8
Vol. 78 No. 2 August 1994 OiUAL AND MAXILLOFACIAL RADIOLOGY Editor: Allan G. Farman Intraoral computed radiography using the Fuji computed radiography imaging plate Correlation between image quality and reading condition Isamu Kashima, DDS, PhD,a Takashi Sakurai, DDS,b Takahiko Matsuki, DDS,b Kouji Makamura, DDS,b Hidehito Aoki, DDS,C and Megumi Ishii, DDS,b Yokosuka Kanagawa, Japan DEPARTMENT OF RADIOLOGY, KANAGAWA DENTAL COLLEGE The quality of images obtained by a new method of intraoral computed radiography using three types of imaging plates and a drum scanner designed exclusively for use with an imaging plate was investigated. The aperture size for reading of radiographic data from the imaging plate and for film recording was 50, 100, or 150 pm2. The results indicated that high resolution type irnaging plate for mammography produced the best image quality. For each imaging plate, it was also found that the image quality improved as the aperture size was decreased. However, the high resolution type IP even at the aperture size of 100 pm* showed diagnostic accuracy as high as the image processing of conventional periapical radiographs. This finding suggests that this new imaging system using an imaging plate at the reading and recording aperture size 100 pm2 has promise for use in intraoral radiography. (ORAL SURC ORAL MED ORAL PATHOL 1994;78:239-46) RadioVisioGraphy1-4 (Trophy Radiologie, Vin- cennes, France) and Sens-A-Ray51 6 (Regam Corpo- ration, Sundsvall, Sweden), both of which use a charge-coupled device, are in clinical use for intraoral digital radiography. These modalities do not provide resolution as high as conventional periapical radio- graphs, which have a resolution of more than 12 lines/mm combined with a relatively wide dynamic range and low noise. However, improved resolution and processing of digital images may facilitate im- provement of direct digital image quality.4 Intraoral and panoramic are the radiographic pro- cedures most commonly used in dentistry today. We previously applied computed radiography (CR) with the imaging plate (IP) to panoramic radiography.“IO More recently we have been engaged in research on the use of CR for intraoral radiography. For this it is necessary to determine whether the conventional CR systems with reading and recording aperture size of “Professor and Chairman. bPostgraduate student. ‘Lecturer. Copyright @ 1994 by Mosby-Year Book, Inc. 0030-4220/94/$3.00 + 0 7/16/54875 1NCOOtR RIAD DRUM 6 0 OCTICAL DISK COATED llLM Fig. 1. Block diagram of drum scanner used to read imaging plates. 239

Intraoral computed radiography using the Fuji computed radiography imaging plate: Correlation between image quality and reading condition

Embed Size (px)

Citation preview

Page 1: Intraoral computed radiography using the Fuji computed radiography imaging plate: Correlation between image quality and reading condition

Vol. 78 No. 2 August 1994

OiUAL AND MAXILLOFACIAL RADIOLOGY Editor: Allan G. Farman

Intraoral computed radiography using the Fuji computed radiography imaging plate Correlation between image quality and reading condition

Isamu Kashima, DDS, PhD,a Takashi Sakurai, DDS,b Takahiko Matsuki, DDS,b Kouji Makamura, DDS,b Hidehito Aoki, DDS,C and Megumi Ishii, DDS,b Yokosuka Kanagawa, Japan DEPARTMENT OF RADIOLOGY, KANAGAWA DENTAL COLLEGE

The quality of images obtained by a new method of intraoral computed radiography using three types of imaging plates and a drum scanner designed exclusively for use with an imaging plate was investigated. The aperture size for reading of radiographic data from the imaging plate and for film recording was 50, 100, or 150 pm2. The results indicated that high resolution type irnaging plate for mammography produced the best image quality. For each imaging plate, it was also found that the image quality improved as the aperture size was decreased. However, the high resolution type IP even at the aperture size of 100 pm* showed diagnostic accuracy as high as the image processing of conventional periapical radiographs. This finding suggests that this new imaging system using an imaging plate at the reading and recording aperture size 100 pm2 has promise for use in intraoral radiography. (ORAL SURC ORAL MED ORAL PATHOL 1994;78:239-46)

RadioVisioGraphy1-4 (Trophy Radiologie, Vin- cennes, France) and Sens-A-Ray51 6 (Regam Corpo- ration, Sundsvall, Sweden), both of which use a charge-coupled device, are in clinical use for intraoral digital radiography. These modalities do not provide resolution as high as conventional periapical radio- graphs, which have a resolution of more than 12 lines/mm combined with a relatively wide dynamic range and low noise. However, improved resolution and processing of digital images may facilitate im- provement of direct digital image quality.4

Intraoral and panoramic are the radiographic pro- cedures most commonly used in dentistry today. We previously applied computed radiography (CR) with the imaging plate (IP) to panoramic radiography.“IO More recently we have been engaged in research on the use of CR for intraoral radiography. For this it is necessary to determine whether the conventional CR systems with reading and recording aperture size of

“Professor and Chairman. bPostgraduate student. ‘Lecturer. Copyright @ 1994 by Mosby-Year Book, Inc. 0030-4220/94/$3.00 + 0 7/16/54875

1NCOOtR RIAD DRUM

6 0 OCTICAL DISK

COATED llLM

Fig. 1. Block diagram of drum scanner used to read imaging plates.

239

Page 2: Intraoral computed radiography using the Fuji computed radiography imaging plate: Correlation between image quality and reading condition

240 Kashima et al. ORAL SURGERY ORAL MEDICINE ORAL PATHOLOGY August I994

0.s 1.0 ¶.O lo.0

-ial mm ( c/mm )

0.1 0.6 1.0 1.0 lo.0

Spatid fnqrrmcll c/mm I

0.1 0.5 I.0 1.0 10.0

Spaini trequ(nrcy ( c,mm !

Fig. 2. Three sets of image-processing conditions varying in gradation, gamma frequency type, and frequency enhancement.

100 pm2 and conventional IP are suitable for intraoral radiography. In this study three types of IP that dif- fered in resolution were evaluated to investigate the relationship between the reading and recording aper- ture size and the radiographic image quality to develop a system of intraoral CR.

MATERIAL AND METHODS Current IP technology

The following three types of IP (Fuji Photo Film Co., Tokyo, Japan) were used: ST III N (ST) for standard CR; HR III N (HR) for mammography, and IP for electron-microscopy (TEM). The reading of radiographic data was performed with the use of a CR-7000 System (Fuji Photo Film Co.), and a CR

drum scanner (Fuji Photo Film Co.) designed to read IP. Fig. 1 shows a block diagram of the CR drum scanner. The CR drum scanner consists of a read drum, a photomultiplier with a reflecting mirror, an analog-to-digital converter, a work station, and an optical filing unit. The IPs set to the read drum are two-dimensionally scanned with a helium-neon laser beam deflected by a mirror. The light signals obtained by scanning are converted by the analog-to-digital converter using a blue filter. The electrical signals are then stored on an optical disk. The stored data are subjected to an image-processing sequence generated by work station commands. The processed, digital ra- diographic information is recorded on single-coated film. The output images are the same size as a

Page 3: Intraoral computed radiography using the Fuji computed radiography imaging plate: Correlation between image quality and reading condition

ORAL SURGERY ORAL MEDICINE ORAL PATHOLOGY Kashima et al. 247 Volume 78, Number 2

Fig. 3. Conventional periapical radiogram (Kodak Ultra Speed Film).

conventional periapical radiograph. The reading of the radiographic data from the IP and its film record- ing was done at the three aperture sizes of 50 ,um2, 100 pm2, or 150 pm2.

X-ray source and conventional film The x-ray generator used was an Asahi Roentogen

Auto 1000 (Asahi Roentogen Co. Tokyo, Japan) with a focal spot size of 0.8 mm2, at 60 kVp, 10 mA, using an exposure time of 0.24 and IP focus distance of 33 cm. Conventional direct exposure periapical film was used for comparison (Ultra Speed Film DF 57, East- man Kodak Co., Rochester, N.Y.).

Resolution and1 signal-to-noise ratio For assessment of resolution, Coltman’s’ 1 conver-

sion equation to compute the total modulation trans- fer function (MTF) from the rectangular response on the test chart was used. For evaluation of the signal- to-noise (S/N) ratio, the number of quanta (NEQ) (a parameter with which the image quality is expressed in terms of the. number of photons) was computed. NEQ was determined from the noise Wiener spec- trum (WS), MTF, gamma (y), density (D), and spa- tial frequency (u) were found using the following equation:

NEQ(D.u) = (log 10 e)2 . y (D)2 MTF(u)~/WS(D.U)

Image processing conditions Fig. 2 shows the three sets of image processing

conditions used. In set A, y is equal to 2.0 and shows linear gradation, the frequency type diminishes the low density area, and the spatial frequency enhance- ment is 2.0 cycles/mm. In set B, the conditions are the

Fig. 4. Ten assessed anatomic features. (1. General mor- phologic structure; 2. periodontal membrane space in the apical region; 3. lamina dura; 4. bone trabecular pattern (gross); 5. bone trabecular pattern (fine); 6. dento enamel border; 7. interdental septal bone level; 8. pulp canal in the apical region; 9. mental foramen; and 10. furcation.)

same as those in set A, except that y is reduced from 2.0 to 1.2. In set C, the gradation is curvilinear, and the frequency is enhanced two to four fold.

Image quality evaluation A dried cadaver mandible that wasvacuum wrapped

in 25 mm expoxy resin was used for an x-ray test phantom. Image quality was evaluated with the use of a subjective visual rating scale in which subject image quabty is compared to conventional periapical radio- graphs (Fig. 3). Each image was rated as follows: 1 = inferior; 2 =: slightly inferior; 3 = equivalent; 4 = slightly superior; or 5 = superior. The 10 ana- tomic features listed in Fig. 4 were assessed. The mean rating scale score and the range of scores (maximum, minimum) was determined for each reading aperture size. Each image was scored by five oral and maxil- lofacial radiologists and five general dentists. Our image assessmem involved five conventional images with different density levels and 27 CR images produced with variable IPs, variable aperture sizes, and variable image processing conditions.

RESlJLTS The MTF of each IP (ST, HR, and TEM) at the

three aperture sizes of 50 pm2, 100 pm2, and 150 pm2 is shown in Fig. 5.. At all aperture sizes, TEM showed the highest total MTF value followed by HR.

The NEQ in a relatively low dose area at 100 pm2 aperture size is shown in Fig. 6. The HR type IP showed the highest value in a spatial frequency area,

Page 4: Intraoral computed radiography using the Fuji computed radiography imaging plate: Correlation between image quality and reading condition

242 Kashima et al. ORAL SURGERY ORAL MEDlCINE ORAL PATHOLOGY August 1994

12345678 0 10

Spatial frequency ( cycle/mm I

n

k E 10’

a

iii

0 1234567

Spatial frequency (cycle/mm ) Spatial frequency (cycle/mm )

Fig. 5. MTF of each IP at three aperture size of 50 ,um2 (A), 100 pm2 (B), and 1.50 pm2 (C).

EXPOSURE 6OKV lOQmR

- ST ““““““““” HR -----.- v EM

IQ” 0 1 2 3 4 5 cycle/mm

Spatial Frequency

Fig. 6. NEQ in low dose area at aperture size of 100 prn2.

and TEM showed the lowest value within 1 - 3 cy- cles/mm.

Fig. 7 shows the image of the dental phantom ob- tained with each IP at the aperture size of 50 pm2 (A), 100 pm2 (B), and 150 pm2 (C) under the image-pro- cessing conditions shown in Fig. 2, A. The mean im- age quality rating score corresponding to each image is shown in Fig. 8. At 50 pm2 aperture size, the mean image quality score for HR and TEM was slightly higher than that of the conventional image, and TEM showed the highest image quality score. At LOO pm2 aperture size, the image quality score for each IP was lower than that at 50 ym2. However, the image qual- ity.of TEM and HR were about the same as that of the conventional image, At 150 pm2 aperture size, the image quality for each IP was less than that of con- ventional film.

Fig. 9 shows the images of the phantom obtained

Page 5: Intraoral computed radiography using the Fuji computed radiography imaging plate: Correlation between image quality and reading condition

??i% SURGERY ORAL MEDICINE ORAL PATHOLOGY Volume 78, Number 2

Kashima et al. 243

Fig. 7. Image of phantom obtained with each IP at the aperture size of 50 pm2 (A), 100 pm2 (B), and 150 pm2 (C), under image-processing conditions shown in Fig. 2, A.

with the HR type IP at 100 pm2 IA] and 150 pm* (B) aperture size under the conditions described in Fig. 2, B and Fig. 2, C, respectively. Each image had higher contrast than the images obtained with the same IP and aperture size shown in Fig. 8. Fig. 10 shows the mean image quality rating score for each IP under the image processing conditions shown in Fig. 2, B. Although the quality of images obtained with ST and TEM was poorer than that of the conventional peri- apical image, the quality for HR at 100 and 150 ym* aperture sizes was improved by the image processing contrast and was somewhat higher than for conven- tional film.

DISCUSSION Changes in the periodontal membrane space, lam-

ina dura, and bone trabecular pattern in the jaws are usually diagnosed with the aid of periapical radio- graphs with resolution of more than 20 cycles/mm

and a high S/N ratio. It was generally considered necessary that the aperture size be less than 100 pm2 to obtain digital periapical radiographs of the same image quality as conventional. Therefore it is uncer- tain whether radiographic information in the high- frequency area of more than 10 cycles/mm improves visual diagnosis. It has been reported that for the di- agnosis of periodontal membrane space and lamina dura, a frequency Itransfer of 7.0 - 8.0 cycles/mm is required. l2 It is ge,nerally argued that the frequency of the components of the periodontal membrane space and the bone trabecular pattern are within the range of 0.5 - 3.0 cycles/mm.13 In fact, the TEM and HR images (Fig. 7, B) processed at the aperture size of 100 pm2 showed the same image quality as the con- ventional image (Fig. 3) with a resolution of more than 20 cycles/mm. If the aperture size is 100 pm2, the theoretical spatial frequency area is as much as 5 cycles/mm.

Page 6: Intraoral computed radiography using the Fuji computed radiography imaging plate: Correlation between image quality and reading condition

ORAL SURGERY ORAL MEDICINE QRAL PATHOLOGY August I994

ST TEM HRCIP)

ZOc/mm , 2. a. 0,

Fig. 8. Mean image quality rating scores of computed radiographs shown in Fig. ‘7. Score for conventional image is indicated by dotted line.

Fig, 9, Images of phantom (A, 100 pm2, B, 150 pm2) obtained with HR type 1P under image processing conditions shown in Fig. 2, B and Fig. 2; G.

Page 7: Intraoral computed radiography using the Fuji computed radiography imaging plate: Correlation between image quality and reading condition

bRAL SURGERY ORAL MEDICINE ORAL PATHOLOGY Kashima et al. 245 Volume 78, Number 2

1.2A. Cl. 2.0c/mm . 2.0 n 1OOpm

ST TEM HRCIP) B

l.PA. Q. 2.0c/mm, 2.0, 150pm

ST TEM HRCIP)

Fig. 10.. Mean image quality rating score for images obtained with each IP under image processing con- ditions shown in Fig. 2, B, A, 100 pm2, B, 150 pmZ).

However, at the aperture size of 150 pm2, the HR images showed the same image quality as the conven- tional film as a result of the image processing (Fig. 10). This finding suggests that radiographic informa- tion that may be visually perceived with conventional images and that is of importance in interpretation is likely to be obtained within the spatial frequency range of 5 cycles/mm or less.

In this study, radiographic information was read and recorded using three types of IP at the aperture sizes of 50 ym2, 100 pm2, and 150 pm2. The IP for TEM that shows the highest degree of resolu- tion contains a larger amount of blue pigment than the other IP. We believe that the MTF values were high because the laser absorption rate was high and prevemed scattering at the laser spots (Fig. 5). The resolution of each IP at 50 pm2 aperture size was proportional to the image quality (Fig. 8, A). However, at 100 gm2, the image quality score for HR was higher than that for TEM regardless of the image processing conditions (Fig. 8 and Fig. 10). Although the MTF value of HR was inferior to that of TEM, its image quality was better. It is gen- erally understood that the noise in the low-dose area has a substantial influence on the visual image diag-

nosis. The findings for visual assessments at 100 ym2 are consistent with those for NEQ in the low-dose area at the same aperture size. As TEM has a high degree of resolution but a low S/N ratio, the noise is visually striking as its images are processed and the image quality is reduced (Fig. 6). Thus IP with high resolution does not necessarily provide images with high diagnostic utility, and the image quality may depend on the equilibration of resolution with the S/N ratio.

This suggests that the image quality of conven- tional film images depends on the S/N ratio rather than on high resolution. We conclude that the image processing of IP for which the resolution is balanced with the S/N ratio may produce computed periapical radiographs with the same image quality as that of conventional periapical radiographs, even at 100 pm2 aperture size.

CONCLUSIONS With the use of ILP of the ST, HR, and TEM types,

images of a phantom were read and recorded at the aperture sizes of 50 pm2, 100 pm2, and 150 prn2, and the relationship of the type of IP to the aperture size and image quality was investigated.

Page 8: Intraoral computed radiography using the Fuji computed radiography imaging plate: Correlation between image quality and reading condition

246 Kashima et al, ORAL SURGERY ORAL MEDICINE ORAL PATHOLOGY August 1994

It was found that: 1.

2.

3.

4.

5.

High-resolution IP does not necessarily provide images with high image quality. The radiographic image quality of the IP de- pends on the equilibration of resolution and the S/N ratio. For both reading and recording of IP, a decrease of the aperture size resulted in improvement of image quality. HR showed the best image quality at the aper- ture size of 100 ym2. Image processing improved the quality of HR images, even at the aperture sizes of 100 pm2 and 150 hm2.

These findings suggest that the use of IP with a good balance between resolution and the S/N ratio facilitates intraoral computed radiography at the ap- erture size of 100 pm2. Such use of IP provides an im- age that is comparable to that of conventional in- traoral radiography.

We thank Shigeru Saotome and Tutomu Arakawa, Fuji Photo Film CO., Ltd., Technology Development Center, for their helpful advice.

REFERENCES

1. Mouyen F, Benz C, Sonnabend E, Lodter JP. Presentation and physical evaluation of RadioVisioGraphy. ORAL SURG OR-AL MED ORAL PATHOL 1989;68:238-42.

2. Horner K, Shearer AC, Walker A, Wilson NHF. RadioVisi- oGraphy: an initial evaluation. Br Dent J 1990;168:244-8.

3. Shearer AC, Horner K, Wilson NHF. RadioVisioGraphy of imaging root canals: an in vitro comparison with conventional radiography Quintessence Int 1990;21:789-94.

4. Benz C, Mouyen F. Evaluation of the new Radio VisioGraphy system image quality. ORAL SURG ORAL MED ORAL PATHQL 1991;72:627-31.

5. Nelvig P, Wing K, We1ander.U. Sens-A-Ray. ORAL SURG ORAL MED ORAL PATHOL 1992;74:818-23.

6. Welander U, Nelvig P, Tronje G, et al. Basic technical prop- erties of a system for direct acquisition of digital intraoral ra- diographs. ORAL SURG ORAL MED ORAL PATHOL 1993; 75:506-16.

7. Kashima I, Kanno M, Higashi T, Takano M. Computed pan- oramic tomography with scanning laser-stimulated lumines- cence. ORAL SURG ORAL MED ORAL PATHOL 1985;60:448-53.

8. Kashima I, Kanno M, Oguro T, et al. Bone trabecular pattern analysis in Down’s syndrome with the use of compnted panoramic tomography with a laser scan system. ORAL SURC ORAL MED ORAL PATHOL 1988;65:366-70.

9. Kashima I, Tajima K, Nishimura K, et al. Diagnostic imaging of diseases affecting the mandible with the use of computed panoramic radiography. ORAL SURG ORAL MED ORAL PATHOL 1990;70:110-6.

10. Kashima I, Bando S, Kanishi D, Miyake K, Yamane R, Ta- kano M. Bone trabecular pattern analysis in Down’s syndrome with the use of computed panoramic radiography. ORAL SURG ORAL MED ORAL PATHOL 1990;70:360-4.

11. Coltman JW. The specification of imaging properties by response to a sine wave input. J Opt Sot Am [A] 1954;44:468- 73.

12. Tutiya T. Visual recognition of spatial frequancy information in diagnostic intraoral roentgenograms. Dental Radiology (Japan) 1988; 28:267-84.

13. Sato S. A study on evaluation of dental films by digital image processing: analysis of alveolar trabecula by means of two di- mensional FFT. Dental Radiology (Japan) 1986;26:242-53.

Reprint requests: Isamu Kashima, DDS, PhD Department of Radiology Kanagawa Dental College 82-Inaoka-cho Yokosuka Kanagawa Japan