The use of endoscopic surgery is growing, in large part because it generally is safer and less expensive than conventional surgery, and patients tend to require less time in a hospital after this type of surgery. Industry experts conservatively estimate that about 4 million minimally invasive procedures were performed in 1996.
Endoscopes are used in a variety of surgical procedures. These include gastrointestinal, obstetrics/gynecology, urology, arthroscopic, and ear, nose, and throat.
The assessment of the optical performance of endoscopes often is difficult in a clinical setting. The surgeon depends on a high-quality image to perform minimally invasive surgery. Yet, assurance of properly functioning equipment not always is straightforward. Many variables in both the patient and the equipment may result in a poor image.
Equipment variables which may degrade image quality include problems with the rigid endoscope, either with the optics or the light transmission. As a result of optical loss from damaged fibers, the light cable is another source of uncertainty. Malfunctions of the charge coupled device (CCD) video camera also contribute to poor image quality.
By the very nature of its working environment, cleanliness of the equipment, especially the lens surfaces on the endoscope (both proximal and distal ends), is a particularly common problem. And, patient factors make the objective assessment of image quality even tougher. Large operative fields and bleeding at the operative site are just two factors that may affect image quality.
The lack of objective performance standards makes the evaluation of endoscopic equipment more difficult. With all these contributing factors, the purchasers of equipment previously made an essentially subjective decision about image quality. To combat this problem, we developed an instrument, the EndoTester, with integrated software to quantify the optical properties of fiber-optic endoscopes.
The EndoTester is a specialized optical bench for quantitative testing of the fiber-optic path and the lens system in rigid endoscopes (Figure 1). The tester consists of a standard PC containing an 8-bit, gray-scale, video digitizing board and custom software for easy acquisition and analysis of an endoscope’s optical properties.
The tester uses LabVIEW® from National Instruments to quantify the optical properties. An IMAQ PCI-1408 image acquisition board from National Instruments digitizes and preprocesses the images.
The optical tester allows you to perform a series of tests on the endoscope and its peripheral devices. Specifically, these tests include relative light loss, reflective symmetry, percent of lighted (good) fibers, geometric distortion, and modulation transfer function (MTF).
Relative Light Loss
The relative optical light-loss measurements quantify the degree of light loss from the light source to the distal tip of the endoscope. The relative light loss will increase with fiber-optic damage. Changes in the intensity of the light source or the condition of the fibers in the fiber-optic cable are normalized out of the relative-loss calculation. The relative light loss is determined directly from the endoscope’s light output and light input.
A simple means to quantitatively measure performance variation with respect to time is desirable for all endoscopes but appears to be particularly valuable for the evaluation of disposable endoscopes. These units may have a rated life of 30 uses, but measurements of performance change under actual operating, and sterilizing conditions are proof-positive.
The relative light loss for the optic fibers of the endoscope-under-test is calculated by:
The light-in value is measured by configuring the output connection end of the fiber-optic cable to a test fixture that holds the cable a fixed 2″ distance from a photometer. The light-out value is measured by the photometer illuminated by the endoscope-under- test.
Reflective symmetry is a measure of light amplitude in the endoscope’s field of view. This value quantifies the effective distribution of light.
By employing five magnitude comparators, the continuous illumination patterns can be transformed into five annular rings of decreasing gray scales (Figure 2). The pattern of gray rings produced by this test should be nearly centered on the image. The pattern is circular for zero-degree endoscope tip angles and sometimes elliptical for angled tips. A histogram graphs the number of pixels from each comparator output (light intensity).
For each ring displayed in the filtered image, you can see how many pixels exist for each band of intensity. To pass this test, the percent bright area must be greater than or equal to 50% of the maximum brightness. This equation defines how percent bright area is calculated; Gray 1 and Gray 2 refer to the two brightest (innermost) rings:
In addition to calculating the percent bright area of the field of view, the bore-sight error also can be measured. The bore-sight error is the difference between the center of the field of view and the center of illumination provided by the endoscope.
A grid of 0.2″ × 0.2″ squares, whose center square is solid black, is positioned in the center of the field of view. Combining this grid with the ring pattern shows the geometrical distance from the center of illumination to the center of the field of view. By counting the 0.2″ square grid, a quantitative measure of the bore-sight error is determined for the endoscope at a 2″ tip distance.
The CCD camera captures a close-in reflection (less than 0.25″ separation) of the distal end of the endoscope from a polished Lucite surface. This provides a record of the pattern of lighted optical fibers for the endoscope-under-test (Figure 3).
The number of lighted pixels will depend on the endoscope’s dimensions, the distal end geometry, and the number of failed optical fibers. New fiber damage to an endoscope will be apparent by comparing the lighted fiber pictures (and histogram profiles) from successive tests. Statistical data also is available to calculate the percentage of working fibers in a given endoscope.
In addition to the 2-D profile of lighted fibers, this pattern (and all other image patterns) also can be displayed as a 3-D contour plot shown in Figure 4. This interactive graph may be viewed from a variety of points so you can vary the elevation, rotation, size, and perspective controls.
The geometric-distortion test quantifies the optical distortion of the rod-lens system. The EndoTester’s video frame grabber captures the image of a square grid pattern.
Distortion is the change of magnification at points around the field of view with respect to the maximum magnification occurring at the center of the field of view. Measuring the diagonal length of the central square of the pattern, with respect to the diagonal length of any other square, determines the geometric distortion. The percent distortion is calculated by:
Figure 5 is the screen that is used to measure the geometric distortion of the endoscope-under-test.
Modulation Transfer Function
Aperture response is a universal criterion for specifying picture definition and other aspects of imaging-system performance. It can be used for film images, camera lenses, television camera imagers, receiver picture tubes, and the human eye. The aperture response is measured as a contrast ratio by square-wave contrast transfer function or with a sine wave using MTF.
The square-wave response data can be converted to equivalent sine-wave data by mathematical manipulation; however, the sine-wave method is the most direct approach. Variable-density film targets now are available as sine transmission targets so a direct measurement of MTF can be performed. Even a single-space frequency measurement can provide a good index of performance for an optical lens system.
The MTF of the lens system is measured at a spatial frequency of six cycles per degree of the apparent field of view. Measurements at this frequency are an accurate indication of good optical instrument performance when the MTF is high. As a result, a single spatial frequency on the test target provides a good quantitative index of the local spot performance of the endoscope’s lens system.
Typically, the center region of a lens has the best optical performance, and the edges of the lens have less visual sharpness. Consequently, the MTF test measures performance at the left and right edges, the top and bottom edges, and the center of the lens (Figure 6). Studies have shown that your perception of image quality can be correlated to high values of MTF.
The EndoTester performs objective measurement of endoscopic performance. While measuring parameters of scope performance can facilitate equipment purchase, the greatest potential for this system is its role in preventive maintenance.
Endoscopes usually are removed from service and sent for repair when they fail in clinical use. The problem is difficult to diagnose and alleviate because an endoscope may be adequate in one procedure but fail in the next. Objective assessment of endoscope function may eliminate some of these problems.
The authors express their gratitude to Hartford Hospital’s Departments of Research, Surgery, and Biomedical Engineering for their support and direction in developing this system. Special thanks go to Curt Youndahl and Dan Gugliotti for their efforts in the design and fabrication of the optical test fixture.
1. Campbell, J., “Snapshot for LabVIEW for Windows,” Viewpoint Software Solutions, 1994.
2. Inglis, A., Video Engineering, McGraw Hill, 1993.
3. Kutzner, J., Hightower, L., Pruitt, C., “Measurement and Testing of CCD Sensors and Cameras,” SMPTE Journal, pp. 325-327.
4. “Measurement of Resolution of Camera Systems,” IEEE Standard 208, 1995.
5. “Surgical Video Systems,” Health Devices, Vol. 24, No. 11, Nov. 1995, pp. 428-457.
6. Walker, B., Optical Engineering Fundamentals, McGraw Hill, 1995.
About the Authors
Eric Rosow is director of the Biomedical Engineering Department at Hartford Hospital and a co-founder of Premise Development. He has more than 12 years experience in biomedical engineering and life-science applications of virtual instrumentation. Mr. Rosow earned a B.S. degree in mechanical engineering from Trinity College and an M.S. degree in biomedical engineering from the Hartford Graduate Center/Rensselaer Polytechnic Institute.
Finton Beatrice is an engineer who volunteers his time to the Biomedical Engineering Department at Hartford Hospital. Before his retirement, Mr. Beatrice was an electrical designer of aircraft control systems and sensors at Hamilton Standard for 30 years. He holds B.S.E.E. and M.S.E. degrees.
Hartford Hospital, 80 Seymour St., Hartford, CT 06102, (860) 545-3915, e-mail: [email protected].
Joseph Adam is a co-founder of Premise Development with more than 12 years experience in software and product development. He received a B.S. degree in mathematics and computer science, coordinated with engineering, from Trinity College; an M.S. degree in computer science from Rensselaer Polytechnic Institute at the Hartford Graduate Center, and advanced degrees in mathematics, computer science, and software engineering. Premise Development, 36 Cambridge St., West Hartford, CT 06110, (860) 673-0484, e-mail: [email protected].
Copyright 1997 Nelson Publishing Inc.