Review Reports - Myocardial T2 Star (T2*) in a Large Healthy Population: Correction Factors for a Segmental Approach Using Commercially Available Software in the Current MRI Era

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

The manuscript (Myocardial T2 in a large healthy population: correction factors for a segmental approach in commercially available software in the current MRI era) addresses an important clinical topic that is worth investigation. However, there are required changes.

The study enrolls only 50 subjects, yet both the title and abstract repeatedly describe this as a "large healthy population." Either justify the descriptor "large" relative specifically to prior work, or replace "large" with more precise language (e.g., "well-stratified cohort" or "sex- and age-stratified sample").

The Methods state subjects were enrolled "between November 2021 and May 2022 and underwent CMR in May 2022." This implies all 50 patients were scanned in a single month after up to 6 months of enrollment. Clarify: were all CMR scans truly performed in a single session/month? If so, explain the logistics. If this is a typographic error, correct it.

The reported correlation between age and T2 is R = –0.29, p = 0.04, a very weak effect. The Discussion claims this is "highly novel" and recommends age-specific normative values based on it. Temper the interpretation. An R of –0.29 explains only ~8% of variance. Acknowledge the weakness explicitly and note that validation in a larger cohort is needed before recommending age-stratified cutoffs as clinical practice. Consider including a scatter plot of age vs. global T2 to let readers assess the relationship visually.

Segment-level correction factors range enormously, from -1.45 to +8.31 Hz. The inferolateral segments, in particular, show correction factors more than 10× higher than those of the septal segments. Provide a dedicated figure or extended discussion mapping the anatomical drivers of these outlier correction values, particularly for segments 4 (basal inferior, 7.29), 5 (basal inferolateral, 8.31), 11 (mid-inferolateral, 8.11), and 16 (apical lateral, 8.18). Simply attributing them to "heart-lung interface" is insufficient.

The manuscript uses both "MEGE" and "MEGRE" interchangeably throughout the text.

The text uses "EKG" (electrocardiogram) while the abbreviation table defines "ECG."

Correlations between T2 values and age, sex, and wall thickness are assessed independently (univariate). There is no multivariate model to assess their combined or confounding effects.

Add a sentence in the Limitations explaining why COVID-specific criteria were applied, and note that future studies may not require them.

Section 3.2 reports a mean global T2 of 34.03 ± 6.65 ms, while Table 4 reports corrected global T2 for "All" as 37.2 ± 3.1 ms. The difference is expected (pre- vs. post-correction), but this is never explicitly explained in the text.

Several grammatical corrections are required.

Keywords are too narrow

Comments on the Quality of English Language

Minor grammatical corrections are required

Author Response

Response: Thank you for your comment.

R: Thank you for this observation. Based on current clinical recommendations (Messroghli et al , JCMR 2017), 20 subjects are considered sufficient to address this issue. Thus 50 subjects could be considered a large population. According to this, we added a dedicated explanation in the discussion and we also specified “well-stratified cohort” as suggested.

R: We enrolled all healthy volunteers between November 2021 and May 2022, but we acquired all of them in May, when we obtained scanner availability. In May 2022, 2-3 healthy volunteers per day underwent CMR. We added these details in the text.

R: Thank you for your valuable comment, we totally agree and rescaled accordingly our discussion and conclusion as follows: “cautiously, age-specific normative values can be considered. Nonetheless, these preliminary reference values require validation in a larger cohort, needed for application of age-specific normative values in clinical practice.” (discussion) and “These exploratory reference values require confirmation in larger cohorts.” (conclusion). You can find attached the scatter plot, but we prefer not to include it in the manuscript due to the low relationship.

R: Thank you for this valuable comment, in the dedicated section of the discussion, we added further explanations as follows: “it is possible to observe that highest values correspond to the basal inferior, basal and mid- inferolateral, and apical lateral segments, suggesting that artifacts are more prevalent in those regions, likely due to pulmonary veins, diaphragm, subdiaphragmatic structures, and geometric artifacts”.

The manuscript uses both "MEGE" and "MEGRE" interchangeably throughout the text.

R: Thank you for this observation, we have corrected the abbreviation throughout the text. We apologize for this inattention.

The text uses "EKG" (electrocardiogram) while the abbreviation table defines "ECG."

R: Thank you for this observation, we have corrected the abbreviation throughout the text. We apologize for this inattention.

Correlations between T2 values and age, sex, and wall thickness are assessed independently (univariate). There is no multivariate model to assess their combined or confounding effects.

R: Thank you for your valuable comment. We added a multivariate analysis to assess the combined effect, using ANCOVA across segments (7–12), with age, sex, and wall thickness as predictors. To mitigate Type I error from multiple comparisons, p-values were Bonferroni-corrected. The multivariate model implemented did not show significative combined effect of the variables (p>0.05). We added this update in the results.

Add a sentence in the Limitations explaining why COVID-specific criteria were applied, and note that future studies may not require them.

R: We thank the reviewer for this comment. We have added in the limitations the following sentence: “Future studies will not require COVID-specific inclusion criteria, applied in current study conducted in the historical era of COVID pandemic”.

R: We thank the reviewer for this comment, accordingly, section 3.2 reports “the mean global native T2* value” while in the text above Table 4 we rephrased “Post-correction T2* global values are reported in Table 4.”

Several grammatical corrections are required.

R: Thank you for your observation. The manuscript was reviewed for the English revision by a C2 grade A Cambridge certified professional, added into the acknowledgements.

Keywords are too narrow

R: Thank you, we provided the new following keyword: reference values

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

The manuscript entitled “Myocardial T2 in a large healthy population: correction factors for a segmental approach in commercially available software in the current MRI era”* addresses an important and clinically relevant topic. The effort to derive segmental myocardial T2* values and correction factors using commercially available software is commendable and represents a potentially useful step toward bridging the gap between research tools and routine clinical practice. The study is generally well organized, includes a reproducibility analysis, and follows a prospective design with stratification by age and sex, which are notable strengths.

However, several important concerns related to the framing, interpretation, and generalizability of the findings limit the strength of the conclusions in its current form. A major issue is the repeated characterization of the study cohort as a “large population,” whereas the study includes only 50 subjects, with approximately five individuals per age group. This description is misleading and should be revised to more accurately reflect the sample size. The limited number of subjects per subgroup also raises concerns regarding the robustness of the derived reference values, particularly when stratified by age and sex.

Related to this, the manuscript appears to overinterpret the findings by suggesting the establishment of normative values and even implying reconsideration of conventional clinical thresholds. Given the relatively small sample size and the absence of external validation, these conclusions are too strong. The results should instead be framed as preliminary or exploratory reference values that require confirmation in larger and independent cohorts. This point is particularly important to ensure that readers do not misinterpret the findings as definitive normative standards.

Another key limitation is the restricted generalizability of the results. The study was conducted at a single center using a single scanner (1.5T Siemens) and a single software platform (cvi42), and all participants were of Caucasian ethnicity. These factors may significantly influence T2* measurements and limit the applicability of the findings to other populations, scanners, or post-processing tools. The manuscript would benefit from a more explicit discussion of these constraints and their implications.

Furthermore, the correction factors derived in this study are presented with a degree of clinical applicability that is not fully supported by the data. While the comparison with previously published correction factors is informative, no independent validation cohort is included. Therefore, the manuscript should clearly state that these correction factors are preliminary and require external validation before they can be considered for clinical use.

From a statistical perspective, the analysis relies primarily on p-values, with limited reporting of confidence intervals or effect sizes. For example, the reported correlation between age and T2* values is statistically significant but weak, and its clinical relevance is uncertain. This should be interpreted more cautiously in the Results and Discussion sections. Including confidence intervals and providing a more balanced interpretation of statistically significant findings would strengthen the manuscript.

The characterization of the study population also appears limited. Although subjects were screened clinically, no biochemical parameters related to iron metabolism or broader cardiovascular risk profiling were included. This restricts the robustness of the definition of “healthy” and should be acknowledged more explicitly as a limitation. Additionally, the assumption that the mid-ventricular septum is relatively free from artifacts and can serve as a reference region for correction factors should be better justified, with appropriate references and discussion of potential limitations.

Author Response

Response: We thank the reviewer for this comment.

R: Thank you for this observation. We have tried to better justify why we could globally consider the population large based on the specific aim of the study. In fact, based on current clinical recommendations (Messroghli et al JCMR 2017), 20 subjects are considered sufficient to address this issue, thus 50 subjects could be considered a large population. According to this, we added a dedicated explanation in the discussion: “Based on current clinical recommendations [11], a healthy population of 20 subjects is needed to obtain normal mapping reference values, but to the extent of our knowledge, no data are available in the literature about the analysis of segmental and global T2* values in a so large and well-stratified cohort of healthy subjects using a commercially available software.”

R: Thank you for this valuable comment. We agree by a statistical point of view, even though by a clinical point of view it is the largest population on this issue and in line with current recommendations (Messroghli et al, JCMR 2017). Nonetheless, following the valuable reviewer’s suggestion, we provided a validation analysis in an independent cohort where the developed correction factors were able to correct for segmental susceptibility and/or geometric artifacts. We have updated the methods, the results, and the discussion. Moreover we added the suggested considerations into Discussion section as follows “these preliminary reference values require validation in a larger cohort, needed for application of age-specific normative values in clinical practice.” and rescaled the conclusions accordingly “These exploratory reference values require confirmation in larger cohorts.”

R: Thank you for this comment. Based on previous studies, it has been clearly demonstrated that T2* values are transferable among different scanners for clinical purposes, we added also another reference supporting this consideration (Tanner et al, Haematologica 2006), while regarding post-processing tool and Caucasian ethnicity, no data are available. These issues have been further commented in the Limitation section.

R: We thank the reviewer for this suggestion. We provided a validation on independent cohort and also added “these preliminary reference values require validation in a larger cohort, needed for application of age-specific normative values in clinical practice.” and rescaled the conclusions accordingly “These exploratory reference values require confirmation in larger cohorts.”

R: We agree with this comment. We rescaled the discussion and the conclusion accordingly as follows: “cautiously, age-specific normative values can be considered. Nonetheless, these preliminary reference values require validation in a larger cohort, needed for application of age-specific normative values in clinical practice .” (discussion) and “These exploratory reference values require confirmation in larger cohorts.” (conclusion). You can find attached the scatter plot, but we prefer not to include it in the manuscript due to the low relationship.

R: Thank you for this comment. For screening our healthy population, we checked also the last biochemical profile available, but the biochemical parameters related to iron metabolism are a second level analysis usually not required to define a healthy population as in previous study on this issue. Moreover, the corrected T2* values and the more sensitive T1 values in this population were negative, indirectly confirming the absence of clinical hemochromatosis. We specified “also based on the last biochemical profile available” in methods section. Furthermore, we had addressed the issue regarding mid-ventricular septum in the discussion, but we have better justified the assumption by adding other references (Carpenter, 2011)

Author Response File: Author Response.pdf

Reviewer 3 Report

Comments and Suggestions for Authors

This is an interesting manuscript regarding a clinically relevant gap in the scientific literature, namely the lack of segm. T2 correction factor for commercially available software, using current-gen equipment with Black Blood MEGE sequences. There are, however, numerous issues that should be corrected before the article is ready to be published:

the authors stated that the subjects "have been prospectively enrolled between November 2021 and May 2022 and underwent CMR without contrast administration in May 2022." The wording is confusing. Were all subjects scanned in May 2022, or did examinations take place throughout the enrollment period? Clarification is needed.
the stratification by decade (20-29, 30-3, etc.) with only 5 subjects per decade/sex is statistically insufficient to detect meaningful age-group differences. The statistical power for inter-group comparisons is severely limited at n=5 /cell
"echo spacing = 2.26 ms" should be verified. With TEs from 2.0 to 20.3 ms over 10 echoes, the spacing would be (20.3-2.0)/9 ≈ 2.03 ms, not 2.26 ms. The exact TE values must be clarified
"echo train length = 18.34 ms" is confusing. Echo train length typically refers to the number of echoes, not a duration in ms. This likely refers to the total acquisition duration of the echo train from first to last echo), in which case it should be reworded
No FOV (field of view) is reported
OIs were manually traced, introducing variability. The authors mention an "offset" to avoid epicardial fat and blood pool but do not specify the magnitude of this offset (in pixels or mm). This information is neededfor reproducibility
The global T2 value is computed as the arithmetic mean of all 16 segments. This is debatable, as basal, mid-ventricular, and apical segments have different areas. An area-weighted mean would be more representative. The authors should justify their choice of simple averaging
The method for computing correction factors is described in summary fashion. The T2 to R2 conversion, deviation from the mid-ventricular septum, and generation of a correction factor are conceptually clear, but the authors do not present the explicit formula. For instance, it is unclear whether the correction factors are expressed in Hz (R2) or in ms (T2). In Table 3, values appear without an explicit unit of measurement.
comparing mean R2 values with those from Positano et al. using a t-test is problematic: the studies have different populations (22 vs. 50 subjects), different equipment, and different sequences. A simple t-test does not control for these differences. Furthermore, it is unclear whether the authors had access to individual-level data from the Positano study or only compared published summary statistics (in which case a specific test for comparison with published reference values would be needed)
No correction for multiple comparisons is mentioned, which is a significant problem with 16 segments tested simultaneously. Without Bonferroni or similar correction, the significance at p < 0.05 for 8 out of 16 segments (Table 3) should be interpreted with great caution
"34.03 ± 6.65 (range: 29.9 - 37.9 ms)." The range is completely incompatible with the SD. If the true range is 29.9-37.9, then the SD cannot be 6.65 but should likely be approximately 2-3 ms. Alternatively, if the SD is truly 6.65, the range should be much wider
table 2 is difficult to read and would benefit from reformatting. The column header "Avg. Differe" (mean difference) should be "Avg. Difference."
The Bland-Altman limits for inter-study reproducibility at segment 11 (mid-ventricular infero-lateral) are extremely wide: "from -10.04 to 11.86," suggesting poor reproducibility in this region. This is not discussed
R = -0.29 with p = 0.04 is a weak correlation, explaining only approximately 8% of the variance. At n=50, this is at the threshold of significance and may be unstable. The authors correctly interpret it as "weak," but they overinterpret the clinical implications in the Discussion section
there are some recent and relevant references missing from the discussion section, which would alter the claims of the authors.
in the limitations section should be added: (a) small sample size per age decade (n=5), (b) use of a single scanner, (c) limited experience of the observers, (d) absence of external validation of the correction factors in patients with iron overload.

Comments on the Quality of English Language

there are many grammar and syntax issues that should be corrected

Author Response

the authors stated that the subjects "have been prospectively enrolled between November 2021 and May 2022 and underwent CMR without contrast administration in May 2022." The wording is confusing. Were all subjects scanned in May 2022, or did examinations take place throughout the enrollment period? Clarification is needed.

Response: Thank you for this comment. We enrolled between November 2021 and May 2022 but we acquired all subjects in May 2022, when we obtained scanner availability. We addedd these details in the text.

the stratification by decade (20-29, 30-3, etc.) with only 5 subjects per decade/sex is statistically insufficient to detect meaningful age-group differences. The statistical power for inter-group comparisons is severely limited at n=5 /cell

R: Thank you for this valuable comment. We agree by a statistical point of view, even though by a clinical point of view it is the largest population on this issue and in line with current recommendations (Messroghli et al, JCMR 2017). Nonetheless, we added into Discussion section that “these preliminary reference values require validation in a larger cohort, needed for application of age-specific normative values in clinical practice.” and rescaled the conclusions accordingly “These exploratory reference values require confirmation in larger cohorts.”

"echo spacing = 2.26 ms" should be verified. With TEs from 2.0 to 20.3 ms over 10 echoes, the spacing would be (20.3-2.0)/9 ≈ 2.03 ms, not 2.26 ms. The exact TE values must be clarified

R: We thank the reviewer for this suggestion. We verified that echo spacing was 2.26 ms, 10 TE are acquired from 2.02-22.36 ms, as reported in the text.

"echo train length = 18.34 ms" is confusing. Echo train length typically refers to the number of echoes, not a duration in ms. This likely refers to the total acquisition duration of the echo train from first to last echo), in which case it should be reworded

R: We thank the reviewer and correct with “echo train length = 10”. We apologize for the inconvenience.

No FOV (field of view) is reported

R: Thank you for this suggestion, we added FOV = 400 mm.

ROIs were manually traced, introducing variability. The authors mention an "offset" to avoid epicardial fat and blood pool but do not specify the magnitude of this offset (in pixels or mm). This information is needed for reproducibility

R: Thank you for this comment, the offset was up to 2 mm on both endocardial and epicardial side. We added this detail in the method section.

The global T2 value is computed as the arithmetic mean of all 16 segments. This is debatable, as basal, mid-ventricular, and apical segments have different areas. An area-weighted mean would be more representative. The authors should justify their choice of simple averaging

R: Thank you for this valuable observation. We know well this debate, that is shareable by a theoric point of view, but according to literature and available software tools, we applied the arithmetic mean. We have futher justified this choice in the methods.

The method for computing correction factors is described in summary fashion. The T2 to R2 conversion, deviation from the mid-ventricular septum, and generation of a correction factor are conceptually clear, but the authors do not present the explicit formula. For instance, it is unclear whether the correction factors are expressed in Hz (R2) or in ms (T2). In Table 3, values appear without an explicit unit of measurement.

R: We thank the reviewer for this suggestion and apologize for the inconvenience. We added the formula in the text and we explicited the unit of the measurement in Table 3.

comparing mean R2 values with those from Positano et al. using a t-test is problematic: the studies have different populations (22 vs. 50 subjects), different equipment, and different sequences. A simple t-test does not control for these differences. Furthermore, it is unclear whether the authors had access to individual-level data from the Positano study or only compared published summary statistics (in which case a specific test for comparison with published reference values would be needed)

R: Thank you for this observation. We know well that population, equipment and sequences are different, and the aim of the current study is to obtain correction factors in the current MRI era. Nonetheless, the transferability of T2* technique across vendors and sequences has been previously demonstrated. We have compared the values with the published summary statistics and we added this detail.

No correction for multiple comparisons is mentioned, which is a significant problem with 16 segments tested simultaneously. Without Bonferroni or similar correction, the significance at p < 0.05 for 8 out of 16 segments (Table 3) should be interpreted with great caution

R: We thank the reviewer and we totally agree with this comment. We added Bonferroni correction results in Table 3.

"34.03 ± 6.65 (range: 29.9 - 37.9 ms)." The range is completely incompatible with the SD.If the true range is 29.9-37.9, then the SD cannot be 6.65 but should likely be approximately 2-3 ms. Alternatively, if the SD is truly 6.65, the range should be much wider

R: We thank the reviewer and apologize for the inconvenience. We corrected the range in the text.

table 2 is difficult to read and would benefit from reformatting. The column header "Avg. Differe" (mean difference) should be "Avg. Difference."

R: Thank you for this suggestion, we reformatted the Table 2.

The Bland-Altman limits for inter-study reproducibility at segment 11 (mid-ventricular infero-lateral) are extremely wide: "from -10.04 to 11.86," suggesting poor reproducibility in this region. This is not discussed

R: We agree with the reviewer about the extremely wide range in the mid-ventricular inferolateral segment, probably suggesting poor reproducibility in this region and a great influence from susceptibility and geometric artifacts. These findings have been reported also in previous studies and we added a comment regarding this issue in discussion section.

R = -0.29 with p = 0.04 is a weak correlation, explaining only approximately 8% of the variance. At n=50, this is at the threshold of significance and may be unstable. The authors correctly interpret it as "weak," but they overinterpret the clinical implications in the Discussion section

there are some recent and relevant references missing from the discussion section, which would alter the claims of the authors.

R: We thank the reviewer for this suggestion. Nonetheless, at the best of our knowledge we did not find relevant references in disagreement with our observation. For example, Ozcan et al. found lower T2* values in older females (Özcan Ç, Yiğit H, Çetin MS, Özcan İ. Analysis of myocardial T1, T2, and T2* values by age, sex, and cardiac segments in normal population: a prospective study. Int J Cardiovasc Imaging. 2024 Nov;40(11):2357-2369. doi: 10.1007/s10554-024-03241-5. Epub 2024 Sep 16. PMID: 39283474.) We added the above-mentioned reference.

in the limitations section should be added: (a) small sample size per age decade (n=5), (b) use of a single scanner, (c) limited experience of the observers, (d) absence of external validation of the correction factors in patients with iron overload.

R: We thank for the observations and we added these details in limitation section.

there are many grammar and syntax issues that should be corrected

R: Thank you for your observation. The manuscript was revised for the English revision by a professional C2 grade A Cambridge certified , added into the acknowledgements.