Investigating Human Torso Asymmetries: An Observational Longitudinal Study of Fluctuating and Directional Asymmetry in the Scoliotic Torso

Round 1

Reviewer 1 Report

The paper investigates human torso asymmetries: an observational longitudinal study of Fluctuating and Directional Asymmetry in the scoliotic torso. The results seem to be new and interesting. However, the methodology is not clearly discussed. There are some places that need to be clarified. Some detailed comments as follows.

The problem of directional and fluctuating asymmetries in the torso of patients with idiopathic scoliosis has been extensively studied in the literature during the last twentyish years. The novelty of the work should be highlighted. It seems the methods used here are basic and primary. The new bit is just the dataset. If this is not the case, the authors should give detailed explanations on the novelty and technical difficulties.

The sample was formed with 24 adolescents with idiopathic scoliosis and 24 control  subjects of the same range of age (10–18) after the approval of the Research Ethics Committee of Canabria Spain was applied with a compliance of 18 out of 22 and all anthropometric data were collected at the  baseline time and six months later, including height, weight and body fat percentage. BMI was then calculated and used for further analysis, and the Cobb angle was measured three–times on different days in all patients. The method is not described in detail. 

A general comment of the work is that the table presentation is not apealing. I suggest using more box plots and line graphs as they are more informative than the limited data given in the several tables. It would be not reproducible in the current format.

Line 230, it is mentioned that the human torso is not perfectly symmetrical and the presence  of directional asymmetry in the control group was not, therefore, a surprise compared to [20]. This is not easy to follow. Perhaps the authors would like to give some more explanations on the rationale.

 

Author Response

The paper investigates human torso asymmetries: an observational longitudinal study of Fluctuating and Directional Asymmetry in the scoliotic torso. The results seem to be new and interesting. However, the methodology is not clearly discussed. There are some places that need to be clarified. Some detailed comments as follows.

Point 1: The problem of directional and fluctuating asymmetries in the torso of patients with idiopathic scoliosis has been extensively studied in the literature during the last twentyish years. The novelty of the work should be highlighted. It seems the methods used here are basic and primary. The new bit is just the dataset. If this is not the case, the authors should give detailed explanations on the novelty and technical difficulties.

Response 1: Thank you very much for your comments regarding the results of our research. You are indeed right with the novelty and the difficulties found during the research process. We have added a paragraph in the discussion to highlight the new possibilities that GMM brings to the study of shape in patients with scoliosis. It is well-known the ability of GMM in morphometric and shape studies, but it has never been applied before to the 3D shape of these patients, which is our strength. To the date, all research about directional and fluctuating asymmetry in patients with scoliosis had been done with traditional morphometrics, angles, linear measurements... Regarding the technical difficulties, we have completed the description of the methods, and also we have expressed the main limit that it could present for the use of clinicians. It is reflected now at the end of the discussion, and it is related with the manually step-by-step GMM analyses done.

Point 2: The sample was formed with 24 adolescents with idiopathic scoliosis and 24 control  subjects of the same range of age (10–18) after the approval of the Research Ethics Committee of Cantabria Spain was applied with a compliance of 18 out of 22 and all anthropometric data were collected at the  baseline time and six months later, including height, weight and body fat percentage. BMI was then calculated and used for further analysis, and the Cobb angle was measured three–times on different days in all patients. The method is not described in detail. 

Response 2: Thank you very much for your comment, it really seems that the explanations of data acquisition methods are scarce in our manuscript. We have added new lines describing in depth materials used in data acquisition, the software used to do some of the measurements and the scanning protocol followed. We hope this new version is more completed than the previous.

Point 3: A general comment of the work is that the table presentation is not apealing. I suggest using more box plots and line graphs as they are more informative than the limited data given in the several tables. It would be not reproducible in the current format.

Response 3: Table 2 has been substituted by figure 1, a box plot that according to your suggestion shows better the significant differences between groups in both variables (FA and DA score).

 

 

Point 4: Line 230, it is mentioned that the human torso is not perfectly symmetrical and the presence of directional asymmetry in the control group was not, therefore, a surprise compared to [20]. This is not easy to follow. Perhaps the authors would like to give some more explanations on the rationale.

 

Response 4: Thank you very much for your comment, we have done a new try to improve our argument that directional asymmetry may be expected in healthy subjects but in a less grade, according to our results. We have added new lines supporting this idea in the discussion and 2 new references that support that argument had been added as well. We hope this time explanations are easier to follow.

Author Response File: Author Response.docx

Reviewer 2 Report

The manuscript Investigating human torso asymmetries: an observational longitudinal study of Fluctuating and Directional Asymmetry in the scoliotic torso is devoted to an interesting and very relevant problem of contemporary humanity. The progress of technology has lead to a general decrease in the physical activity of mankind, which negatively affects many anthropological parameters. The number of children and adolescents with scoliosis increases year by year, so such studies are not only of theoretical, but also of practical importance. I had great hopes when I decided to take a look at the study, and I believe that the authors' endeavor is commendable. I have some questions and comments concerning both the essentials of the study and the presentation of the results.

The volume of the material is very moderate, but considering the difficulties in obtaining the data, it is acceptable. The small sample size can be neutralized by a careful selection of the subjects and detailed data analysis.

1. At the age of 11–13 (especially, 12) the sexual dimorphism in the body size between boys and girls reaches its peak, with girls measurements being higher. Note that BMI difference is significant between boys and girls, and there are also differences between the age groups. The authors studied scoliosis in precisely that age group, so one can expect a high variability of individual parameters. In that case, the data averaged for both sexes seem unreliable. There are ways to counterbalance this drawback. First, by presenting the individual data demonstrating that the individual variability is lower than the group one in both groups.
2. Second, because in the scoliosis group there was only two boys, the control group should have had the same sex ratio. Can the authors argue that the lower BMI in the control group stems from the absence of scoliosis and not from the almost equal sex ratio in that group? During the period of active growth this parameter in boys is lower than in girls.
3. The decrease in numbers in both groups during the observations appears insignificant at first glance — only 5 individuals in the scoliosis group and 2 in the control, but considering the small sample size it is serious: a decrease of 21% and 8%, respectively. To what extent these changes affected the results?
4. All the data on height, body mass, and BMI are averaged, but no significance of differences between the groups is presented anywhere. Judging by the average values and standard deviations, statistically these groups show almost no differences due to high scatter of the values of all the parameters, except, perhaps, the Cobb angle.
5. The title of the Table 1. Anthropometric, radiological and personal data of the AIS and control groups classified according to the time of observation (initial or 6 months later). We can see anthropometrical and radiological data, but not personal.
6. All the abbreviations in the text should be used only after the first use of the full term, even for well-known terms, such as Body mass index. The reader should not be made to guess what the authors meant.

Note that all my comments refer to the secondary aspects of the study. The analysis of the data on manifestations of both directional and fluctuating asymmetry looks fairly convincing, and many previously conducted studies have proven that these manifestations do not depend on the sex of the individual.

On the whole, the work is very interesting and I wish the authors good luck.

Comments for author File: Comments.docx

Author Response

The manuscript Investigating human torso asymmetries: an observational longitudinal study of Fluctuating and Directional Asymmetry in the scoliotic torso is devoted to an interesting and very relevant problem of contemporary humanity. The progress of technology has lead to a general decrease in the physical activity of mankind, which negatively affects many anthropological parameters. The number of children and adolescents with scoliosis increases year by year, so such studies are not only of theoretical, but also of practical importance. I had great hopes when I decided to take a look at the study, and I believe that the authors' endeavor is commendable. I have some questions and comments concerning both the essentials of the study and the presentation of the results.

The volume of the material is very moderate, but considering the difficulties in obtaining the data, it is acceptable. The small sample size can be neutralized by a careful selection of the subjects and detailed data analysis.

Point 1: At the age of 11–13 (especially, 12) the sexual dimorphism in the body size between boys and girls reaches its peak, with girls measurements being higher. Note that BMI difference is significant between boys and girls, and there are also differences between the age groups. The authors studied scoliosis in precisely that age group, so one can expect a high variability of individual parameters. In that case, the data averaged for both sexes seem unreliable. There are ways to counterbalance this drawback. First, by presenting the individual data demonstrating that the individual variability is lower than the group one in both groups.

 

Response 1: Thank you very much for your words about our work. In my role of main investigator and responsible of the sample recruitment I will try to explain further in this reply about the sample characteristics, which show no differences between scoliosis and control group and also between sexes regarding BMI despite you are completely right with your comment. First difficult of the study was to get a sex-balanced control group, because we all know that scoliosis mainly affects girls and this was reflected in our study. Voluntarily recruitment of control sample made it been composed of 54% of boys. I personally agree with you that in the mean age reported, BMI should be expected higher in girls but some boys of the control group aged > 14 (6 subjects), and present “adult bodies” where BMI exceed that reported in girls. Additionally, many girls (3) show overweight and obesity in the scoliosis group, acting as a counter balanced of the older boys of the control group. There were also some missing data (some subjects were not weighted on the TANITA due to mechanical failures) that do not affect results of objective 1, because we have correlated available data with Cobb angle and fluctuating asymmetry score. Anyway, I reviewed carefully data about BMI and did between groups analysis using study group and sex as classifiers.

 

Tests for equal means

 

BMI SCOLIOSIS     BMI CONTROL

N:        37        N:        44

Mean:  19.96   Mean:  19.147

95% conf.:      (19.135 20.784)          95% conf.:      (18.365 19.929)

Variance:        6.1152 Variance:        6.6093

 

Difference between means:   0.81264

95% conf. interval (parametric):        (-0.30916 1.9345)

95% conf. interval (bootstrap):          (-0.24371 1.9326)

 

t :        1.4419 p (same mean):           0.15328           Critical t value (p=0.05):        1.9905

Uneq. var. t :  1.4468 p (same mean):           0.15198

Monte Carlo permutation:      p (same mean):           0.1519

 

Response figure 1: Blue is scoliosis distribution for BMI and red for control

 

Tests for equal means

 

BMI FEMALES        BMI MALES

N:        54        N:        27

Mean:  19.31   Mean:  19.934

95% conf.:      (18.66 19.96)  95% conf.:      (18.809 21.059)

Variance:        5.6673 Variance:        8.0863

 

Difference between means:   0.62411

95% conf. interval (parametric):        (-0.56863 1.8168)

95% conf. interval (bootstrap):          (-0.61452 1.8112)

 

t :        1.0415 p (same mean):           0.30082           Critical t value (p=0.05):        1.9905

Uneq. var. t :  0.98136           p (same mean):           0.33169

Monte Carlo permutation:      p (same mean):           0.3087

 

Response figure 2: Blue is female distribution for BMI and red for male

 

 

Point 2: Second, because in the scoliosis group there was only two boys, the control group should have had the same sex ratio. Can the authors argue that the lower BMI in the control group stems from the absence of scoliosis and not from the almost equal sex ratio in that group? During the period of active growth this parameter in boys is lower than in girls.

 

Response 2: We agree with the idea of the differences in BMI made by different composition by sex in both groups. In fact, in response figure 2 it is clearly evidenced differences in BMI between sexes. Thus, the lower BMI in the control group may be caused by the higher amount of males than the scoliosis group. Anyway, regarding objective 1 of our study, only BMI of scoliosis groups matters because the correlation with Cobb angle could not be done in control group (because they don´t have scoliosis and due to ethical reasons that not allowed us to take a X-ray in healthy subjects). All comparisons of BMI between scoliosis and controls were just informative in table 1, and clearly biased (in agreement with your comments) by the sex distribution and the age of some of the males of the control group (see response 1)

 

Point 3: The decrease in numbers in both groups during the observations appears insignificant at first glance — only 5 individuals in the scoliosis group and 2 in the control, but considering the small sample size it is serious: a decrease of 21% and 8%, respectively. To what extent these changes affected the results?

 

Response 3: It is true. That sample loss in the second observation was important and clearly affect longitudinal observations. In hypothesis 1, the comparison of fluctuating and directional asymmetry between groups was not affected, because we have used scoliosis and control group (including first and second measurement available). For the same reason, comparison regarding sex was not affected as well, and only results about longitudinal observation may be biased. Being our hypothesis related with differences between scoliosis and control this problem was not a major concern, but we thought that the results would have been different if we had had the possibility of a 2^nd measurement of the whole sample. In hypothesis 2, the loss of 5 individuals in scoliosis group reduced statistical power of multivariate regression surely, but fortunately we found correlation between both independent factors combined and Cobb angle. Regarding objective 1, it has been widely explained in response 1 and 2; and, finally, regressions of shape on symmetric and asymmetric components (objective 2) would have seen their statistical power reduced as well but it did not affect the methodology, which was our target. We aimed to develop a predictive methodology of the 3D shape of the torso influenced by changes in fluctuating and directional asymmetries of scoliosis patients and we achieved that target despite the sample loss.

 

Point 4: All the data on height, body mass, and BMI are averaged, but no significance of differences between the groups is presented anywhere. Judging by the average values and standard deviations, statistically these groups show almost no differences due to high scatter of the values of all the parameters, except, perhaps, the Cobb angle.

 

Response 4: You are right and statistics of the table 1 have been included. We have observed differences only in body fat percentage, which was measured with bioimpedance and it is independent from height and weight. A higher body fat composition has been found in scoliosis group, due to its majority of females, which are more susceptible to show higher levels of fat than males in the age gap analyzed 

 

Point 5: The title of the Table 1. Anthropometric, radiological and personal data of the AIS and control groups classified according to the time of observation (initial or 6 months later). We can see anthropometrical and radiological data, but not personal.

 

Response 5: We are strongly grateful with this correction, which was a mistake from us. Effectively there is no intention to reflect personal data in table 1 so the title has now been corrected

 

Point 6: All the abbreviations in the text should be used only after the first use of the full term, even for well-known terms, such as Body mass index. The reader should not be made to guess what the authors meant.

 

Response 6: Thank you very much for this brief comment. We have now reviewed all abbreviations in the text and they are all full-written the first time they appeared in the manuscript

Note that all my comments refer to the secondary aspects of the study. The analysis of the data on manifestations of both directional and fluctuating asymmetry looks fairly convincing, and many previously conducted studies have proven that these manifestations do not depend on the sex of the individual.

On the whole, the work is very interesting and I wish the authors good luck.

Author Response File: Author Response.docx

Round 2

Reviewer 1 Report

The paper has been improved. I recommend publication.