Retrospective Analysis of IOL Power Calculation by Ray Tracing in Eyes with Previous Radial Keratotomy

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

I have reviewed the manuscript and consider it a solid and clearly presented study. The methodological approach is appropriate for the research question, and the results are consistent with the current understanding of IOL power calculation in post-RK eyes. The manuscript is coherent, and the data are presented in a clear and accessible way.

I only have a few minor suggestions that may improve clarity:

• Since PK did not add benefit in your series, whereas it has shown value in post-LASIK eyes, a brief remark in the “Discussion" explaining this difference may help readers interpret the finding.

 

• In the “Methods", a short clarification on the rationale for applying AL adjustment would be useful for readers who are less familiar with this step.

 

• Likewise, expanding slightly on why eyes with a pinhole IOL were excluded (i.e., how this lens design may mask refractive error) could improve transparency.

These are small editorial points and do not affect the validity of the study. I recommend Minor Revision. 

Author Response

Please see the attachment

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

It is an interesting article. With high clinical relevance. 

Well-defined objective and clinically relevant population (post-RK eyes.

Use of a modern ray-tracing platform with appropriate comparator formulas.

Complete reporting of refractive accuracy metrics (PE, MAE, MedAE, threshold percentages). Balanced conclusion that does not overclaim superiority

Key points for improvement

1. Objective wording (important)

“Accuracy of IOL power calculation” is commonly flagged by reviewers.

Recommendation: use “predictive accuracy” or “refractive prediction accuracy”.

Why: The IOL power itself is not “accurate” or “inaccurate”; the refractive outcome is.

2. Methods clarity and concision

“A consecutive series of patients” → outcomes are analyzed per eye.

“The ray tracing software of the MS-39” sounds slightly awkward

Suggestions

Replace “patients” with “eyes”.

Use “embedded ray-tracing software” or “built-in ray-tracing software”.

 

3. Formula classification

 

The conclusion refers to “paraxial formulas,” which may be questioned.

Barrett True-K, EVO, and PEARL-DGS are often described as theoretical or hybrid formulas, not strictly paraxial.

Recommendation: replace “paraxial formulas” with “modern theoretical formulas.”

 

4. Statistical interpretation

You correctly state that differences were not statistically significant.

However, “compared to 65% with ray tracing” slightly overstates the finding.

Recommendation

Frame this as a numerical trend, not better performance.

5. Consistency and precision

Percentages: 62.50% vs 65% → choose one and be consistent.

Minor stylistic improvement: avoid repetition of “with a PE within”.

Conclusions need to be reformulate

Author Response

Please see the attachment

Author Response File: Author Response.pdf

Reviewer 3 Report

Comments and Suggestions for Authors

IOL calculation after refractive corneal procedures is always challenging. The topic of this study is very interesting as it presents some real life results about an IOL calculation formula used in a particular device for eyes that had previously undergone radial keratotomy surgery. However, the study has some fundamental flaws in Materials and Methods (including statistical analysis) and results presentation.

Please find my comments below:

1. Introduction: Since the authors also discuss about LASIK and PRK and compare those refractive surgeries to RK in terms of IOL calculation accuracy, they should include a sentence or sentences to say which formulas are being used for eyes that have undergone LASIK and PRK.

2.Line 64: How does the MS-39 consider both surfaces? The authors should explain because MS-39 is not using a Scheimpflug technology and a reader that is not familiar with the device could be confused. Points like this are essential for the understanding of the background for this study.

3. Line 85: "AS-OCT combined with the MS-39 (software version 4.1.4)" MS-39 device includes OCT. In this sentence it seems like the authors have used a separate AS-OCT - not clear; authors need to clarify this information.

4. Section 2.2. It is needed to be clarified how were the AL and the pupil diameter were calculated? Which device was used? Even it MS-39 was used for those measurements, the authors need to specify this, as readers might not be familiar with the exact capabilities/functions of this device.

5. Line 136: "calculated a minimum sample size for paired comparisons of 18 eyes." The meaning is not clear here.

6. It is fundamental to mention the statistical analysis tests that were used i.e. ANOVA,  t-test etc.

7. Major point: Although the authors mention in Statistical analysis section that they used p-value<0.05 to define statistical significance, they do not mention in the Results section any p-value of comparison. Instead, statistical significance is mentioned for first time in Discussion without justification.

8. Table 1: Similarly to one of my previous comments, the authors mention that MS-39 evaluates bot corneal surfaces; why the authors decided to include IOL Master measurements in the table then?

9. Lines 209-215. Limitations: 

An important limitation that needs to be mentioned is the fact eight different IOL models were used.

Another limitation is that only 3 pupil diameters are available (2.0, 2.5 and 3.0 mm) for selection.

Author Response

Please see the attachment

Author Response File: Author Response.pdf

Round 2

Reviewer 2 Report

Comments and Suggestions for Authors

It's an interesting article. I have some suggestions for improvement.

Major Points for Improvement

1. Language and Grammar

Several typographical and grammatical issues reduce clarity:

“patients eyes” → patients’ eyes

“Pre-operatively” → Preoperatively

angeding” → ranging

 

Sentence fragmentation and line-break artifacts

These should be corrected for journal submission.

2. Study Design Clarity

The study is retrospective, but this is not emphasized in the title or conclusion.

The inclusion criteria (number of RK incisions, time since RK, IOL type) are not mentioned.

Suggestion:

Briefly clarify key inclusion characteristics if space allows.

 

3. Statistical Interpretation

 

The comparison with other formulas is described as “not statistically significant,” but:

No statistical test is mentioned

 

Sample size for comparison dropped to 20 eyes

 

Suggestion:

Mention the test used and acknowledge the limited power.

 

4. Interpretation of Accuracy

The ray tracing method achieved 65% within ±0.50 D, which is reasonable in RK eyes but still below modern benchmarks in virgin eyes.

The conclusion is accurate but could be slightly more conservative.

5. Conclusion Balance

The final statement appropriately acknowledges similar performance of paraxial formulas—this is a strength.

However, calling ray tracing “relatively accurate” should be contextualized within the difficulty of RK eyes.

Author Response

Reviewer 2

It's an interesting article. I have some suggestions for improvement.

Major Points for Improvement

1. Language and Grammar

 

Several typographical and grammatical issues reduce clarity:

“patients eyes” → patients’ eyes

“Pre-operatively” → Preoperatively

angeding” → ranging

Reply: thanks for the suggestions. The 3 above-mentioned errors are due to the conversion from Word to pdf of the manuscript with tracked changes and are no longer present. However, we checked grammar once more and made the following corrections:

• Line 89: we deleted “the anterior”, since it was a repetition.
• Line 114: we corrected “an” into “and”
• Table 1, Posterior Keratometry Steep: we added the signus “-“ in front of the maximum value (3.82)
• Line 262: we added “by” between “approved” and “Comitato Etico”
• Line 275: “has received” has been changed into “have received”
• Line 286: Ophthalmoloogy has been changed into Ophthalmology

Sentence fragmentation and line-break artifacts.

Reply: again, we cannot see any line-break artifact and believe these may be due to the pdf conversion.

These should be corrected for journal submission.

 

1. Study Design Clarity

 

The study is retrospective, but this is not emphasized in the title or conclusion.

Reply: we changed the title by including the term “Retrospective analysis”.

 

The inclusion criteria (number of RK incisions, time since RK, IOL type) are not mentioned.

Suggestion:

Briefly clarify key inclusion characteristics if space allows.

Reply: although the word count is already above the limit, we added the following sentence: “Eyes were not excluded based on the number of radial incisions or the time since RK.” Since we already stated that the inclusion of multiple IOL models was a limitation, we did not add any formula.

 

1. Statistical Interpretation

 

The comparison with other formulas is described as “not statistically significant,” but:

No statistical test is mentioned

Sample size for comparison dropped to 20 eyes

Suggestion:

Mention the test used and acknowledge the limited power.

Reply:

• Thanks for the suggestion. As requested also by Reviewer #3, we added (lines 177-180) the following sentence to specify the tests that were used, i.e. the 2-sample robust t-test and Cochran Q test: “More specifically, according to the two-sample robust t-test, the adjusted p-value ranged between 0.58 and 1 for trueness, between 0.46 and 1 for precision and between 0.12 and 1 for accuracy. According to Cochran Q test, the adjusted p-values ranged between 0.27 (for the PE threshold of ±0.75 D) and 0.93 (for the threshold of ±1.00 D).” Since Cochran Q test had not been mentioned in the previous submission, we included a sentence about it in the Methods (lines 143-144).
• We agree that the sample size of this study is limited and, in fact, we already pointed this issue out as the first of the limitations in the Discussion section.

 

1. Interpretation of Accuracy

 

The ray tracing method achieved 65% within ±0.50 D, which is reasonable in RK eyes but still below modern benchmarks in virgin eyes.

The conclusion is accurate but could be slightly more conservative.

Reply: thanks for the suggestion. In the Conclusions of the Abstract (lines 34-35) we added the following sentence: “The refractive outcomes of IOL power calculation in post-RK eyes are still below modern benchmarks for virgin eyes.”

 

1. Conclusion Balance

The final statement appropriately acknowledges similar performance of paraxial formulas—this is a strength.

However, calling ray tracing “relatively accurate” should be contextualized within the difficulty of RK eyes.

Reply: in addition to the sentence added in the Conclusions of the Abstract, we modified the sentence in the Conclusions of the manuscript as follows: “the refractive accuracy is still lower than in eyes with previous laser vision correction as well as in unoperated eyes”.

 

Reviewer 3 Report

Comments and Suggestions for Authors

The authors have responded to each comment and revised the manuscript accordingly.

Response to Comment 4: The authors state that “as regards AL, the legend of Table 1 already reports that biometric measurements were obtained with the IOLMaster 700.” This information is essential and should be included in the main text, not only in the table legend. Table legends are meant to clarify the content of the table and should not be used to introduce new methodological details that are not described elsewhere in the manuscript.

Responses to Comments 5–7: Based on the authors’ replies, there appears to be a lack of understanding of the fundamental statistical analyses required for a study of this nature. In my initial review, I explicitly noted that the statistical analysis required improvement. However, the authors have not adequately addressed which specific statistical tests were used. While they refer to the Eyetemis online tool, the manuscript also mentions the use of MedCalc software for statistical analysis, yet no details are provided regarding the statistical tests applied for comparisons (e.g., ANOVA, independent or paired t-tests) and correlations analysis. Instead, the authors state that the term “paired comparisons” was used in error, without clarifying the actual analyses performed.

Given this ongoing lack of clarity and apparent misunderstanding regarding the statistical methodology, despite a full round of revision, there are serious concerns about the reliability of the reported results and the appropriateness of the statistical methods employed.

 

Author Response

Reviewer 3

The authors have responded to each comment and revised the manuscript accordingly.

Response to Comment 4: The authors state that “as regards AL, the legend of Table 1 already reports that biometric measurements were obtained with the IOLMaster 700.” This information is essential and should be included in the main text, not only in the table legend. Table legends are meant to clarify the content of the table and should not be used to introduce new methodological details that are not described elsewhere in the manuscript.

Reply: thanks for the suggestion. We added 2 statements to clarify that the AL measurement obtained with the IOLMaster 700 was used with ray tracing. First, we wrote (line 109) “The AL measured by the IOLMaster 700 was used.” Second, we wrote (line 110) “the AL value obtained by optical biometry was adjusted according to the polynomial equation”.

Moreover (line 118), we specified that all paraxial formulas were “based on the IOLMaster 700 measurements”.

Responses to Comments 5–7: Based on the authors’ replies, there appears to be a lack of understanding of the fundamental statistical analyses required for a study of this nature. In my initial review, I explicitly noted that the statistical analysis required improvement. However, the authors have not adequately addressed which specific statistical tests were used. While they refer to the Eyetemis online tool, the manuscript also mentions the use of MedCalc software for statistical analysis, yet no details are provided regarding the statistical tests applied for comparisons (e.g., ANOVA, independent or paired t-tests) and correlations analysis. Instead, the authors state that the term “paired comparisons” was used in error, without clarifying the actual analyses performed.

Given this ongoing lack of clarity and apparent misunderstanding regarding the statistical methodology, despite a full round of revision, there are serious concerns about the reliability of the reported results and the appropriateness of the statistical methods employed.

Reply: we are sorry for the lack of clarity and did our best to improve the explanation of the statistical analysis.

As regards MedCalc, this was used to perform the correlation analysis between the prediction error of ray tracing and the preoperative variables, as previously stated (lines 136-137). To help readers to better understand the results, we added the values of the correlation coefficient r (and the corresponding p-value) for each preopeative parameter in the Results (lines 171-172). With regard to pupil diameter, we reported only the results of correlation for the photopic pupil, since refraction was measured under photopic conditions.

As regards Eyetemis, we added (lines 177-180) the following sentence to specify the tests that were used, i.e. the 2-sample robust t-test and Cochran Q test: “More specifically, according to the two-sample robust t-test, the adjusted p-value ranged between 0.58 and 1 for trueness, between 0.46 and 1 for precision and between 0.12 and 1 for accuracy. According to Cochran Q test, the adjusted p-values ranged between 0.27 (for the PE threshold of ±0.75 D) and 0.93 (for the threshold of ±1.00 D).” Since Cochran Q test had not been mentioned in the previous submission, we included a sentence about it in the Methods (lines 143-144).

We hope these corrections are sufficient, but may add the graphics generated by Eyetemis, although they are not very interesting due to the lack of statistically significant differences.