Valeria De Luca

Hybrid Ultrasound/Magnetic Resonance Simultaneous Acquisition and Image Fusion for Motion Monitoring in the Upper Abdomen

ObjectivesThe combination of ultrasound (US) and magnetic resonance imaging (MRI) may provide a complementary description of the investigated anatomy, together with improved guidance and assessment of image-guided therapies. The aim of the present study was to integrate a clinical setup for simultaneous US and magnetic resonance (MR) acquisition to obtain synchronized monitoring of liver motion. The feasibility of this hybrid imaging and the precision of image fusion were evaluated. Materials and MethodsUltrasound imaging was achieved using a clinical US scanner modified to be MR compatible, whereas MRI was achieved on 1.5- and 3-T clinical scanners. Multimodal registration was performed between a high-resolution T1 3-dimensional (3D) gradient echo (volume interpolated gradient echo) during breath-hold and a simultaneously acquired 2D US image, or equivalent, retrospective registration of US imaging probe in the coordinate frame of MRI. A preliminary phantom study was followed by 4 healthy volunteer acquisitions, performing simultaneous 4D MRI and 2D US harmonic imaging (Fo = 2.2 MHz) under free breathing. ResultsNo characterized radiofrequency mutual interferences were detected under the tested conditions with commonly used MR sequences in clinical routine, during simultaneous US/MRI acquisition. Accurate spatial matching between the 2D US and the corresponding MRI plane was obtained during breath-hold. In situ fused images were delivered. Our 4D MRI sequence permitted the dynamic reconstruction of the intra-abdominal motion and the calculation of high temporal resolution motion field vectors. ConclusionsThis study demonstrates that, truly, simultaneous US/MR dynamic acquisition in the abdomen is achievable using clinical instruments. A potential application is the US/MR hybrid guidance of high-intensity focused US therapy in the liver.

Model-guided respiratory organ motion prediction of the liver from 2D ultrasound

With the availability of new and more accurate tumour treatment modalities such as high-intensity focused ultrasound or proton therapy, accurate target location prediction has become a key issue. Various approaches for diverse application scenarios have been proposed over the last decade. Whereas external surrogate markers such as a breathing belt work to some extent, knowledge about the internal motion of the organs inherently provides more accurate results. In this paper, we combine a population-based statistical motion model and information from 2d ultrasound sequences in order to predict the respiratory motion of the right liver lobe. For this, the motion model is fitted to a 3d exhalation breath-hold scan of the liver acquired before prediction. Anatomical landmarks tracked in the ultrasound images together with the model are then used to reconstruct the complete organ position over time. The prediction is both spatial and temporal, can be computed in real-time and is evaluated on ground truth over long time scales (5.5 min). The method is quantitatively validated on eight volunteers where the ultrasound images are synchronously acquired with 4D-MRI, which provides ground-truth motion. With an average spatial prediction accuracy of 2.4 mm, we can predict tumour locations within clinically acceptable margins.

Modeling the detectability of vesicoureteral reflux using microwave radiometry

We present the modeling efforts on antenna design, frequency selection and receiver sensitivity estimation to detect vesicoureteral reflux (VUR) using microwave (MW) radiometry as warm urine from the bladder maintained at fever range temperature using a MW hyperthermia device reflows into the kidneys. The radiometer center frequency (f(c)), frequency band (Deltaf) and aperture radius (r(a)) of the physical antenna for kidney temperature monitoring are determined using a simplified universal antenna model with a circular aperture. Anatomical information extracted from the computed tomography (CT) images of children aged 4-6 years is used to construct a layered 3D tissue model. Radiometric antenna efficiency is evaluated in terms of the ratio of the power collected from the target at depth to the total power received by the antenna (eta). The power ratio of the theoretical antenna is used to design a microstrip log spiral antenna with directional radiation pattern over f(c) +/- Deltaf/2. Power received by the log spiral from the deep target is enhanced using a thin low-loss dielectric matching layer. A cylindrical metal cup is proposed to shield the antenna from electromagnetic interference (EMI). Transient thermal simulations are carried out to determine the minimum detectable change in the antenna brightness temperature (deltaT(B)) for 15-25 mL urine refluxes at 40-42 degrees C located 35 mm from the skin surface. Theoretical antenna simulations indicate maximum eta over 1.1-1.6 GHz for r(a) = 30-40 mm. Simulations of the 35 mm radius tapered log spiral yielded a higher power ratio over f(c) +/- Deltaf/2 for the 35-40 mm deep targets in the presence of an optimal matching layer. Radiometric temperature calculations indicate deltaT(B) 0.1 K for the 15 mL urine at 40 degrees C and 35 mm depth. Higher eta and deltaT(B) were observed for the antenna and matching layer inside the metal cup. Reflection measurements of the log spiral in a saline phantom are in agreement with the simulation data. The numerical study suggests that a radiometer with f(c) = 1.35 GHz, Deltaf = 500 MHz and detector sensitivity better than 0.1 K would be the appropriate tool to noninvasively detect VUR using the log spiral antenna.

SwissCheese at SemEval-2016 Task 4: Sentiment Classification Using an Ensemble of Convolutional Neural Networks with Distant Supervision

In this paper, we propose a classifier for predicting message-level sentiments of English micro-blog messages from Twitter. Our method builds upon the convolutional sentence embedding approach proposed by (Severyn and Moschitti, 2015a; Severyn and Moschitti, 2015b). We leverage large amounts of data with distant supervision to train an ensemble of 2-layer convolutional neural networks whose predictions are combined using a random forest classifier. Our approach was evaluated on the datasets of the SemEval-2016 competition (Task 4) outperforming all other approaches for the Message Polarity Classification task.

Leveraging Large Amounts of Weakly Supervised Data for Multi-Language Sentiment Classification

This paper presents a novel approach for multi-lingual sentiment classification in short texts. This is a challenging task as the amount of training data in languages other than English is very limited. Previously proposed multi-lingual approaches typically require to establish a correspondence to English for which powerful classifiers are already available. In contrast, our method does not require such supervision. We leverage large amounts of weakly-supervised data in various languages to train a multi-layer convolutional network and demonstrate the importance of using pre-training of such networks. We thoroughly evaluate our approach on various multi-lingual datasets, including the recent SemEval-2016 sentiment prediction benchmark (Task 4), where we achieved state-of-the-art performance. We also compare the performance of our model trained individually for each language to a variant trained for all languages at once. We show that the latter model reaches slightly worse - but still acceptable - performance when compared to the single language model, while benefiting from better generalization properties across languages.

A learning-based approach for fast and robust vessel tracking in long ultrasound sequences.

We propose a learning-based method for robust tracking in long ultrasound sequences for image guidance applications. The framework is based on a scale-adaptive block-matching and temporal realignment driven by the image appearance learned from an initial training phase. The latter is introduced to avoid error accumulation over long sequences. The vessel tracking performance is assessed on long 2D ultrasound sequences of the liver of 9 volunteers under free breathing. We achieve a mean tracking accuracy of 0.96 mm. Without learning, the error increases significantly (2.19 mm, p<0.001).

Estimation of Large-Scale Organ Motion in B-Mode Ultrasound Image Sequences: A Survey

Reviewed here are methods developed for following (i.e., tracking) structures in medical B-mode ultrasound time sequences during large-scale motion. The resulting motion estimation problem and its key components are defined. The main tracking approaches are described, and their strengths and weaknesses are discussed. Existing motion estimation methods, tested on multiple in vivo sequences, are categorized with respect to their clinical applications, namely, cardiac, respiratory and muscular motion. A large number of works in this field had to be discarded as thorough validation of the results was missing. The remaining relevant works identified indicate the possibility of reaching an average tracking accuracy up to 1-2 mm. Real-time performance can be achieved using several methods. Yet only very few of these have progressed to clinical practice. The latest trends include incorporation of complementary and prior information. Advances are expected from common evaluation databases and 4-D ultrasound scanning technologies.

Speeding-up image registration for repetitive motion scenarios

We propose a novel approach for real-time image registration for image sequences of organs subject to repetitive movement, such as breathing. The method exploits the redundancy within the images and consists of a training and an application phase. During training, the images are registered and then the relationship between the image appearance and the spatial transformation is learned by employing dimensionality reduction to the images and storage of the corresponding displacements. For each image in the application phase, the most similar images in the training set are selected for predicting the associated displacements. Registration and update of the training data is only performed for outliers. The method is assessed on 2D sequences (4 MRI, 1 ultrasound) of the liver during free breathing. The performance is evaluated on manually selected landmarks, such as vessel centers and the distal point of the inferior segment. The proposed algorithm is real-time (9 ms per frame) and the prediction error is on average 1.2 mm for both MRI and ultrasound.

Shaping and resizing of multifed slot radiators used in conformal microwave antenna arrays for hyperthermia treatment of large superficial diseases

It has been recently shown that chestwall recurrence of breast cancer and many other superficial diseases can be successfully treated with the combination of radiation, chemotherapy and hyperthermia. Conformal microwave antenna array for hyperthermia treatment of large area superficial diseases can significantly increase patient comfort while at the same time facilitate treatment of larger and more irregularly shaped disease. A large number of small efficient antennas is preferable for improved control of heating, as the disease can be more accurately contoured and the lower power requirement correlates with system reliability, linearity and reduced cost. Thus, starting from the initially proposed square slot antennas, we investigated new designs for multi-fed slot antennas of several shapes that maximize slot perimeter while reducing radiating area, thus increasing antenna efficiency. Simulations were performed with commercial electromagnetic simulation software packages (Ansoft HFSS) to demonstrate that the antenna size reduction method is effective for several dual concentric conductor (DCC) aperture shapes and operating frequencies. The theoretical simulations allowed the development of a set of design rules for multi-fed DCC slot antennas that facilitate conformal heat treatments of irregular size and shape disease with large multi-element arrays. Independently on the shape, it is shown that the perimeter of 10cm at 915 MHz delivers optimal radiation pattern and efficiency. While the maximum radiation is obtained for a circular pattern the rectangular shape is the one that feels more efficiently the array space.

Gated-tracking: Estimation of Respiratory Motion with Confidence

Image-guided radiation therapy during free-breathing requires estimation of the target position and compensation for its motion. Estimation of the observed motion during therapy needs to be reliable and accurate. In this paper we propose a novel, image sequence-specific confidence measure to predict the reliability of the tracking results. The sequence-specific statistical relationship between the image similarities and the feature displacements is learned from the first breathing cycles. A confidence measure is then assigned to the tracking results during the real-time application phase based on the relative closeness to the expected values. The proposed confidence was tested on the results of a learning-based tracking algorithm. The method was assessed on 9 2D B-mode ultrasound sequences of healthy volunteers under free-breathing. Results were evaluated on a total of 15 selected vessel centers in the liver, achieving a mean tracking accuracy of 0.9 mm. When considering only highly-confident results, the mean (95th percentile) tracking error on the test data was reduced by 12% (16%) while duty cycle remained sufficient (60%), achieving a 95% accuracy below 3 mm, which is clinically acceptable. A similar performance was obtained on 10 2D liver MR sequences, showing the applicability of the method to a different image modality.