Alfredo Illanes

Active contours extension and similarity indicators for improved 3D segmentation of thyroid ultrasound images

Thyroid segmentation in tracked 2D ultrasound (US) using active contours has a low segmentation accuracy mainly due to the fact that smaller structures cannot be efficiently recognized and segmented. To address this issue, we propose a new similarity indicator with the main objective to provide information to the active contour algorithm concerning the regions that the active contour should continue to expand or should stop. First, a preprocessing step is carried out in order to attenuate the noise present in the US image and to increase its contrast, using histogram equalization and a median filter. In the second step, active contours are used to segment the thyroid in each 2D image of the dataset. After performing a first segmentation, two similarity indicators (ratio of mean square error, MSE and correlation between histograms) are computed at each contour point of the initial segmented thyroid between rectangles located inside and outside the obtained contour. A threshold is used on a final indicator computed from the other two indicators to find the probable regions for further segmentation using active contours. This process is repeated until no new segmentation region is identified. Finally, all the segmented thyroid images passed through a 3D reconstruction algorithm to obtain a 3D volume segmented thyroid. The results showed that including similarity indicators based on histogram equalization and MSE between inside and outside regions of the contour can help to segment difficult areas that active contours have problem to segment.

Novel clinical device tracking and tissue event characterization using proximally placed audio signal acquisition and processing

We propose a new and complementary approach to image guidance for monitoring medical interventional devices (MID) with human tissue interaction and surgery augmentation by acquiring acoustic emission data from the proximal end of the MID outside the patient to extract dynamical characteristics of the interaction between the distal tip and the tissue touched or penetrated by the MID. We conducted phantom based experiments (n = 955) to show dynamic tool/tissue interaction during tissue needle passage (a) and vessel perforation caused by guide wire artery perforation (b). We use time-varying auto-regressive (TV-AR) modelling to characterize the dynamic changes and time-varying maximal energy pole (TV-MEP) to compute subsequent analysis of MID/tissue interaction characterization patterns. Qualitative and quantitative analysis showed that the TV-AR spectrum and the TV-MEP indicated the time instants of the needle path through different phantom objects (a) and clearly showed a perforation versus other generated artefacts (b). We demonstrated that audio signals acquired from the proximal part of an MID could provide valuable additional information to surgeons during minimally invasive procedures.

A study of concentration changes of Protoporphyrin IX and Coproporphyrin III in mixed samples mimicking conditions inside cancer cells for Photodynamic Therapy

Photodynamic Therapy (PDT) using Aminolevulinic acid (ALA) could be an effective and minimally invasively applicable way to treat many different types of tumors without radiation and large incisions by just applying a light pulse. However the PDT process is difficult to observe, control and optimize and the dynamical relationships between the variables involved in the process is complex and still hardly understood. One of the main variables affecting the outcome of the process is the determination of the interval of time between ALA inoculation and starting of light delivery. This interval, better known as drug-light interval, should ensure that enough Protoporphyrin IX (PPIX) is located in the vicinity of functional structures inside the cells for the greatest damage during the PDT procedure. One route to better estimate this time interval would be by predicting PPIX from the dynamical changes of its precursors. For that purpose, in this work a novel optical setup (OS) is proposed for differentiating fluorescence emitted by Coproporphyrin III (CPIII) and PPIX itself in samples composed of mixed solutions. The OS is tested using samples with different concentrations in mixed solutions of PPIX and the precursor CPIII as well as with a Polymethyl methacrylate test sample as additional reference. Results show that emitted fluorescence of the whole process can be measured independently for PPIX and its precursor, which can enable future developments on PPIX prediction from the dynamical changes of its precursor for subject-dependent drug-light interval assessment.

In-room ultrasound fusion combined with fully compatible 3D-printed holding arm – rethinking interventional MRI

There is no real need to discuss the potential advantages – mainly the excellent soft tissue contrast, nonionizing radiation, flow, and molecular information – of magnetic resonance imaging (MRI) as an intraoperative diagnosis and therapy system particularly for neurological applications and oncological therapies. Difficult patient access in conventional horizontal-field superconductive magnets, very high investment and operational expenses, and the need for special nonferromagnetic therapy tools have however prevented the widespread use of MRI as imaging and guidance tool for therapy purposes. The interventional use of MRI systems follows for the last 20+ years the strategy to use standard diagnostic systems and add more or less complicated and expensive components (eg, MRI-compatible robotic systems, specially shielded in-room monitors, dedicated tools and devices made from low-susceptibility materials, etc) to overcome the difficulties in the therapy process. We are proposing to rethink that approach using an in-room portable ultrasound (US) system that can be safely operated till 1 m away from the opening of a 3T imaging system. The live US images can be tracked using an optical inside–out approach adding a camera to the US probe in combination with optical reference markers to allow direct fusion with the MRI images inside the MRI suite. This leads to a comfortable US-guided intervention and excellent patient access directly on the MRI patient bed. This was combined with an entirely mechanical MRI-compatible 7 degrees of freedom holding arm concept, which shows that this test environment is a different way to create a cost-efficient and effective setup that combines the advantages of MRI and US by largely avoiding the drawbacks of current interventional MRI concepts.

Evaluation of Commonly Used Algorithms for Thyroid Ultrasound Images Segmentation and Improvement Using Machine Learning Approaches

The thyroid is one of the largest endocrine glands in the human body, which is involved in several body mechanisms like controlling protein synthesis and the bodys sensitivity to other hormones and use of energy sources. Hence, it is of prime importance to track the shape and size of thyroid over time in order to evaluate its state. Thyroid segmentation and volume computation are important tools that can be used for thyroid state tracking assessment. Most of the proposed approaches are not automatic and require long time to correctly segment the thyroid. In this work, we compare three different nonautomatic segmentation algorithms (i.e., active contours without edges, graph cut, and pixel-based classifier) in freehand three-dimensional ultrasound imaging in terms of accuracy, robustness, ease of use, level of human interaction required, and computation time. We figured out that these methods lack automation and machine intelligence and are not highly accurate. Hence, we implemented two machine learning approaches (i.e., random forest and convolutional neural network) to improve the accuracy of segmentation as well as provide automation. This comparative study intends to discuss and analyse the advantages and disadvantages of different algorithms. In the last step, the volume of the thyroid is computed using the segmentation results, and the performance analysis of all the algorithms is carried out by comparing the segmentation results with the ground truth.

Proximally placed signal acquisition sensoric for robotic tissue tool interactions

Abstract Robotic surgeries are still limited with respect to the surgeon’s natural senses. The tactile sense is exceptional important in conventional clinical procedures. To identify critical structures inside the tissue, palpation is a commonly used technique in conventional open surgeries. The underlying organ or pathological structures conditions (healthy, abnormally hard or soft) can for example be localized and assessed through this process. Palpation needs a tactile sense; however, that is commonly not available or limited in robotic surgeries. The palpation need was already addressed by several research groups that integrated complex sensor-feedback-systems into prototype surgical instruments for robotic systems. We propose a new technique to acquire data of the tissue tool interaction of the surgical instruments. The structure borne transmission path is used to measure acoustic emission (AE) at the outpatient (proximal) end of the instruments with the help of different sensors attached to the surface of the surgical tool. Initial tests were performed using a microphone in combination with a stethoscope. This setup showed promising results and a more integrated prototype was subsequently designed. A piezoelectric charge accelerometer was used as vibration sensor and compared to a MEMS microphone. A signal acquisition system was developed to acquire signals from both sensors in parallel. The sensors were then attached onto the shaft of a daVinci Prograsp Forceps instrument. According to the surgery observation, a series of simulated experiments was conducted. The tip of the grasper was swiped manually over a human subject’s dorsal and palmar hand side, lateral side of neck and over the carotid artery. Additionally, contact with soft tissue and other instruments were evaluated since these are events of interest during surgery. Advanced signal processing techniques allowed the identification and characterization of significant events such as palpation dynamics, contact and pulsation. Signals acquired by the MEMS microphone showed the most promising results. This approach will now be used to build a prototype for further evaluation in a clinical setup. The paper presents the first results that show that this novel technique can provide valuable information about the tool-tissue interaction in robotic surgery that typically can only be obtained through advanced distal sensor systems or actual human touch.

Foetal heart rate signal spectral analysis by using time-varying autoregressive modelling

Abstract During labour, foetal monitoring enables clinicians to prevent potential adverse outcomes, whose surveillance procedure is commonly based on analysis of cardiotocographic (CTG) signals. Unfortunately, this procedure is difficult because it involves human interpretation of highly complex signals. In order to improve the CTG assessment, different approaches based on signal processing techniques have been proposed. However, most of them do not consider the progression of the foetal response over time. In this work, we propose to study such progression along the foetal heart rate (FHR) signal by using spectral analysis based on time-varying autoregressive modelling. The main idea is to investigate if a particular FHR signal episode in the time-domain reflects dynamical changes in the frequency-domain that can help to assess the foetal condition. Results show that each FHR deceleration leaves a particular time-varying frequency signature described by the spectral energy components which could help to distinguish between a normal and a pathological foetus.

Vascular pattern detection and recognition in endoscopic imaging of the vocal folds

Abstract At present transoral laryngeal interventions are mainly observed and controlled by an external two dimensional direct microscopic view. This modality provides an overall view on the surgery situs in a straight line of sight. For treatment planning and appropriate documentation, an endoscopic inspection is mandatory prior to surgery. Nowadays a detailed endoscopic work-up of laryngeal lesions can be performed by contact endoscopy in combination with structure enhancement like Narrow Band Imaging. High resolution and magnification of up to 150 times provide detailed visualization of vascular structures and pathological changes of the tissue surface. In these procedures it is difficult however to localize the evaluated areas on large scale scenes like the microscopic view used for surgery. To provide a fast and easy image matching an automated vessel pattern recognition and allocation is presented. Endoscopic images depicting representative vessel structures of the vocal folds are selected out of contact endoscopy video scenes. These images are pre-processed for background homogenization. A Frangi Vessel Segmentation filter and morphological operations are used to extract the vessel structure and match it to the microscopic image. Using this method 4 detailed contact endoscopy images could be allocated in different scenes of the microscope video. This method can be used to simplify treatment planning and to prepare image data for documentation.

Cyclostationary analysis of ECG signals acquired inside an ultra-high field MRI scanner

In this paper, a strategy is proposed to estimate the R-peaks in ECG signals recorded inside a 7 T magnetic resonance imaging (MRI) scanner in order to reduce the disturbances due to the magnetohydrodynamic (MHD) effect and to finally obtain high quality cardiovascular magnetic resonance (CMR) images. We first show that the cyclostationarity property of the ECG signal disturbed by the MHD effect can be quantified by means of cyclic spectral analysis. Then, this information is forwarded as input to a cyclostationary source extraction algorithm applied to a set of ECG recordings acquired inside the MRI scanner in a Feet first (Ff) and a Head first (Hf) positions. Finally, detection of the R-peaks in the estimated cyclostationary signal completes the proposed procedure. Validation of the method is performed by comparing the estimated with clinical R-peaks annotations provided with the real world dataset. The obtained results are promising and future research directions are discussed.

Intravascular endoscopy improvement through narrow-band imaging

PurposeRecent advances in endoscopy have led to new technologies with significant optical imaging improvements. Since its development a few years ago, narrow-band imaging (NBI) has already been proved useful in detecting malignant lesions and carcinoma in clinical settings of urology, gastroenterology and ENT. The potential of this technology for imaging applications of the arterial vessel wall has not been properly analysed yet, but with the observed benefits could prove valuable for this clinical use as well.MethodsIn order to assess the efficacy of NBI, defects such as burns and mechanical tears were created on the walls of an arterial vessel sample. Ex vivo imaging using NBI and white light imaging (WLI) were performed with rigid and flexible fibre endoscopes.ResultsA thorough comparison of the images proved that NBI enhances the visualisation of lesions and defects on the artery walls compared to normal WLI.ConclusionWLI provides a direct image of the vessel lumen and its anatomical shape. It is suitable for observation and documentation of intravascular therapies. NBI images are more distinct and have more contrast. This helps to detect even small defects or changes on the inner vessel wall that could provide additional information and lead to more precise and personalised therapies.