Neuroimaging-based classification of PTSD using data-driven computational approaches: A multisite big data study from the ENIGMA-PGC PTSD consortium.

BACKGROUND: Recent advances in data-driven computational approaches have been helpful in devising tools to objectively diagnose psychiatric disorders. However, current machine learning studies limited to small homogeneous samples, different methodologies, and different imaging collection protocols, limit the ability to directly compare and generalize their results. Here we aimed to classify individuals with PTSD versus controls and assess the generalizability using a large heterogeneous brain datasets from the ENIGMA-PGC PTSD Working group. METHODS: We analyzed brain MRI data from 3,477 structural-MRI; 2,495 resting state-fMRI; and 1,952 diffusion-MRI. First, we identified the brain features that best distinguish individuals with PTSD from controls using traditional machine learning methods. Second, we assessed the utility of the denoising variational autoencoder (DVAE) and evaluated its classification performance. Third, we assessed the generalizability and reproducibility of both models using leave-one-site-out cross-validation procedure for each modality. RESULTS: We found lower performance in classifying PTSD vs. controls with data from over 20 sites (60 % test AUC for s-MRI, 59 % for rs-fMRI and 56 % for d-MRI), as compared to other studies run on single-site data. The performance increased when classifying PTSD from HC without trauma history in each modality (75 % AUC). The classification performance remained intact when applying the DVAE framework, which reduced the number of features. Finally, we found that the DVAE framework achieved better generalization to unseen datasets compared with the traditional machine learning frameworks, albeit performance was slightly above chance. CONCLUSION: These results have the potential to provide a baseline classification performance for PTSD when using large scale neuroimaging datasets. Our findings show that the control group used can heavily affect classification performance. The DVAE framework provided better generalizability for the multi-site data. This may be more significant in clinical practice since the neuroimaging-based diagnostic DVAE classification models are much less site-specific, rendering them more generalizable.

Correction to: Towards clinical grating-interferometry mammography.

The article Towards clinical grating-interferometry mammography, written by Carolina Arboleda, Zhentian Wang, Konstantins Jefimovs, Thomas Koehler, Udo Van Stevendaal, Norbert Kuhn, Bernd David, Sven Prevrhal, Kristina Lång, Serafino Forte, Rahel Antonia Kubik-Huch, Cornelia Leo.

Towards clinical grating-interferometry mammography.

OBJECTIVES: Grating-interferometry-based mammography (GIM) might facilitate breast cancer detection, as several research works have demonstrated in a pre-clinical setting, since it is able to provide attenuation, differential phase contrast, and scattering images simultaneously. In order to translate this technique to the clinics, it has to be adapted to cover a large field-of-view within a clinically acceptable exposure time and radiation dose. METHODS: We set up a grating interferometer that fits into a standard mammography system and fulfilled the aforementioned conditions. Here, we present the first mastectomy images acquired with this experimental device. RESULTS AND CONCLUSION: Our system performs at a mean glandular dose of 1.6 mGy for a 5-cm-thick, 18%-dense breast, and a field-of-view of 26 × 21 cm2. It seems to be well-suited as basis for a clinical-environment device. Further, dark-field signals seem to support an improved lesion visualization. Evidently, the effective impact of such indications must be evaluated and quantified within the context of a proper reader study. KEY POINTS: • Grating-interferometry-based mammography (GIM) might facilitate breast cancer detection, since it is sensitive to refraction and scattering and thus provides additional tissue information. • The most straightforward way to do grating-interferometry in the clinics is to modify a standard mammography device. • In a first approximation, the doses given with this technique seem to be similar to those of conventional mammography.

In vivo evaluation and proof of radiofrequency safety of a novel diagnostic MR-electrophysiology catheter.

An MR-electrophysiology (EP) catheter is presented that provides full diagnostic EP functionality and a high level of radiofrequency safety achieved by custom-designed transmission lines. Highly resistive wires transmit intracardiac electrograms and currents for intracardiac pacing. A transformer cable transmits the localization signal of a tip coil. Specific absorption rate simulations and temperature measurements at 1.5 T demonstrate that a wire resistance > 3 kΩ/m limits dielectric heating to a physiologically irrelevant level. Additional wires do not increase tip specific absorption rate significantly, which is important because some clinical catheters require up to 20 electrodes. It is further demonstrated that radiofrequency-induced and pacing-induced resistive heating of the wires is negligible under clinical conditions. The MR-EP catheters provided uncompromised recording of electrograms and cardiac pacing in combination with a standard EP recorder in MR-guided in vivo EP studies, and the tip coil enabled fast and robust catheter localization. In vivo temperature measurements during such a study did not detect any device-related heating, which confirms the high level of safety of the catheter, whereas unacceptable heating was found with a standard EP catheter. The presented concept for the first time enables catheters with full diagnostic EP functionality and active tracking and at the same time a sufficient level of radiofrequency safety for MRI without specific absorption rate-related limitations.

Microbubbles as a contrast agent in grating interferometry mammography: an ex vivo proof-of-mechanism study.

Grating interferometry mammography (GIM) is an experimental breast imaging method at the edge of being clinically implemented. Besides attenuation, GIM can measure the refraction and scattering of x-rays resulting in differential phase contrast (DPC) and dark-field (DF) images. In this exploratory study, we assessed the feasibility of using microbubbles as a contrast agent in GIM. Two millilitres of microbubbles and iodine were respectively injected into ex vivo breast phantoms, consisting of fresh chicken breasts. Native and postcontrast images were acquired with a clinically compatible GIM setup, operated at 38 kVp, 14-s acquisition time, and with a dose of 1.3 mGy. The visibility of the contrast agents was analysed in a side-by-side comparison by three radiologists. The contrast-to-noise-ratio (CNR) was calculated for each contrast agent. We found that both contrast agents were judged to be visible by the readers. The mean CNR was 3.1 ± 1.9 for microbubbles in DF and 24.2 ± 6.5 for iodine in attenuation. In conclusion, this is a first proof-of-mechanism study that microbubbles could be used as a contrast agent in clinically compatible GIM, due to their scattering properties, which implies the potential use of a contrast agent with a high safety profile in x-ray-based breast imaging.

Perspectives on clinical magnetic particle imaging.

After realizing the worlds' first preclinical magnetic particle imaging (MPI) demonstrator, Philips is now realizing the worlds' first whole-body clinical prototype to prove the feasibility of MPI for clinical imaging. After a brief introduction of the basic MPI imaging process, this contribution presents an overview on the determining factors for key properties, i.e., spatial resolution, acquisition speed, sensitivity, and quantitativeness, and how these properties are influenced by scaling up from preclinical to clinical instrumentation. Furthermore, it is discussed how this scale up affects the physiological compatibility of the method as well as hardware parameters such as power requirements for drive field generation, selection and focus field generation, and the design of the receive chain of the MPI device.

Fundamentals and applications of magnetic particle imaging.

Magnetic particle imaging (MPI) is a new medical imaging technique which performs a direct measurement of magnetic nanoparticles, also known as superparamagnetic iron oxide. MPI can acquire quantitative images of the local distribution of the magnetic material with high spatial and temporal resolution. Its sensitivity is well above that of other methods used for the detection and quantification of magnetic materials, for example, magnetic resonance imaging. On the basis of an intravenous injection of magnetic particles, MPI has the potential to play an important role in medical application areas such as cardiovascular, oncology, and also in exploratory fields such as cell labeling and tracking. Here, we present an introduction to the basic function principle of MPI, together with an estimation of the spatial resolution and the detection limit. Furthermore, the above-mentioned medical applications are discussed with respect to an applicability of MPI.