Dale A. Miles BA, DDS, MS, FRCD(C)
Diplomate, ABOMR

Introduction
Life was simple in 1974 when I graduated from dental school. Film was the only receptor I could use and I didn't need computers. "DQ" or Dairy Queen was a wildly popular soft ice cream and a spot where I could take my sons just for the treat without prizes, rides or game pieces to make our lives more exciting. The excitement was the product - a plain, soft, vanilla ice cream in a plain, beige cone. WOW, HAVE THINGS EVER CHANGED! Imaging choices have certainly changed, not the least of which is the way we evaluate an imaging receptor or imaging system. Let me explain.

"DQE" (Detector Quantum Efficiency" and "System Performance"
If you've been following along on my site, you know that the sensor is NOT the only important piece of a digital x-ray system. In my last article (please read it first ("The Digital Imaging Chain"), I described several important pieces of the electronic imaging systems that you MUST consider to "Go Digital". In a film-based system, there were "parts" as well, but they basically were the film and the processing, the latter being the part that often was poor in our offices. In a digital system, we have many more "parts" as I described previously. Before I continue, let me explain the concept of "Signal-to-Noise Ratio". Don't worry, I'm a simple guy and I like to make my explanations simple too!

SNR (Signal-to-Noise Ratio)
For any given diagnostic imaging task, we are trying to detect the "signal". For example, if we're looking for an interproximal carious lesion or cavity…THAT is the signal. All of the rest of the image information is the "noise". The noise can be "structured', such as anatomy (the crown, the root, the pulp canal) or "unstructured" - random x-ray photon exposure that degrades the image and clouds the picture. The "unstructured noise" is unwanted. We like to see detailed anatomy. But in reality when looking at bitewing x-rays, for cavities between teeth, we are looking at a very small region and trying to find a very faint, low contrast shadow called the "early carious lesion". This is NOT an easy task! High contrast film would help, and has, but we also want to see bony detail like alveolar crest change and subtle PDL widening. These are very different tasks, but we expect our x-ray film to show it all to us. The old, standard measure we used to determine if a system was "good" enough to show these carious lesion was the spatial resolution - usually expressed in line pairs/millimeter (LP/mm). And sophisticated radiographic studies looked at the MTF (Modulation Transfer Function - don't even ask). For digital systems, these standards are not good enough. What we want in a system is a good SNR.

Signal = Useful image information
Noise      Random information

Thus, a high signal-to-noise ratio (high SNR) results in low system noise. This is the best way to capture the useful, diagnostic information. In a digital system, the best way to compensate for a low or poor SNR is to increase the exposure time and thus the patient's dose. This is NOT good.

DQE measures digital x-ray image quality as the "combined effect of noise and contrast performance of an imaging system, expressed as a function of object detail". GE Medical Systems

If a system has very low noise and excellent contrast performance, the system will then be capable of detecting low contrast objects such as earl, interproximal carious lesions. According to the web site author(s) at GE medical Systems, DQE is more important to detecting small objects than is spatial resolution for many imaging tasks. As dentists, we deal with "small". We're always thinking in terms of millimeters and fractions of millimeters, even microns in many instances.

We need good digital systems with superior DQE to help us detect the disease entities we treat such as caries, early periodontal bone loss and periapical changes.

Figure 1
Low contrast image on left makes it difficult to see carious lesions on teeth #29 and #30. High contrast image in image on the right allows us to see the caries spread along the DEJ on the distal of #30. The signal was improved just by altering contrast alone. However we degraded the area on the mesial and distal root surfaces of several teeth such as #4, 5, 28 and 29 and made the "cervical burnout" look like root caries. This is the "trade-off" in order to detect the signal called the cavity. A good detector with high DQE and wide scale of contrast would solve this processing problem.

Please click on link to go to GE Medical Imaging site to see
Figure 2 - an image of a contrast-detail phantom

Conclusion
If the DQE of a system is the ultimate measure of imaging quality of the digital imaging sensor, we need to see those specifications in the manufacturer's product monographs, and we need to see some studies from oral and maxillofacial radiologists to compare these imaging systems using measures of spatial resolution, contrast resolution, MTF and DQE so that we can make more definitive choices or even suggestions to the manufacturers to maximize their system performance. Boy, we've come a long way from "soft ice cream" in imaging haven't we?
Hope this info helps you as you "go Digital"!

Reference
Boone JM: "Spectral modeling and compilation of quantum fluence in radiography and mammography," in Physics of Medical Imaging, James T. Dobbins III, John M. Boone, Editors, Proceedings of SPIE Vol. 3336, 592-596, (1988).