Repeatability and reproducibility of multiparametric magnetic resonance imaging of the liver.

As the burden of liver disease reaches epidemic levels, there is a high unmet medical need to develop robust, accurate and reproducible non-invasive methods to quantify liver tissue characteristics for use in clinical development and ultimately in clinical practice. This prospective cross-sectional study systematically examines the repeatability and reproducibility of iron-corrected T1 (cT1), T2*, and hepatic proton density fat fraction (PDFF) quantification with multiparametric MRI across different field strengths, scanner manufacturers and models. 61 adult participants with mixed liver disease aetiology and those without any history of liver disease underwent multiparametric MRI on combinations of 5 scanner models from two manufacturers (Siemens and Philips) at different field strengths (1.5T and 3T). We report high repeatability and reproducibility across different field strengths, manufacturers, and scanner models in standardized cT1 (repeatability CoV: 1.7%, bias -7.5ms, 95% LoA of -53.6 ms to 38.5 ms; reproducibility CoV 3.3%, bias 6.5 ms, 95% LoA of -76.3 to 89.2 ms) and T2* (repeatability CoV: 5.5%, bias -0.18 ms, 95% LoA -5.41 to 5.05 ms; reproducibility CoV 6.6%, bias -1.7 ms, 95% LoA -6.61 to 3.15 ms) in human measurements. PDFF repeatability (0.8%) and reproducibility (0.75%) coefficients showed high precision of this metric. Similar precision was observed in phantom measurements. Inspection of the ICC model indicated that most of the variance in cT1 could be accounted for by study participants (ICC = 0.91), with minimal contribution from technical differences. We demonstrate that multiparametric MRI is a non-invasive, repeatable and reproducible method for quantifying liver tissue characteristics across manufacturers (Philips and Siemens) and field strengths (1.5T and 3T).

Utility and variability of three non-invasive liver fibrosis imaging modalities to evaluate efficacy of GR-MD-02 in subjects with NASH and bridging fibrosis during a phase-2 randomized clinical trial.

BACKGROUND: Given the worldwide prevalence of NAFLD and NASH, there is a need to develop treatments to slow or reverse disease progression. GR-MD-02 (galactoarabino-rhamnogalaturonate) has been shown to reduce hepatic fibrosis in animal studies, and lower serum biomarkers of NASH fibrogenesis in humans. The primary aim of this study was to determine the difference between four-months of treatment with GR-MD-02 or placebo in liver inflammation and fibrosis as measured by iron-corrected T1 (cT1) mapping, a non-invasive magnetic resonance imaging (MRI) biomarker that correlates with the extent of hepatic fibro-inflammatory disease. The secondary aims were to determine change in liver stiffness as measured by magnetic resonance elastography (MRE) and shear-wave ultrasonic elastography (LSM), and to explore test-retest repeatability of the three biomarkers. MATERIALS AND METHODS: Thirty subjects (13 females, 46-71 years) with NASH and advanced fibrosis were recruited. Subjects were randomized to receive 8 mg.kg-1 GR-MD-02 (via IV infusion) or placebo, administered biweekly over a 16-week period. Therapeutic efficacy was examined using cT1, MRE, and LSM. Statistical analyses on group differences in the biomarkers were performed using robust ANCOVA models adjusting for baseline measurement and additional covariates. RESULTS: There was no significant difference in cT1 (p = 0.16) between GR-MD-02 and placebo groups following a 16-week intervention. There was also no significant difference in liver stiffness, measured by MRE (p = 0.80) or LSM (p = 0.63), between groups. Examination of repeatability of the cT1, MRE and LSM revealed coefficient of variations of 3.1%, 11% and 40% respectively. CONCLUSIONS: 8 mg.kg-1 of GR-MD-02 had no significant effect on non-invasive biomarkers of liver inflammation or fibrosis over a 4-month period. Histological confirmation was not available in this study. The high reproducibility of the primary outcome measure suggests that cT1 could be utilized for monitoring longitudinal change in patients with NASH.

Multiparametric magnetic resonance imaging for quantitation of liver disease: a two-centre cross-sectional observational study.

LiverMultiScan is an emerging diagnostic tool using multiparametric MRI to quantify liver disease. In a two-centre prospective validation study, 161 consecutive adult patients who had clinically-indicated liver biopsies underwent contemporaneous non-contrast multiparametric MRI at 3.0 tesla (proton density fat fraction (PDFF), T1 and T2* mapping), transient elastography (TE) and Enhanced Liver Fibrosis (ELF) test. Non-invasive liver tests were correlated with gold standard histothological measures. Reproducibility of LiverMultiScan was investigated in 22 healthy volunteers. Iron-corrected T1 (cT1), TE, and ELF demonstrated a positive correlation with hepatic collagen proportionate area (all p < 0·001). TE was superior to ELF and cT1 for predicting fibrosis stage. cT1 maintained good predictive accuracy for diagnosing significant fibrosis in cases with indeterminate ELF, but not for cases with indeterminate TE values. PDFF had high predictive accuracy for individual steatosis grades, with AUROCs ranging from 0.90-0.94. T2* mapping diagnosed iron accumulation with AUROC of 0.79 (95% CI: 0.67-0.92) and negative predictive value of 96%. LiverMultiScan showed excellent test/re-test reliability (coefficients of variation ranging from 1.4% to 2.8% for cT1). Overall failure rates for LiverMultiScan, ELF and TE were 4.3%, 1.9% and 15%, respectively. LiverMultiScan is an emerging point-of-care diagnostic tool that is comparable with the established non-invasive tests for assessment of liver fibrosis, whilst at the same time offering a superior technical success rate and contemporaneous measurement of liver steatosis and iron accumulation.

The anti-malarial atovaquone increases radiosensitivity by alleviating tumour hypoxia.

Tumour hypoxia renders cancer cells resistant to cancer therapy, resulting in markedly worse clinical outcomes. To find clinical candidate compounds that reduce hypoxia in tumours, we conduct a high-throughput screen for oxygen consumption rate (OCR) reduction and identify a number of drugs with this property. For this study we focus on the anti-malarial, atovaquone. Atovaquone rapidly decreases the OCR by more than 80% in a wide range of cancer cell lines at pharmacological concentrations. In addition, atovaquone eradicates hypoxia in FaDu, HCT116 and H1299 spheroids. Similarly, it reduces hypoxia in FaDu and HCT116 xenografts in nude mice, and causes a significant tumour growth delay when combined with radiation. Atovaquone is a ubiquinone analogue, and decreases the OCR by inhibiting mitochondrial complex III. We are now undertaking clinical studies to assess whether atovaquone reduces tumour hypoxia in patients, thereby increasing the efficacy of radiotherapy.

Multiparametric magnetic resonance imaging predicts clinical outcomes in patients with chronic liver disease.

BACKGROUND & AIMS: Multiparametric magnetic resonance (MR) imaging has been demonstrated to quantify hepatic fibrosis, iron, and steatosis. The aim of this study was to determine if MR can be used to predict negative clinical outcomes in liver disease patients. METHODS: Patients with chronic liver disease (n=112) were recruited for MR imaging and data on the development of liver related clinical events were collected by medical records review. The median follow-up was 27months. MR data were analysed blinded for the Liver Inflammation and Fibrosis score (LIF; <1, 1-1.99, 2-2.99, and ⩾3 representing normal, mild, moderate, and severe liver disease, respectively), T2∗ for liver iron content and proportion of liver fat. Baseline liver biopsy was performed in 102 patients. RESULTS: Liver disease aetiologies included non-alcoholic fatty liver disease (35%) and chronic viral hepatitis (30%). Histologically, fibrosis was mild in 54 (48%), moderate in 17 (15%), and severe in 31 (28%) patients. Overall mortality was 5%. Ten patients (11%) developed at least one liver related clinical event. The negative predictive value of LIF<2 was 100%. Two patients with LIF 2-2.99 and eight with LIF⩾3 had a clinical event. Patients with LIF⩾3 had a higher cumulative risk for developing clinical events, compared to those with LIF<1 (p=0.02) and LIF 1-1.99 (p=0.03). Cox regression analysis including all 3 variables (fat, iron, LIF) resulted in an enhanced LIF predictive value. CONCLUSIONS: Non-invasive standardised multiparametric MR technology may be used to predict clinical outcomes in patients with chronic liver disease.

Improved Methods to Generate Spheroid Cultures from Tumor Cells, Tumor Cells & Fibroblasts or Tumor-Fragments: Microenvironment, Microvesicles and MiRNA.

Diagnostic and prognostic indicators are key components to achieve the goal of personalized cancer therapy. Two distinct approaches to this goal include predicting response by genetic analysis and direct testing of possible therapies using cultures derived from biopsy specimens. Optimally, the latter method requires a rapid assessment, but growing xenograft tumors or developing patient-derived cell lines can involve a great deal of time and expense. Furthermore, tumor cells have much different responses when grown in 2D versus 3D tissue environments. Using a modification of existing methods, we show that it is possible to make tumor-fragment (TF) spheroids in only 2-3 days. TF spheroids appear to closely model characteristics of the original tumor and may be used to assess critical therapy-modulating features of the microenvironment such as hypoxia. A similar method allows the reproducible development of spheroids from mixed tumor cells and fibroblasts (mixed-cell spheroids). Prior literature reports have shown highly variable development and properties of mixed-cell spheroids and this has hampered the detailed study of how individual tumor-cell components interact. In this study, we illustrate this approach and describe similarities and differences using two tumor models (U87 glioma and SQ20B squamous-cell carcinoma) with supporting data from additional cell lines. We show that U87 and SQ20B spheroids predict a key microenvironmental factor in tumors (hypoxia) and that SQ20B cells and spheroids generate similar numbers of microvesicles. We also present pilot data for miRNA expression under conditions of cells, tumors, and TF spheroids.

Glucose-lactate metabolic cooperation in cancer: insights from a spatial mathematical model and implications for targeted therapy.

A recent study has hypothesised a glucose-lactate metabolic symbiosis between adjacent hypoxic and oxygenated regions of a developing tumour, and proposed a treatment strategy to target this symbiosis. However, in vivo experimental support remains inconclusive. Here we develop a minimal spatial mathematical model of glucose-lactate metabolism to examine, in principle, whether metabolic symbiosis is plausible in human tumours, and to assess the potential impact of inhibiting it. We find that symbiosis is a robust feature of our model system-although on the length scale at which oxygen supply is diffusion-limited, its occurrence requires very high cellular metabolic activity-and that necrosis in the tumour core is reduced in the presence of symbiosis. Upon simulating therapeutic inhibition of lactate uptake, we predict that targeted treatment increases the extent of tissue oxygenation without increasing core necrosis. The oxygenation effect is correlated strongly with the extent of wild-type hypoxia and only weakly with wild-type symbiotic behaviour, and therefore may be promising for radiosensitisation of hypoxic, lactate-consuming tumours even if they do not exhibit a spatially well-defined symbiosis. Finally, we conduct in vitro experiments on the U87 glioblastoma cell line to facilitate preliminary speculation as to where highly malignant tumours might fall in our parameter space, and find that these experiments suggest a weakly symbiotic regime for U87 cells, thus raising the new question of what relationship might exist between symbiosis and tumour malignancy.

Regulation of O2 consumption by the PI3K and mTOR pathways contributes to tumor hypoxia.

BACKGROUND: Inhibitors of the phosphatidylinositol 3-kinase (PI3K) and the mammalian target of rapamycin (mTOR) pathway are currently in clinical trials. In addition to antiproliferative and proapoptotic effects, these agents also diminish tumor hypoxia. Since hypoxia is a major cause of resistance to radiotherapy, we sought to understand how it is regulated by PI3K/mTOR inhibition. METHODS: Whole cell, mitochondrial, coupled and uncoupled oxygen consumption were measured in cancer cells after inhibition of PI3K (Class I) and mTOR by pharmacological means or by RNAi. Mitochondrial composition was assessed by immunoblotting. Hypoxia was measured in spheroids, in tumor xenografts and predicted with mathematical modeling. RESULTS: Inhibition of PI3K and mTOR reduced oxygen consumption by cancer cell lines is predominantly due to reduction of mitochondrial respiration coupled to ATP production. Hypoxia in tumor spheroids was reduced, but returned after removal of the drug. Murine tumors had increased oxygenation even in the absence of average perfusion changes or tumor necrosis. CONCLUSIONS: Targeting the PI3K/mTOR pathway substantially reduces mitochondrial oxygen consumption thereby reducing tumor hypoxia. These alterations in tumor hypoxia should be considered in the design of clinical trials using PI3K/mTOR inhibitors, particularly in conjunction with radiotherapy.

3D tumour spheroids as a model to assess the suitability of [18F]FDG-PET as an early indicator of response to PI3K inhibition.

BACKGROUND: [18F] Fluorodeoxyglucose Positron Emission Tomography ([18F]FDG-PET) is widely used to monitor response to therapy in the clinic and has, more recently, been proposed as an early marker of long term response. This relies on the assumption that a change in glucose consumption parallels a reduction in viability and long term growth potential. However, cells may utilise substrates other than glucose and as many therapeutics interfere with glucose metabolism directly, it is entirely plausible that a positive [18F]FDG-PET response may be unrelated to long term growth. Furthermore, changes in metabolism and proliferation may take place on different temporal scales, thus restricting the time window in which [18F]FDG-PET is predictive. The PI3K oncogenic signalling pathway is a master regulator of multiple cellular processes including glucose metabolism, proliferation and cell survival. Inhibition of PI3K has been shown to reduce [18F]FDG uptake in several tumour types but the relative influence of this pathway on glucose metabolism and proliferation is not fully established. AIM: We proposed to (i) assess the suitability of [18F]FDG as a tracer for measuring response to PI3K inhibition and (ii) determine the optimum imaging schedule, in vitro. We used multicellular tumour spheroids, an excellent 3D in vitro model of avascular tumours, to investigate the effects of the PI3K inhibitors, NVP-BKM120 and NVP-BEZ235, on [18F]FDG uptake and its relation to 3D growth. METHODS: Spheroids were prepared from two cell lines with a constitutively active PI3K/Akt pathway, EMT6 (highly proliferative mouse mammary) and FaDu (moderately proliferate human nasopharyngeal). Treatment consisted of a 24h exposure to either inhibitor, and growth was monitored over the following 7 days. To mimic potential imaging regimens with [18F]FDG-PET, average [18F]FDG uptake per viable cell was measured (a) directly following the 24h exposure, (b) following an additional 24h recovery period, or (c) following a 48 h exposure. RESULTS: Growth was restricted significantly (p<0.0001) in a dose-dependent fashion in spheroids from both cell lines treated with either inhibitor. In the highly proliferative cell line EMT6, [18F]FDG uptake was significantly reduced at all concentrations of inhibitor. For the moderately proliferative cell line FaDu, [18F]FDG was affected in a dose-dependent fashion, but to lesser degree. To assess the predictivity of [18F]FDG uptake for long term growth restriction, Pearson correlation coefficients were calculated for each imaging regimen. These indicated that the optimal imaging schedules differed between cell lines. CONCLUSION: This study suggests that [18F]FDG may be a suitable marker of response to PI3K inhibition in the cell lines that we have studied. Our data support the hypothesis that imaging schedules should be optimised on a tumour type-specific basis.

A model to simulate tumour oxygenation and dynamic [18F]-Fmiso PET data.

The microenvironment of a tumour, in particular its hypoxic status, is a crucial factor in its response to radiotherapy. Conventional techniques for measuring hypoxia are either invasive or follow surgical intervention, and thus not ideal. Positron emission tomography allows the non-invasive pre-surgical assessment of oxygen status by measuring the spatiotemporal distribution of hypoxia-specific tracers. However, the relationship between levels of uptake and the underlying oxygen tension are yet to be elucidated. Furthermore, it is not fully understood how changes in the underlying physiology affect the appearance of uptake. This paper presents a modular simulation of the tumour microenvironment, underpinned by a probability density function (PDF) to model the vasculature. The model is solved numerically, to simulate both the steady-state oxygenation of a tumour and the spatiotemporal distribution of the hypoxia-specific tracer, [18F]-fluoromisonidazole (Fmiso), in a 2D environment. The results show that using a PDF to represent the vasculature effectively captures the 'hypoxic island' appearance of oxygen-deficient tissues seen ex vivo. Simulated tissue activity curves (TACs) demonstrate the general two-stage trend of empirical data, with an initial perfusion-dominated uptake, followed by hypoxia-specific binding. In well-perfused tissue, activity follows plasma levels in early stages, with binding of Fmiso only becoming apparent at a later stage. In structurally hypoxic tissue, a more gradual initial increase in activity is observed, followed by the same accumulation slope. We demonstrate the utility of theoretical modelling of tracer uptake, by quantifying the changes in TAC structure that arise as a result of altering key physiological characteristics. For example, by decreasing either the proximity of tissue to the vasculature, or the effective diffusion coefficient of Fmiso, we can observe a shift of TAC structure from corresponding to well-perfused to avascular regions, despite wholly different underlying causes.