Ryo Haraguchi

We evaluated the new electronic medical record system (EMR) of Kyoto Medical Center for usefulness. Change in Gefitinib side-effect records was used as an index for evaluation. Gefitinib is used for the treatment of lung cancer, but its side effects have attracted much attention, and records of such effects provide important information which needs to be shared among medical staff to ensure drug safety. Medical records of 18 inpatients treated with Gefitinib were evaluated, and the numbers and input times of records on side effects were compared between the pre-introduction period and every 3 months after introduction of this system. A year after this system was introduced, the number of side-effect records per patient was increased from 0.25 per day to 0.97. The number of all records per patient was increased from 0.50 per day to 4.66. However, the percentage of side-effect records entered during regular working hours (8:30-17:15) decreased. In conclusion, although the EMR increases workload, it increases the number of side-effect records and is useful in improving safe medical care by ensuring proper use of medications.

Recently it has become easy to use medical electronic record systems thanks to the improvement in the hardware performance and user interface. However, in order to further medical electronic record systems, it is necessary to improve user interfaces so that it can take advantage of computerization and support operations reflecting the user's intension. We therefore developed an effective interface for graphical finding reporting in the cardiac catheterization using hand-written sketches. In addition, to cooperate with other systems, we developed an algorithm that can extract semantic information from the graphical representation and store it in XML format. In this system, the user can easily record the position and degree of a stenosis on a template of coronary schema. The system can also generate a table in the format specified in AHA Committee Report. This system is useful not only as a tool for efficiently generating finding reports but also as an effective explanation tool for patients.

In this study, we present a new method for constructing panorama images from a series of coronary cineangiograms. The obtained panorama images enable us to observe a complete coronary at a glance. In order for medical staff to take angiocardiographic images, one often has to move a patient's bed in order to catch the vessels to be visualized in the viewing field after injecting a contrast medium. Moreover, those vessels are usually inconsistently contrasted because it takes time for the contrast medium to reach the end of every vessel. In this study, we present a method that reduces such inconvenience and constructs useful coronary images for clinical diagnosis. First, we roughly estimate the heartbeat cycle and the bed movement using a spatiotemporal image generated from a series of original coronary cineangiograms. Next, we group each angiocardiogram frame based on heartbeat phase and estimate the bed movement spatiotemporally. Panorama images can then be obtained by merging a few frames in the same heartbeat phase with a composite background image. This makes it possible to easily observe a contrasted, complete coronary in a still viewing field.

For the purpose of electronic processing of medical records, we have developed a pen-based interface to realize the flexibility of conventional hand-written patient records in a natural form. This pen-based interface has effective functions not only to support the data entry but also to survey the input data. In fact, however, we had a serious problem because intact hand-written characters input by a pen device are too hard to be utilized for secondary data use. In this paper, we have discovered the possibility of using secondary data from our pen-based interface by using advanced functions of a box-free character recognition and a hand-written query search. With a box-free character recognition function, we can easily transfer hand-written characters into text data whenever the need arises. Moreover, after inputting a series of hand-written data, the text data, which can be used on a secondary level, is also made possible by the process of conversion in bulk. Additionally, we can find the targeting characters in the whole of the hand-written characters with the function of the hand-written query search, even if it does not make a specific database structure.

Some criticisms, such as input difficulty, security, and so on, have been pointed out as flaws in systems of electronic medical records. We have been long engaged in development of the pen-based interface to improve these flaws and promote the use of electronic medical records. However, the evaluation of effectiveness of this interface has not been performed thoroughly enough yet. We did conduct a survey using questionnaires, which asked 244 people about this interface in seminar. Responses were obtained from 82 people. Of these people, 67(81.7%) agreed that the assistant functions in this interface were useful. Particularly, the pen ID function scored higher than all other assistant functions. As well, 57(69.5%) people answered that our pen-based interface would promote the use of electronic medical records. Wilcoxon's signed ranks test showed that the pen-based interface is significantly easier to use than a keyboard interface, especially for outpatient consultation (p<0.01). It turned out that most of the respondents expected that the pen-based interface could have potential to improve shortcomings in the systems of electronic medical records. These results suggest that the pen-based interface will be effective in popularizing electronic medical records.

Multielectrode cardiac mapping is used for clarifying myocardial activation sequences in an experimental study for cardiac arrhythmias or even in such a clinical examination. To visualize the activation propagation as an isochronal map or movie, the determination in manual of a series of activation time in every electrode is indispensable, however it is an extremely hard work, and moreover it is often difficult to be executed objectively and accurately, especially in visualizing the complicated arrhythmias such as atrial fibrillation. We developed a new algorism to automatically visualize the activation propagation from the multielectrode mapping data. Utilizing this algorism enabled us to clearly visualize even the complicated propagation, and also to clearly visualize the abnormal conduction, which becomes an important clue for clarifying the arrhythmia mechanisms. Clinically, this algorithm might possibly help the ablation therapy by quick identification of the origin of the arrhythmias.

We present an electrophysiological heart simulator with schematic 3-D modeling interface. It has been tedious and troublesome interactive work to create a shape of heart for using it in the simulator. In our developing system, as soon as changing the shape of heart easily and quickly, we can execute the simulation and the visualization of the result. Our system increases the facility of computer simulation.

This paper deals with a method of registration and superimposition of a coronary arterial tree on a myocardial SPECT (Single Photon Emission Computed Tomography) image. We can grasp the myocardial function more easily in connection with the shape of thc. coronary arterial tree. The superimposed image is easily obtainable through some manual pointing on coronary angiograms (CAG) followed by an automatic matching method: First, a rough shape, model of left ventricle is estimated by using SPECT data, This 3-D left ventricular model is projected on a pair of bi-plane CAG images. We can obtain two 2-D coronary images on bull's eye map by scanning the left ventricular surface projected on CAG. By maximizing a matching degree between two 2-D coronary images, registration between CAG and SPECT is performed. Finally the superimposed image is obtained by integrating two 2-D coronary images and bull's eye image of SPECT, We validated our method by numerical experiments with artificial data set and also applied it to two clinical data sets.

We have reported[1] at the former CiC conference about a method of superimposing a coronary arterial tree on a bull's eye map of SPECT In the method, the left ventricular epicardial surface model (LV surface model) was constructed from the SPECT data. Coronary images corresponding to the bull's eye map was obtained by mapping bi-plane coronary angiograms (CAG) onto the LV surface model. This paper reports that we improve the matching degree and apply the Powell method for parameters estimation. Furthermore, we improve the procedure for generating resultant images. These improvements enable us to automate several processes that required interactive procedures in the former work. We validated the improved method by numerical experiments with artificial data set and applied the method to clinical data set successfully.

This paper deals with a method of generating a CAT's Eye Image, in which the coronary arterial tree is superimposed on bull's eye map of SPECT. We can grasp the myocardial function more easily in connection with the shape of the coronary artery. The CAT's eye image is easily obtainable through some manual pointing on coronary angiograms (CAG) followed by an automatic matching method: First, a rough shape of the left ventricle (LV) is estimated by using SPECT data. This 3-D LV model is projected on a pair of bi-plane CAG images. By scanning the LV surface projected on CAG, we can obtain two temporal 2-D coronary images on a bull's eye map. By maximizing a matching degree between two temporal images, registration between CAG and SPECT is performed. Finally the CAT's eye image is obtained by integrating two temporal 2-D coronary images and bull's eye image of SPECT. Our method has been successfully applied to one artificial data set and several clinical cases.

To assist in understanding a relationship between stenotic change in coronary arteries and change in myocardial function, we are developing a method of registration and superimposed display of the coronary arterial tree on a myocardial SPECT(Single Photon Emission Tomography). Firstly we introduce a fully 3-dimensional registration and display method. A 3-dimensionally reconstructed coronary arterial tree data is also transfered to a bull's eye map representation. We also introduce a semi 3-dimensional method. By this method, original 2-dimensional coronary angiograms are directly, i.e. without 3-dimensional reconstruction, transfered to a bull's eye map representation by referring a manually defined left ventricular long ads on angiograms and referring a epicardial surface data obtained from SPECT.