Bone Chemical Composition Assessment with Multi-Wavelength Photoacoustic Analysis
, Yunhao Zhu 3,4, Kenneth M. Kozloff 4, Basma Khoury 4, Yejing Xie 1, Xueding Wang 4, Meng Cao 3, Jie Yuan 3,*, Dean Ta 5 and Qian Cheng 2,*Round 1
Reviewer 1 Report
The authors of the submitted manuscript photoacoustically evaluated the bone marrow density (BMD). However, the authors should address the below concerns before consideration of the publication.
- This study’s analysis seems to be incomplete. Especially, the lack of the gold standard measurement of the contents may be a critical flaw for this study. As claimed by authors, we can “EXPECT” EDTA treatment may change the content of BMD, but that is not a scientific approach. Please provide the true content volume changes and compare it with your results.
- As authors mentioned in the manuscript “For the in vivo study, the human bone is large which leads to high light attenuation in it. Therefore, the compensation of light attenuation should be considered in future work.”, the light penetration into the bone caused critical feasibility issue as the non-invasive technology. I have no idea how authors can overcome the inherent problem of light. Analysis on the short axis of the bone is not a non-invasive model if authors want to develop the non-invasive bone analysis tool.
Author Response
Dear Reviewer,
We want to appreciate your great efforts reviewing our manuscript entitled “Quantification of Bone Chemical Composition with Multi-wavelength Photoacoustic Analysis” (Manuscript ID: applsci-938563) and your thoughtful comments. The manuscript has been revised accordingly, with point-by-point responses to reviewers’ comments listed below.
The authors of the submitted manuscript photoacoustically evaluated the bone marrow density (BMD). However, the authors should address the below concerns before consideration of the publication.
1.This study’s analysis seems to be incomplete. Especially, the lack of the gold standard measurement of the contents may be a critical flaw for this study. As claimed by authors, we can “EXPECT” EDTA treatment may change the content of BMD, but that is not a scientific approach. Please provide the true content volume changes and compare it with your results.
Answer: We have conducted the X-ray images by Faxitron Microradiography on each bone to confirm the decalcification in the bones after being treated with EDTA (see Figures 2 (a)-(c)). The X-ray images were further quantified using Matlab software to evaluate the mean grayscale value for each sample as an indicator of the BMD in the bone.
2. As authors mentioned in the manuscript “For the in vivo study, the human bone is large which leads to high light attenuation in it. Therefore, the compensation of light attenuation should be considered in future work.”, the light penetration into the bone caused critical feasibility issue as the non-invasive technology. I have no idea how authors can overcome the inherent problem of light. Analysis on the short axis of the bone is not a non-invasive model if authors want to develop the non-invasive bone analysis tool.
Answer: For the in vivo study, it is possible to compensate the light attenuation in the bone by using photoacoustic and ultrasound dual-modality bone assessment system which is working in transmission mode. First, we compensate the ultrasound attenuation in the bone by using the ultrasound attenuation coefficient obtained from the ultrasound bone assessment system. Second, we can get the effective optical attenuation coefficient μeff by exponential fitting in the time domain for the PA signal after the compensation of ultrasound attenuation. Third, with the known effective light attenuation coefficient μeff, the light attenuation in the bone can be compensated.
Besides, we have studied the feasibility of the bone assessment in vivo and non-invasively based on photoacoustic and ultrasound dual-modality system. More details can be found in the paper shown below https://spj.sciencemag.org/journals/bmef/aip/1081540/.
Author Response File: 
Author Response.pdf
Reviewer 2 Report
This manuscript reports the application of using multi-wavelength PA to quantify the bone composition. It’s an interesting application and valuable. However, I have the following concerns.
1. Please include the scale bar in Fig. 2a.
2. How do you use the transducer with black tape to calibrate PA signal? Does the black tape has similar absorption within the wavelength range? Why don’t you just use the photodiode to calibrate the pulse energy?
3. Please specify how many wavelengths are used within the range.
4. As mentioned in the manuscript, most PA signal detected is from the surface. But the ultrasound transducer is placed perpendicular
5. To do the spectral unmixing from the PA signal, is the Grüneisen parameter also known?
6. Figure 5 is very confusing to me. How do you quantify the contribution in percentage? Some wavelengths show the total percentage larger than 100%. What does that mean?
Author Response
Dear Reviewer,
We want to appreciate your great efforts reviewing our manuscript entitled “Quantification of Bone Chemical Composition with Multi-wavelength Photoacoustic Analysis” (Manuscript ID: applsci-938563) and your thoughtful comments. The manuscript has been revised accordingly, with point-by-point responses to reviewers’ comments listed below.
This manuscript reports the application of using multi-wavelength PA to quantify the bone composition. It’s an interesting application and valuable. However, I have the following concerns.
1. Please include the scale bar in Fig. 2a.
Answer: It has been updated in the manuscript.
2. How do you use the transducer with black tape to calibrate PA signal? Does the black tape has similar absorption within the wavelength range? Why don’t you just use the photodiode to calibrate the pulse energy?
Answer:
(1) We used the black rubber to calibrate the PA signal in this study. The black rubber has a uniform absorption value within the wavelength range of 680nm-950nm. After passing through a beam splitter, 10% of the laser energy was guided to a black rubber and 90% of the laser energy was guided to the bone sample to generate the PA signal of bone. We recorded the PA signal generated by black rubber and bone samples by using different ultrasonic transducer simultaneously. Then, we use the peak to peak value of the PA signal generated by the black rubber as the recorded energy which can be used for later calibration of the PA signal from bone.
(2) We have considered these two similar methods. Using the photodiode is more straightforward to record the laser energy for each laser pulse. Comparing with the photodiode, using the black rubber can set up the identical signal path with the bone signal due to both signals are acoustic signals, including the transducer and amplifier.
3. Please specify how many wavelengths are used within the range.
Answer: 28 wavelengths are used within the range. The additional information has been added in the manuscript.
4. As mentioned in the manuscript, most PA signal detected is from the surface. But the ultrasound transducer is placed perpendicular
Answer: It needs much more water to couple the ultrasound signal if we put the ultrasound transducer above the bone. However, the water will affect the optical absorption of the bone samples, especially for the wavelength larger than 900nm.
5. To do the spectral unmixing from the PA signal, is the Grüneisen parameter also known?
Answer: The Grüneisen parameter is unknown, however, the Grüneisen parameter is independent with the wavelength, then it can be removed by normalizing the PA spectral as shown in equation (4). Based on the normalized PA spectral which is independent with the Grüneisen parameter, we can get the unmixing from the PA signal without known the Grüneisen parameter.
6. Figure 5 is very confusing to me. How do you quantify the contribution in percentage? Some wavelengths show the total percentage larger than 100%. What does that mean?
Answer: Figures 5 (a)(b)(c) and (e)(f)(g) shows the relative content of mineral, Hb, HbO2, lipid, collagen and water. The total percentage of those six chemical components at all wavelength are equal to 100%. For example, for the fresh bone, the percentage value of each components is about: 52%, 24%, 0%, 0%, 24% and 0% at the 680nm. Figure 5 (d) and (f) shows the percentage of whole blood (total of Hb and HbO2), while Figure 5(h) shows the total content of lipid, collagen and water.
Author Response File: Author Response.pdf
Reviewer 3 Report
See attached document
Comments for author File: Comments.pdf
Author Response
1. I find the title somewhat misleading by using “Quantification” when the authors mean assessing the relative content of the chemical components. Quantification of bone chemical composition suggest that the absolute amount of material is determined. The manuscript does not address this issue. In addition, when quantifying something uncertainties are required otherwise it does not make sense.
Answer: We totally agree with you. We find the title somewhat misleading by using “Quantification”, what we truly mean is assessing the relative content of the chemical components. Therefore, the title has been changed to “Bone Chemical Composition Assessment with Multi-wavelength Photoacoustic Analysis”.
2. There are no error bars in the figures, only standard deviations.
Answer: The error bars have been added in the revised manuscript.
3. Did the relative spectral features change when other levels of fluence where used? In photoacoustics the interaction between different materials can vary depending on the level of excitation.
Answer: In this study, since the parameter about the level of fluence is removed in normalized PA spectral curve, the relative spectral features is independent on the level of the fluence, as shown in equation (3) and (4). Besides, the levels of fluence was controlled under ANSI safety (<20 mJ/cm2) in the experiment. At this range, the interaction between different materials will not varied.
4. In line 128 the standard deviation must have the same number of digits as the value measured.
Answer: It has been updated in the revised manuscript.
5. In relation to performing measurements on the bone, it would be interesting to know if the position of the laser beam spot is important. I assume the bone is not homogenous in terms of chemical content across its diameter.
Answer: The position of the laser beam spot has been added in Figure 1(d). The diameter of the laser beam is about 4 mm while the thickness of the trabeculae in the bone is about 0.1~0.2 mm, therefore, the bone could be considered as a macro-homogeneous tissue in this study. Besides, to improve the stability and reduce the measurement error, as well as the effect of inhomogeneity of bone, the PA spectra for each bone sample was detected at 3 different positions and averaged for further analysis.
6. A reference related to eq. 1 would be useful.
Answer: The reference has been added in the revised manuscript.
7. Why is the standard deviation at the long wavelength end so small for Group 3 relative to the two other groups?
Answer: The main absorption components at the long wavelength are blood, lipid, and water. For the Group3, all the bone samples have been placed in EDTA solution for about 26 hrs, leading to the loss of most blood. Therefore, the content of absorber among bone samples in Group 3 is reduced, and the difference between the absorption curve for each bone in Group 3 is also decreased. It is the reason that the standard deviation at the long wavelength end small for Group3 relative to other groups.
8. In relation to Fig. 5, it would be useful to have error bars for the contributions determined and in the text some discussion about the reproducibility of the measurements.
Answer: The error bars have been added in Figure 5 in the revised manuscript. The discussion about the reproducibility of the measurements has been added in line 329-336.
9. In the conclusion, there is some discussion about the limitations of the study. It would be useful to include discussion about the foreseen practical use of the technique where a bone is not available in the way treated in the manuscript.
Answer: For practical use, the bone assessment technique based on the PA detecting method described in our manuscript maybe not available and should be improved with better design. First, the mode of detection should be improved based on the target bone in vivo. For example, we used the transmission PA mode for multi-wavelength PA measurement of bone in vivo in our study published recently (https://spj.sciencemag.org/journals/bmef/2020/1081540/). Second, the optical penetration depth at different wavelengths and the ultrasound penetration depth at different frequencies should be studied. The optimized range of wavelengths and the center frequency of the transducer should be used. This has been added in the Discussion part of the revised manuscript.
Round 2
Reviewer 1 Report
- We have conducted the X-ray images by Faxitron Microradiography on each bone to confirm the decalcification in the bones after being treated with EDTA (see Figures 2 (a)-(c)). The X-ray images were further quantified using Matlab software to evaluate the mean grayscale value for each sample as an indicator of the BMD in the bone.
- Yes, but it shows just trend, not quantificational analysis. What we only know is decalcification may be happened. As shown in Figure 4, Authors should provide the reference numbers (as a gold standard) or relative percentage by measuring other modality to evaluate accuracy of the proposed multi-spectral analysis. As shown in the title, quantification of bone chemical composition is the main goal of the study so that authors should provide ground truth.
- For the in vivo study, it is possible to compensate the light attenuation in the bone by using photoacoustic and ultrasound dual-modality bone assessment system which is working in transmission mode. First, we compensate the ultrasound attenuation in the bone by using the ultrasound attenuation coefficient obtained from the ultrasound bone assessment system. Second, we can get the effective optical attenuation coefficient μeff by exponential fitting in the time domain for the PA signal after the compensation of ultrasound attenuation. Third, with the known effective light attenuation coefficient μeff, the light attenuation in the bone can be compensated.
Besides, we have studied the feasibility of the bone assessment in vivo and non-invasively based on photoacoustic and ultrasound dual-modality system. More details can be found in the paper shown below https://spj.sciencemag.org/journals/bmef/aip/1081540/.
- As shown in the cited reference, only low frequency ultrasound (lower than 1MHz) may penetrate bone (normal transcranial ultrasound utilizes 0.7-0.5MHz). The setup with the single transducer centered at 5MHz, may not be available for in vivo experiment. To claim availability of in vivo experiment, authors should provide the result with lower frequency transducer, or effect and deep discussion about this issue.
Author Response
(2nd round)1: Yes, but it shows just trend, not quantificational analysis. What we only know is decalcification may be happened. As shown in Figure 4, Authors should provide the reference numbers (as a gold standard) or relative percentage by measuring other modality to evaluate accuracy of the proposed multi-spectral analysis. As shown in the title, quantification of bone chemical composition is the main goal of the study so that authors should provide ground truth.
Answer (2nd round): We totally agree with you. We find the title somewhat misleading by using “Quantification”, what we truly mean is assessing the relative content of the chemical components. In this work, the goal is accessing the relative contents of chemical components in bone. And X-ray method employed in the study can provide 2D image of the bone and we can thusly obtain the quantified grayscale value of the X-ray image to derive the trend of bone mineral content. However, as you said, we only provided the trend of the changes in different chemical components but not the arcuate value. Therefore, the title has been changed to “Bone Chemical Composition Assessment with Multi-wavelength Photoacoustic Analysis”. Besides, the relative percentage of the quantified X-ray images as the gold standard have been added in Figure 4 (a).
(2nd round)2: As shown in the cited reference, only low frequency ultrasound (lower than 1MHz) may penetrate bone (normal transcranial ultrasound utilizes 0.7-0.5MHz). The setup with the single transducer centered at 5MHz, may not be available for in vivo experiment. To claim availability of in vivo experiment, authors should provide the result with lower frequency transducer, or effect and deep discussion about this issue.
Answer (2nd round): The initial photoacoustic signal generated by the light absorption in bone tissue is a broadband signal. For ex vivo study, since we hoped to acquire the PA signal with the information as much as possible, we used a relative high-frequency transducer with a center frequency of 5MHz. However, due to the strong acoustic attenuation in the bone, especially for the high-frequency signals, most of the PA signals received by the transducer in vivo are lower than 1 MHz. Therefore, to enhance the sensitivity in detecting PA signals from bone, the transducer used for the in vivo study should have a relatively low center frequency.
For the homogenous tissue, since the profile of the multi-wavelength PA curve is independent on the frequency of PA signal, therefore, the PA spectral curves should be similar by using the transducers with different center frequencies. In this work, the bone could be considered as the macro-homogenous tissue, then the profile of the multi-wavelength PA curve is independent with the frequency of PA signal also, as shown in equation (3) and (4). However, if the center frequency of the transducer is much higher than the frequency of the PA signal generated from bone, the bone tissue cannot be considered as the macro-homogenous tissue. Hence, in the future, the effect from the center frequency of the transducer to the profile of the multi-wavelength PA curve should be further studied with more factors considered. For example, the relationship between the PA frequency and size of the different chemical components in the bone. Besides, one of the focuses in future work would be studying the PA spectral curves as functions of the center frequency of the transducer. This information has been added in the Discussion section (line 324-329) of the revised manuscript.
Reviewer 2 Report
I think the author has addressed all of my previous concerns. Thus, I reccommend it for publication.
Author Response
Thank you so much for your good suggestions and recommendation!
