Are Quantitative Errors Reduced with Time-of-Flight Reconstruction When Using Imperfect MR-Based Attenuation Maps for ¹⁸F-FDG PET/MR Neuroimaging?

Round 1

Reviewer 1 Report

The work presented by the authors is very interesting and novel since the improvement of TOF-reconstruction in neuroimaging was not thoroughly investigated. The authors have shown improvements in TOF-images even for a timing resolution of 525 ps, so it is expected that further improvements will appear with better timing resolutions. This result encourages to continue in reducing the timing resolution of the new PET devices. The main limitation to the study is presented in the work and is well addressed. The methodology is clear and the results are sound. Therefore, we recommend this paper for publication.

There are small corrections:

-line 24: we've --> we have

-line 324: it's --> it is.

Best regards,

Author Response

Please see the attachment

Author Response File: Author Response.docx

Reviewer 2 Report

This paper presents a comparison of time-of-flight (TOF) reconstruction versus non-time-of-flight reconstruction for neuroimaging using a hybrid PET/MR scanner and applying attenuation correction of varying accuracy. Reconstructed PET images using three variants of MR-derived attenuation correction (two-class or three-class tissue segmentation) are compared to CT-derived attenuation correction. For whole-body PET/MR scans a number of studies have demonstrated the benefit of using TOF reconstruction to reduce bias from imperfect attenuation correction, and the current work confirms this overall behavior for brain imaging. Unfortunately a direct comparison between non-TOF and TOF reconstruction has not been possible due to limitations in the applied software. Nevertheless the work can be considered as a useful contribution to the important challenge with improvement of MR-derived attenuation correction.

In general the manuscript is well written, and the results are clearly presented, however, several mistakes must be corrected.

The work is recommended for publication after minor revision.

Specific comments:

Line 31-33:

It should be added that the uPMR790 with 480 ps TOF from United Imaging Healthcare has been available since 2018. Suggested reference: Chen et al, Eur. J. Nucl. Med. 2021, 48:1726–1735

Figure 2 caption:

The red symbols ("+") should be explained. Is it outliers? Have they been omitted during computation of the sample median, and the first and third quartiles?

Figure 3:

Clearly the units are incorrect. Probably kBq/ml should be replaced by Bq/ml (x 10e4).

For both two-class MRAC methods using non-TOF reconstruction it is found that MRAC-PET is slightly higher than CTAC-PET. This is clearly in disagreement with figure 2 and 4.

Completely removal of figure 3 should be considered, since it does not add much information compared to figure 4.

Table 1: The coefficients for 2-class MRAC should be corrected (see commetn above).

Author Response

Please see the attachment

Author Response File: Author Response.docx

Reviewer 3 Report

1. A couple of abbreviations are used in the article without being translated beforehand.
2. There are times when the article's language and flow are confusing and difficult to follow. Please improve the style and flow of the article in order to improve its general readability.
3. The major claim in the paper is that “Using TOF reconstruction is beneficial in reducing the quantitative errors due to imperfections in MRAC also in the head region.” The paper provides explained variance for the linear fits for multiple MRAC methods between TOF and non-TOF. However, there is no statistical analysis conducted to compare if the differences in the regression slopes are significant between TOF and non-TOF across MRAC methods. The paper shows qualitative plots in figures 5,6 and 7. Similarly, no quantification of differences shown in these figures is carried which limits any conclusive inference from this study. The paper should clearly describe their hypothesis and test it with statistical analysis. Consequently, no conclusive scientific conclusions can be drawn from the results presented in this paper.

Author Response

Please see the attachment

Author Response File: Author Response.docx

Round 2

Reviewer 3 Report

The paper now includes a slightly improved statistical analysis. The paper's statistical analyses, however, remain flawed:

1. The paper reports now that their primary claim is not valid when comparing TOF and non-TOF medians. This finding should be reported clearly in the paper and the authors should adjust their main claim accordingly.
2. In addition, the statistical validity of their finding is still questionable. In Table 1, the paper reports p-values for the explained variance, but fails to report the significance of the slope parameters. In order to test the hypothesis, the significance of each slope parameter must be determined.
3. The quantitative differences between the data shown in Figure 1 and the explained variance in Table 1 are almost nonexistent.

This raises suspicions about the claims made in the paper.

Author Response

Please see the attachment

Author Response File: Author Response.docx

Round 3

Reviewer 3 Report

No further suggestions.