Ultrasound grayscale image quality comparison between a 2D intracavitary transducer and a 3D intracavitary transducer used in 2D mode: A phantom study

Abstract Purpose It is unclear if a 3D transducer with the special design of mechanical swing or 2D array could provide acceptable 2D grayscale image quality for the general diagnosis purpose. The aim of this study is to compare the 2D image quality of a 3D intracavitary transducer with a conventional 2D intracavitary transducer using clinically relevant phantom experiments. Methods All measurements were performed on a GE Logiq E9 scanner with both a 2D (IC5‐9‐D) and a 3D (RIC5‐9‐D) transducer used in 2D mode. Selection of phantom targets and acquisition parameters were determined from analysis of 33 clinical pelvic exams. Depth of penetration (DOP), contrast response, contrast of anechoic cylinders (diameter: 6.7 mm) at 1.5 and 4.5 cm depths in transverse planes, and in‐plane resolution represented by full‐width half‐maximum of pin targets at multiple depths were measured with transmit frequencies of 7 and 8 MHz. Spherical signal‐noise‐ratio (SNR) (diameter: 4 and 2 mm) at multiple depths were measured at 8 MHz. Results RIC5‐9‐D demonstrated <8% decrease in DOP for both transmit frequencies (7 MHz: 69.7 ± 8.2 mm; 8 MHz: 64.3 ± 7.8 mm) compared with those from IC5‐9‐D (7 MHz: 73.9 ± 4.4 mm; 8 MHz: 69.4 ± 7.8 mm). A decreased anechoic contrast was observed with a 4.5 cm depth for RIC5‐9‐D (7 MHz: 23.2 ± 1.8 dB, P > 0.05; 8 MHz: 17.7 ± 0.9 dB, P < 0.01) compared with IC5‐9‐D (7 MHz: 25.9 ± 1.2 dB; 8 MHz: 21.5 ± 0.8 dB). The contrast response and spatial resolution performance were comparable between the two transducers. RIC5‐9‐D showed comparable SNR of anechoic spheres compared to IC5‐9‐D. Conclusions 2D images from a 3D probe exhibited comparable overall image quality for routine clinical pelvic imaging.


| INTRODUCTION
Transvaginal ultrasound has been widely used as a routine diagnostic tool in pelvic imaging for decades. 1,2 Conventional two-dimensional (2D) ultrasound transducers are commonly used in clinical practice for gynecology exams. Previous studies have demonstrated that volumetric three-dimensional (3D) images could provide unique benefits of diagnosing abnormalities such as unicornuate uterus. 3,4 However, the freehand sweeping of a traditional 2D transducer to form 3D images is heavily dependent on operator's skill and is susceptible to measurement errors as the actual transducer position during the sweeping is not known by the scanner.
Alternatively, the 3D transducer techniques have been developed using either a motorized mechanism or a 2D array approach. 2 The programmed volumetric acquisition by the 3D transducer is relatively independent of operator's skill and enables an extensive visualization of pelvic structures. 5,6 The clinical practice in our Radiology department uses 2D intracavitary transducers for routine diagnosis. A 3D acquisition may be requested by the referring physicians or radiologists based on findings from 2D images. In this case, we must either switch intracavitary transducers during the exam or schedule a second exam specifically with a 3D transducer. Both alternatives can inevitably lower the clinical efficiency and pose inconvenience for patients. It has been proposed to use 3D transducers for all transvaginal ultrasound imaging in our practice. As such, during a single visit without switching transducers, standard 2D images would be acquired with the 3D transducer. In addition, a 3D acquisition would be obtained and reconstructed using the same 3D transducer for further evaluation for each patient. Previous studies comparing 2D and 3D transducers were focused on the specific diagnostic tasks, such as the accuracy of uterus volume measurements 7 and the visualization of intrauterine devices and deep infiltrating endometriosis. 8,9 However, it is still unclear if a 3D transducer with the special design of mechanical swing or 2D array could provide acceptable 2D grayscale image quality for the general diagnosis purpose. To accomplish comprehensive image quality assessments, it is desirable to incorporate the scanning variations in gynecology exams.
The purpose of this study is to compare the 2D basic image quality of a 3D intracavitary transducer with a conventional 2D intracavitary transducer using a clinically relevant phantom study.

2.A | Clinical exam survey
To ensure the clinical relevance of acquisition parameters and phantom targets used for performance measurement, 33 clinical exams using the 2D transducer model on the GE Logiq E9 ultrasound system (GE Healthcare, Milwaukee, WI, USA) for pelvic imaging from our clinical practice were randomly identified using a customized informatics toolbox. 10 A total of 188 images from these exams were reviewed by a sonographer with more than 20 years of experience.
Acquisition parameters and the characteristics of cysts (including echogenicity, size and depth), as one of the most common pathological targets, 11 were used to determine the characteristics of targets to be measured and corresponding acquisition parameters (Table 1), including the exclusive use of harmonic imaging mode in our tests.
Discussion with the radiologists also indicated that it is necessary to mimic and investigate the clinical task of searching for and counting follicles. The image depth and signal appearance information from review also indicated that the section of the phantom with 0.5 dB/ cm/MHz attenuation coefficient was most appropriate for our measurements. Two acquisition modes, survey and detail characterization, were identified. The survey acquisitions were applied to search for possible occult findings or pathologies in a global manner while detail characterizations were applied for the optimal appearance of specific targets upon detection with adjusted parameters, for example, focus zone depth and gain.

2.B | Phantom studies
All phantom image acquisitions were made with a GE Logiq E9 ultrasound system, the same system used for clinical exams in our prac- setting of focal zone and gain was applied; while with the detail characterization mode, focal zone and gain were adjustable to achieve the optimal definition of the target at a specific depth. Table 2 illustrates the acquisition parameters for survey mode used in this study. All other acquisition parameters such as frame averaging were equivalent between the two transducers. Five repeated measurements with the CIRS phantom were acquired for each of the following tasks: DOP: DOP was calculated from a pair of images (one with background base material of the phantom, the other in air) with the same acquisition settings. 12 Gray Contrast Response: Contrast response was measured with the survey mode as the gray level value per dB, using the cylindrical targets with different echogenicities (−9, −6, −3, +3, and +6 dB). 13 The echogenicity of the background base material (0 dB) was also included in the calculation.
Anechoic Contrast: Anechoic cylinders with 6.0 mm diameter at 1.5 and 4.5 cm depth were measured in the transverse plane for both survey and detail characterization acquisitions. 14 The gray level difference between the anechoic cylinder and background regions were measured and converted to dB scale using the averaged gray contrast response values (measured above, in gray level per dB) for each probe and transmit frequency.   paired T-test was conducted to compare the measurement differences between the two transducers. Statistical significance was considered with P < 0.05.

| RESULTS
The 2D images from the RIC5-9-D demonstrated a significantly smaller (P < 0.001) DOP for both transmit frequencies (7 MHz: T A B L E 2 Acquisition parameters for 2D image performance comparisons between the 2D (IC5-9-D) and 3D (RIC5-9-D)

Performance
Scanner control IC5-9-D RIC5-9-D The listed parameters were applied for the survey mode acquisitions. Gain values were adjusted to avoid any pixel value saturation for the corresponding transmit frequency which were separated by a slash (/). Controls of acquisition parameters with an asterisk (*) were optimized for each specific target when assessing performance in detailed characterization mode.

| DISCUSSION
It is important to assess and compare basic image quality before exclusively using the 3D transducer in clinical practice. 18 Objective measurements have been used for ultrasound quality control programs. 13,19 Other scenarios such as acceptance tests and equipment purchase evaluations could also benefit from the objective assessment with appropriate tissue-mimicking phantoms. Objective Our data showed, in general, that the RIC5-9-D provided comparable imaging performance to the IC5-9-D. The approximately 6-7% or <5 mm decrease in DOP for RIC5-9-D compared to IC5-9-D could be due to the oil in typical 3D transducer designs which couples the array with the scanning window. 22 As our clinical exam survey results showed, the majority of clinical interest regions or targets for pelvic intracavitary exams are superficial (<5.0 cm depth).
At both transmit frequencies, the RIC5-9-D could still provide sufficient signals with meaningful echo information for clinical intracavitary exams. Therefore, the slight decrease in DOP would not pose a substantial impediment for using the RIC5-9-D.
In this study, we found the RIC5-9-D had a slightly worse contrast of anechoic cylinders and similar SNR of anechoic spheres, compared to the IC5-9-D. Several reasons could explain the observed discrepancies between contrast and SNR results. First, contrast measurements do not take target size into calculation, while SNR does. Second, the SNR of spherical targets were comprehensively affected by the anechoic contrast as well as the spatial resolution performances. 23 In addition, all the SNR results were above the visual detection threshold that the small differences in contrast might not be clinically important. 24 This study established the framework of utilizing physics tests for obtained with both probe models in the same patients should also be reviewed. This step is currently underway in our practice. The final decision on which probe to use in routine practice will also be dependent on several other factors, including the cost of replacing all 2D transducers and adding 3D imaging capability upgrades to more scanners. Finally, the image quality and measurement accuracy using the 3D mode should be evaluated as the next step.

| CONCLUSION
The 2D grayscale image quality of a conventional 2D intracavitary and 3D intracavitary transducer was compared in this study. We demonstrated that the 3D RIC5-9-D was comparable to IC5-9-D in terms of 2D image quality. With further confirmation from patient image comparison, RIC5-9-D intracavitary transducers could routinely be used in all pelvic exams in the practice. F I G . 6. Measurements of lesion signal-noise-ratio (SNR) for the 2D (IC5-9-D) and 3D (RIC5-9-D) transducers from anechoic spheres with 4.0 mm (a) and 2.0 mm (b) diameter. Mean ± SD with * and ** indicating statistical significances P < 0.05 and P < 0.01, respectively.