In vitro assessment of periapical lesions created in sheep mandibles by using high resolution ultrasonography and cone beam computed tomography

Abstract

Objectives:

Providing ultrasound images of periapical lesions may be problematic depending on the thickness of the overlying cortical bone. Clinically, it is crucial to determine the cut-off value of overlaying bone thickness in terms of interference with ultrasound imaging in conjunction with assessment of changes in periapical jaw bone lesions. Our aim was to determine the minimum amount of overlaying buccal bone thickness of artificial periapical lesions in order to be visible by ultrasound imaging and to compare width, height, depth, surface area and volume measurements of detectable periapical lesions obtained from ultrasound and CBCT images.

Methods:

Periapical lesions were created in 16 molar teeth of sheep mandibles. Cavities were enlarged until the borders of lesions were visible on 14 MHz hockey probe ultrasound imaging. CBCT and ultrasound images were obtained simultaneously after drilling and enlarging each size of the cavities and replacing the teeth in their sockets. two observers separately assessed images twice within 2 weeks of interval. By using CBCT and ultrasound images, buccal bone thickness, maximum width, height, depth, surface area and volume of periapical lesions were measured. Intraclass correlation coefficient (ICC) was utilized and significance level was set at p < 0.05.

Results:

The mean buccal bone thickness ranged between 1.21 mm and 1.31 mm for both imaging techniques. For the measurement of buccal bone thickness, periapical lesion width, height, depth, surface area, and volume excellent ICC values were found in terms of intrarater (ranging between 0.907 and 1) and inter-rater (ranging between 0.864 and 1) reliability for both observers and their readings. There were no statistically significant differences for both observers and for their two readings between ultrasound and CBCT measurements of buccal bone thickness, and periapical lesion width and height (p > 0.05).

Conclusions:

We suggested that a buccal thickness of approximately 1.28 mm might be accepted as a cut-off value for the detection of periapical lesions with 14 MHz hockey probe ultrasound. High resolution ultrasound provided accurate information for the measurement of buccal bone thickness and lesion width and height.

Introduction

Diagnosis and assessment of periapical lesions are of paramount importance in order to render accurate dental treatment. In routine clinical practice, intraoral periapical radiography and panoramic radiography are the most preferred modality of choice for the assessment of periapical lesions, higher spatial resolution offered by the former technique.¹ Authors compared the accuracy of film and digital periapical radiography in detecting apical periodontitis by using gold-standard histopathological findings and found that the diagnostic accuracy of single digital periapical radiography was significantly better than single film periapical radiography. Moreover, use of two additional periapical radiographs taken by using mesial and distal horizontal angulations was suggested for improved detection of periapical lesions.² Another study assessed the diagnostic potential of intraoral digital and conventional film and two different CBCT units in the detection of chemically created periapical lesions. No difference was found between film and digital intraoral radiographic techniques tested which performed worse than the two CBCT units.³ Two-dimensional techniques offer diagnostically useful information for the assessment of periapical lesions, however; due to their inherent limitations it is not possible to determine the exact location, extent, area and volume of the periapical lesions by using two-dimensional techniques. Intraosseous cavities created by odontogenic and non-odontogenic lesions of the jaws are imaged by using advanced diagnostic techniques when conventional two-dimensional radiographic methods are insufficient to reveal sufficient and accurate information.⁴ The ability to provide high-resolution and accurate multiplanar reformatted images with a lower dose than multidetector computed tomography (MDCT) made cone beam computed tomography (CBCT) the imaging modality of choice for assessing intraosseous lesions of the jaws.⁵ Reformatted CBCT images provide useful information for the diagnosis of a variety of jaw-related entities, including cysts, granulomas and abscess of periapical structures, benign and malign lesions of the jaws. However, patient is subjected to ionizing radiation and accurate assessment of soft tissue structures is not possible with CBCT due to low soft tissue contrast.⁶ Use of ultrasonography in the dentomaxillofacial region became popular in recent years owing to increasing concerns regarding radiation dose and economic limitations. In dentomaxillofacial imaging, ultrasound provides several advantages such as; presence of non-ionizing radiation, portability, possibility of dynamic and repeated real-time examinations and low cost.^7,8 In general terms, ultrasound is now utilized for the assessment of salivary glands,⁹ orofacial soft tissue swelling,¹⁰ examination of muscles¹¹ and cervical lymphadenopathy,¹² detection of caries,¹³ dental and maxillofacial fracture and cracks,¹⁴ assessment of periodontal and gingival tissues,^15,16 examination of temporomandibular joint,¹⁷ evaluation of intraosseous lesions,¹⁸ and differentiation^19–22 and follow-up^23,24 of periapical lesions.

In diagnostic ultrasound, high-frequency (30–20 KHz) sound waves are created and transmitted into the body by a transducer that contains piezoelectric crystals, which transform electrical signals into mechanical vibrations and vice versa. Echoes from the interface between tissues are detected, processed, and translated into grayscale images, which are displayed on a computer screen.^7,17,24 Main drawbacks of ultrasound include limited penetration into bone and gas-filled structures, less spatial resolution at deep tissues and dependency on expertise. If the material is solid, the particles are denser and sonographic waves are reflected more. Therefore, solid materials transmit fewer sound waves than fluids and less ultrasound waves are reflected back from fluids.^17–25 Authors of an earlier in vitro study, assessed artificially created bony defects of varying sizes in bovine scapula by using ultrasound and reported that failure to detect central lesions occurred in cortical bone more than 1.1 mm in thickness.²⁵ In addition to the amount of remaining cortical bone, other factors that may affect the visibility of periapical lesions by ultrasound imaging include the ultrasound probe and settings selected, operator performance, and patient-related factors. In view of recent advancements in ultrasound technology, including the development of high-resolution and high-definition definition ultrasound devices, further research is required to determine the cut-off value of overlaying bone thickness in terms of interference with ultrasound imaging.^25–27

Providing ultrasound images of periapical lesions may be problematic depending on the thickness of the overlying cortical bone. Clinically, it is crucial to determine the minimum overlaying bone thickness in terms of interference with ultrasound imaging in conjunction with assessment of changes in periapical jaw bone lesions.^25–27 To our knowledge, no previous study assessed changes in periapical jaw bone lesions by comparing ultrasound and CBCT techniques along with perforation at the cortical bone which occurs due to lesion expansion that facilitates the detection of pathosis by ultrasound imaging. Therefore, the aim of the present study was twofold: (1) to determine the cut-off value of overlaying buccal bone thickness in terms of interference with high resolution ultrasound in artificial intraosseous periapical lesions created in sheep mandibles in comparison to CBCT; (2) to compare width, height, depth, surface area and volume measurements of detectable periapical lesions obtained from ultrasound and CBCT images.^25–27

Methods and materials

Ethical approval was obtained from ------- University, Faculty of Dentistry, Research and Ethics Committee (No:20.11.2019-13.03). Six sheep heads and mandibles with soft tissues, obtained from Animal Research Institute, were stored in 10% formalin solution during the study. After pilot studies, we concluded that molar teeth were easy to extract and large enough to prepare appropriate periapical lesions. Therefore, after preliminary radiographic and visual examination, 20 molar teeth were selected. Teeth without fracture, resorption or bone defects were included in our study.

Preparation of the experimental lesions

All surgical procedures were performed by the same experienced periodontologist (ÖÇ). Intracrevicular incisions with a scalpel (Hu-Friedy 10-130-07K) (Hu-Friedy 10-256-15) were made buccally. The incisions were made in a mesiodistal direction in the mandibular molar region. The periosteal flap was elevated by using a periosteal elevator (Hu-Friedy P24GSP) in order to approach buccal cortical plate. Before imaging, the flap was placed in its original position in order to mimic real clinical conditions. Soft tissue flap was repositioned and teeth were extracted by forceps, replaced in the sockets and again imaged by using periapical radiography and if deemed necessary by CBCT to ensure absence of any bone, root fractures, or defects, which could mislead observers. During extraction, excessive damage to the alveolar bone occurred in two teeth and two teeth roots broke. These teeth were excluded and finally our study comprised 16 mandibular molars. Periapical lesions were created by extracting each tooth with minimal force and progressively enlarging with the help of round burs (ISO 0.6 mm, and 0.8 mm) by drilling to the full-depth under ×2.5 magnification at the available periapical locations of the mesial roots. Buccally oriented lesions were created and buccal cortical thicknesses were measured as buccal periapical lesions were easier to detect by using ultrasound²⁷ and buccal lesions were easier and more appropriate to create in the molar regions due to bone thickness.²⁸ Lesions were created cautiously and almost repeatably by our periodontologist (ÖÇ). No perforation occurred during the procedures. Samples were numbered and then marked with a 0.5 mm red wax in order to delineate the superior and inferior borders of the cavities in the buccolingual direction. CBCT and ultrasound images were obtained simultaneously after drilling and enlarging each size of the cavities and replacing the teeth in their sockets. Therefore, identical lesions were assessed by using both techniques. Figure 1 shows a representative experimental sheep head and mandible sample for imaging.

Figure 1.

a-b: Experimental sample for the preparation of lesions and imaging. a: Representative experimental sheep head and mandible for lesion preparation. b: Experimental sample located for CBCT imaging.

CBCT image acquisition

CBCT images of samples were taken by using a Planmeca Promax 3D max CBCT unit (Planmeca, Helsinki, Finland) at 96 kVp, 5mA, and a 0.2 mm voxel size at 140 × 90 mm field of view (FOV) with an exposure time of 10 sec. Volumetric data were reconstructed to provide serial cross-sectional views.

Uultrasound image acquisition

Ultrasound examinations were performed by using an ACUSON S 2000 (Siemens, Munich, Germany) high-resolution ultrasonography by the same researchers who conducted CBCT measurements. The ACUSON S 2000 is an ultrasound system that includes an ultrasound scanner and a special stationary device with transducer attached to a mechanical arm. The 14L5BV transducer (maximum frequency 14 MHz, average scanning frequency 10 MHz, width of 15.4 cm, 768 piezoelectric elements) receives the image with a maximum depth of up to 6 cm with 0.09 mm axial and lateral 0.16 mm resolution. It has dynamic range display from 30 to 90 dB and Dynamic TCE^™ (tissue contrast enhancement) technology. A gain collection algorithm analyzes the data and adjusts for the brightness variation artifacts caused by transducer channel to channel effects. Ultrasound examinations were performed by using a 14 MHz hockey (stick) probe covered with ultrasonography gel and sterile sheath on the longitudinal and transversal planes, with the probe position changed constantly to obtain sufficient longitudinal and transversal scan images on the monitor. The transducer was positioned perpendicular to teeth in order to enable cross-sectional assessment of the lesion site.

Image analysis

Gold-standard buccal cortical thickness measurements were conducted after relocating the teeth after each size of cavity enlargement from CBCT images. Cross-sectional images were created from the arch that was drawn in the lesion region. Two oral and maxillofacial radiologists (with 5 years and 10 years of experience of clinical radiology) acted as observers and separately assessed images twice within 2 weeks of interval time. Prior to image analysis, a calibration session was conducted in order to carry out pilot measurements on 2 CBCT and 2 ultrasound images those were not included in the study. For each lesion, buccal cortical thickness measurements were performed perpendicular to alveolar crest at the largest diameter of the lesion on cross-sectional CBCT images by using measurement tools in mm. Average of two measurements were noted for each site. Detection of perforation was determined by considering CBCT assessment as the reference modality. We did not detect any perforation during the procedures. For the image interpretation protocol, dedicated software of the CBCT system was used. Images were viewed on a 21.3-inch medical diagnostic monitor at 2048 × 1536 resolution (NEC, Tokyo, Japan) in a dimly lit room. Cavities were enlarged until the borders of lesions were visible on ultrasound imaging.

For the dimensional evaluation with ultrasound, buccal cortical thickness seen as hyperechogenic with anechogenic shadow was used as the reference point on the ultrasound image. For each cavity site, measurement of cortical bone thickness in cm was performed at the same two points, which were also calculated on CBCT images. Average of two measurements were noted for each site. Representative buccal cortical bone thickness measurements of the same periapical cavity region conducted by CBCT (Romexis Software, Planmeca Promax 3D, Planmeca, Helsinki, Finland) in mm and ultrasound ACUSON S 2000 (Siemens, Munich, Germany) in cm are shown in Figure 2.

Figure 2.

a–f: Representative buccal cortical bone thickness measurements of the same periapical cavity region conducted by both CBCT (Romexis Software, Planmeca Promax 3D, Planmeca, Helsinki, Finland) in mm and ultrasoundS ACUSON S 2000 (Siemens, Munich, Germany) in cm. Red arrows show wax material. a: Cross-sectional CBCT image shows 1.61 mm initial buccal cortical bone thickness. b: Longitudinal ultrasound image of the same region inFigure 2a shows 0.16 cm initial buccal cortical thickness. c: Cross-sectional CBCT image shows 1.41 mm buccal cortical bone thickness after enlargement of the cavity. d: Ultrasound image of the same region in Figure 2c shows 0.14 cm buccal cortical bone thickness. e: Cross-sectional CBCT image shows 1.28 mm buccal cortical bone thickness after enlargement of the cavity with the final drill. f: Ultrasound image of same region in Figure 2e shows 0.12 cm buccal cortical bone thickness. CBCT, cone beam CT.

On ultrasound, images where lesion borders were visible on cross-sectional views, maximum height, maximum (buccolingual) depth, surface area and volume of the periapical lesions were measured using the built-in software. In addition, maximum mesiodistal width was measured by placing the probe transversally. All measurements were performed by the same researchers twice between 2 weeks of intervals and then compared with those obtained from CBCT images. CBCT maximum width, height and depth measurements were obtained from multiplanar reformatted reconstructions using the built-in software. Representative width, height and depth measurements of the same periapical cavity region conducted by both CBCT (Romexis Software, Planmeca Promax 3D, Planmeca, Helsinki, Finland) in mm and ultrasound ACUSON S 2000 (Siemens, Munich, Germany) in cm. are shown in Figure 3. Lesion surface area and volume were calculated by exporting CBCT images to 3D-DOCTOR (Able Software Corp., Lexington, MA). 3D-DOCTOR, a volumetric rendering software that uses vector-based segmentation technology to perform segmentation of lesions on consecutive axial slices, enabling visualization at each level. Lesion borders were segmented manually on each slice using a mouse and the “trace boundary” tool to create a turquoise border. Representative surface area and volume measurements of the same periapical cavity region conducted by both 3D-DOCTOR (Able Software Corp., Lexington, MA) in mm² and mm³ and ultrasound ACUSON S 2000 (Siemens, Munich, Germany) in cm² and cm³ are shown in Figure 4. CBCT measurements were considered as the reference as it is the preferred modality and proved to be the most accurate clinical method for the assessment of periapical bone lesions. Total duration of the study was approximately 3 months for data collection including specimen preparation, creation and imaging of lesions sequentially, assessment of images and critical analysis.

Figure 3.

a–d: Representative width, height and depth measurements of the same periapical cavity region conducted by both CBCT (Romexis Software, Planmeca Promax 3D, Planmeca, Helsinki, Finland) in mm and ultrasound ACUSON S 2000 (Siemens, Munich, Germany) in cm. a: Axial CBCT image shows 3.20 mm width measurement. b: Ultrasound image of the same regıon ın Fıgure 3a obtained by placing the probe transversally shows 0.32 cm width measurement. c: Cross-sectional CBCT image shows 3.81 mm depth and 5.00 mm heıght measurement. d: Ultrasound image of the same region in Figure 3c shows 0.38 cm depth and 0.5 cm heıght measurement. CBCT, cone beam CT.

Figure 4.

a–d: Representative surface area and volume measurements of the same periapical cavity region conducted by both 3D-DOCTOR (Able Software Corp., Lexington, MA) in mm² and mm³ and ultrasound ACUSON S 2000 (Siemens, Munich, Germany) in cm² and cm³. a: Volumetric and surface area measurement of artificial periapical lesion from exported CBCT images by using 3D-DOCTOR. b: Ultrasound image shows volumetric measurement of the artificial lesion by using build-in software. c: Ultrasound image shows surface area measurement of the same artificial lesion in Figure 4b by using build-in software. CBCT, cone beam CT.

Statistical analysis

Cohen κ coefficient was calculated. Intraclass correlation coefficient (ICC) was used for the assessment of intrar- and interrater reliability for the observers. ICC was also used to determine agreement between ultrasound and CBCT measurements. Statistical significance level was set at p < 0.05.

Results

Table 1 shows the mean, minimum, and maximum buccal bone thickness measurements obtained by both imaging techniques after the preparation of final cavities in which lesions were detectable by ultrasound. The mean buccal bone thickness was approximately 1.29 mm ranging between 1.21 mm and 1.31 mm for both imaging techniques. In consideration to percentiles we suggested that a buccal thickness of approximately 1.28 mm or less might be accepted as a cut-off value for the detection of intraosseous periapical lesions with high resolution ultrasound. Lesions were visible on ultrasound images when there was approximately 1.28 mm or less buccal cortical bone around artificially created lesions on CBCT images. For the measurement of buccal bone thickness around artificial lesions, almost excellent ICC values were found in terms of intra- and interrater reliability for the observers and their readings. Table 2 shows Cohen κ coefficients for both observers. Table 3 shows ICC values obtained from both imaging techniques for the two readings of both observers for buccal bone thickness, periapical lesion width and periapical height measurements. Ultrasound measurements also highly correlated with those of CBCT measurements for buccal bone thickness measurements. According to t test, there was no statistically significant difference between ultrasound and CBCT measurements for both readings of Observer 1 (Reading 1: p = 0.491; and Reading 2: p = 0.454) and Observer 2 (Reading 1: p = 0.226; and Reading 2: p = 0.226) for buccal bone thickness.

Table 1.

Mean, minimum, maximum, and standard deviation (std) buccal bone thickness measurements obtained by both imaging techniques after the preparation of final cavities

	Observer 1				Observer 2
	Reading 1		Reading 2		Reading 1		Reading 2
	CBCT_OBS1_ READING1	USG_OBS1_ READING1	CBCT_OBS1_ READING2	USG_OBS1_ READING2	CBCT_OBS2_ READING1	USG_OBS2_ READING1	CBCT_OBS2_ READING2	USG_OBS2_ READING2
N	16	16	16	16	16	16	16	16
Mean	1.2858	1.2833	1.2892	1.2850	1.2892	1.2842	1.2900	1.2850
Std. deviation	0.02429	0.02570	0.02392	0.02680	0.02275	0.02610	0.01706	0.02236
Minimum	1.22	1.21	1.22	1.21	1.22	1.21	1.24	1.22
Maximum	1.31	1.30	1.31	1.31	1.30	1.30	1.30	1.30
Percentiles 75th	1.2825	1.2725	1.2825	1.2800	1.2900	1.2725	1.2900	1.2825
Percentiles 50th	1.2900	1.2900	1.2950	1.2900	1.3000	1.2950	1.2950	1.2900
Percentiles 25th	1.3000	1.3000	1.3000	1.3000	1.3000	1.3000	1.3000	1.3000

Table 2.

Cohen κ Coefficient for intra- and interrater reliability

	Intrarater reliability (κ)				Interrater reliability (κ)				CBCT-Ultrasonographyd (κ)
	Observer 1		Observer 2		Reading 1		Reading 2		Observer 1		Observer 2
	CBCT	USG	CBCT	USG	CBCT	USG	CBCT	USG	Reading 1	Reading 2	Reading 1	Reading 2
κ	0,556	0,824	0,400	0,800	0,751	0.911	0,751	0,636	0,402	0,636	0,571	0,751
p value	0,054	0,004	0,166	0,005	0,007	0,001	0,007	0,018	0,157	0,018	0,028	0,007

CBCT, cone beam CT; USG, ultrasonography.

Table 3.

ICC values obtained from CBCT and ultrasound imaging techniques for the two readings of both observers for buccal bone thickness, periapical lesion width, and periapical lesion height measurements

Intrarater reliability (ICC) (t test p value)				Interrater reliability (ICC) (t test p value)				CBCT-ultrasound (ICC) (t test p value)
Observer 1		Observer 2		Reading 1		Reading 2		Observer 1		Observer 2
CBCT	USG	CBCT	USG	CBCT	USG	CBCT	USG	Reading 1	Reading 2	Reading 1	Reading 2
Bone thickness
0.943	0.976	0.907	0.965	0.864	0.994	0.932	0.945	0.886	0.955	0.945	0.917
Observer 1		Observer 2		Reading 1		Reading 2		Observer 1		Observer 2
CBCT	USG	CBCT	USG	CBCT	USG	CBCT	USG	Reading 1	Reading 2	Reading 1	Reading 2
0.166	0.339	0.754	0.674	0.368	0.339	0.723	1.000	0.491	0.454	0.226	0.226
Intrarater reliability (ICC) (t test p value)				Interrater reliability (ICC) (t test p value)				CBCT-ultrasound (ICC) (t test p value)
Observer 1		Observer 2		Reading 1		Reading 2		Observer 1		Observer 2
Lesion Width
CBCT	USG	CBCT	USG	CBCT	USG	CBCT	USG	Reading 1	Reading 2	Reading 1	Reading 2
0.997	0.997	0.998	0.997	0.996	0.999	0.997	1	0.931	0.920	0.951	0.946
Observer 1		Observer 2		Reading 1		Reading 2		Observer 1		Observer 2
CBCT	USG	CBCT	USG	CBCT	USG	CBCT	USG	Reading 1	Reading 2	Reading 1	Reading 2
0,2800	0.0020	0.8310	0.0010	0.1400	0.9470	0.674	0.602	0.958	0.701	0.737	0.715
Intrarater reliability (ICC) (t test p value)				Interrater reliability (ICC) (t test p value)				CBCT-US (ICC) (t test p value)
Observer 1		Observer 2		Reading 1		Reading 2		Observer 1		Observer 2
Lesion height
CBCT	USG	CBCT	USG	CBCT	USG	CBCT	USG	Reading 1	Reading 2	Reading 1	Reading 2
0.995	0.994	0.996	0.995	0.997	0.997	0.997	0.999	0.936	0.928	0.934	0.931
Observer 1		Observer 2		Reading 1		Reading 2		Observer 1		Observer 2
CBCT	USG	CBCT	USG	CBCT	USG	CBCT	USG	Reading 1	Reading 2	Reading 1	Reading 2
0.564	0.342	0.650	0.150	0.165	0.865	0.787	0.567	0.858	0.802	0.675	0.614

CBCT, cone beam CT; ICC, Intraclass correlation coefficient;USG, ultrasonography.

For the measurement of periapical lesion width, height, depth, surface area, and volume excellent ICC values were found in terms of intra- and interrater reliability for the observers and their readings. Ultrasound measurements also highly correlated with those of CBCT measurements for periapical lesion width. Considering periapical lesion width measurements, there was no statistically significant difference between ultrasound and CBCT measurements for both readings of Observer 1 (Reading 1: p = 0.958; and Reading 2: p = 0.701) and Observer 2 (Reading 1: p = 0.737; and Reading 2: p = 0.715). Mean lesion width measurements ranged between 4.34 and 4.39 mm (min 3.02 and max 5.62) for ultrasound and 4.33 and 4.37 mm (min 3.02 and max 5.68) for CBCT. There was no statistically significant difference between ultrasound and CBCT for periapical lesion height measurements for both readings of Observer 1 (Reading 1: p = 0.858; and Reading 2: p = 0.802) and Observer 2 (Reading 1: p = 0.675; and Reading 2: p = 0.614). Mean periapical lesion height measurements ranged between 3.33 mm and 4.01 mm (min 3.03 and max 5.02) for ultrasound and 3.31 mm and 4.07 mm (min 3.01 mm and max 5.08 mm) for CBCT.

Table 4 shows ICC values obtained from both imaging techniques for the two readings of both observers for periapical lesion depth, surface area and volume measurements. There was a statistically significant difference between ultrasound and CBCT measurements for both readings of Observer 1 (Reading 1: p = 0.007; and Reading 2: p = 0.010) and Observer 2 (Reading 1: p = 0.010; and Reading 2: p = 0.012) for lesion depth. Mean lesion depth measurements ranged between 4.39 and 4.41 mm (min 3.39 and max 5.99) for ultrasound and 4.43 and 4.44 mm (min 3.45 and max 5.99) for CBCT. For lesion surface area measurements, there was a statistically significant difference between ultrasound and CBCT measurements for both readings of Observer 1 (Reading 1: p < 0.001; and Reading 2: p = 0.001) and Observer 2 (Reading 1: p < 0.001; and Reading 2: p = 0.001). Mean lesion surface area measurements ranged between 132.91 and 134.16 mm² (min 33 mm² and max 222 mm²) for ultrasound and 153 and 155.83 mm² (min 63 mm² and max 246 mm²) for CBCT. In consideration to lesion volume measurements, again there was a statistically significant difference between ultrasound and CBCT measurements for both readings of Observer 1 (Reading 1: p = 0.005; and Reading 2: p = 0.005) and Observer 2 (Reading 1: p = 0.006; and Reading 2: p = 0.007). Mean lesion volume measurements ranged between 92.83 and 93.33 mm³ (min 40 mm³ and max 195 mm³) for ultrasound and 106.58 and 107.75 mm³ (min 46 mm³ and max 218 mm³) for CBCT.

Table 4.

ICC values obtained from CBCT and ultrasound imaging techniques for the two readings of both observers for periapical lesion depth, periapical lesion surface area, and periapical lesion volume measurements

Intrarater reliability (ICC) (t test p value)				Interrater reliability (ICC) (t test p value)				CBCT-ultrasound (ICC) (t test p value)
Observer 1		Observer 2		Reading 1		Reading 2		Observer 1		Observer 2
CBCT	USG	CBCT	USG	CBCT	USG	CBCT	USG	Reading 1	Reading 2	Reading 1	Reading 2
0.999	1	0.999	1	1	0.999	1	0.999	0.996	0.996	0.996	0.998
Observer 1		Observer 2		Reading 1		Reading 2		Observer 1		Observer 2
CBCT	USG	CBCT	USG	CBCT	USG	CBCT	USG	Reading 1	Reading 2	Reading 1	Reading 2
Lesion depth
0.355	0.049	0.892	0.024	0.658	0.830	0.131	0.623	0.007	0.010	0.010	0.012
Intrarater reliability (ICC) (t test p value)				Interrater reliability (ICC) (t test p value)				CBCT-ultrasound (ICC) (t test p value)
Observer 1		Observer 2		Reading 1		Reading 2		Observer 1		Observer 2
CBCT	USG	CBCT	USG	CBCT	USG	CBCT	USG	Reading 1	Reading 2	Reading 1	Reading 2
Lesion surface area
0.996	0.999	0.998	0.997	0.998	0.999	0.999	0.998	0.887	0.902	0.883	0.897
Observer 1		Observer 2		Reading 1		Reading 2		Observer 1		Observer 2
CBCT	USG	CBCT	USG	CBCT	USG	CBCT	USG	Reading 1	Reading 2	Reading 1	Reading 2
0,0270	1	0.4180	0.2920	0.4170	0.2830	0.077	0.6330	<0.001	0.001	<0.001	0.001
Intrarater reliability (ICC) (t test p value)				Interrater reliability (ICC) (t test p value)				CBCT-ultrasound (ICC) (t test p value)
Observer 1		Observer 2		Reading 1		Reading 2		Observer 1		Observer 2
CBCT	USG	CBCT	USG	CBCT	USG	CBCT	USG	Reading 1	Reading 2	Reading 1	Reading 2
Lesion Volume
0.996	0.998	0.999	0.998	1	1	0.998	0.999	0.890	0.907	0.886	0.905
Observer 1		Observer 2		Reading 1		Reading 2		Observer 1		Observer 2
CBCT	USG	CBCT	USG	CBCT	USG	CBCT	USG	Reading 1	Reading 2	Reading 1	Reading 2
0.679	0.905	0.041	0.815	0.131	0.166	0.931	0.339	0.005	0.005	0.006	0.007

CBCT, cone beam CT; USG, ultrasonography.

intraclass correlation coefficient (ICC)

Discussion

To our knowledge, the present study was the first to compare experimentally created periapical lesion cavities by using high resolution ultrasonography in comparison to CBCT in sheep mandibles without cortical bone perforation. The pros of our study were the comparison of CBCT and ultrasonography findings and the use of periosteal flaps in sheep mandibles for oral soft tissue simulation. This procedure also allowed us to see visually whether there was perforation in the buccal cortical bone or not. Accurate assessment of periapical pathosis is of paramount importance in enabling the clinician to provide immediate and appropriate dental intervention. We evaluated the diagnostic potential of high-resolution ultrasound that provided real-time rapid images, at a low cost and without ionizing radiation in assessing artificial periapical lesions and cortical buccal bone around those lesions. The use of ultrasound in inflammatory soft tissue conditions and superficial tissue disorders of the head and neck region was already established.⁷ In spite of all limitations, use of ultrasound for monitoring bone healing and detection of intrabony lesions showed promising results.^18,27

Tissues can be categorized as either echogenic or anechogenic in ultrasound assessment. In echogenic, tissue is reflected with high intensity to produce light images whereas in anechogenic, tissue is reflected with low intensity to produce dark screen images. Reverberation artifact occurs when an ultrasound beam encounters two strong parallel reflectors and objects beyond these interfaces mostly cannot be imaged through ultrasound.^7,17,26,27 In our samples, the cortical bone was intact, so this artifact could be observed in all blocks as hyperechoic interrupted lines. However, in those sites, where the cortex was thinned, hypoechoic shadows of the lesions could be seen and the hyperechoic areas among them demonstrated bony septa between the cavities.^7,17,26,27 Bones reflect more sound waves than fluid and they display bright images. Since ultrasound waves cannot transmit through compact bones, a black acoustic shadow is present behind them. Besides, air is a robust ultrasound beam reflector which makes it difficult for the operator to visualize structures. In ultrasound imaging, the borders of the lesions may vary according to bone thickness and continuity and acoustic impedance. The term “the poorly or well-defined margin” and posterior acoustic enhancement used for periapical lesions should be considered for diagnosis and measurement. In particular, image quality may be negatively affected by acoustic impedance, which may vary with patient weight, gender and age in real clinical situations.^7,17,26,27 Further studies with larger sample sizes are needed to better understand the factors influencing the effectiveness and measurement accuracy of ultrasound to assess periapical lesions. It is important to keep in mind that ultrasound is an operator- and technique-dependent imaging modality and this may lead to manual errors and therefore rigorous training of oral radiologists is required. Incorporation and implementation of ultrasound in case of widespread lesions, such as; malignancy, cellulitis, etc is an important point to be considered.

According to the results of our study, bone cavities could be observed in ultrasound when the cortical buccal bone thickness around the periapical cavities in molar region was 1.28 mm and below in sheep mandibles. So, it is expected that, in human samples, a thicker cortical bone (>1.28 mm) would reflect the ultrasound waves completely and therefore camouflage the underlying periapical lesions. Authors of an earlier in vitro study, assessed ultrasound in the detection of artificially created bony defects of varying sizes in bovine scapula and reported that failure to detect central lesions occurred in cortical bone more than 1.1 mm in thickness.²⁵ In a recent study, the boundary of the lesion could not be detected by using innovative and non-invasive Optical Coherence Tomographic images when there was residual bone thickness beyond 1.18 mm.²⁹ These values are slightly lower than ours and this difference may be attributable to, sample and bone region related factors, ultrasound software, probe and settings used, observer performance, and difference in technique used. Another in vivo study,²⁶ reported that ultrasound could establish the presence or absence of a lesion, erosion of the buccal cortical plate, and identification of associated soft tissue involvement in all cases studied (sensitivity and specificity, 100 and 100%, respectively; area under the receiver operating characteristic curve, 1.0; p  <  .001).²⁶ However, unlike our study authors of the mentioned study did not suggest a minimum cut-off value required to detect lesions with ultrasound. A recent in vivo study,²⁷ found that high-resolution ultrasound with color Doppler provided useful information for the differentiation and assessment of granulomas and radicular cysts. Although lesion depth, surface area and volume were underestimated, lesion width and pathology as well as treatment outcomes were accurately assessed by using ultrasound.²⁷ Analogous to our results, the findings of the mentioned study showed significantly lower measurements for depth (p = 0.004), surface area (p < 0.001) and volume (p < 0.001) with ultrasound as compared to CBCT. We found excellent ICC values for the measurement of buccal bone thickness around artificial lesions in terms of intra- and interrater reliability for the observers and their two readings. Moreover, our ultrasound measurements also highly correlated with those of CBCT measurements for buccal bone thickness measurements. We did not perform histological and direct measurements as our methodology and CBCT measurements were proved to be accurate previously. CBCT measurements could be considered as reference measurements for periapical bone lesions without soft tissue involvement.²⁷

For the measurement of lesion width, height, depth, surface area, and volume, we found excellent ICC values in terms of intra- and interrater reliability for the observers and their two readings. Ultrasound measurements also highly correlated with those of CBCT measurements for lesion width and height measurements. In the present study, we did not use Doppler to diagnose and differentiate lesions since our main objectives were to assess the necessary minimum bone thickness to detect periapical lesions by ultrasound along with 2D and 3D measurement accuracy of ultrasound in comparison to CBCT. Previous studies which assessed the accuracy of ultrasound in the detection of periapical lesions mostly reported ultrasound imaging to be successful in the determination of lesion origin and nature,^30–32 however; with some degree of underestimation in lesion dimensions.^30,31 In a previous study,³¹ authors assessed 15 intraosseous jaw lesions and found that size values were in almost complete agreement with CBCT in 12 cases The size of three lesions could not be measured due to the thickness of buccal cortical plate. Authors suggested that ultrasound might be a useful tool in estimating the size of intraosseous jaw lesions with little underestimation. They also confirmed that ultrasound imaging could provide significant diagnostic information regarding the content of jaw bone lesions where there was thin enough buccal bone. However; unlike our study authors were unable to reveal a specific minimum bone thickness value for the detection of lesions by ultrasound. In another study,³² a total of 123 lesions were evaluated; 74 (60.2%) were cysts and 49 (39.8%) were tumors or tumor-like lesions. The ultrasound and histopathological findings on the content of the lesions highly correlated with each other (p < 0.001). Statistically significant differences were reported in superoinferior measurements of jaw lesions by using ultrasound images and CBCT (p < 00.1); however, no statistically significant differences were reported in mesiodistal (p = 0.700) or buccolingual (p = 0.572) measurements.³²

Similar to ours, authors of a recent study,³³ evaluated the accuracy of ultrasound examination (USE) for the detection of artificial bone defects in bovine mandibles in the absence of complete erosion of the cortical bone plate and to determine the minimum cortical thickness that constitutes a barrier for ultrasound waves. The USE was performed with a Toshiba Medical‐Aplio i700, (Toshiba, Tokyo, Japan) by using a high‐frequency linear transducer where the frequency was 4.5 Mhz. USE showed high sensitivity (97.3%), negative predictive value (89%), and accuracy (97.8%) along with perfect score for specificity and positive predictive value. Analogous to our findings, authors found very high intra‐ and interobserver agreement.³³ Authors, demonstrated that artificial bony lesions in bovine mandible sections were detectable with USE independent of the diameter, thickness and presence/absence of perforation in the buccal cortical plate. USE was highly accurate and reliable for the detection of artificial lesions within bovine mandibles, regardless of the thickness or presence of the cortical plate. Authors also suggested that further studies comparing different ultrasound systems with CBCT units should be conducted on humans to investigate lesions of smaller diameters as conducted in our present research.³³ Different findings found in various studies might be attributable to different materials used, experimental design, devices and settings utilized.

Imaging settings and software used and operator experience along with selected sample may be detrimental for the assessment of periapical lesions with ultrasound and CBCT. We utilized 14 MHz hockey (stick) probe for ultrasound examination and CBCT images taken at 96 kVp, 5mA, and a 0.2 mm voxel size. The validity and concordance of ultrasound and CBCT settings and build-in and third-party softwares used in the present study were previously demonstrated.²⁷ Obviously, further research is recommended to investigate the effect of bone thickness on the capability of ultrasound to show periapical lesions. It is recommended that future studies utilizing various ultrasonographic machines, probes, and settings are conducted.

Conclusion

We suggested that a buccal thickness of approximately 1.28 mm or less might be accepted as a cut-off value for the detection of intraosseous periapical lesions with 14 MHz high resolution ultrasound probe. We obtained excellent ICC values for intra- and interrater agreement values for all measurements conducted. Buccal bone thickness and lesion width and height measurements obtained by ultrasound highly correlated with those of CBCT measurements. However, we found statistically significant differences between measurements obtained by ultrasound and CBCT for lesion depth, surface area and volume.