INTRODUCTION

Digital radiography has been reported to provide the same diagnostic quality as conventional film techniques.1–4 Two types of digital imaging techniques are known: 1) direct radiography with CCD—(Charge Coupled Device) and CMOS—(Complementary Metal Oxide Semiconductor) detectors and 2) semi-direct radiography with storage phosphor image plates (SPIP), also known as photo-stimulable phospor (PSP) image plates.5–6

Direct digital radiography systems use solid-state sensors. In the early years of digital radiography, CCD-sensors were introduced.7–8 In recent direct digital radiography, the efficiency of CCD sensors was improved through the conversion of x-rays by a scintillation layer into light, which is detected by an optical sensor and converted into an electronic signal. The technology of CCD and CMOS sensors is basically the same, and each pixel can be addressed individually. Yet, in CMOS chips, controlling conversion of the photon energy into an electronic signal following conversion into the digital signal is incorporated into the sensor, which means that the image is processed in the sensor where every pixel has its own analog-to-digital converter. Whereas in CCD technology, the electronic signal of each pixel leaves the sensor and is then digitized in a processor to create the digital image.59

Reusable photo-stimulable phosphor-coated imaging plates (SPIP) store the latent image as a distribution of electron charges from x-ray emission. The energy of the electron charges is emitted as light by stimulation with a scanning helium-neon laser beam. The light is directed to a photomultiplier tube and the output electrical signal is converted into a digital image.10

Under clinical conditions, the visibility of contrast differences is essential.911–15 For example, the visibility of an endodontic file and of the filling material in the root canal or the assessment of approximal caries are of clinical interest.16–29 Recently, in addition to metal root posts, fiber-reinforced composite root posts are used, which should be radiographically visible.30–34

Digital image systems and conventional x-ray films should provide a high contrast resolution to enhance radiological diagnostics. The better the contrast resolution, the better the adjacent structures of different density and thickness that are displayed by appropriate gray shades. Therefore, the bit depth of the radiographic images is decisive. The bit depth quantifies how many unique shades are available. Sixteen bit images offer 65,536 gray shades compared to the 256 gray shades of 8 bit images. The higher the bit depth, the more gray shades are available for storage.35–41

The current study compared the gray value differences between dentin and root posts made of titanium or FRC in anterior and posterior teeth radiographed using six digital intraoral systems and conventional x-ray films. The hypothesis was that the gray value differences of root posts may be affected by the material of the root posts and tooth location. Furthermore, different digital intraoral systems and conventional x-ray films are expected to provide comparable gray value differences of root posts.

METHODS AND MATERIALS

Ten human teeth, five molars from the mandible and five upper incisors, extracted due to periodontal disease with progressive bone loss, were endodontically-treated at the Periodontal Department (University of Hamburg) and used for each radiographic system. A post-space preparation was performed in the canals of the incisors and the distal canals of the molars using the Erlangen Post System (ER-System, Komet/Brasseler, Lemgo, Germany) to a length of 9 mm and a diameter of ISO size 90. The root posts had a diameter of 0.689 mm at the tip and an angle of elevation of 2.1°. Titanium or fiber-reinforced composite (FRC) root posts (ER-System, Komet/Brasseler) were placed into the post spaces without cement.

The sets of five incisors and five molars were each positioned in a row in a block of silicone (Silagum-Putty Standard, DMG, Hamburg, Germany), which was used as a soft and hard tissue equivalent.715 The silicone blocks were placed on a bench 10 cm from the end of an x-ray tube with the long axis in the middle of the bucco-oral distance of the radiographed tooth. The position of the silicone blocks and each tooth were marked to allow for reproducible images. The radiographic sensors, SPIP or films (Table 1) were placed in a slot approximately 3 centimeters behind the middle of the bucco-oral distance of the radiographed tooth. The alignment of the entire appliance is displayed in Figure 1.A0.4 mm-thick (10 cm x 10 cm) aluminum shield was fixed in the appliance 6 cm from the end of the tube as a soft tissue equivalent. The combination of silicone block and aluminum shield was pretested to be equivalent to the effect of human soft and hard tissues on the contrast. Additionally, a step-wedge of the aluminum alloy AlMgSi1 was fixed in front of the sensor slot,42 so that it was displayed above the respectively radiographed tooth.42 This allowed for the calibration of a constant reference gray shade on all images. The alignment of the entire appliance fulfilled the following requirements: the tubes were fixed at a right angle to the imaging devices. The central x-ray went through the middle of the root of the respective tooth using a parallel technique. The appliance allowed for a reproducible arrangement of various radiographic generators, teeth and imaging devices for repeatable images with different post materials in the root canals.

View Figure

• Download as Powerpoint
• Download Figure

Table 1 Radiographic Systems, X-ray Generators and Exposure Times

View Fullscreen

Six digital radiographic systems, four sensor systems, two SPIP systems and a conventional dental F-speed film (control group) were used in the current study (Table 1). Three x-ray generators had to be utilized due to the fixed setup of some systems recommended by the manufacturers. Heliodent DS was used with 7 mA radiographic current and 60 kV tube voltage and included the following parameters: 0.7 mm focal point, 2 mm Al-filtration, 203 mm between the focal point and the end of the tube, and 331 mm between the focal point and the detector. Gendex 765 DC was used with 7 mA radiographic current and 65 kV tube voltage; the following parameters were given: 0.4 mm focal point, 2 mm Al-filtration, 300 mm between the focal point and end of the tube, 428 mm between the focal point and the detector. Philips Oralix U3-DC was used with 10 mA radiographic current and 60 kV tube voltage; the following parameters were fixed: 0.8 mm focal point, 1.5 mm Al-filtration, 245 mm between the focal point and end of the tube, and 373 mm between the focal point and detector. The direct radiographic systems were mounted in separate rooms according to the manufacturers' recommendations: RVG 6100 with Gendex generator, Visualix eHD with Philips Oralix and XIOS, as well as Sidexis with Heliodent DS.28 The different x-ray generators were used in order to compare different radiographic systems under practical conditions.

The exposure times were determined by recommendations given by the manufacturers and previous experiments comparing image quality of the same object from low to high exposure settings.8 The Kodak RVG 6100 (Trophy Dicom 6.0.3.1, Eastman Kodak Company, Rochester, NY, USA) and Dürr (DBSWin V.3.3, Dürr Dental, Bietigheim-Bissingen, Germany) software of the systems offered a function that indicated the ideal exposure time.

Each of the 10 teeth was radiographed with titanium and FRC root posts and seven radiographic systems, resulting in 140 radiographs (Table 2). The storage plates were scanned at the highest given resolution mode. This was 40 lp/mm for the VistaScan System (Dürr Dental) and 600 dpi for the DenOptix System (KaVo Dental, Gendex Dental Systems, Des Plaines, IL, USA). After scanning, the SPIP were erased with Dürr ReSetter (Dürr Dental), which optically and acoustically signaled the end of the process.

Table 2 Study Design (Titanium and FRC Posts in Incisors or Molars Radiographed with Seven Radiographic Systems)

View Fullscreen

The images were exported in uncompressed TIF-Format (Tagged Image File-Format) for evaluation. The systems RVG 6100, Sidexis, Visualix eHD and DenOptix allowed the export of 8-bit images. XIOS and VistaScan images were exported in 16-bit format. The conventional dental films were processed by a fully automatic developing device (Dürr XR 24-II, Dürr Dental) using standard chemicals and the recommended time and temperature settings. The conventional x-ray films were then scanned using a scanner with a transparency unit (Epson Expression 1680 pro, Seiko Epson Corp, Nagano, Japan) and Silver Fast Ai v6.0.1r16 software (LaserSoft Imaging, Kiel, Germany). The conventional x-ray films were saved in 16-bit TIF-format with a 1600 dpi resolution. This resolution corresponded to the spatial resolution of 22 lp/mm of the dental films (product information by Eastman Kodak Company). Theoretically, the spatial resolution could be higher, but it was not measurable. In the literature, a spatial resolution of >20 lp/mm is described for the Kodak system. 1536

For each system, the gray shades of the stepwedge on each image were compared to ensure similar contrast. Only the gray shade of the 3-mm thick step of the aluminum stepwedge was used for calibration. The absolute gray shade value of this aluminum step was assessed by a graphics editing program (Adobe Photoshop 7.0, Adobe Systems Incorporated, San Jose, CA, USA) on all 10 images of the respective radiographic system. When different, the gray shades of the stepwedge were numerically set in accordance with a referential image of the third molar in line. The third molar with no post in the endodontically-treated root canal was radiographed in a pretest with each of the radiographical systems and functioned as a referential image. Calibration of the images was a means of checking the reproducibility of the image contrast. Images usually had the same gray value for the steps. In very few cases, gray values had to be adjusted, and if so, they were in 1–5 gray value differences and no visual change in contrast was affected. Then, the same referential stepwedge area of all images of the same radiographic system had the same numerical gray shade value. In previous trials, it was found that not all radiographic systems could be brought to the same referential gray shade value, because the variations of the rest of the images would have been too extreme and unrealistic for image analysis. Prior to any further evaluation, every image was checked by an examiner to provide clinically acceptable diagnostic quality.

The calibrated images were saved as TIF images. The scanned film, Dürr VistaScan and XIOS images, were converted into 8-bit images by the same graphics editing program (Adobe Photoshop 7.0) to provide the same 256 gray scales as the other systems. Additionally, the images were inspected by three radiographically-experienced examiners, and their subjective opinions regarding the image quality were summarized. All the images could be used for the computer-assessed measurements. Moreover, their subjective impressions were compared with the objectively assessed values in order to determine which measurable gray level difference was visible to the human eye.

The region of interest (ROI) of the current study was the root post and the surrounding dentin. The absolute gray values of the root posts in six different heights (six measurements were evenly distributed in the long axis over the part of the post 9mmin length positioned in the root canal and in six corresponding heights of the surrounding dentin on the same horizontal level in the middle of the dentin between the root canal wall and the outer wall of the root, Figure 2) were measured on the calibrated images using the ImageJ 1.37 v software (Wayne Rasband, National Institutes of Health, Bethesda, MD, USA). The gray level values of single pixels were estimated at these locations. The measurements were repeated three times, hitting different pixels, and the six corresponding measurements offered a large number of data to aid in avoiding failures. Furthermore, wrong measurements would have been found during the plausibility check or by large standard deviations during the final calculations. Neither occurred.

View Figure

• Download as Powerpoint
• Download Figure

The calculation of gray value differences was computer-assisted and came from the six corresponding value pairs of each radiograph. If the gray value difference was positive, the root post material was more radiopaque than dentin; and if the gray value difference was negative, it was less radiopaque than dentin. From 140 images, 840 gray value differences were assessed.

Statistical analyses of the results were performed with a 7 (radiographic systems) by 2 (post materials) by 2 (tooth location) three-way analysis of variance (ANOVA, α=0.05) to evaluate interaction effects among these independent variables. One-way ANOVA and Bonferroni-Dunn's multiple comparison post-hoc analyses of the gray value differences were conducted for the test groups (α=0.05).

RESULTS

Three-way ANOVA revealed significant effects for radiographic systems, post material, tooth location and the interactions of “radiographic system x post material,” “radiographic system x tooth location,” “post material x tooth location” and “radiographic system x post material x tooth location” on the gray value differences (Table 3) (p<0.0001, p=0.0053). Therefore, radiographic systems, post material, tooth location or combinations of these independent variables showed a significant effect on the gray value differences.

Table 3 Three-way ANOVA for Effects and Interactions of Radiographic Systems, Post Materials and Tooth Location

View Fullscreen

Bonferroni-Dunn's post-hoc analysis showed statistical differences of gray value differences between groups (Figures 3–4). Comparisons in these figures are not significant unless the corresponding p-value is less than 0.0001.

View Figure

• Download as Powerpoint
• Download Figure

View Figure

• Download as Powerpoint
• Download Figure

Titanium root posts showed positive gray value differences. The mean values ranged from28 to 85 (Figure 3); whereas for FRC-posts, the mean gray value differences of 4 to 27 were obtained (Figure 4).

Significantly higher gray value differences of titanium and FRC posts were found in anterior teeth, when compared with molars, for XIOS, Sidexis and Visualix digital intraoral systems (p<0.0001, Figures 3–4). The radiographic systems of RVG (p=0.0018, p=0.0219), DenOptix (p=0.0037, p=0.014) and conventional x-ray film (p=0.0131, p=0.17) showed no significant differences between anterior and posterior teeth for titanium and FRC posts. For the VistaScan digital intraoral system, a significant difference was obtained for titanium posts (p<0.0001, Figure 3); whereas for FRC posts, no significant difference was found for anterior teeth when compared with molars (p=0.0409, Figure 4).

With the exception of DenOptix with incisors (p=0.021), x-ray films (control group) showed significantly lower gray value differences of titanium posts in incisors and molars compared to the corresponding digital radiographs (Figure 3). Significantly higher gray value differences were observed for FRC posts in incisors radiographed with XIOS, Sidexis and Visualix systems when compared to conventional x-ray films (p<0.0001, Figure 4). No significant differences were noted between FRC posts in molars radiographed with conventional x-ray films or with digital systems (p=0.0145 and p=0.6359, respectively).

XIOS, Sidexis and Visualix showed significantly higher gray value differences of titanium posts in incisors than did the other digital image systems (p<0.0001). Gray value differences of Visualix were significantly higher than that of Sidexis (p<0.0001), but exhibited similar gray value differences as the XIOS system (p=0.0212). The XIOS system with titanium posts in molars showed significantly higher gray value differences than the other digital intraoral systems (p<0.0001).

Gray value differences of FRC posts in incisors radiographed with VistaScan were significantly lower than from the XIOS and Visualix systems (p<0.0001). No significant differences were noted among the other digital systems. For FRC posts in molars, no significant differences were obtained for the tested digital systems.

All of the tested digital systems and conventional films showed significantly higher gray value difference for titanium posts compared to FRC posts, independent of tooth location (incisors or molars) (p<0.0001, Figures 3–4).

The subjective ability to differentiate root posts from dentin could not be determined with a special gray value difference value. In general, posts were visible in the root canal when gray value differences exceeded 40. Posts with gray value differences lower than 40 were not visible on all radiographic images. Posts with the same gray value differences below 40 were visible on some images but not on others.

DISCUSSION

Based on the current results, the hypothesis that the material of the root posts may affect gray value differences compared to the surrounding dentin was accepted for all of the tested digital intraoral systems and conventional x-ray film. The hypothesis that gray value differences of root posts is affected by tooth location was accepted for XIOS, Sidexis and Visualix systems for titanium and FRC posts and VistaScan with titanium posts and rejected for RVG, DenOptix and conventional x-ray film with titanium and FRC posts and for VistaScan with FRC posts. The hypothesis that different digital intraoral systems and conventional x-ray films are expected to provide comparable gray value differences of root posts was accepted for DenOptix (titanium posts, incisors), DenOptix and VistaScan (titanium posts, molars), RVG, DenOptix and VistaScan (FRC posts, incisors) and all of the tested digital systems with FRC posts in molars when compared to the corresponding conventional films. The hypothesis of similar gray value differences was rejected for RVG, XIOS, Sidexis and Visualix (titanium posts, incisors and molars), VistaScan (titanium posts, incisors) and XIOS, Sidexis, Visualix (FRC posts, incisors) compared to conventional x-ray films.

In general, significant gray value differences depended on the post material, rather than on the radiographic system or tooth location.

Theoretically, the x-ray generator might have had an influence on the image quality for digital radiography;1 however, different x-ray generators are used together with their respective radiographic system in dental practice.28 No effect regarding the different x-ray generators was noticed, although comparing this was not the aim of the current study. Further evaluation may address the influence of the x-ray generators.

Digital radiography has to fulfill the requirement of revealing a root canal filling or a root post in the root canal, especially in Endodontics. Root posts are made from various materials and, recently, the FRC material is favored in some studies.33–34 FRC material is composed of glass-fibers and a resin matrix, thereby, the radiopacity of FRC posts was found to be lower than that of titanium posts.2931–32 The radiopacity of FRC root posts has been found to depend on the FRC materials.19–20

Instead of measuring the radiopacity of the respective material, the current study compared the gray value differences to dentin of different post materials with different digital radiographic systems, simulating the clinical setting. Previous studies used similar test assemblies with dried mandibles110 or with single teeth restored with composite restorations with overhangs and voids.17–18 The advantage of such a test design was to guarantee the reproducibility of the setup. Furthermore, simulation of the clinical setup was used, because the various radiographic systems had to prove their performance under clinical conditions independent of technical differences. A clinical setup includes modifying factors, such as hard and soft tissues, tooth anatomy and the radiopacity of materials.

The current study revealed that a gray shade difference as the subtraction of the gray values of the FRC post material and the dentin does not ensure visibility of the root post. Therefore, it has been suggested that a gray level difference of at least 40 is required to guarantee visibility. The only comparable study by Soares and others,31 using a similar method, confirmed the current results but gave no gray level difference to confirm visibility.

The type of radiographic system had no effect on post visibility. Similar findings were described by Tagger and Katz,19 who found that a certain value equivalent of aluminum did not guarantee the radio visibility in a clinical setting.

The results of the scanned conventional x-ray film might be worse than the evaluation of the actual film on a light projector. Some authors accepted scanning of the films to gain digital information;192326 whereas Goga and others found that the clarity and diagnostic quality of the conventional radiographs were superior to the scanned images.12 Theoretically, scanned images of conventional film offer the same information,5 but clinicians argued that the “plasticity” is missing during observation.36 The diagnostic quality of scanned conventional films is generally accepted when high quality medical scanners are used.192326

In accordance with Rasimick and others, the present results showed that gray value differences depended on the radiographic system used.24 For the XIOS, Sidexis and Visualix systems, significantly higher gray value differences were obtained with titanium and FRC posts in incisors compared to conventional x-ray films. With the exception of DenOptix and VistaScan, digital radiographic systems showed significantly higher gray value differences of titanium posts in molars when compared to x-ray films; whereas similar results were found for all digital systems and conventional x-ray films with FRC posts in molars.

In accordance with Farman and others, significant gray value differences depended on the post material, rather than on the radiographic system or tooth location.27 This could be explained by the fact that radiographic visibility was more influenced by the material than by the radiographic system. Therefore, materials with a low radiopacity might result in low gray shade differences from one radiographic system and no gray shade differences from another radiographic system. In the current study, the RVG system was the only system that showed the space between the root post and the root dentin, which made FRC posts with low radiopacity more visible. It might be assumed that this result is an effect of the automatic image processing of the RVG system that could not be switched off. Otherwise, the reason might be the high spatial resolution of RVG, which is higher than that of the other systems, making the narrow cement gap of about 35 μm visible. Farman and others also measured a high spatial resolution for the Kodak RVG 6000 system (>20 lp/mm).36

The contrast difference between SPIP and conventional film was found to be weaker than that of the solid-state detectors, which might be explained by the different image technique regarding SPIP and the scanning process regarding the film; whereas no significant differences were found between CCD and CMOS detectors, yet the results within the CMOS group (RVG and XIOS) varied.9

CONCLUSIONS

Considering the current results, the following can be concluded:

1. The current study demonstrates that all of the tested digital systems and conventional films showed significantly higher gray value differences for titanium posts when compared to FRC posts, independent of the tooth location;

2. In anterior teeth, when compared with molars, significantly higher gray value differences of titanium and FRC posts were found for XIOS, Sidexis and Visualix digital intraoral systems but not for RVG, DenOptix and VistaScan (FRC posts). Yet, the gray shade differences between FRC post and dentin were found to be too low;

3. Except for the FRC posts in molars, XIOS, Sidexis and Visualix digital radiographic systems showed significantly higher gray value differences than conventional x-ray films.