A Review of the Vibration Arthrography Technique Applied to the Knee Diagnostics

Round 1

Reviewer 1 Report

1. It is better to say, “the knee is one of the most complex joints”, because shoulders are also complex.
2. Reference order should be ascending. For example, in figure1 [2,13,12,14] should be [2, 12,13,14], and [5,7,17,18-23,25] should be [5,7,17-23,25]. Please make sure the format for other references in text is consistent.
3. In Figure1, the latest reference is 2018. As a review paper, is there any study using this VAG technique from 2019-2021?
4. In Figure2, the authors mentioned the experimental protocol is sitting on a highchair with both legs freely suspended in the air. Is this the only one protocol used for VAG technique? If not, as a review paper, the other protocols should be included.
5. The references/examples in the section of signal analysis are not up to date.
6. The authors should keep the format consistent for the whole paper, such as indentation for each paragraph.
7. In Figure3 and Figure4, reference [5] was cited for (A). If (B) and (C) in both figures were from the same reference [5]? If so, as a review paper, why did the authors only choose [5] as an example?
8. The number of subsection “time-frequency” was not correct. It should be 3.2.3.
9. In Table1, the year of reference [21] is not 2008. What is the sequence to list all these studies (e.g., by year or by author’s name)? It seems that the authors listed these references randomly.
10. For the section4 (Classification), it is not clear in this paper. Should it be listed in a separate section, or included in section3?
11. The authors mentioned “the Fourier and Wavelet transform”, the relative references are needed.
12. The reference [28] was from 1985. As a review paper, is there any study from 1985 until 2021 focusing on the improvement of biomechanical modeling?

Author Response

Article :

A review of the vibration arthrography technique applied to the knee diagnostics

Response to reviewer no.1

Thank you for helping to improve the manuscript. Below are the responses to your remarks. The areas changed in the review are marked in yellow highlight.

1.  It is better to say, “the knee is one of the most complex joints”, because shoulders are also complex.

The text was adapted.

1. Reference order should be ascending. For example, in figure1 [2,13,12,14] should be [2, 12,13,14], and [5,7,17,18-23,25] should be [5,7,17-23,25]. Please make sure the format for other references in text is consistent.

Reference order was checked.

1. In Figure1, the latest reference is 2018. As a review paper, is there any study using this VAG technique from 2019-2021?

The newest references were included in the Figure 1.

1. In Figure2, the authors mentioned the experimental protocol is sitting on a highchair with both legs freely suspended in the air. Is this the only one protocol used for VAG technique? If not, as a review paper, the other protocols should be included.

Following text was added:

“Some other studies propose different measurement setup, though. Bolus et al. [27] propose a measurement an inverted squat-like position with the knee loaded by a given force. A similar setup with traditional squats was used by Olowiana et al. [30].  Kalo et al. used either sit-ups in [9] or climbing up or going down the stairs in [31].”

1. The references/examples in the section of signal analysis are not up to date.

This paper from 2020 has been added to the time-domain analysis part, in order to give more recent references :

Similarly, Olowiana et al. [30] have estimated the variability of the VAG signal by computing the variance of the mean-squared values and the signal amplitude.

 

Moreover, the article [26] from Ching chiu Tuan has been written in 2019, we assumed it was up to date. Indeed, there are not so many recently written papers about vibration arthrography.

1. The authors should keep the format consistent for the whole paper, such as indentation for each paragraph.

The paragraphs have been aligned.

1. In Figure3 and Figure4, reference [5] was cited for (A). If (B) and (C) in both figures were from the same reference [5]? If so, as a review paper, why did the authors only choose [5] as an example?

According to another reviewer the whole illustration of the chapter 3 was done differently. The first figure shows the time domain, the second the Fourier domain and the third illustrates the wavelet usage.

1. The number of subsection “time-frequency” was not correct. It should be 3.2.3.

Section number has been corrected.

1. In Table1, the year of reference [21] is not 2008. What is the sequence to list all these studies (e.g., by year or by author’s name)? It seems that the authors listed these references randomly.

The date has been corrected. The table 1 is listed by the year of publication.

1. For the section4 (Classification), it is not clear in this paper. Should it be listed in a separate section, or included in section3?

The paragraph about classification was included in section 3.3 are rewritten to link better to part 3.

Classification is the last step in the post-processing method described above. Its aim is to predict a pathology from measured signals. Typically, the classification model is trained with two datasets. One belongs to the normal and the other to abnormal knees.

A range of parameters and statistics discussed in Section 3.2. may be used for classification. Nevertheless, the parameter selection is essential and it should include highly discriminant features to increase the predictive accuracy [2]. The so-called Supervised Classification has therefore been used recently and rely on chosen features that will affect the presence or absence of a knee pathology. Befrui et al. [25] performed a Supervised Classification to distinguish asymptomatic versus OA knees, thanks to a Linear Support Vector machine. Their results were conclusive as specificity and sensitivity reached 80% and 75%, respectively. This accuracy is as high as existing diagnostic tests. Similarly, Kim et al. classified with an average accuracy of 91.4% normal and abnormal VAG signals from four parameters that are mentioned above in the time-frequency analysis part [14]. Furthermore, normal and OA knees have been classified using a discriminant function based on four other parameters [LIN W.]. Classification results showed a sensitivity of 89.52%, a specificity of 67.50% and a total accuracy rate of 81.52%. Moreover, Yang et al. [37] performed the Least-Squares Support Vector Machine (LS-SVM) and the Bayesian Decision Rule (BDR) to classify knee joint vibroarthrographic signals. The VAG signal is represented by a two-dimensional feature vector composed of a fractal scaling index and the average envelope amplitude. Figure 7 illustrates the classification result of the two techniques. It can be observed that the least-square support vector machine and the Bayesian decision rule  provide two different nonlinear decision boundaries and produce an overall accuracy rate of 82.67% and 88%, respectively. Other classification studies are listed in Table 2.

 

1. The authors mentioned “the Fourier and Wavelet transform”, the relative references are needed.

References [32] and [33] have been added.

1. The reference [28] was from 1985. As a review paper, is there any study from 1985 until 2021 focusing on the improvement of biomechanical modeling?

References 34-36 were added. However, the up-to-date models do not provide dynamical simulations for the VAG phenomena. The text was adapted:

“Therefore, future research could focus on a biomechanical modeling of knee to come to further understanding of mechanisms behind VAG signals. Although numerous studies dealt with biomechanical modelling of the knee joint [34-36], only a handful tried to explain the nature of VAG signals.”

Author Response File: Author Response.docx

Reviewer 2 Report

Thanks for giving me the possibility to review this paper. Finding new non-invasive solutions to assess knee disorders is mandatory nowadays, as TKA requests are progressively increasing worldwide. Although the paper is well written and structured, it would be more suitable for a book chapter instead of a journal article. Narrative reviews are fascinating but provide low quality of evidence in the international literature. If authors want to submit the article to a journal, I suggest performing a systematic review of the literature on this topic and use the discussion to better understand the feasibility and the validity of this technique. Moreover, I suggest reducing the introduction, as it is a bit redundant, focusing on the problem (necessity to diagnose knee disorders with non-invasive methods) and the solution (VAG). 

Therefore, I am sorry, but I have to reject the paper

Best regards

Author Response

Response to reviewer no.2:

We understand that the chapter in the book may be more adequate. We try to do better next time.

The text was adapted according to remarks of other reviewers. The new or modified parts are highlighted in yellow.

Thank you for your time reading the manuscript.

Reviewer 3 Report

This is an interesting paper which will bring great value to the readers as it compiles a series of publications on knee diagnostics using vibration arthrography technique. The following recommendations are required for the improvement of the paper:

1) the advantages and disadvantages of each diagnostic technique (introduction - paragraph 3) should be summarised into a table with more references.

2) as this is a review paper, it will be good to illustrate how the sensors are attached to the knee, referring to line 96.

3) the authors should also discuss about the type of sensors used, including bandwidth, sensitivity, etc as these are important for accurate signal acquisitions. 

4) the authors illustrated 3 waveforms in Figure 3. more examples of signals are required to show the variation of signals acquired. These signals should also be correlated with the types of sensors used.

5) The authors explained about 3.2.2 Frequency domain analysis and 3.2.1 Time-Frequency domain analysis but illustrations should also be included. 
6) should be 3.2.2 for Time-Frequency domain analysis.

7) examples of waveforms or data are required  to highlight the classification of the signals and correlation to the type of conditions.

8) the accuracy of the methods based on different types of analyses should be included in the conclusion. 

Author Response

Thank you for helping to improve the manuscript. Below are the responses to your remarks. The areas changed in the review are marked in yellow highlight.

Response to reviewer no.3

This is an interesting paper which will bring great value to the readers as it compiles a series of publications on knee diagnostics using vibration arthrography technique. The following recommendations are required for the improvement of the paper:

1) the advantages and disadvantages of each diagnostic technique (introduction - paragraph 3) should be summarized into a table with more references.

Table 1 was included in the introduction to list the major techniques in current use.

2) as this is a review paper, it will be good to illustrate how the sensors are attached to the knee, referring to line 96.

Figure 2 has been added showing the two most common sensors attached.

3) the authors should also discuss about the type of sensors used, including bandwidth, sensitivity, etc as these are important for accurate signal acquisitions. 

The choice of an accelerometer depends on parameters such as its frequency response and its sensitivity. The typical bandwidth is up to 10kHz while their sensitivity ranges from 100 to 600mV/g [11, 14, 23, 25, 27]. For obvious reasons, accelerometers are often very small in size, thus enabling their placement upon skin.

4) the authors illustrated 3 waveforms in Figure 3. more examples of signals are required to show the variation of signals acquired. These signals should also be correlated with the types of sensors used.

Figure 5 was added showing waveforms and corresponding spectrograms.

5) The authors explained about 3.2.2 Frequency domain analysis and 3.2.1 Time-Frequency domain analysis but illustrations should also be included. 

Figure 5 was added showing spectrograms and Figure 6 showing a detailed view of a VAG signal and Daubechies wavelets.

6) should be 3.2.2 for Time-Frequency domain analysis.

Section number was updated.

7) examples of waveforms or data are required to highlight the classification of the signals and correlation to the type of conditions.

Figure 7 was added showing an example of SVM technique applied to the classification of VAG signals.

8) the accuracy of the methods based on different types of analyses should be included in the conclusion. 

Following text was added in the conclusion:

Classification achieved impressive results with accuracy ranging from 81.52% to 91,4% according to the signal analysis techniques used [2, 14, 22, 25, 37, 38].

 

 

Author Response File: Author Response.docx

Round 2

Reviewer 2 Report

I appreciate the revision of the authors. However, my concerns about the quality of evidence and the suitability of this paper in the journal still remain.