Attenuation and dispersion in cancellous bone: implication 1 in developing ultrasound- guided spinal navigation

25 Background: The purpose of this study was to understand the acoustic properties of 26 vertebral cancellous bone by combining micro-CT scan data and use this as the 27 theoretical basis for ultrasonic navigation with posterior pedicle screw fixation during 28 spinal fusion surgery. 29 Results: Two bovine spinal cancellous bone blocks and two human spinal allograft 30 bone blocks simulated the condition of cancellous bone in the pedicle screw channel. 31 Transmission experiments were performed on them, including amplitude, attenuation 32 and sound velocity tests. Four unfocused wideband ultrasonic transducers were used 33 for the detection, with the central frequencies of 2.2MHz, 2.5MHz, 3MHz and 34 12MHz respectively. The results were favorable and stable. The amplitude of the 35 signal decreased with depth penetration (p< 0.05). With the increase of frequency, 36 sound attenuation and sound velocity increased (p< 0.05). 37 Conclusions: In summary, the conclusions lay a theoretical foundation for the 38 ultrasonic navigation system. However, how ultrasound navigation will facilitate 39 pedicle screw insertion in spine surgery remains to be determined. Therefore, 40 ultrasonic guided pedicle screw implantation is effective and promising in theory. 41


138
Micro-CT scanning 139 By micro-CT scanning of four different bone blocks, we obtained multiple 140 parameters for two bovine and two human allograft bones (Fig. 1). The BMD of the two 141 bovine bones had no significant difference and was higher than that of human allograft, 142 while BMD of high density allograft was higher than that of low density allograft 143 human bone (Table 1). BV/TV of low density allograft human bone (26.11±0.03) was 144 higher than that of low density human allograft bone (14.21±0.02). Tb. Th and Th. N of 145 human allograft bone (high density) were higher than human allograft bone (low 146 density). Th. Sp and SMI of human allograft bone (high density) were lower than these 147 of human allograft bone (low density). Nonetheless, there was no significant difference 148 in the above indexes. However, the above parameters were not significantly different in Table 1 The characteristics of each bone block   This phenomenon is particularly evident in relatively low-density bone samples.

160
Acoustic attenuation and thickness of bone samples 161 In this experiment, the sound attenuation coefficient increased with the thickness 162 of each bone sample (Fig. 3). The reason is that the trabeculae in cancellous bone are 163 anisotropic. When the thickness of bone sample is relatively large, the attenuation of 164 ultrasonic signal is relatively large, and the acoustic attenuation coefficient measured by 165 experiment is relatively large. As the thickness of cancellous bone sample decreased, Under the same thickness condition, the BMD of bovine bone was higher than that  structure. If the trabecular structure is mainly lamellar structure, then SMI is close to 0.

304
On the other hand, SMI is closer to 3 if it is predominantly rod-shaped trabeculae. In 305 osteoporosis, the trabecular bone changes from plate to rod, and the value increases.

306
For bone samples, the higher the bone density, the lower the penetration amplitude.

307
The bone density of cattle was higher than that of allogeneic bone, and the penetration 308 amplitude of cattle bone was lower than that of allogeneic bone. There was no successively from 8.00mm to 6.50mm, 5.00mm, 3.00mm and 2.00mm.

358
A custom-designed ultrasonic immersion system for through transmission 359 measurements was applied in the current investigation and is shown in Fig. 7(A). The

360
water tank was filled with deionized-degassed water maintained by a canister filter 361 (Liqui-Cel® membrane 2.5x 8X-50 Fiber) and a custom-designed integrated filtration system and all experiments were carried out in degassed water at room temperature 363 around 25℃, and the speed of sound is 1497m/s. The air bubbles in the water will 364 scatter the ultrasonic waves; it will greatly affect the accuracy and reliability of the 365 experimental results. Therefore, before the experiment, the pulse-echo measurements of 366 the water medium was carried out to confirm that there was no millimeter-sized air 367 bubbles into the water tank through the filtration system.

368
The cancellous bone chips were fixed on a four-dimensional control platform, and 369 could perform X, Y, and Z translations, and rotation about the Z-axis along with the 370 platform. As indicated in Fig. 7(B), the transmitting transducer (custom-designed) and Needle Hydrophone and displayed on the oscilloscope. The RF data was saved as a .csv 397 file for subsequent data processing.

398
The standard substitution technique was used to measure the ultrasonic 399 characteristics of cancellous bone. As Fig. 8 shows, the principle is that when a  When the ultrasonic wave is vertically incident to a large enough smooth flat 421 interface, a reflected wave in the first medium will be generated in the opposite 422 direction of the incident wave, and a transmitted wave in the second medium will be 423 generated in the same direction as the incident wave.

424
The distances between the sample and transducers and the radius of the transducers 425 are very important information. The size of the bone sample used in the experiment was 426 greater than 20mm, while the maximum size of the ultrasonic transducer used in the 427 experiment was less than 5mm. In order to ensure the accuracy and reliability of the 428 experimental test data, the pulse transmission method is adopted in the experiment. In 429 addition, the transducer should be as close to the bone sample as possible while the 430 echo signal and the excitation pulse can be distinguished in the data analysis (Fig. 9).

431
The standard substitution technique was used to measure the ultrasonic 432 characteristics of cancellous bone. The characteristics of cancellous bone at different ultrasonic frequencies were obtained by calculating the time and amplitude changes 434 between the reference signal and the transmitted signal (Fig. 10).
variance, or Dunnett T3 with heterogeneity of variance for multiple comparisons.    534 All the authors of the paper approved the publication of the article.

535
Availability of data and materials 536 Data related to the current study are available from the corresponding author on 537 reasonable request.

538
Competing interests 539 The authors declare that they have no competing interests.