distinguishing benign from malignant solid renal masses and to compare model performance with three abdominal radiologists. METHODS: Patients who underwent intra-operative ultrasound during a partial nephrectomy were identified within our institutional database, and those who had pre-operative contrast-enhanced CT examinations were selected. The renal masses were segmented from the CT images and radiomics features were derived from the segmentations. The pathology of each mass was identified; masses were labeled as either benign [oncocytoma or angiomyolipoma (AML)] or malignant [clear cell, papillary, or chromophobe renal cell carcinoma (RCC)] depending on the pathology. The data were parsed into a 70/30 train/test split and a random forest machine learning model was developed to distinguish benign from malignant lesions. Three radiologists assessed the cohort of masses and labeled cases as benign or malignant. RESULTS: 148 masses were identified from the cohort, including 50 benign lesions (23 AMLs, 27 oncocytomas) and 98 malignant lesions (23 clear cell RCC, 44 papillary RCC, and 31 chromophobe RCCs). The machine learning algorithm yielded an overall accuracy of 0.82 for distinguishing benign from malignant lesions, with an area under the receiver operating curve of 0.80. In comparison, the three radiologists had significantly lower accuracies (p = 0.02) ranging from 0.67 to 0.75. CONCLUSION: A machine learning model trained with CT-based radiomics features can provide superior accuracy for distinguishing benign from malignant solid renal masses compared to abdominal radiologists.

Maternal malnutrition and anaemia in India: dysregulations leading to the 'thin-fat' phenotype in newborns.

Maternal and child malnutrition and anaemia remain the leading factors for health loss in India. Low birth weight (LBW) offspring of women suffering from chronic malnutrition and anaemia often exhibit insulin resistance and infantile stunting and wasting, together with increased risk of developing cardiometabolic disorders in adulthood. The resulting self-perpetuating and highly multifactorial disease burden cannot be remedied through uniform dietary recommendations alone. To inform approaches likely to alleviate this disease burden, we implemented a systems-analytical approach that had already proven its efficacy in multiple published studies. We utilised previously published qualitative and quantitative analytical results of rural and urban field studies addressing maternal and infantile metabolic and nutritional parameters to precisely define the range of pathological phenotypes encountered and their individual biological characteristics. These characteristics were then integrated, via extensive literature searches, into metabolic and physiological mechanisms to identify the maternal and foetal metabolic dysregulations most likely to underpin the 'thin-fat' phenotype in LBW infants and its associated pathological consequences. Our analyses reveal hitherto poorly understood maternal nutrition-dependent mechanisms most likely to promote and sustain the self-perpetuating high disease burden, especially in the Indian population. This work suggests that it most probably is the metabolic consequence of 'ill-nutrition' - the recent and rapid dietary shifts to high salt, high saturated fats and high sugar but low micronutrient diets - over an adaptation to 'thrifty metabolism' which must be addressed in interventions aiming to significantly alleviate the leading risk factors for health deterioration in India.

Magnetic particle imaging of islet transplantation in the liver and under the kidney capsule in mouse models.

BACKGROUND: Islet transplantation (Tx) represents the most promising therapy to restore normoglycemia in type 1 diabetes (T1D) patients to date. As significant islet loss has been observed after the procedure, there is an urgent need for developing strategies for monitoring transplanted islet grafts. In this report we describe for the first time the application of magnetic particle imaging (MPI) for monitoring transplanted islets in the liver and under the kidney capsule in experimental animals. METHODS: Pancreatic islets isolated from Papio hamadryas were labeled with superparamagnetic iron oxides (SPIOs) and used for either islet phantoms or Tx in the liver or under the kidney capsule of NOD scid mice. MPI was used to image and quantify islet phantoms and islet transplanted experimental animals post-mortem at 1 and 14 days after Tx. Magnetic resonance imaging (MRI) was used to confirm the presence of labeled islets in the liver and under the kidney capsule 1 day after Tx. RESULTS: MPI of labeled islet phantoms confirmed linear correlation between the number of islets and the MPI signal (R2=0.988). Post-mortem MPI performed on day 1 after Tx showed high signal contrast in the liver and under the kidney capsule. Quantitation of the signal supports islet loss over time, which is normally observed 2 weeks after Tx. No MPI signal was observed in control animals. In vivo MRI confirmed the presence of labeled islets/islet clusters in liver parenchyma and under the kidney capsule one day after Tx. CONCLUSIONS: Here we demonstrate that MPI can be used for quantitative detection of labeled pancreatic islets in the liver and under the kidney capsule of experimental animals. We believe that MPI, a modality with no depth attenuation and zero background tissue signal could be a suitable method for imaging transplanted islet grafts.

T1ρ and T2 -based characterization of regional variations in intervertebral discs to detect early degenerative changes.

Lower back pain is one of the main contributors to morbidity and chronic disability in the United States. Despite the significance of the problem, it is still not well understood. There is a clear need for objective, non-invasive biomarkers to localize specific pain generators and identify early stage changes to enable reliable diagnosis and treatment. In this study we focus on intervertebral disc degeneration as a source of lower back pain. Quantitative imaging markers T1ρ and T2 have been shown to be promising techniques for in vivo diagnosis of biochemical degeneration in discs due to their sensitivity to macromolecular changes in proteoglycan content and collagen integrity. We describe a semi-automated technique for quantifying T1ρ and T2 relaxation time maps in the nucleus pulposus (NP) and the annulus fibrosus (AF) of the lumbar intervertebral discs. Compositional changes within the NP and AF associated with degeneration occur much earlier than the visually observable structural changes. The proposed technique rigorously quantifies these biochemical changes taking into account subtle regional variations to allow interpretation of early degenerative changes that are difficult to interpret with traditional MRI techniques and clinical subjective grading scores. T1ρ and T2 relaxation times in the NP decrease with degenerative severity in the disc. Moreover, standard deviation and texture measurements of these values show sharper and more significant changes during early degeneration compared to later degenerative stages. Our results suggest that future prospective studies should include automated T1ρ and T2 metrics as early biomarkers for disc degeneration-induced lower back pain. © 2016 Orthopaedic Research Society. Published by Wiley Periodicals, Inc. J Orthop Res 34:1373-1381, 2016.

Accelerated T1ρ acquisition for knee cartilage quantification using compressed sensing and data-driven parallel imaging: A feasibility study.

PURPOSE: Quantitative T1ρ imaging is beneficial for early detection for osteoarthritis but has seen limited clinical use due to long scan times. In this study, we evaluated the feasibility of accelerated T1ρ mapping for knee cartilage quantification using a combination of compressed sensing (CS) and data-driven parallel imaging (ARC-Autocalibrating Reconstruction for Cartesian sampling). METHODS: A sequential combination of ARC and CS, both during data acquisition and reconstruction, was used to accelerate the acquisition of T1ρ maps. Phantom, ex vivo (porcine knee), and in vivo (human knee) imaging was performed on a GE 3T MR750 scanner. T1ρ quantification after CS-accelerated acquisition was compared with non CS-accelerated acquisition for various cartilage compartments. RESULTS: Accelerating image acquisition using CS did not introduce major deviations in quantification. The coefficient of variation for the root mean squared error increased with increasing acceleration, but for in vivo measurements, it stayed under 5% for a net acceleration factor up to 2, where the acquisition was 25% faster than the reference (only ARC). CONCLUSION: To the best of our knowledge, this is the first implementation of CS for in vivo T1ρ quantification. These early results show that this technique holds great promise in making quantitative imaging techniques more accessible for clinical applications.

Accelerating T1ρ cartilage imaging using compressed sensing with iterative locally adapted support detection and JSENSE.

PURPOSE: To accelerate T1ρ quantification in cartilage imaging using combined compressed sensing with iterative locally adaptive support detection and JSENSE. METHODS: To reconstruct T1ρ images from accelerated acquisition at different time of spin-lock (TSLs), we propose an approach to combine an advanced compressed sensing (CS) based reconstruction technique, LAISD (locally adaptive iterative support detection), and an advanced parallel imaging technique, JSENSE. Specifically, the reconstruction process alternates iteratively among local support detection in the domain of principal component analysis, compressed sensing reconstruction of the image sequence, and sensitivity estimation with JSENSE. T1ρ quantification results from accelerated scans using the proposed method are evaluated using in vivo knee cartilage data from bilateral scans of three healthy volunteers. RESULTS: T1ρ maps obtained from accelerated scans (acceleration factors of 3 and 3.5) using the proposed method showed results comparable to conventional full scans. The T1ρ errors in all compartments are below 1%, which is well below the in vivo reproducibility of cartilage T1ρ reported from previous studies. CONCLUSION: The proposed method can significantly accelerate the acquisition process of T1ρ quantification on human cartilage imaging without sacrificing accuracy, which will greatly facilitate the clinical translation of quantitative cartilage MRI.

Caspase-responsive smart gadolinium-based contrast agent for magnetic resonance imaging of drug-induced apoptosis.

Non-invasive detection of caspase-3/7 activity in vivo has provided invaluable predictive information regarding tumor therapeutic efficacy and anti-tumor drug selection. Although a number of caspase-3/7 targeted fluorescence and positron emission tomography (PET) imaging probes have been developed, there is still a lack of gadolinium (Gd)-based magnetic resonance imaging (MRI) probes that enable high spatial resolution detection of caspase-3/7 activity in vivo. Here we employ a self-assembly approach and develop a caspase-3/7 activatable Gd-based MRI probe for monitoring tumor apoptosis in mice. Upon reduction and caspase-3/7 activation, the caspase-sensitive nano-aggregation MR probe (C-SNAM: 1) undergoes biocompatible intramolecular cyclization and subsequent self-assembly into Gd-nanoparticles (GdNPs). This results in enhanced r1 relaxivity-19.0 (post-activation) vs. 10.2 mM-1 s-1 (pre-activation) at 1 T in solution-and prolonged accumulation in chemotherapy-induced apoptotic cells and tumors that express active caspase-3/7. We demonstrate that C-SNAM reports caspase-3/7 activity by generating a significantly brighter T1-weighted MR signal compared to non-treated tumors following intravenous administration of C-SNAM, providing great potential for high-resolution imaging of tumor apoptosis in vivo.

Redox-triggered self-assembly of gadolinium-based MRI probes for sensing reducing environment.

Controlled self-assembly of small molecule gadolinium (Gd) complexes into nanoparticles (GdNPs) is emerging as an effective approach to design activatable magnetic resonance imaging (MRI) probes and amplify the r1 relaxivity. Herein, we employ a reduction-controlled macrocyclization reaction and self-assembly to develop a redox activated Gd-based MRI probe for sensing a reducing environment. Upon disulfide reduction at physiological conditions, an acyclic contrast agent 1 containing dual Gd-chelates undergoes intramolecular macrocyclization to form rigid and hydrophobic macrocycles, which subsequently self-assemble into GdNPs, resulting in a ∼60% increase in r1 relaxivity at 0.5 T. Probe 1 has high r1 relaxivity (up to 34.2 mM(-1) s(-1) per molecule at 0.5 T) upon activation, and also shows a high sensitivity and specificity for MR detection of thiol-containing biomolecules.

The utility of micro-CT and MRI in the assessment of longitudinal growth of liver metastases in a preclinical model of colon carcinoma.

RATIONALE AND OBJECTIVES: Liver is a common site for distal metastases in colon and rectal cancer. Numerous clinical studies have analyzed the relative merits of different imaging modalities for detection of liver metastases. Several exciting new therapies are being investigated in preclinical models. But, technical challenges in preclinical imaging make it difficult to translate conclusions from clinical studies to the preclinical environment. This study addresses the technical challenges of preclinical magnetic resonance imaging (MRI) and micro-computed tomography (CT) to enable comparison of state-of-the-art methods for following metastatic liver disease. MATERIALS AND METHODS: We optimized two promising preclinical protocols to enable a parallel longitudinal study tracking metastatic human colon carcinoma growth in a mouse model: T2-weighted MRI using two-shot PROPELLER (Periodically Rotated Overlapping ParallEL Lines with Enhanced Reconstruction) and contrast-enhanced micro-CT using a liposomal contrast agent. Both methods were tailored for high throughput with attention to animal support and anesthesia to limit biological stress. RESULTS AND CONCLUSIONS: Each modality has its strengths. Micro-CT permitted more rapid acquisition (<10 minutes) with the highest spatial resolution (88-micron isotropic resolution). But detection of metastatic lesions requires the use of a blood pool contrast agent, which could introduce a confound in the evaluation of new therapies. MRI was slower (30 minutes) and had lower anisotropic spatial resolution. But MRI eliminates the need for a contrast agent and the contrast-to-noise between tumor and normal parenchyma was higher, making earlier detection of small lesions possible. Both methods supported a relatively high-throughput, longitudinal study of the development of metastatic lesions.

Efficient Bloch-Siegert B1 (+) mapping using spiral and echo-planar readouts.

The Bloch-Siegert (B-S) B1 (+) mapping technique is a fast, phase-based method that is highly SAR limited especially at 7T, necessitating the use of long repetition times. Spiral and echo-planar readouts were incorporated in a gradient-echo based B-S sequence to reduce specific absoprtion rate (SAR) and improve its scan efficiency. A novel, numerically optimized 4 ms B-S off-resonant pulse at + 1960 Hz was used to increase sensitivity and further reduce SAR compared with the conventional 6 ms Fermi B-S pulse. Using echo-planar and spiral readouts, scan time reductions of 8-16 were achieved. By reducing the B-S pulse width by a factor of 1.5, SAR was reduced by a factor of 1.5 and overall sensitivity was increased by a factor of 1.33 due to the nearly halved resonance offset of the new B-S pulse. This was validated on phantoms and volunteers at 7 T.

Reduction of artifacts in T2 -weighted PROPELLER in high-field preclinical imaging.

A simple technique is implemented for correction of artifacts arising from nonuniform T(2) -weighting of k-space data in fast spin echo-based PROPELLER (periodically rotated overlapping parallel lines with enhanced reconstruction). An additional blade with no phase-encoding gradients is acquired to generate the scaling factor used for the correction. Results from simulations and phantom experiments, as well as in vivo experiments in free-breathing mice, demonstrate the advantages of the proposed method. This technique is developed specifically for high-field imaging applications where T(2) decay is rapid.

Multishot PROPELLER for high-field preclinical MRI.

With the development of numerous mouse models of cancer, there is a tremendous need for an appropriate imaging technique to study the disease evolution. High-field T(2)-weighted imaging using PROPELLER (Periodically Rotated Overlapping ParallEL Lines with Enhanced Reconstruction) MRI meets this need. The two-shot PROPELLER technique presented here provides (a) high spatial resolution, (b) high contrast resolution, and (c) rapid and noninvasive imaging, which enables high-throughput, longitudinal studies in free-breathing mice. Unique data collection and reconstruction makes this method robust against motion artifacts. The two-shot modification introduced here retains more high-frequency information and provides higher signal-to-noise ratio than conventional single-shot PROPELLER, making this sequence feasible at high fields, where signal loss is rapid. Results are shown in a liver metastases model to demonstrate the utility of this technique in one of the more challenging regions of the mouse, which is the abdomen.