Comparison Between Differently Priced Devices for Digital Capture of X-Ray Films Using Computed Tomography as a Gold Standard: A Multireader-Multicase Receiver Operating Characteristic Curve Study

Abstract

Objective: Film digitizers are a specialized technology that is available for scanning X-ray radiographs; however, their cost makes them unaffordable for developing countries. Thus, less expensive alternatives are used. The purpose of this study was to compare three devices for digital capture of X-ray films: a film digitizer (US $15,000), a flatbed scanner (US $1800), and a 10-megapixel digital camera (US $450), in terms of diagnostic accuracy, defined as the area under the receiver operating characteristic curves and computed tomography as the gold standard. Materials and Methods: The sample included 136 chest X-ray cases with computed tomography confirmation of the presence or absence of pneumothorax, interstitial opacities, or nodules. The readers were six radiologists who made observations of eight variables for each digital capture of the X-ray films: three main variables to determine the accuracy in the detection of the above-mentioned pathologies, four secondary variables to categorize other pathological classifications, and one variable regarding digital image quality. Results: The receiver operating characteristic curves for each device and pathology were very similar. For the main variables, there was no significant statistical difference in diagnostic accuracy between the devices. For the secondary variables, > 84% of cases were correctly classified, even those that were classified with the lowest image quality. High accuracy was determined for the three main variables (0.75 to 0.96), indicating good performance for all tested devices, despite their very different prices. Conclusions: Choosing a device for a teleradiology service should involve additional factors, such as capture time, maintenance concerns, and training requirements.

Comparison between Different Cost Devices for Digital Capture of X-ray Films: An Image Characteristics Detection Approach

“…Evaluation of the same devices [21] in a clinical ROC curve statistical study did not find statistical significant differences in diagnostic accuracy. However, this study was more expensive and time-consuming: The study length was 8 months and included six radiologists, with a factorial design that required 2,448 observations which were analyzed through receiver operating characteristic curves [22,23] and statistical analysis of variance [24], which are highly complex epidemiological and statistical models to perform.…”

Compensation of the non-uniformity of back-lighting sources when digitising X-ray films for video-densitometric measures

“…This methodology has been tested on one only hardware: also flat-bed scanners and film digitizers could be considered; however other authors demonstrated how the here-used acquisition set up performs similarly to the others, having the advantage of being less expensive [9,10] . The last aspect is significant, considering minor healthcare centres which could take benefit from teleradiology implementation [20] .…”

“…An important evolution came in the seventies when radiographs have started to be acquired by means of a video-camera [7] , and successively with the advent of solid-state slide or image scanners, until nowadays when digitised radiographs are directly available. Nevertheless x ray film digitisation remains necessary when retrospective analysed are being performed [8] or electronic archives are being built, or for implementing tele-radiology in places where conventional radiology is still used routinely as in rural healthcare centres of the developing world [9] .…”

“…Different acquisition systems can be used: camera based ones, flatbed scanner, x-ray digitisers; up to now, none seems to have definitively overcome the other ones [9,10] , and all are affected by noise produced by the non-uniformity of the back-lighting source: Wenzel and Sewerin well outlined out the main sources of noise related to radiograph digitisation process [11] , while Audenino et al demonstrated the importance of the non-uniformity of the back-lighting source to the end of results repeatability: it produces the dependency of measured grey levels on the position and orientation of the image on the back-lighting system [12] . This phenomenon has been seldom considered, but may not be insignificant:…”

Digitizing an Analog Radiography Teaching File Under Time Constraint: Trade-Offs in Efficiency and Image Quality

“…High quality digital images of radiographs preserve diagnostic quality compared to the original hard copies [4][5][6] and are comparable to more expensive flatbed scanners and film digitizers [7]. More recent studies have shown that even smartphone camera pictures of radiographs are of sufficient quality for detection of the presence or absence of major abnormalities, and frequently allow for a specific diagnosis [8].…”