Clinical Quality Control of MRI Total Kidney Volume Measurements in Autosomal Dominant Polycystic Kidney Disease

Round 1

Reviewer 1 Report

General comments:

This study evaluates the data consistency and reproducibility in TKV measurements among various MRI sequences for AKPKD patients. The usefulness and pitfalls of MRI-TKV measurements are discussed. This text is well written and understandable.

Specific comments:

Material and methods:

MRI acquisition

1. “All MRI examinations were … [Table S1].” What does the [Table S1] show?

2. Please explain the standardized protocols in detail (e.g. slice thickness).

Results:

Study participants

3. The median TKV values in the text and the table 1 are different.

Inter-observer variations

4. (Table 2) What does “Model” mean? Please describe in the text.

Acquisition errors

5. About “Breathing during 2D scan” and “Different breath-hold positions in multi-acquisitions”, the text and tables are reversed, making it difficult to read.

Discussion:

6. Please provide the reference numbers.

7. Please provide a brief conclusion in the final paragraph.

Author Response

We appreciate the positive feedback regarding the clarity and comprehensibility of the manuscript.

Specific comments:

Material and Methods:

1. “"All MRI examinations were … [Table S1]. " What does the [Table S1] show? ” We have corrected the typo, which should be [Supplementary Table S2] which summarize the MR protocols. Supplementary Table S1 continues to be mentioned in the introduction and summarizes prior publications on repeatability of TKV measurements.
2. “Please explain the standardized protocols in detail (e.g. slice thickness).” The standardized protocol details, including slice thickness, gap/spacing, matrix, TR, TE, bandwidth, and flip angle, are now described in the supplementary materials as Supplementary Table S2: Scanners utilized for training and reproducibility data acquisition.

Results:

3. “The median TKV values in the text and the table 1 are different.” We have corrected the discrepancy between the median TKV values in the text and Table 1.
4. “(Table 2) What does “Model” mean? Please describe in the text.” The meaning of "Model" in Table 2, referring to our auto-segmentation algorithm, is now clarified in the text under Table 2.
5. “About “Breathing during 2D scan” and “Different breath-hold positions in multi-acquisitions”, the text and tables are reversed, making it difficult to read.” We have rearranged the text and tables to ensure consistent and readable presentation of the points related to "Breathing during 2D scan" and "Different breath-hold positions in multi-acquisitions".

Discussion:

6. “Please provide the reference numbers.” Reference numbers are now provided.
7. “Please provide a brief conclusion in the final paragraph. ” We have now added a brief one-sentence concluding final paragraph.

Reviewer 2 Report

Summary:

This manuscript has determined the biases in measuring TKV by MRI, relying on robust data acquisition techniques.  The cohort study has discussed the main source of errors, including breathing during scan, breathing position, and overlapping slices acquisitions. Overall, the authors have provided strong evidence that MRI images should be carefully acquired and interpreted to avoid any misguidance to the patients’ monitoring and treatment.  

Minor:

The authors have given multiple factors to maintain the clinical quality control of MRI TKV. However, they have missed the more obvious and even greater factors that might be included to assure quality control of MRI.  These two factors are 1) human errors 2) difference between MRI scanners.  Obviously, interpretation of MRI images can vary from one person to another. Yet, differences between MRI scanners (calibrations, models, and software) can greatly impact the quality of MRI acquisition (Richard Herzog et al. 2017).  In Table S2 the authors have listed multiple MRI scanners but did not discuss if the data were acquired by multiple or single MRI scanner, nor did they mention the specification of the MRI scanner used in this study over in the material and methods section.

Author Response

We appreciate the positive comments acknowledging the robust data acquisition techniques used to determine biases in TKV measurements by MRI.

We have addressed the reviewer's concern regarding human errors and differences between MRI scanners as important factors for quality control with an additional paragraph in the discussion, see the last paragraph on page 13. Supplementary Table S2 now includes the list of all MRI scanners used in the study, along with their specifications. We also cite the reference by Herzog et al. 2017, highlighting the impact of MRI scanner differences. The inter-sequence coefficient of variation in TKV from different scanners is presented in Supplementary Figure 5, ranging from 3.5% to 5.2%. We mention the observation of human error and the development of an automatic extraction method for volume data from organ segmentation.