Search

International audience; ObjectivesThe aim of this study is to evaluate a deep learning method designed to increase the contrast-to-noise ratio in contrast-enhanced gradient echo T1-weighted brain MRI acquisitions. The processed images are quantitatively evaluated in terms of lesion detection performance.Materials and methodsA total of 250 multiparametric brain MRIs, acquired between November 2019 and March 2021 at Gustave Roussy Cancer Campus (Villejuif, France), were considered for inclusion in this retrospective monocentric study. Independent training (107 cases, age 55y±14, 58 women) and test (79 cases, age 59y±14, 41 women) samples were defined. Patients had glioma, brain metastasis, meningioma, or no enhancing lesion. Gradient echo and turbo spin echo with variable flip angles postcontrast T1 sequences were acquired in all cases. For the cases that formed the training sample, “low-dose” postcontrast gradient echo T1 images using 0.025mmol/kg injections of contrast agent were also acquired. A deep neural network was trained to synthetically enhance the low-dose T1 acquisitions, taking standard-dose T1 MRI as reference. Once trained, the contrast enhancement network was used to process the test gradient echo T1 images. A read was then performed by two experienced neuroradiologists to evaluate the original and processed T1 MRI sequences in terms of contrast enhancement and lesion detection performance, taking the turbo spin echo sequences as reference.ResultsThe processed images were superior to the original gradient echo and reference turbo spin echo T1 sequences in terms of contrast-to-noise ratio (44.5 versus 9.1 and 16.8, p.99). The same effect was observed when considering all lesions larger than 5mm: sensitivity increased from 70% to 85% (p

International audience; We report a macrocyclic ligand (H) based on a 3,6,10,13-tetraaza-1,8(2,6)-dipyridinacyclotetradecaphane platform containing three acetate pendant arms and a benzyl group attached to the fourth nitrogen atom of the macrocycle. The X-ray structures of the Y and Tb complexes reveal nine coordination of the ligand to the metal ions through the six nitrogen atoms of the macrocycle and three oxygen atoms of the carboxylate pendants. A combination of NMR spectroscopic studies (H, C, and Y) and DFT calculations indicated that the structure of the Y complex in the solid state is maintained in an aqueous solution. The detailed study of the emission spectra of the Eu and Tb complexes revealed Ln-centered emission with quantum yields of 7.0 and 60%, respectively. Emission lifetime measurements indicate that the ligand offers good protection of the metal ions from surrounding water molecules, preventing the coordination of water molecules. The Y complex is remarkably inert with respect to complex dissociation, with a lifetime of 1.7 h in 1 M HCl. On the other hand, complex formation is fast (∼1 min at pH 5.4, 2 × 10 M). Studies using the Y-nuclide confirmed fast radiolabeling since [Y]Y is nearly quantitatively formed (radiochemical yield (RCY) > 95) in a short time over a broad range of pH values from ca. 2.4 to 9.0. Challenging experiments in the presence of excess ethylenediaminetetraacetic acid (EDTA) and in human serum revealed good stability of the [Y]Y complex. All of these experiments combined suggest the potential application of H derivatives as Y-based radiopharmaceuticals.

Pyclen-dipicolinate chelates proved to be very efficient chelators for the radiolabeling with β–-emitters such as 90Y. In this study, a pyclen-dipicolinate ligand functionalized with additional C12 alkyl chains was synthesized. The radiolabeling with 90Y proved that the addition of saturated carbon chains does not affect the efficiency of the radiolabeling, whereas a notable increase in lipophilicity of the resulting 90Y radiocomplex was observed. As a result, the compound could be extracted in Lipiodol® and encapsulated in biodegrable pegylated poly(malic acid) nanoparticles demonstrating the potential of lipophilic pyclen-dipicolinate derivatives as platforms for the design of radiopharmaceuticals for the treatment of liver or brain cancers by internal radiotherapy.

International audience; Objectives: This study proposes and evaluates a deep learning method that predicts surrogate images for contrast-enhanced T1 from multiparametric magnetic resonance imaging (MRI) acquired using only a quarter of the standard 0.1 mmol/kg dose of gadolinium-based contrast agent. In particular, the predicted images are quantitatively evaluated in terms of lesion detection performance.Materials and methods : This monocentric retrospective study leveraged 200 multiparametric brain MRIs acquired between November 2019 and February 2020 at Gustave Roussy Cancer Campus (Villejuif, France). A total of 145 patients were included: 107 formed the training sample (age 55y±14, 58 women) and 38 the separate test sample (age 62±12, 22 women). Patients had glioma, brain metastases, meningioma, or no enhancing lesion. T1, T2-Flair, DWI, low-dose and standard-dose postcontrast T1 sequences were acquired. A deep network was trained to process the precontrast and low-dose sequences in order to predict “virtual” surrogate images for contrast-enhanced T1. Once trained, the deep learning method was evaluated on the test sample. The discrepancies between the predicted virtual images and the standard-dose MRIs were qualitatively and quantitatively evaluated using both automated voxel-wisemetrics and a reader study, where two radiologists graded image qualities and marked all visible enhancing lesions.Results : The automated analysis of the test brain MRIs computed a structural similarity index of 87.1% (±4.8) between the predicted virtual sequences and the reference contrast-enhanced T1 MRIs, a peak signalto-noise ratio of 31.6dB (±2.0), and an area under the curve of 96.4% (±3.1). At Youden’s operating point, the voxel-wise sensitivity and specificity were 96.4% and 94.8% respectively. The reader study found that virtual images were preferred to standard-dose MRI in terms of image quality (p=.008). A total of 91 reference lesions were identified in the 38 test T1 sequences enhanced with full dose of contrast agent. On average across readers, the brain lesion sensitivity of the virtual images was 83% for lesions larger than 10mm (n=42), and the associated false detection rate was 0.08 lesion/patient. The corresponding positive predictive value of detected lesions was 92%, and the F1-score 88%. Lesion detection performance however dropped when smaller lesions were included: average sensitivity was 67% for lesions larger than 5mm (n=74), and 56% with all lesions included regardless of their size. The false detection rate remained below 0.50 lesion/patient in all cases, and the positive predictive value above 73%. The composite F1 score was 63% at worst.Conclusion : The proposed deep learning method for virtual contrast-enhanced T1 brain MRI prediction showed very high quantitative performance when evaluated with standard voxel-wise metrics. The reader study demonstrated that for lesions larger than 10mm, good detection performance could be maintained despite a 4-fold division in contrast agent usage, unveiling a promising avenue for reducing the gadolinium exposure of returning patients. Small lesions proved however difficult to handle for the deep network, showing that full-dose injections remain essential for accurate first-line diagnosis in neuro-oncology.

International audience; Contrast-enhanced medical images offer vital insights for the accurate diagnosis, characterization and treatment of tumors, and are routinely used worldwide. Acquiring such images requires to inject the patient intravenously with a gadolinium-based contrast agent (GBCA). Although GBCAs are considered safe, recent concerns about their accumulation in the body tilted the medical consensus towards a more parsimonious usage. Focusing on the case of brain magnetic resonance imaging, this paper proposes a deep learning method that synthesizes virtual contrast-enhanced T1 images as if they had been acquired after the injection of a standard 0.100 mmol/kg dose of GBCA, taking as inputs complementary imaging modalities obtained either after a reduced injection at 0.025 mmol/kg or without any GBCA involved. The method achieves a competitive structural similarity index of 94.2%. Its asymptotic performance is estimated, and the most important input modalities are identified.

International audience; The aim of this study was to investigate the presence and chemical forms of residual gadolinium (Gd) in rat brain after a single dose of Gd-based contrast agent. METHODS: Four groups of healthy rats (2 sacrifice time-points, n = 10/group, 80 rats in total) were randomized to receive a single intravenous injection of 1 of the 3 Gd-based contrast agents (GBCAs) (gadoterate meglumine, gadobenate dimeglumine, or gadodiamide) or the same volume of 0.9% saline solution. The injected concentration was 0.6 mmol/kg, corresponding to a concentration of 0.1 mmol/kg in humans after body surface normalization between rats and humans (according to the US Food and Drug Administration recommendations). Animals were sacrificed at 2 washout times: 1 (M1) and 5 (M5) months after the injection. Total Gd concentrations were determined in cerebellum by inductively coupled plasma mass spectrometry. Gadolinium speciation was analyzed by size-exclusion chromatography coupled to inductively coupled plasma mass spectrometry after extraction from cerebellum. RESULTS: A single injection of a clinically relevant dose of GBCA resulted in the detectable presence of Gd in the cerebellum 1 and 5 months after injection. The cerebellar total Gd concentrations after administration of the least stable GBCA (gadodiamide) were significantly higher at both time-points (M1: 0.280 ± 0.060 nmol/g; M5: 0.193 ± 0.023 nmol/g) than those observed for macrocyclic gadoterate (M1: 0.019 ± 0.004 nmol/g, M5: 0.004 ± 0.002 nmol/g; P < 0.0001). Gadolinium concentrations after injection of gadobenate were significantly lower at both time-points (M1: 0.093 ± 0.020 nmol/g; M5: 0.067 ± 0.013 nmol/g; P < 0.05) than the Gd concentration measured after injection of gadodiamide. At the 5-month time-point, the Gd concentration in the gadoterate group was also significantly lower than the Gd concentration in the gadobenate group (P < 0.05). Gadolinium speciation analysis of the water-soluble fraction showed that, after injection of the macrocyclic gadoterate, Gd was still detected only in its intact, chelated form 5 months after injection. In contrast, after a single dose of linear GBCAs (gadobenate and gadodiamide), 2 different forms were detected: intact GBCA and Gd bound to soluble macromolecules (above 80 kDa). Elimination of the intact GBCA form was also observed between the first and fifth month, whereas the amount of Gd present in the macromolecular fraction remained constant 5 months after injection. CONCLUSIONS: A single injection of a clinically relevant dose of GBCA is sufficient to investigate long-term Gd retention in the cerebellar parenchyma. Administration of linear GBCAs (gadodiamide and gadobenate) resulted in higher residual Gd concentrations than administration of the macrocyclic gadoterate. Speciation analysis of the water-soluble fraction of cerebellum confirmed washout of intact GBCA over time. The quantity of Gd bound to macromolecules, observed only with linear GBCAs, remained constant 5 months after injection and is likely to represent a permanent deposition.

[Abstract] Thanks to a smart regiospecific N-functionalization, a pyclen based ligand bearing one picolinate and two acetate arms organized in a dissymmetric manner was synthesized for Y3+ complexation, and compared to its symmetric analogue. The nature of the capping bonds around the metal coordination environment has a dramatic effect on the properties of the chelate, the natY3+ and 90Y3+ dissymmetric derivatives presenting enhanced thermodynamic stability and kinetic inertness. Hungarian Scientific Research Fund; NKFIH K-120224

Molecular magnetic resonance imaging targeted to an endothelial integrin involved in neoangiogenesis was compared to DCE-US and immunochemistry to assess the early response of three different therapeutic agents in renal cell carcinoma. Human A498 renal cells carcinoma was subcutaneously inoculated into 24 nude mice. Mice received either phosphate-buffered saline solution, sunitinib, everolimus, or bevacizumab during 4 days. DCE-US and molecular MRI targetingαvβ3 were performed at baseline and 4 days after treatment initiation. PI, AUC, relaxation rate variationsΔR2⁎, and percentage of vessels area quantified on CD31-stained microvessels were compared. Significant decreases were observed for PI and AUC parameters measured by DCE-US for bevacizumab group as early as 4 days, whereas molecularαvβ3-targeted MRI was able to detect significant changes in both bevacizumab and everolimus groups. Percentage of CD31-stained microvessels was significantly correlated with DCE-US parameters, PI (R=0.87,p=0.0003) and AUC (R=0.81,p=0.0013). The percentage of vessel tissue area was significantly reduced (p<0.01) in both sunitinib and bevacizumab groups. We report an early detection of neoangiogenesis modification after induction of targeted therapies, using DCE-US orαvβ3-targeted MRI. We consider these outcomes should encourage clinical trial developments to further evaluate the potential of this molecular MRI technique.

Human gastric mucin MUC5AC is secreted in the colonic mucus of cancer patients and is a specific marker of precancerous lesions called aberrant crypt foci. Using MUC5AC as a specific marker can improve sensitivity in the detection of early colorectal cancer. Here we demonstrated that the accumulation of MUC5AC in xenograft and mouse stomach can be detected by magnetic resonance imaging (MRI). We used ultrasmall particles of iron oxide (USPIOs) conjugated with disulfide constrained heptapeptide that were identified using a screening phage display. To accomplish this, we employed positive selection of the phage display library on MUC5AC purified from fresh human colonic adenomas in combination with negative selection of the phage library on purified human MUC2, which is predominantly found in normal colorectal tissues. This conjugate was tested on human colorectal cancer cell lines that were either able or unable to secrete MUC5AC, both in vitro and in vivo. MUC5AC-USPIO contrast agent and USPIOs alone were not detected in cell lines unable to secrete MUC5AC. A combination of MRI and microscopy studies was performed to detect a specific accumulation of the contrast agent in vivo. Thus, the MUC5AC contrast agent enabled non-invasive detection of precancerous lesions and colorectal cancer, highlighting its potential use in diagnostics, in the early detection of colorectal cancer recurrences after treatment and in mechanistic studies implicating MUC5AC. Copyright © 2016 John Wiley & Sons, Ltd.

Delivery of therapeutic or diagnostic agents to the brain is majorly hindered by the blood–brain barrier (BBB). Recently, many studies have demonstrated local and transient disruption of the BBB using low power ultrasound sonication combined with intravascular microbubbles. However, BBB opening and closure mechanisms are poorly understood, especially the maximum gap that may be safely generated between endothelial cells and the duration of opening of the BBB. Here, we studied BBB opening and closure under magnetic resonance (MR) guidance in a rat model. First, MR contrast agents (CA) of different hydrodynamic diameters (1 to 65 nm) were employed to estimate the largest molecular size permissible across the cerebral tissues. Second, to estimate the duration of the BBB opening, the CA were injected at various times post-BBB disruption (12 minutes to 24 hours). A T1 mapping strategy was developed to assess CA concentration at the ultrasound (US) focal point. Based on our experimental data and BBB closure modeling, a calibration curve was obtained to compute the half closure time as a function of CA hydrodynamic diameter. These findings and the model provide an invaluable basis for optimal design and delivery of nanoparticles to the brain.

International audience; The long term outcome of nanoparticles in the organism is one of the most important concerns raised by the development of nanotechnology and nanomedicine. Little is known on the way taken by cells to process and degrade nanoparticles over time. In this context, iron oxide superparamagnetic nanoparticles benefit from a privileged status, because they show a very good tolerance profile, allowing their clinical use for MRI diagnosis. It is generally assumed that the specialized metabolism which regulates iron in the organism can also handle iron oxide nanoparticles. However the biotransformation of iron oxide nanoparticles is still not elucidated. Here we propose a multiscale approach to study the fate of nanomagnets in the organism. Ferromagnetic resonance and SQUID magnetization measurements are used to quantify iron oxide nanoparticles and follow the evolution of their magnetic properties. A nanoscale structural analysis by electron microscopy complements the magnetic follow-up of nanoparticles injected to mice. We evidence the biotransformation of superparamagnetic maghemite nanoparticles into poorly-magnetic iron species probably stored into ferritin proteins over a period of three months. A putative mechanism is proposed for the biotransformation of iron-oxide nanoparticles

International audience; In the present work, the NMR properties of perfluorooctylbromide are revisited to derive a high-sensitivity fluorine MRI strategy. It is shown that the harmful effects of J-coupling can be eliminated by carefully choosing the bandwidth of the 180 degrees pulses in a spin-echo sequence. The T(2) of the CF(3) resonance of the molecule is measured using a multispin-echo sequence and shown to dramatically depend on the interpulse delay. Following these observations, an optimized multispin-echo imaging sequence is derived and compared with short TE/pulse repetition time gradient echo and chemical shift imaging sequences. The unparalleled sensitivity yielded by the multispin-echo sequence is promising for future applications, in particular for targeted contrast agents such as perfluorooctylbromide nanoparticles.

International audience; A novel mechanism of MRI contrast enhancement, based on the detection by a balanced steady-state free precession (SSFP) sequence of the proton resonance frequency shift induced by bulk magnetic susceptibility (BMS) contrast agents, was investigated. The potential for this contrast mechanism to image blood vessels was explored. The relaxation time and the frequency shift effects of gadolinium- and dysprosium-DOTA on SSFP signal was first simulated and evaluated on a water phantom at 1.5 T. In vitro, a 5-mM concentration in contrast agent induced a 20-Hz frequency shift, leading to a signal increase of 92% for Dy-DOTA, and a 10-Hz frequency shift, leading to a signal increase of 58% for Gd-DOTA at the reference frequency, taking into account the nonlinear SSFP signal response on frequency offset. The concept was then evaluated in vivo on anesthetized rabbits. Low doses of dysprosium-DOTA were injected in their vascular system, and imaging was performed at the level of neck vessels. Following a bolus injection, mean signal changes of 31%, 20% and 14% were observed in the carotid arteries, the vertebral veins and the jugular veins, respectively. The bolus peak times in arteries and veins were consistent with the rabbit vascular circulation. This frequency-shift based contrast mechanism presents interesting potential for contrast-enhanced MR angiography (CE-MRA) compared to usual relaxation-based contrast, but further investigations on reproducibility will be necessary.