Assessment of Ultra-Early-Stage Liver Fibrosis in Human Non-Alcoholic Fatty Liver Disease by Second-Harmonic Generation Microscopy

Round 1

Reviewer 1 Report

The manuscript "Assessment of ultra-early-stage liver fibrosis in human non-alcoholic fatty liver disease by second-harmonic generation microscopy" by T. Minamikawa et.al., deals to the investigation of liver pathology by using second harmonic generation imaging and quantification. The manuscript might be of some interest to the scientific community after some aspects which I express in the following are clarified and corrected. 

1. My main concern regarding this manuscript is related  to the limited number of patients enrolled and the overstatement of some conclusions based on such a limited number of patients. The authors also recognize this limitation of their manuscript but I consider that some conclusions should weighted based on this limited number of patients. For example I find the -sentence in lines 147-149 to be an overstatement; the results of a study on only 4 patients can hardly be applied and even compared to an existing known procedure such as Sirius Red assessment of fibrosis even if the latter has some known limitations. In the same time the explanation in lines 261-263 is to general as drown from only 1 field of view per case. The statement in lines 346-348 is also exaggerated in my opinion. If a pathologist can be currently excluded from assessing liver fibrosis, this can be imagined to be done only based on the factors which can be currently computed and are also mentioned in the manuscript. This is but not the case. Moreover, how can the authors demonstrate that their results is not due to interpatient variability. Just to give an example and to speculate, in fig 2e for thick fibers the results are correlated with patient age. This might not be the case, but I think this is the trap one might fall when using limited sized samples.
2. I think in line 113 NASH should be instead of NAFLD.
3. The fraction analysis is not explained in the manuscript. Please consult and if appropriate also cite "Hristu, R., Eftimie, L. G., Stanciu, S. G., Tranca, D. E., Paun, B., Sajin, M., & Stanciu, G. A. (2018). Quantitative second harmonic generation microscopy for the structural characterization of capsular collagen in thyroid neoplasms. Biomedical Optics Express, 9(8), 3923-3936. " where the authors use similar metrics such as total collagen ratio (TC-ratio) and Average value of SHG in significant areas (S-mean). The authors should comment on the similarity of their metrics to there previously used. 
4. The explanation of the results in lines 208-210 is confusing since the authors claim that "Patient 1 with an ultra-early stage of NAFLD is significantly lower than that of the other patients (p < 0.01)." In the same time, patient 2 diagnosed also with ultra-early stage of NAFLD, had no differences compared to the patients with advanced stages. Please clarify.
5. Line 251 states 5 um thick sections while in the Methods section 2 um thickness is mentioned.
6. In line 334 the authors describe fluorescence microscopy as a method for fibrosis assessment without any references. If they are referring to two-photon fluorescence they should also include in the reference list the following paper: "Hsiao, C. Y., Teng, X., Su, T. H., Lee, P. H., Kao, J. H., & Huang, K. W. (2021). Improved second harmonic generation and two-photon excitation fluorescence microscopy-based quantitative assessments of liver fibrosis through auto-correction and optimal sampling. Quantitative Imaging in Medicine and Surgery, 11(1), 351." Otherwise, is classical fluorescence microscopy is targeted, please include a proper reference.
7. In line 350 the authors claim that SHG microscopy is less invasive to the human body. To my knowledge, currently the only clinically approved SHG instrument is JenLab's MPTflex or MPTcompact and only for skin use, hence outside the body. SHG microscopy or endoscopy is only envisioned to be used in the future for such clinical applications. I think SHG use is not that straightforward and the authors should balance their statements. Moreover for in vivo use, SHG microscopy/endoscopy will first have to deal with the Glisson's capsule. If the authors are targeting such in vivo studies an interesting paper with results in this area should be mentioned: Xu, S., Kang, C.H., Gou, X., Peng, Q., Yan, J., Zhuo, S., Cheng, C.L., He, Y., Kang, Y., Xia, W. and So, P.T., 2016. Quantification of liver fibrosis via second harmonic imaging of the Glisson's capsule from liver surface. Journal of biophotonics, 9(4), pp.351-363.
8. In the sample preparation the authors mention that the tissue was cleared. Why was this step necessary since clearing is useful in thick tissue samples and here they used tissue sections?
9. Three sections were prepared: one H&E, one Sirius Red and one unstained. I assume that the unstained one was used for SHG. This is but not clear from the manuscript. Moreover it was currently demonstrated that on tissue slides, SHG quantitative microscopy might provide similar results either on unstained or H&E-stained tissue sections: "Hristu, R., Stanciu, S. G., Dumitru, A., Paun, B., Floroiu, I., Costache, M., & Stanciu, G. A. (2021). Influence of hematoxylin and eosin staining on the quantitative analysis of second harmonic generation imaging of fixed tissue sections. Biomedical Optics Express, 12(9), 5829-5843." The authors should comment on the use of an extra step as compared to standard histology protocol. Why not use directly the H&E-stained section?
10. In the introduction the authors mention tricrom Mason staining for standard fibrosis assessment, but use Sirius Red staining. Please explain why.
11. For FSHG imaging the authors use a condenser with an NA lower than the one of the illumination objective. This is unusual, please see for example Chen, X., Nadiarynkh, O., Plotnikov, S., & Campagnola, P. J. (2012). Second harmonic generation microscopy for quantitative analysis of collagen fibrillar structure. Nature protocols, 7(4), 654-669. In the current manuscript situation a lower intensity SHG might have been collected. What is the rationale behind using such an objective-condenser pair?
12. Based on the pixel dwell time, image resolution and the number of images recorded for a mosaic, I estimate a time of more that 15 min to acquire images required for a mosaic. Is the system stable enough not to get any thermal or mechanical drift during such a prolonged period of time? Is the mosaic not including stitching artefacts?
13. More on the above and on studying images included in the manuscript I am under the impression that only an image mosaic per sample was used. How many biopsies were used per patient? How many images per biopsy? Are the results relevant if such a limited dataset is recorded and analyzed?
14. The protocol for statistical analysis is not cleat. In line 440 the authors claim that "The averaged data over a 500 μm×500 μm area (N=6) was used". Were there 6 such areas used? The SD was computed per area or over the 6 instances. The authors should clearly explain how the quantitative results were computed. Moreover for F/B SHG ratio "50 data were used". What does data mean: pixels, ROIs? What is the ROI size?
15. The authors should also pay more attention to figures and figure captions. Line 138, final (c) should be (d). Line 159, Figures 3(e)-(h) should be Figures 2(e)-(h). Line 413 Figure 1 should be Fig. 5
16. Some English corrections are mandatory. To give just some examples: line 72 "is collagen molecules", line 133 "were seemed". Please also revise sentences in lines 197-200 which is hard to follow. The authors should also rephrase: "propagation behaviour of elastic waves propagating through the liver"

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 2 Report

This is an interesting and very nice paper describing an effective implementation of the second harmonic generation (SHG) microscopy to reveal the properties of fibrous tissues composed mainly by collagen. The main point of novelty is the evaluation of ultra-early stage of liver fibrosis in non-alcoholic fatty liver disease (NAFLD) patients from a histology and fibrosis maturation point of view.

--------

In the Introduction section: “Tissue sections stained with dyes, such as haematoxylin and eosin (HE) for histological analysis, …” (line 62-63), I may erase the abbreviation (HE) since it does not appear again on the main manuscript body.

Then, the manuscript mentions many times the resolution achieved by SHG microscopy (EXAMPLES) but without specifying this data. I may include in the Introduction section a statement like “SHG microscopy is a powerful tool which allows acquire images with lateral resolution upon 200-500 nm relying on the excitation wavelength and after scanning structured illumination process [Yeh, C-H. et al. Biomed. Opt. Express 2018 9, (12), 6081-6090. Doi: 10.1364/BOE.9.006081]”.

--------

The manuscript is generally well-written. Here, some recommendations are provided in order to improve the English quality:

“Interestingly, the stronger SHG intensity image visualised not only the histologically well-matured and aggregated fibrous regions but also some of the fine fibrous regions of which the most regions were visualised as the weak SHG intensity region.” (lines 197-200). Please, replace the second verb “visualised” by a synonym like “observed/noticed”.

“This method is well developed and is a gold standard for the definitive diagnosis of liver fibrosis. However, this method is ….” (lines 329-330). Please, change “However, this method is” by “However, this approach is”.

“Fluorescence microscopy evaluates fibrosis by detecting the fluorescence emitting from the fluorescent dye selectively stained at the fibrous tissues” (lines 334-336) and “Since the dark-field nature of fluorescence microscopy, a bright fluorescent signal is detected only form the fluorescent dye, enabling the sensitive detection of fibrous tissues even with fine structures” (lines 336-337). I may modify the second sentence (lines 336-337) to prevent redundancy.

--------

Figure 2 insets (A and B for (a), (b), (c) and (d), respectively) lack the scale bar. Please, add this information on the respective images or in the figure footnote legend. Same comment for Figure 3 and Figure 4.

In Figure 5 caption, I may suggest to introduce the terms and respective acronyms by alphabetic order (e.g. BPF, D.F., G.M., HWP, L, N.D., P, PMT and QWT).

--------

Other remarkable aspect is that the key outcomes of the submitted work can be extended for the general understanding of some general aspects of NAFLD which might be of clinical interest for early diagnosis and disease prevention. Finally, I would also suggest to the authors to include a short statement as Conclusions section that eventually aid to the reader to better understand the relevance of the submitted manuscript. Please, note that this new Conclusions section must be listed in the manuscript text as “4. Conclusions” and then, move the Material and methods section afterwards like “5. Materials and methods”; “5.1. Patients”, “5.2. ,,,,”.

--------

References format must be consistent. It must be changed the number of pages style of those bibliography that does not match with the fully extension. For example, reference numbers 6 and 7 are good (pages 793-801 and 202-209, respectively). On the other hand, reference numbers 1 and 2 are not in agreement with the aforementioned (pages 1221-31 and 483-95, respectively). In these examples, I may change the above describe page numbers by 1221-1231 and 486-495 for reference 1 and ref. 2, respectively. Please, note that reference numbers 1, 2, 10, 11, 15, 16, 23, 24, 31, 36, 38, 39, 41, 43, 50, 54, 55, 57, 58, 60 and 61 must be checked.

--------

Interestingly, one of my recommendations (for future work) is in agreement with the authors (in accordance to lines 362-364). It would be interesting to enrol more participants in order to enhance the statistical robustness.

To sum up, I see the enclosed manuscript appropriate for publication in the International Journal of Molecular Sciences (IJMS) after carrying out the proposed minor remarks.

Comments for author File: Comments.pdf

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

The authors have now improved the manuscript and have aligned the conclusions to the reduced patient number and obtained results. The manuscript is now technically sound and demonstrates a possible use of SHG microscopy in liver pathology. Further extended studies are requires before being certain about future possible use.

If the editor agrees studies with such limited patient numbers to be published in IJMS, I find the manuscript ready to be published in the current form.