Intraoperative PRO Score Assessment of Actinic Keratosis with FCF Fast Green-Enhanced Ex Vivo Confocal Microscopy

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

In this manuscript, the authors introduced ex vivo confocal laser microscopy to improve the PRO scoring of for dermatologists to diagnose actinic keratoses. Most aspects of the methods and results are well described. The entire manuscript is well written. There are some little issues to be resolved before the publication.

 

1.     In the image analysis step, two examiners were involved to review the images. The authors didn’t clarify whether one of them review only the AO channel and the other review only the FCF channel. Or both reviewed both channels.

If it was the first case, how did the authors prevent the personal biases affecting the difference in the PRO scores developed with two independent channels? If they exchange the channels after the first round of review, another type of biases might be introduced by the difference in the order of channels reviewed.

If it was the second case, these biases were still there. In addition to the problems of biases. The PRO scoring with AO and the PRO scoring with FCF were not independent anyway in this case. It compromised the soundness of the conclusion that FCF improved the PRO scoring.

 

2.     The morphological features are totally based on human descriptions. It is OK to apply human descriptions. However, when it came to two examiners, how did the author deal with the variability of descriptions from two sources?

3.     The collagen fiber quantification was not well designed. 6x6 is a too small patch and there were only 3 locations picked for each image. The authors didn’t explain how they found the “representative” locations. This processing inevitably introduced a lot of biases.

4.     One suggestion: to compare PRO scoring with AO and FCF, the authors could provide confusion tables.

Author Response

We are sincerely grateful for the thoughtful and constructive comments provided by the reviewer. We highly value the time and expertise invested in the detailed review of our manuscript. In the following sections, we have carefully addressed each of your concerns - changes are marked in yellow in the revised manuscript: 

1. In the image analysis step, two examiners were involved to review the images. The authors didn’t clarify whether one of them review only the AO channel and the other review only the FCF channel. Or both reviewed both channels.

We appreciate the reviewer's query about the image analysis step All examiners, in fact, reviewed both AO and FCF images for all cases in a randomized order. Therefore all examiners were blinded to the clinical diagnosis and staining protocol. The blinding process and resolution of disagreements by a third senior dermatologist were integral components of maintaining rigor and impartiality in the evaluation. We added this aspect to the manuscript.

2. If it was the first case, how did the authors prevent the personal biases affecting the difference in the PRO scores developed with two independent channels? If they exchange the channels after the first round of review, another type of biases might be introduced by the difference in the order of channels reviewed

See next question

3. If it was the second case, these biases were still there. In addition to the problems of biases. The PRO scoring with AO and the PRO scoring with FCF were not independent anyway in this case. It compromised the soundness of the conclusion that FCF improved the PRO scoring.

The reviewer rightly points out the potential biases in PRO scoring and their impact on the study's conclusions. We want to express our gratitude for this insightful observation. To mitigate biases, all reviewers were blinded to clinical diagnosis and staining protocols. The randomization of images further aimed to prevent any systematic bias, ensuring a robust and unbiased assessment. Importantly, FCF staining had no discernible impact on PRO score assesment compared to the standard AO staining, as evidenced by nearly similar agreement rates with histology (95,6% vs. 96%). Notwithstanding, FCF allowed precise collagen imaging which may give a better understanding of UV-induced skin alterations.

4. The morphological features are totally based on human descriptions. It is OK to apply human descriptions. However, when it came to two examiners, how did the author deal with the variability of descriptions from two sources?

We appreciate the reviewer's concern about the variability in morphological descriptions, especially when involving two examiners. Since ECVM image interpretation is based on visual pattern recognition, we used a intuitively accessible “human” description. Recognizing the inherent subjectivity, we adopted a predefined system with three patterns: "regular," "blurred lines," and "mashed potato." This intuitive approach aimed to maintain consistency in interpretation between examiners, and we have noted the importance of this consideration in the revised manuscript.

5. The collagen fiber quantification was not well designed. 6x6 is a too small patch and there were only 3 locations picked for each image. The authors didn’t explain how they found the “representative” locations. This processing inevitably introduced a lot of biases.

The reviewer rightly raises concerns about the collagen fiber quantification methodology. We appreciate the diligence in pointing out potential biases in the selected representative locations. A thin collagen fiber or a keratinocyte nucleus as reference point fills out about 6x6 pixels, if we broaden our region of interest the results become more variable since the fiber/nucleus signal is mixed with backround signal. Three representative locations were chosen as there was usually some variance. However, the brightness of the collagen fibers overall was very similar at most sites. As a result, the selected representative sites encompassed the majority of the other sites.
When the nuclei of the cells were selected as reference points, they had also mostly the same brightness throughout the image, so there was not a large selection to be made. In the case of artefacts in the image, there were sometimes large changes, so these areas were not included. We acknowledge the limitations and have incorporated this explanation into the revised manuscript.

6. One suggestion: to compare PRO scoring with AO and FCF, the authors could provide confusion tables.

We appreciate your suggestion. Recognizing the value of this addition, we have incorporated two confusion tables, providing a clearer presentation of the comparison.

Thank you for your time and consideration. We believe that these revisions address the concerns raised by the reviewer and strengthen the overall contribution of our work.

Reviewer 2 Report

Comments and Suggestions for Authors

This is a well-written and well-conducted study titled "Bedside histology for PRO Score assessment of actinic keratosis with FCF Fast Green enhanced ex vivo confocal microscopy." The authors perform ex vivo confocal microscopy on 48 lesional skin samples of AK and 32 healthy controls and compare the histopathologic and dermal appearances in these samples using these techniques. The authors illustrate the study well and provide good examples of images illustrating the findings. The authors also address the limitations of their study.

The questions/comments that I have for the authors are the following:

1. In the participants' population, was there any variation in the skin tones? I know the Fitzpatrick skin type isn't a perfect system, but can you list the Fitzpatrick skin type for all participants so it would be clear if there is any difference in the performance of confocal microscopy in different skin tones, given the novel staining technique?

2. Can you please elaborate on the methods of how the tissue was embedded to obtain the images for confocal microscopy/dye application (i.e., was the sample examined from the side or only from the top, and the images generated?)

3. In the case selection, were there cases suspected to be AK clinically, and it turned out to be a different lesion on confocal and histopathologic examination? If so, can you document how many of those events occurred?

4. Can you list the discordant PRO score in the results of the two cases that EVCM misclassified? I can see that both were in the PRO II category by histology, but what was the designated category by EVCM?

5. Do the authors have a better descriptor than "mashed potato" for the damaged collagen bundles in solar elastosis? I honestly do not see the analogy.

6. Does "bedside" have to be in the title? I wouldn't imagine performing this technique would be of much value on the bedside, especially since it is "ex vivo," where the sample is no longer in the patient to perform all the dyes and imaging.

 

Author Response

We extend our sincere gratitude to the reviewer for their meticulous review and insightful queries, which have significantly contributed to the refinement of our manuscript. We highly appreciate the time and expertise invested in providing constructive feedback. In the following sections, we have carefully addressed each of your concerns - changes are marked in yellow in the revised manuscript: 

1. In the participants' population, was there any variation in the skin tones? I know the Fitzpatrick skin type isn't a perfect system, but can you list the Fitzpatrick skin type for all participants so it would be clear if there is any difference in the performance of confocal microscopy in different skin tones, given the novel staining technique?

We acknowledge the reviewer's consideration regarding skin tones and Fitzpatrick skin types. Patients included in our study had skin types varied from Fitzpatrick type I to III. Since actinic keratosis primarily occur in individuals with a fair skin tone as well as the fact that our patient population includes fewer people with darker skin. Regarding PRO scoring or collagen fiber assessment we saw no significant differences visible between different skin tones comparing FCF to acridine orange staining. We added this aspect to the revised manuscript.

2. Can you please elaborate on the methods of how the tissue was embedded to obtain the images for confocal microscopy/dye application (i.e., was the sample examined from the side or only from the top, and the images generated?)

We express our gratitude for the regardful observation and thank the reviewer for prompting additional clarification.

Prior to imaging, one cohort of the skin samples underwent the new staining protocol containing incubation in ethanol (0.7 mmol/l), acridine orange (AO)(0,04 mmol/l, Sigma-Aldrich, St. Louis/MO, USA), FCF Fast Green (0,067 mmol/l, Sigma-Aldrich, St. Louis/MO, USA) and NaCl (0.09 mmol/l) for 30 seconds each, while the other cohort underwent previous established staining protocol using acridine orange. All samples were fully immersed in the staining solutions, covering the entire cut surface.

Ex vivo confocal laser scanning microscopy (EVCM) imaging was performed using the Vivascope 2500 G-4 device (Vivascope, Munich, Germany) with two laser wavelengths at 488 nm (blue) and 638 nm (red). Therefore, the tissue probes were placed on object slides (R. Langenbrinck, Emmendingen, Germany), mounted with sponges as well as magnets, and sectioned in vertical mode to reveal all skin layers according to standard histological procedures. Afterwards, the tissue was fixed in formalin and underwent conventional histopathological examination.

The inclusion of this information has been added to the manuscript.

3. In the case selection, were there cases suspected to be AK clinically, and it turned out to be a different lesion on confocal and histopathologic examination? If so, can you document how many of those events occurred.

We appreciate the reviewer's interest in the clinical suspicion of AK cases and any discrepancies between clinical, confocal, and histopathological examinations. A total of 97 lesions in total were acquired. 32 lesions were healthy skin control, 4 lesions were excluded due to poor image quality, resulting in 61 samples which were clinically graded as AK. Beneath this, 13 cases eventually proved to be Bowens disease, invasive squamous cell carcinoma or basal cell carcinoma in histology and EVCM. To maintain the focus on actinic keratoses, these lesions were excluded from our study, ensuring that only cases with histologically confirmed AK (=48 lesions) were included.

4. Can you list the discordant PRO score in the results of the two cases that EVCM misclassified? I can see that both were in the PRO II category by histology, but what was the designated category by EVCM?

Both lesions, clinically graded as PRO II, were overclassified as PRO III in EVCM examination. The histological classification for both cases remained PRO II.

5. Do the authors have a better descriptor than "mashed potato" for the damaged collagen bundles in solar elastosis? I honestly do not see the analogy.

We express our gratitude for the suggestion to reconsider the descriptor "mashed potato" for damaged collagen bundles in solar elastosis. In response to the reviewer's feedback, we have revised the wording to a more histologically precise term, now described as "degeneration."

6. Does "bedside" have to be in the title? I wouldn't imagine performing this technique would be of much value on the bedside, especially since it is "ex vivo," where the sample is no longer in the patient to perform all the dyes and imaging.

We appreciate the reviewer's consideration of the term "bedside" in the title. In clinical suspicious lesions EVCM can differ between invasive squamous cell carcinoma and actinic keratoses within minutes, which is crucial for the right selection of the treatment (see also Vladimirova G et al, J Biophotonics, 2022; doi: 10.1002/jbio.202100372). For this purpose the device can be placed in the operation ward directly near to the patients bed, and the diagnostic assessment time from excision to image is <5 min. We acknowledge the potential confusion and have modified the term to "intraoperative" to better reflect the nature of the technique, which allows for rapid diagnostic assessment directly near the patient.

Thank you for your time and consideration. We believe that these revisions address the concerns raised and strengthen the overall contribution of our work.

Reviewer 3 Report

Comments and Suggestions for Authors

Well written manuscript on an actual, interesting and clinically significant applied sciences topic.

Introduction is sound, M&M clear, Results comprehensive with beautiful and instructive image material, Concluisons are clear and plausible - albeit somewhat redundant in the last section.

2nd half of discussion could be substantially condensed and shortened (since the relevant point of standardization of protocols is reiterated from the same angles repeatedly).

Comments on the Quality of English Language

line 287: fields insted of field

line 299: during instead of while

line 317: compensate for instead of substract out

 

Author Response

We sincerely appreciate the time and expertise you invested in reviewing our manuscript, and we are grateful for the insightful comments provided. In the following sections, we have carefully addressed your concerns - changes are marked in yellow in the revised manuscript: 

1. 2nd half of discussion could be substantially condensed and shortened (since the relevant point of standardization of protocols is reiterated from the same angles repeatedly).

In our revised submission, we have appropriately compressed the discussion part.

We believe that these revisions substantially improve the manuscript and address the concerns you raised.

Reviewer 4 Report

Comments and Suggestions for Authors

The authors report about the opportunity to reveal "bedside histology" in patients with actinic keratosis using ex vivo confocal microscopy.

The method is already described elsewhere in detail. 

In the introductional part the authors discuss the grading systems for actinic keratosis (ak), namely KIN and PRO score. It should be recognized that the methods do not reveal a good reproducibility, are characterized by a high interobserver variability and, finally, that the clinical relevance is still questionable.

Did the authors find differences in PROscore depending on the localization of ak?

However, in general, it has to be questioned what the advantages of this technique are. It is an ex-vivo-method which needs an invasive procedure. To my opinion, especially in the discussion section, the impact of the method in diagnostic procedure is largely overestimated by the authors (263-274). There is, in general, no need to obtain any informations about the stages of ak within minutes. To establish a diagnosis cryosections stained with H&E are available in minutes as well.

The discussion is cumbersome and contains general considerations which are redundant (334-350)

Author Response

We sincerely appreciate the time and expertise the reviewer invested in reviewing our manuscript, and we are grateful for the insightful comments provided. In the following sections, we have carefully addressed each of your concerns - changes are marked in yellow in the revised manuscript: 

1. The authors report about the opportunity to reveal "bedside histology" in patients with actinic keratosis using ex vivo confocal microscopy. The method is already described elsewhere in detail. 

We thank the reviewer for noting the detailed assessment of actinic keratoses (AKs) with PRO scoring and the use of FCF Fast Green dye in ex vivo confocal microscopy (EVCM). We appreciate the acknowledgment of previous publications on the use of EVCM in AKs (see also Vladimirova G et al, J Biophotonics, 2022; doi: 10.1002/jbio.202100372). However, we emphasize that our study contributes a novel approach, since a detailed assessment of AKs with PRO scoring and the combined use of the new FCF Fast Green dye in EVCM has not yet been performed.

2. In the introductional part the authors discuss the grading systems for actinic keratosis (ak), namely KIN and PRO score. It should be recognized that the methods do not reveal a good reproducibility, are characterized by a high interobserver variability and, finally, that the clinical relevance is still questionable.

We express our gratitude for the reviewer's attention to the controversial issue on grading systems for AK. The concerns raised about reproducibility and interobserver variability are duly acknowledged. We believe that our choice of the PRO scoring system is well-founded based on the available evidence. In a study conducted by Schmitz L et al (Eur Acad Dermatol Venereol; doi: 10.1111/jdv.14512) which involved the assessment by 21 dermatologists and pathologists, the PRO scoring system demonstrated robust interrater reliability. Notwithstanding we have incorporated this important information into the revised manuscript to provide more clarity.

3. Did the authors find differences in PRO score depending on the localization of ak?

We appreciate the insightful comment on the localization of AK lesions and its potential impact on PRO scoring. Localization of AK lesions were on the capillitium, face and forearms. We have revisited the study data again and found that there were no significant differences in PRO grading based on the localization of AK lesions. This has been incorporated into the results section for clarity.

4. However, in general, it has to be questioned what the advantages of this technique are. It is an ex-vivo-method which needs an invasive procedure. To my opinion, especially in the discussion section, the impact of the method in diagnostic procedure is largely overestimated by the authors (263-274). There is, in general, no need to obtain any informations about the stages of ak within minutes. To establish a diagnosis cryosections stained with H&E are available in minutes as well.

The reviewer rightly points out the need for a balanced discussion on the advantages of the technique. We acknowledge the concerns about the impact of the method in the diagnostic procedure and have revised the discussion section to better reflect the considerations. We appreciate the provided context regarding the urgency in recognizing AKs. Regarding BCCs and SCCs EVCM proved to be a valid adjunction to fresh frozen sectioning since it is fast and easy to handle even in rural areas, where not every clinic or practitioner has a laboratory or pathologist for cryosections nearby. Since NMSC is frequently surrounded by UV induced skin alterations like AKs it is therefore important to also evaluate the diagnostic workflow of these precancerous lesions for a simple and fast risk stratification.

5. The discussion is cumbersome and contains general considerations which are redundant (334-350)

We express our gratitude for the constructive feedback on the discussion section's length and redundancy. In response, we have diligently compressed the discussion to eliminate redundancies and improve the overall flow, addressing the reviewer's concerns.

We thank the reviewer for the diligent evaluation and constructive suggestions. We believe that these revisions substantially improve the manuscript and address the concerns raised.

Round 2

Reviewer 4 Report

Comments and Suggestions for Authors

The revised manuscript is much improved but there still remains the question of any benefit for patients. The authors should mainly focuss on the technical and scientific knowledge which is gained by their study.

Author Response

Dear Reviewer,

we are grateful, that we could improve our manuscript with your constructive suggestions. Regarding the question of the benefit for patients, we genuinely value this perspective. In response, we adjusted the passages in the discussion and conclusion more focused on the technical and scientific aspects, providing a clearer emphasis on the contributions to the field.

Thank you for your thorough reviews throughout the publication process.

Best Regards

Round 3

Reviewer 4 Report

Comments and Suggestions for Authors

Discussion: It is completely unrealistic that this expensive technique will be used in "rural areas" by dermatologist/practioners for investigating actinic keratoses instead of histopathological techniques, so far.

Just state:  "it may be a valid adjunction..."

Author Response

Dear Reviewer,

we understand the concern raised about the practicality of employing EVCM technique in rural areas. While we acknowledge the potential challenges associated with the cost of the technique, we would like to emphasize that our intention was to highlight its diagnostic capabilities, especially in settings where immediate access to histopathological techniques may be limited.

We are grateful for this constructive feedback, and in our revision, we changed the wording according to your recommendations. We therefore hope to enhance the clarity and applicability of our manuscript.