Deep learning-based attenuation correction for whole-body PET - a multi-tracer study with 18F-FDG, 68 Ga-DOTATATE, and 18F-Fluciclovine.

A novel deep learning (DL)-based attenuation correction (AC) framework was applied to clinical whole-body oncology studies using 18F-FDG, 68 Ga-DOTATATE, and 18F-Fluciclovine. The framework used activity (λ-MLAA) and attenuation (µ-MLAA) maps estimated by the maximum likelihood reconstruction of activity and attenuation (MLAA) algorithm as inputs to a modified U-net neural network with a novel imaging physics-based loss function to learn a CT-derived attenuation map (µ-CT). METHODS: Clinical whole-body PET/CT datasets of 18F-FDG (N = 113), 68 Ga-DOTATATE (N = 76), and 18F-Fluciclovine (N = 90) were used to train and test tracer-specific neural networks. For each tracer, forty subjects were used to train the neural network to predict attenuation maps (µ-DL). µ-DL and µ-MLAA were compared to the gold-standard µ-CT. PET images reconstructed using the OSEM algorithm with µ-DL (OSEMDL) and µ-MLAA (OSEMMLAA) were compared to the CT-based reconstruction (OSEMCT). Tumor regions of interest were segmented by two radiologists and tumor SUV and volume measures were reported, as well as evaluation using conventional image analysis metrics. RESULTS: µ-DL yielded high resolution and fine detail recovery of the attenuation map, which was superior in quality as compared to µ-MLAA in all metrics for all tracers. Using OSEMCT as the gold-standard, OSEMDL provided more accurate tumor quantification than OSEMMLAA for all three tracers, e.g., error in SUVmax for OSEMMLAA vs. OSEMDL: - 3.6 ± 4.4% vs. - 1.7 ± 4.5% for 18F-FDG (N = 152), - 4.3 ± 5.1% vs. 0.4 ± 2.8% for 68 Ga-DOTATATE (N = 70), and - 7.3 ± 2.9% vs. - 2.8 ± 2.3% for 18F-Fluciclovine (N = 44). OSEMDL also yielded more accurate tumor volume measures than OSEMMLAA, i.e., - 8.4 ± 14.5% (OSEMMLAA) vs. - 3.0 ± 15.0% for 18F-FDG, - 14.1 ± 19.7% vs. 1.8 ± 11.6% for 68 Ga-DOTATATE, and - 15.9 ± 9.1% vs. - 6.4 ± 6.4% for 18F-Fluciclovine. CONCLUSIONS: The proposed framework provides accurate and robust attenuation correction for whole-body 18F-FDG, 68 Ga-DOTATATE and 18F-Fluciclovine in tumor SUV measures as well as tumor volume estimation. The proposed method provides clinically equivalent quality as compared to CT in attenuation correction for the three tracers.

Single-walled carbon nanotube thermopile for broadband light detection.

We designed a thermopile based on a PN doping profile engineered in a suspended film of single-walled carbon nanotubes (SWNTs). Using estimates of the film local Seebeck coefficients, the SWNT thermopile was optimized in situ through depositions of potassium dopants. The overall performances of the thermopile were found to be comparable to state-of-the-art SWNT bolometers. The device is characterized at room temperature by a time response of 36 ms, typical of thermal detectors, and an optimum spectral detectivity of 2 × 10(6) cm Hz(1/2)/W in the visible and near-infrared. This paper presents the first thermopile made of a suspended SWNT film and paves the way to new applications such as broadband light (including THz) detection and thermoelectric power generation.

PET Image Denoising using a Deep-Learning Method for Extremely Obese Patients.

The image quality in clinical PET scan can be severely degraded due to high noise levels in extremely obese patients. Our work aimed to reduce the noise in clinical PET images of extremely obese subjects to the noise level of lean subject images, to ensure consistent imaging quality. The noise level was measured by normalized standard deviation (NSTD) derived from a liver region of interest. A deep learning-based noise reduction method with a fully 3D patch-based U-Net was used. Two U-Nets, U-Nets A and B, were trained on datasets with 40% and 10% count levels derived from 100 lean subjects, respectively. The clinical PET images of 10 extremely obese subjects were denoised using the two U-Nets. The results showed the noise levels of the images with 40% counts of lean subjects were consistent with those of the extremely obese subjects. U-Net A effectively reduced the noise in the images of the extremely obese patients while preserving the fine structures. The liver NSTD improved from 0.13±0.04 to 0.08±0.03 after noise reduction (p = 0.01). After denoising, the image noise level of extremely obese subjects was similar to that of lean subjects, in terms of liver NSTD (0.08±0.03 vs. 0.08±0.02, p = 0.74). In contrast, U-Net B over-smoothed the images of extremely obese patients, resulting in blurred fine structures. In a pilot reader study comparing extremely obese patients without and with U-Net A, the difference was not significant. In conclusion, the U-Net trained by datasets from lean subjects with matched count level can provide promising denoising performance for extremely obese subjects while maintaining image resolution, though further clinical evaluation is needed.

Determining the minimal ultra-low dose CT for reliable attenuation correction of 18F-FDG PET-CT: a phantom study.

To investigate minimal required sub milli-Sievert (mSv) ultra-low dose CT and corresponding tube current and voltage for reliable attenuation correction and semi- quantitation in 18F-FDG PET-CT in an effort for radiation dose reduction. Methods: We performed a PET-CT investigational study using a NEMA torso phantom containing six spheres (diameter: 10, 13, 17, 22, 28, 37 mm) filled with a fixed concentration of 60 kBq/ml and a background of 15 kBq/ml of 18F-FDG. Two sets of PET images, separated by 2 hours, were acquired for 3 minutes in a single bed position using 3-D mode with and without time-of-flight in a GE D-690 scanner. Several sets of CT images were acquired for attenuation correction with different combinations of tube voltage (80, 100, 120 kVp) and effective mAs (tube current-time product divided by pitch), using the maximum beam collimation (64 x 0.625 mm). The lowest CT acquisition technique available on this scanner is 10 mA, 0.4 s and 1.375 for the tube current, tube rotation time and pitch, respectively. The CT radiation dose was estimated based on the computed tomography dose index volume (CTDIvol) measurements performed following the standard methodology and the Imaging Performance Assessment of CT Scanners (ImPACT) calculator. Each of the CT techniques was used for attenuation correction to the same PET acquisition, using ordered-subset expectation maximum (OSEM) algorithm with 24 subsets and 2 iterations. The maximal and average radioactivity (kBq/ml) and standardized uptake values (SUV) of the spheres were measured. The minimal ultra-low dose CT for attenuation correction was determined by reproducible SUV measurements (±10%) compared to our reference CT protocol of 100 kVp and 80 mA for 0.5 s rotation. Results: The minimal ultra-low dose of CT for reproducible quantification in all spheres (<10% relative difference) was determined to be 0.3 mSv for a combination of 100 kVp and 10 mA at 0.5 s rotation, 0.984 helical pitch (0.26 mGy measured CTDIvol) . Based on these results we could confidently determine the CT parameters for reliable attenuation correction of PET images while significantly reducing the associated radiation dose. Conclusion: Our phantom study provided guidance in using ultra-low dose CT for precise attenuation correction and semi-quantification of 18F-FDG PET imaging, which can further reduce CT dose and radiation exposure to patients in clinical PET-CT studies. Clinical application: Based on the data, we can further reduce the radiation dose to sub-mSv using an ultra-low dose CT protocol for reliable attenuation correction in clinical 18F-FDG PET-CT studies.

Position sensitive photothermoelectric effect in suspended single-walled carbon nanotube films.

The photovoltage properties of suspended single-walled carbon nanotube (SWNT) films were measured in high vacuum. Experiments with localized illumination showed significant photovoltage amplitudes of up to 0.36 mV at 1.2 mW intensity. The photoresponse dependence upon the laser position was explained by a thermal mechanism that is independent of the nanotube-metal barrier. The response was also found to depend on doping heterogeneities of the film. A model was developed to deduce from the data the spatial variation of the local Seebeck coefficient for a given photovoltage profile.

Electronic Transport in the Biopigment Sepia Melanin.

Eumelanin is the most common form of the pigment melanin in the human body, with diverse functions including photoprotection, antioxidant behavior, metal chelation, and free radical scavenging. Melanin also plays a role in melanoma skin cancer and Parkinson's disease. Sepia melanin is a natural eumelanin extracted from the ink sac of cuttlefish. Eumelanin is an ideal candidate to eco-design technologies based on abundant, biosourced, and biodegradable organic electronic materials to alleviate the environmental footprint of the electronics sector. Herein, the focus is on the reversible electrical resistive switching in dry and wet Sepia eumelanin pellets, pointing to the possibility of predominant electronic transport satisfying conditio sine qua non to develop melanin-based electronic devices. These findings shed light on the possibility to describe the transport physics of dry eumelanin using the amorphous semiconductor model. Results are of tremendous importance for the development of sustainable organic electronics.

Design of a compensated signal rod for low magnetic moment sample measurements with a vibrating sample magnetometer.

A zero-signal sample holder is proposed for the measurement of weak magnetic signals with vibrating sample magnetometers. With proper shape of the support rod, a nearly vanishing signal can be obtained as a function of the magnetic field and the temperature. In particular, it is shown that the addition of an extra part to a standard glass sample holder can reduce the diamagnetic signal by more than three orders of magnitude with no noise increase. The proposed method is applicable to field, temperature, and angular measurements; it is also ideally suited to direct measurement of nanometer thick magnetic layers deposited on much thicker diamagnetic substrates.

Growth, structure, and properties of BiFeO3/-BiCrO3 films obtained by dual cross beam PLD.

The properties of epitaxial Bi(2)FeCrO(6) thin films, recently synthesized by pulsed laser deposition, have partially confirmed the theoretical predictions (i.e., a magnetic moment of 2 micro(B) per formula unit and a polarization of approximately 80 microC/cm(2) at 0 K). The existence of magnetic ordering at room temperature for this material is an unexpected, but very promising, result that needs to be further investigated. Because magnetism is assumed to arise from the exchange interaction between the Fe and Cr cations, the magnetic behavior is strongly dependent on both their ordering and the distance between them. We present here the successful synthesis of epitaxial Bi(2)Fe(x)CryO(6) (BFCO x/y) films grown on SrTiO3 substrates using dual crossed-beam, pulsed-laser deposition. The crystal structure of the films has different types of (111)-oriented superstructures, depending on the deposition conditions. The multiferroic character of BFCO (x/y) films is proven by the presence of both ferroelectric and magnetic hysteresis at room temperature. The oxidation state of Fe and Cr ions in the films is shown to be 3+ only, and the difference in macroscopic magnetization with Fe/Cr ratio composition could only be due to ordering of the Cr(3+) and Fe(3+) cations to the modification of the exchange interaction between them.

The influence of iron oxide nanoparticles upon the adsorption of organic matter on magnetic powdered activated carbon.

Combining powdered activated carbon (PAC) with magnetic iron oxides has been proposed in the past to produce adsorbents for natural organic matter (NOM) removal that can be easily separated using a magnetic field. However, the trade-off between the iron oxides' benefits and the reduced carbon content, porosity, and surface area has not yet been investigated systematically. We produced 3 magnetic powdered activated carbons (MPAC) with mass fractions of 10%, 38% and 54% maghemite nanoparticles and compared them to bare PAC and pure nanoparticles with respect to NOM adsorption kinetics and isotherms. While adsorption kinetics were not influenced by the presence of the iron oxide nanoparticles (IONP), as shown by calculated diffusion coefficients from the homogeneous surface diffusion model, nanoparticles reduced the adsorption capacity of NOM due to their lower adsorption capacity. Although the nanoparticles added mesoporosity to the composite materials they blocked intrinsic PAC mesopores at mass fractions >38% as measured by N2-adsorption isotherms. Below this mass fraction, the adsorption capacity was mainly dependent on the carbon content in MPAC and mesopore blocking was negligible. If NOM adsorption with MPAC is desired, a highly mesoporous PAC and a low IONP mass fraction should be chosen during MPAC synthesis.

Epitaxially stabilized thin films of ε-Fe2O3 (001) grown on YSZ (100).

Epsilon ferrite (ε-Fe2O3) is a metastable phase of iron(III) oxide, intermediate between maghemite and hematite. It has recently attracted interest because of its magnetocrystalline anisotropy, which distinguishes it from the other polymorphs, and results in a gigantic coercive field and a natural ferromagnetic resonance frequency in the THz range. Moreover, it possesses a polar crystal structure, making it a potential ferroelectric, hence a potential multiferroic. Due to the need of size confinement to stabilize the metastable phase, ε-Fe2O3 has been synthesized mainly as nanoparticles. However, to favor integration in devices, and take advantage of its unique functional properties, synthesis as epitaxial thin films is desirable. In this paper, we report the growth of ε-Fe2O3 as epitaxial thin films on (100)-oriented yttrium-stabilized zirconia substrates. Structural characterization outlined the formation of multiple in-plane twins, with two different epitaxial relations to the substrate. Transmission electron microscopy showed how such twins develop in a pillar-like structure from the interface to the surface. Magnetic characterization confirmed the high magnetocrystalline anisotropy of our film and revealed the presence of a secondary phase which was identified as the well-known magnetite. Finally, angular analysis of the magnetic properties revealed how the presence of twins impacts their azimuthal dependence.

Determining the Minimal Required Radioactivity of 18F-FDG for Reliable Semiquantification in PET/CT Imaging: A Phantom Study.

UNLABELLED: In pursuit of as-low-as-reasonably-achievable (ALARA) doses, this study investigated the minimal required radioactivity and corresponding imaging time for reliable semiquantification in PET/CT imaging. METHODS: Using a phantom containing spheres of various diameters (3.4, 2.1, 1.5, 1.2, and 1.0 cm) filled with a fixed (18)F-FDG concentration of 165 kBq/mL and a background concentration of 23.3 kBq/mL, we performed PET/CT at multiple time points over 20 h of radioactive decay. The images were acquired for 10 min at a single bed position for each of 10 half-lives of decay using 3-dimensional list mode and were reconstructed into 1-, 2-, 3-, 4-, 5-, and 10-min acquisitions per bed position using an ordered-subsets expectation maximum algorithm with 24 subsets and 2 iterations and a gaussian 2-mm filter. SUVmax and SUVavg were measured for each sphere. RESULTS: The minimal required activity (±10%) for precise SUVmax semiquantification in the spheres was 1.8 kBq/mL for an acquisition of 10 min, 3.7 kBq/mL for 3-5 min, 7.9 kBq/mL for 2 min, and 17.4 kBq/mL for 1 min. The minimal required activity concentration-acquisition time product per bed position was 10-15 kBq/mL⋅min for reproducible SUV measurements within the spheres without overestimation. Using the total radioactivity and counting rate from the entire phantom, we found that the minimal required total activity-time product was 17 MBq⋅min and the minimal required counting rate-time product was 100 kcps⋅min. CONCLUSION: Our phantom study determined a threshold for minimal radioactivity and acquisition time for precise semiquantification in (18)F-FDG PET imaging that can serve as a guide in pursuit of achieving ALARA doses.

Performance of biological magnetic powdered activated carbon for drinking water purification.

Combining the high adsorption capacity of powdered activated carbon (PAC) with magnetic properties of iron oxide nanoparticles (NPs) leads to a promising composite material, magnetic PAC or MPAC, which can be separated from water using magnetic separators. We propose MPAC as an alternative adsorbent in the biological hybrid membrane process and demonstrate that PAC covered with magnetic NPs is suitable as growth support for heterotrophic and nitrifying bacteria. MPAC with mass fractions of 0; 23; 38 and 54% maghemite was colonized in small bioreactors for over 90 days. Although the bacterial community composition (16s rRNA analysis) was different on MPAC compared to PAC, NPs neither inhibited dissolved organic carbon and ammonia biological removals nor contributed to significant adsorption of these compounds. The same amount of active heterotrophic biomass (48 µg C/cm(3)) developed on MPAC with a mass fraction of 54% NPs as on the non-magnetic PAC control. While X-ray diffraction confirmed that size and type of iron oxides did not change over the study period, a loss in magnetization between 10% and 34% was recorded.