Collaboration:Shriner

Computer Assisted, Self Regulated, Embedded Microscope for in vivo cardiac observation

Members

• Shingo Yasuhara, M.D., Ph.D. (Shriners Hospital and Massachusetts General Hospital)
• Nobuhiko Hata, Ph.D. (Brigham and Women's Hospital)
• Junichi Tokuda, Ph.D. (Brigham and Women's Hospital)

Interaction with NA-MIC

This program is support by image processing tools kits available from National Alliance for Medical Image Computing (PI: Kikinis). NA-MIC investigators, in particular Dr. Steve Piper, Ron Kikinis, provides expert consultation on the applicatoin of Slicer and its associated Engineering tools for development.The software developed in this project is available as BSD licensed open source software.

Abstract

Millions of people die each year in the US from heart attacks, arrhythmias, and other cardiac diseases. Despite its potential scientific value, microscopic observation of the beating heart in live animals is still difficult due to the vigorous nature of the cardiac movement. This grant proposal is for a pilot project that will ultimately extend into the construction of a completely new modality of in vivo microscopy that incorporates opto-mechanical feedback regulation and image processing to capture motion-corrected microscopic video images of the heart. Previously, we constructed an ultra-video-rate scanning in vivo confocal microscope (“Antarctica-I”), to investigate the pathophysiological mechanisms of skeletal muscle diseases. These studies demonstrated that in vivo microscopy can provide information that was missing from conventional post-mortem histological analyses and thus fill the gap between cell biology approaches and animal experiments.

If one can observe the living heart under the microscope in a similar manner, it will allow detailed microscopic observation of the progress of coronary arteriosclerosis, post-ischemic arrhythmia, and post-infarction remodeling process, among many other potential applications. It will also improve the feasibility of observing other moving organs including the lung and the gastrointestinal tract in animal studies. Technology developed in this proposal will also be clinically applied to the construction of a prototype self-regulated endoscopic microscopy, and to the monitoring of patients’ cardiopulmonary microcirculation during surgeries. This proposal brings together two research groups with unique strengths in (1) optics, electronics, microscopy and anesthesiology, and in (2) mechanical engineering, image processing, and robotics.

Specific Aims

Specific Aim 1 will examine whether fluorescent staining of the sternomastoid muscles of live mice will provide microscopic images with jiggle-motion at similar frequency to that of the heart beat. In this pilot study, we will primarily use in vivo images from sternomastoid muscles (murine equivalent of sternocleidomastoid muscles), though cardiac muscle observation will be pursued in parallel. The reason for selecting this muscle is, (i) we have already developed expertise in staining skeletal muscles by various fluorescent biomarkers and (ii) sternomastoid muscles show jiggling motion even under anesthesia due to respiratory and pulsatory effects. (iii) The frequency and extent of jiggle motion of sternomastoid muscles can be easily controlled by adjusting the tidal volume and the speed of mechanical ventilation as well as the pulse rate. Images thus captured from this muscle will serve as fundamental map when designing the feedback algorithm in the following specific aims.

Specific Aim-2 will examine (1) whether an off-line image processing method can extract parameters of jiggle-motion of the sternomastoid muscles and thus (2) whether the video image can be stabilized. Based on our previously established algorithm developed for machine vision-guided robotic needle control, a new scheme for muscle target detection and motion guidance will be designed by level-set-based segmentation and active contour model with specific modifications to suit the movement of sternomastoid muscles.

Specific Aim-3 will test the accuracy of the motion-correction by automated quantification software. We have developed an automated computer-assisted image analysis method of muscle microcirculation using the custom-developed software, KEIO-IS2. This method, however, is susceptible to jiggle-movement of the tissue. Therefore, we will feed the motion-corrected images from Specific Aim-2 into KEIO-IS2, and evaluate the accuracy of quantification of microcirculatory parameter by comparing to manual counting methods.

Specific Aim-4 will test the feasibility of using digital signal processor (DSP) for a pre-programmed movement of the servo mirror system in Antarctica-I. By using the fundamental architecture of Antarctica-I, we will install additional servo mirror and a prototype DSP board in the system. A pre-programmed test signal will be generated by the DSP and fed into the drivers of the X- and Y- servo mirrors to move the field of view within the target sample. This simple experiment will confirm the mechanical motion speed and dynamic motion range required for image-correction algorithm generated in Specific Aim-3.