Gernoth Grunst

Augmented reality simulator for training in two-dimensional echocardiography

In two-dimensional echocardiography the sonographer must synthesize multiple tomographic slices into a mental three-dimensional (3D) model of the heart. Computer graphics and virtual reality environments are ideal to visualize complex 3D spatial relationships. In augmented reality (AR) applications, real and virtual image data are linked, to increase the information content. In the presented AR simulator a 3D surface model of the human heart is linked with echocardiographic volume data sets. The 3D echocardiographic data sets are registered with the heart model to establish spatial and temporal congruence. The heart model, together with an animated ultrasound sector represents a reference scenario, which displays the currently selected two-dimensional echocardiographic cutting plane calculated from the volume data set. Modifications of the cutting plane within the echocardiographic data are transferred and visualized simultaneously and in real time within the reference scenario. The trainee can interactively explore the 3D heart model and the registered 3D echocardiographic data sets by an animated ultrasound probe, whose position is controlled by an electromagnetic tracking system. The tracking system is attached to a dummy transducer and placed on a plastic puppet to give a realistic impression of a two-dimensional echocardiographic examination.

EchoComTEE - a simulator for transoesophageal echocardiography.

Transoesophageal echocardiography (TOE) requires extensive hands‐on training, and it is for this purpose we have designed EchoComTEE, a simulator for TOE. It consists of a manikin and dummy probe; according to the position of the dummy probe (tracked by an electromagnetic sensor), two‐dimensional (2D) images are calculated from three‐dimensional (3D) data sets. Echocardiographic images are presented side‐by‐side with a virtual scene consisting of a 3D heart, probe tip and image plane. In this way the trainee is provided with visual feed‐back of the relationship between echocardiogram and image plane position. We evaluated the simulator using a standardised questionnaire. Twenty‐five experts and 31 novice users participated in the study. Most experts graded the simulator as realistic and all recommended its use for training. Most novice users felt the simulator supported spatial orientation during TOE and, as anaesthetists often do not have training in transthoracic echocardiography, in this group the TOE simulator might be particularly useful.

Iterative neuronavigation using 3D ultrasound. A feasibility study.

Abstract Intra-operative ultrasound (iUS) can generate 2D images in real-time as well as near real-time 3D datasets of the current situation during an intervention. Tracked ultrasound can locate the images in 3D space and relate them to patient, devices, and pre-operative planning data. Therefore, tracked US is an efficient means for controlling the validity of pre-operative planning, recognition of changes (brain shift) during the intervention, replanning of the operational path due to situational changes (iterative navigation), and finally, controlling the results (residual tumor). This paper describes a neuronavigation system exploiting this potential of interventional tracked US for permanent control of intervention progress and iterative adaptation of the planned procedure to the current situation.

Intelligent training system integrated in an echocardiography simulator.

Computer simulators play an important role in medical education. We have extended our simulator EchoComJ with an intelligent training system (ITS) to support trainees adjusting echocardiographic standard views. EchoComJ is an augmented reality application that combines real three-dimensional ultrasound data with a virtual heart model enabling one to simulate an echocardiographic examination. The ITS analyzes the image planes according to their position, orientation and the visualization of anatomical landmarks using fuzzy rules. An adaptive feedback is provided that colors the specific anatomic landmarks within the contours of the virtual model based on the quality of the image plane.

Iterative Multimodal Computer-Assisted Navigation Using Pre- and Intraoperative Image Data

Summary Computer-assisted navigation shall reliably guide the surgeon to a planned target position. Various medical image modalities are required for a safe diagnosis and intervention planning. For treatment forms like LITT, therapy simulation should be part of a planning scenario. In soft tissues, the intervention usually will render preoperative images inappropriate if tissue shifts occur. In these cases intraoperative images are necessary to show the current state, to update the preoperative data, and to register preoperative images to the current situation. The update is an iterative process triggered by situational changes during the ongoing intervention. This paper presents our concept of iterative multimodal navigation that has been implemented in the LOCALITE Navigator for interventional MR and ultrasound.

Cooperative Feature Selection in Personalized Medicine

The chapter discusses a research support system to identify diagnostic result patterns that characterise pertinent patient groups for personalized medicine. Example disease is breast cancer. The approach integrates established clinical findings with systems biology analyses. In this respect it is related to personalized medicine as well as translational research. Technically the system is a computer based support environment that links machine learning algorithms for classification with an interface for the medical domain expert. The involvement of the clinician has two reasons. On the one hand the intention is to impart an in-depth understanding of potentially relevant ‘omics’ findings from systems biology (e.g. genomics, transcriptomics, proteomics, and metabolomics) for actual patients in the context of clinical diagnoses. On the other hand the medical expert is indispensable for the process to rationally constrict the pertinent features towards a manageable selection of diagnostic findings. Without the suitable incorporation of domain expert knowledge machine based selections are often polluted by noise or irrelevant but massive variations. Selecting a subset of features is necessary in order to tackle the problem that for statistical reasons the amount of features has to be in an appropriate relationship to the number of cases that are available in a study (curse of dimensionality). The cooperative selection process is iterative. Interim results of analyses based on automatic temporary feature selections have to be graspable and criticisable by the medical expert. In order to support the understanding of machine learning results a prototype based approach is followed. The case type related documentation is in accordance with the way the human expert is cognitively structuring experienced cases. As the features for patient description are heterogeneous in their type and nature, the machine learning based feature selection has to handle different kinds of pertinent dissimilarities for the features and integrate them into a holistic representation.

Iterative neuronavigation using 3D ultrasound - a feasibility study.

Intra-operative ultrasound (iUS) can generate 2D images in real-time as well as near real-time 3D datasets of the current situation during an intervention. Tracked ultrasound can locate the images in 3D space and relate them to patient, devices, andpre-operative planning data. Therefore, tracked US is an efficient means for controlling the validity of pre-operative planning, recognition of changes (brain shift) during the intervention, replanning of the operational path due to situational changes (iterative navigation), and finally, controlling the results (residual tumor). This paper describes a neuronavigation system exploiting this potential of interventional tracked US for permanent control of intervention progress and iterative adaptation of the planned procedure to the current situation.