Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

Comments for author File: Comments.pdf

Author Response

Response to Reviewer 1:

 

The authors examined choroidal thickness in age, sex and intraocular pressure-matched groups of healthy volunteers and patients with early-onset Parkinson's disease or multiple sclerosis. They used a relatively difficult image analyzation program. They found the choroid thinner in the neurodegenerative diseases. This fulfils their aim to learn more about changes in the choroid in neurodegenerative diseases. The authors conclude that the significance of this finding is that choroidal thinning may be a biomarker in the early stages of neurodegenerative diseases.

In order to make choroidal thickness a biomarker, we should be able to predict the normal choroidal thickness. However, this is controversial. It would be worthwhile to look not only at the absolute thickness/volume, but also at the ratio of choroidal layers to each other.

First, I wanted to thank you for your effort for having dedicated time to review this manuscript and help us to improve it.

Regarding this comment and considering the feedback provided by the other reviewer, we should recognize that we have not effectively convey our results and the intended contributions to the scientific community. We sincerely apologize for this oversight.

In our study, it's important to clarify that we did not measure choroidal thickness using our image processing code. Instead, our initial processing step focused on delineating the choroidal area, as this data was fundamental for showcasing our intended findings we decided that we should show the results. Subsequently, we conducted an analysis of the image's semantic properties, specifically in terms of optical density. To encapsulate this result, we introduced a parameter called COID. To enhance comprehension, we have included an explanatory paragraph in the discussion section, which is provided below:

“To gain a better understanding of COID, we can draw a comparison with X-ray imaging. Tissues vary in their density, offering varying resistance to the passage of light, analogous to how bones appear opaque in X-ray images. In our analysis, we capture the reconstructed image of the alterations that occur when an infrared light beam through different ocular media. The speed and direction of light propagation change at each interface, leading to detectable modifications in the interference pattern upon its return (as it exits the eye). These changes manifest as distinct properties in each pixel of the image, which might not be discernible to the human eye but are evident through analysis. We extract and categorize this information based on its semantic characteristics and scrutinize subtle variations. Consequently, we observe that this "obstruction" to light propagation differs based on the eye's condition, whether healthy or afflicted by pathology”.

 

Introduction

Information about OCT technology is very basic. Concentrate on the problem of the investigation of choroid, describe the pitfalls of measuring choroidal thickness, how the choroid can be visualized and why swept source OCT is superior to the other OCT technologies. What is the normal value of choroidal thickness (controversial literature!)?

Thank you for your valuable comments. While we omitted details about the device's operating system, our intention was to simplify the readability of the manuscript for a broader audience. As you noted, there is considerable debate surrounding choroidal thickness, particularly in the context of patients with Parkinson's disease. Existing studies predominantly compare healthy eyes with those affected by pathologies, neurodegenerative or not, but a standardized database with an adequate number of cases is lacking. This challenge has led our research group to shift our focus away from the term "thickness" and explore what insights this device can provide into the interactions of infrared light as it propagates across different structures.

 

 

Methods

How has the effect of diurnal rhythm on choroidal thickness been eliminated?

An entirely fitting comment. We specifically chose images acquired in the morning during a two-hour interval, spanning from 10 to 12. Consequently, due to these criteria, along with our other requirements to ensure sample uniformity and image quality, our total number of processed images is relatively low. We added this information in methodology to clarify this step in line 129.

Which algorithm was used to obtain scans?

Thanks for the detail, we had not provided all information in the manuscript. In lines 123 and 124 the examination protocol was previously explained and we have now added the information of the b-scan used for the image processing in line 137.

L181: How was the „interested” area was selected?

Step 3 explains how O2 reviews the areas selected by the algorithm, verifying that they correspond to the area segmented as choroidal tissue, which is the area of interest where the results calculation is subsequently applied (line 174).

 

Discussion

In order for choroidal thickness to become a biomarker, further studies are needed (determination of normal choroidal thickness, determination of normal values of the relative proportions of choroidal layers.

We completely agree with your perspective. While conducting a histological study of this tissue would be highly valuable, it is currently beyond the means of most researchers. As we've aimed to convey, our approach with the current imaging techniques allows for preliminary assessments, which we aspire to validate with more substantial data in the near future. In this study, our focus is not on choroidal thickness, rather, on analyzing the semantic properties of the images within the definable tissue choroidal area, which exhibit connections to tissue characteristics.

 

Reviewer 2 Report

Comments and Suggestions for Authors

The authors evaluated alterations of the choroidal layers of the eye in multiple sclerosis (MS) patients, Parkinson disease (PD) patients, and healthy subjects using optical coherence tomography (OCT) B-scan images. The read out were choroidal optical image density (COID, standard mean gray level), and the total choroidal area (CA, px2). Additionally, the macular thickness was measured.

The study partly re-evaluates known facts (MS and AD patients using OCT software shows thinning versus healthy subjects) with a new supervised superpixel-based algorithm. The authors claimed “This new technique can detect gradual changes in optical density in OCT images".

Although the manuscript is well written and gives some new insights I have some concerns and questions: 

General: in the title, words like “choroidal layers of the eye” should appear.

Questions: What is the biological reason for changes of CA and CD in PD and MS? Did CA and CD correlate?

 

Neurol Neuroimmunol Neuroinflamm. 2023 Oct 9;10(6):e200169.

IEEE J Biomed Health Inform. 2023 Sep 8;PP.

 

Haller and Sattler layers of the choroid should be explained

 

Do you only evaluate the choroidal layer near the macula? What about the rest of the bulbus, forming the largest part of the choroidal layer?

 

How do you explain the irregular border of the choroidal layer to the sclera (not seen in histological sections)?

 

Please give a graphic presentation of significant results of statistical testing into Fig 6. Please also insert what the hight of the columns concretely mean. The figure legend gives information, that is not part of the figure (i.e., MS, PD).

 

Line 241: What is “central choroid thickness”?

 

Line 277: please cite correctly 46-50

 

Lines 306ff: you are wright: It would be beneficial to have histological studies of healthy subjects’ and Neurodegenerative Disease patients’ eyes to test these hypotheses and to better understand the pathophysiological mechanisms affecting tissues, in particular vascular tissues such as the choroid.

Questions: What are the biological reason for changes of CA and CD in PD and MS? Did CA and CD correlate? These points should be more intensively discussed.

Are there overlaps to your papers: Neurol Neuroimmunol Neuroinflamm. 2023 Oct 9;10(6):e200169; IEEE J Biomed Health Inform. 2023 Sep 8;PP.

Please say in a “Conclusion” very clearly, what is really new in your study (from the standpoint of results) and what is the progress of knowledge by using your new supervised superpixel-based algorithm.

Author Response

Response to Reviewer 2:

The authors evaluated alterations of the choroidal layers of the eye in multiple sclerosis (MS) patients, Parkinson disease (PD) patients, and healthy subjects using optical coherence tomography (OCT) B-scan images. The read out were choroidal optical image density (COID, standard mean gray level), and the total choroidal area (CA, px2). Additionally, the macular thickness was measured.

The study partly re-evaluates known facts (MS and AD patients using OCT software shows thinning versus healthy subjects) with a new supervised superpixel-based algorithm. The authors claimed “This new technique can detect gradual changes in optical density in OCT images".

Although the manuscript is well written and gives some new insights I have some concerns and questions: 

First, I wanted to thank you for your effort for having dedicated time to review this manuscript and help us to improve it.

Comment 1)

General: in the title, words like “choroidal layers of the eye” should appear.

We appreciate your suggestion regarding the title. Wi would modify it to include “choroidal layer of the eye” 

Comment 2)

Haller and Sattler layers of the choroid should be explained

 We will also make sure to clarify and explain the Haller and Sattler Layers of the choroid in the manuscript. Within the choroidal tissue, these two layers can be distingue based on the diameter of their blood vessels. The Haller layer, positioned externally, have larger-caliber vessels, while Sattler layer, situated internally, features vessel of smaller caliber. We have added:

Line 77: …, that is the separation between the external area of large-caliber vessels versus the internal area of small-caliber vessels, ….

Comment 3)

Do you only evaluate the choroidal layer near the macula? What about the rest of the bulbus, forming the largest part of the choroidal layer?

It would be highly intriguing to have access to equipment featuring wide-field optical systems that could cover a larger expanse of the retina. However, as of now, such technology has not yet been developed. OCT is a device designed for evaluating specific areas of the retina, limited by its scan length. In this instance, we utilized the examination mode offering the greatest coverage, yet even with this capability, we can only scan an area approximately 30º surrounding the fovea, situated at the center of the macula.

Comment 4)

How do you explain the irregular border of the choroidal layer to the sclera (not seen in histological sections)?

You are correct in this aspect. The information we derive from light interferometry presents tissues with less distinct boundaries, which differ from what's observable in histological sections. Nonetheless, the ability to detect such transitions in vivo offers a significant advantage. Our designed code identifies variations in the data obtained from the signal at the intersection, and due to the increasing blurriness in deeper areas, we encounter these irregular margins. It's plausible that upcoming OCT technologies under development may help alleviate this effect, enabling us to capture images that define these contours more accurately.

 

Comment 5)

Please give a graphic presentation of significant results of statistical testing into Fig 6. Please also insert what the hight of the columns concretely mean. The figure legend gives information, that is not part of the figure (i.e., MS, PD).

Thank you very much for the appreciation of the error in the legend and in the results presentation. We are preparing to correct it and we added table 1. in which shows results

 

Comment 6)

Line 241: What is “central choroid thickness”?

The interferometry technique used by OCT allows the device's software to calculate distances based on the behavior of the laser light according to the optical path traveled in the tissues it passes through. This distance corresponds to the thickness of the tissue, a term that has been deeply studied and for which there are standardized normal values. Therefore, it is used to investigate variations in tissue thickness between individuals with pathologies and their healthy counterparts.

We have included the clarification: Line 241: "average thickness of the choroidal layer in the sub-macular zone."

Comment 7)

Line 277: please cite correctly 46-50

Our apologies for this confusion

 

Comment 8)

Lines 306ff: you are wright: It would be beneficial to have histological studies of healthy subjects’ and Neurodegenerative Disease patients’ eyes to test these hypotheses and to better understand the pathophysiological mechanisms affecting tissues, in particular vascular tissues such as the choroid.

Thank you very much for your comment. It would be a great advance for many fields of research

Comment 9)

Questions: What are the biological reason for changes of CA and CD in PD and MS? Did CA and CD correlate? These points should be more intensively discussed.

>> In the case of the choroidal area, which is the product of the scan width and choroidal thickness distances, significant alterations have been observed in various studies comparing structural changes in the retinas of individuals with neurodegenerative diseases to those of healthy subjects. For this reason, this parameter is observed altered as in other studies.

However, the precise pathophysiology remains elusive. It is postulated that in individuals with multiple sclerosis (MS), the lower nutritional demand due to neurosensory tissue atrophy may lead to a reduction in vascular tissue. Additionally, the inflammatory processes associated with the disease could potentially affect the vascular tissue. Nevertheless, these remain unverified hypotheses to date. As previously mentioned, the opportunity to work with ocular histological tissues could shed light on this pathophysiology. In the case of Parkinson's disease (PD), it is closely linked to dopamine level disruptions, which subsequently impact blood perfusion. Progress in this research field requires collaborative efforts from various groups. Our objective, although somewhat ambitious, is to contribute new insights, with the hope of inspiring colleagues from diverse specialties to recognize the relevance of investigating these structural changes.

>> Choroidal density, a term we have coined, serves as a representation of the outcomes derived from our code-based calculations. It might be more informative to elucidate the concept of COID, or choroidal optical image density, as CD results from the calculation of COID within a unit area of the choroid.

To gain a better understanding of COID, we can draw a comparison with X-ray imaging. Tissues vary in their density, offering varying resistance to the passage of light, analogous to how bones appear opaque in X-ray images. In our analysis, we capture the reconstructed image of the alterations that occur when an infrared light beam through different ocular media. The speed and direction of light propagation change at each interface, leading to detectable modifications in the interference pattern upon its return (as it exits the eye). These changes manifest as distinct properties in each pixel of the image, which might not be discernible to the human eye but are evident through analysis. We extract and categorize this information based on its semantic characteristics and scrutinize subtle variations. Consequently, we observe that this "obstruction" to light propagation differs based on the eye's condition, whether healthy or afflicted by pathology.

We have added this last paragraph to the manuscript for your better understanding. Lines 286 to 297.

Comment 10)

Are there overlaps to your papers: 

Neurol Neuroimmunol Neuroinflamm. 2023 Oct 9;10(6):e200169; 

IEEE J Biomed Health Inform. 2023 Sep 8; PP.

Thank you for your comments, and I'm relieved to confirm that there are none.

One of the authors of this manuscript was involved in the study mentioned in the first reference. The foundation of this work was to establish a committee aimed at standardizing minimum quality protocols for working with images acquired through a specific type of OCT based on angiography and artificial intelligence. AI is an immensely significant and rapidly expanding field, thanks to its myriad beneficial applications. It's crucial to maintain rigor when drawing conclusions from these studies.

Regarding the second document, I have thoroughly reviewed the publications in that journal's issue and found no relationships or overlaps either.

Please say in a “Conclusion” very clearly, what is really new in your study (from the standpoint of results) and what is the progress of knowledge by using your new supervised superpixel-based algorithm.

Thanks for this recommendation, we added this paragraph (line 423 to 431):

>> Our calculation of choroidal optical image density, achieved through semantic processing, it is shown of the COID parameter. This demonstrates variances between the eyes of healthy individuals and those with ocular diseases, suggesting it could serve as a potential biomarker for structural changes in choroidal vascular tissue. The use of non-invasive OCT techniques for in vivo measurements provides an avenue to enhance our understanding of vascular tissue and to identify early abnormalities. To delve deeper, further research involving patients with vascular diseases or ocular vascular pathologies is essential. We anticipate that this work will pave the way for future collaborations and the development of this promising research line.  

 

Round 2

Reviewer 2 Report

Comments and Suggestions for Authors

Dear authors,

you adressed my comments properly. Thank you.