DTI Abnormalities Related to Glioblastoma: A Prospective Comparative Study with Metastasis and Healthy Subjects

Round 1

Reviewer 1 Report

General Comments:

Your paper is fairly well written with good introduction and method of your study. It is not clear whether your data is new or confirming others' results. Your conclusion is weak with general message rather than specific to GBM for the differential diagnosis.

Specific Comments:

1. Is the Gd-contrast enhancement correlated with aCCV of FA data?
2. How about the correlation with PET/CT using amino acid or thymidine?
3. line 303:....determine.------>....determined.

Author Response

Response to Reviewer 1 Comments

Point 1: Your paper is fairly well written with good introduction and method of your study. It is not clear whether your data is new or confirming others' results. Your conclusion is weak with general message rather than specific to GBM for the differential diagnosis.

Specific Comments:

1. Is the Gd-contrast enhancement correlated with aCCV of FA data?
2. How about the correlation with PET/CT using amino acid or thymidine?
3. line 303:....determine.------>....determined.

We thank the reviewer for the constructive comments.

Response 1: The Gd-contrast enhanced area was included in the macroscopic volume (MV; see Introduction; thus the MV was the gross tumor volume, GTV, of radiotherapy). We did not intend to study specifically the Gd-contrast enhanced area versus the other volumes. Further studies are mandatory, now that we explored aCCV.

Response 2: Futher studies comparing metabolic PET/CT images and MRI volumes should very likely be of interest enbling to broaden the understanding of tumor. This relys on carefull co-registration of image data sets (and consequently of the MV, BAT and aCCV). Multimodal approach will be undoubltly of interest.

Response 3: line 303:....determine.------>....determined. Thank you / Corrected in the manuscript (red)

Curroncol-1653836.

On behalf of the authors, Youssef-El Ouadih, Jean-Jacques Lemaire

Clermont-Ferrand, Monday, April 11, 2022

Reviewer 2 Report

The authors are presenting a paper entitled: DTI abnormalities related to Glioblastoma: A Prospective comparative Study with Metastasis and Healthy Subjects.

Authors focused on MRI 15 abnormalities observable in the tumor and the brain adjacent to the lesion, up to the contralateral hemisphere, with a special interest in tensor diffusion imaging informing on white matter architecture. They aimed to study several related parameters of MRI datasets of two series of patients suffering from GBM and metastasis, and of a control group of healthy subjects.

The paper is of interest, is well written and clear.

I recommend publication because in such a deadly disease any information can bring a piece to the puzzle that we are trying to build. I have however some minor comments:

1. Authors should describe the healthy population in more detail, as to why they went through an MRI for example
2. The comparison with a patient with brain metastasis was to be able to have a second term of comparison or the two lesions might be misdiagnosed sometimes?
3. Auhtos are reporting that: “The growing interest in artificial intelligence methods enabling to detect tumor on medical imaging, and the importance of optimal topographically directed molecular analysis of tissue and of planning and follow-up strategies of treatments, impose to refine our understanding of the macroscopic phenotypic definition of GBM and its invasiveness”. This is a very crucial point that should be connected to the results obtained in this work.
4. What is the clinical translatability of these results. Besides the importance of having the possibility of using imaging to detect alteration of brains structure, authors should stress more the future perspectives of these findings.
5. In line 37 authors are claiming to study the population prospectively, in which sense?
6. Tumors were all primary GB, what about their molecular characteristics?
7. Where is metastasis coming from?
8. Healthy subjects had only four values, are those so relevant then? explain this better.
9. It would more clear to present data as histograms too.

Author Response

Response to Reviewer 2 Comments

The authors are presenting a paper entitled: DTI abnormalities related to Glioblastoma: A Prospective comparative Study with Metastasis and Healthy Subjects.

Authors focused on MRI 15 abnormalities observable in the tumor and the brain adjacent to the lesion, up to the contralateral hemisphere, with a special interest in tensor diffusion imaging informing on white matter architecture. They aimed to study several related parameters of MRI datasets of two series of patients suffering from GBM and metastasis, and of a control group of healthy subjects.

The paper is of interest, is well written and clear.

I recommend publication because in such a deadly disease any information can bring a piece to the puzzle that we are trying to build. I have however some minor comments:

1. Authors should describe the healthy population in more detail, as to why they went through an MRI for example
2. The comparison with a patient with brain metastasis was to be able to have a second term of comparison or the two lesions might be misdiagnosed sometimes?
3. Authors are reporting that: “The growing interest in artificial intelligence methods enabling to detect tumor on medical imaging, and the importance of optimal topographically directed molecular analysis of tissue and of planning and follow-up strategies of treatments, impose to refine our understanding of the macroscopic phenotypic definition of GBM and its invasiveness”. This is a very crucial point that should be connected to the results obtained in this work.
4. What is the clinical translatability of these results. Besides the importance of having the possibility of using imaging to detect alteration of brains structure, authors should stress more the future perspectives of these findings.
5. In line 37 authors are claiming to study the population prospectively, in which sense?
6. Tumors were all primary GB, what about their molecular characteristics?
7. Where is metastasis coming from?
8. Healthy subjects had only four values, are those so relevant then? explain this better.
9. It would more clear to present data as histograms too.

We thank the reviewer for the constructive comments.

Response 1: The control group of healthy people meant that they did not suffer of neurologic nor psychiatric disease, and the observation of the MRI data set showed, so far, the absence of lesions, as also confirmed by the normal values ( Table 2). Precisions were added in the section § population.

 Response 2: The metastasis group was used to compare the two different invasive profiles (see § introduction: “The metastatic lesions were chosen for comparison with GBM because most are re-strained with millimetric invasion [27,28], and as such is frequently used as the oppo-site model of infiltration of gliomas (see above).”). By the way, yes the clinical context and images can be confusing, at least before histopathology. Following the reviewer’s comment we have added a sentence in the section § conclusion.

Response 3: This citation and its related references are placed at the end of the introduction (lines 68-72) highlithing the importance to go beyond our current understanding of phenotypes. This is fully connected to our results. However, following the reveiwer’s comment (plus an editorial suggestion), there is a new section § Conclusion, in which we placed the last sentences of the discussion. Doing so, our findings and conlusive remarks fit with the introduction.

Response 4: The clinical translatability relies on machine/deep learning approaches according to MV, BAT and aCCV volumes, with supervised or un-supervised algorythms of segmentation, enabling  the analysis of volumes, e.g. to guide the determination of pertinant biopsy sites (surgical or “radiomical” biopsies), or to study anticancer treatments. These aspects are specified in the section § conclusion.

Response 5: We would be happy to clarify this point if directed (so far, there is no claim on prospective study in the line 37).

 Response 6: We did not collect the IDH mutations and MGMT promoter methylation for the GBM series for this work (outside the scope and the ethical approval authorization). We focused on pretherapeutic MRI aspects, “naïve”. Further studies on mean-term aspects following radiotherapy-chemotherapy, such as temozolomide, will include these mollecular characteristics with reliable, relative, therapeutic pronotic value (e.g. Śledzińska, P. and al. Prognostic and Predictive Biomarkers in Gliomas. Int J Mol Sci 2021, 22, 10373).

Response 7: The primary cancers are specified table 1.

Response 8: The “four values” were relevant because linked to homologuous locations of ROIs in individual MRI data sets of patients. We considered that comparisons between pathologic (patients) and non-pathologic (heatlhy subject) volumes, were precluded by the impossibility to specify homologuous volumes. The consequences would have been dramatic notably for the analysis of anisotopic parameters, sensitive to the architecture of WM.

Response 9: We thank the reviewer, but we preferred to report the results in a single table (plus complementary data in the text), following attempts with different graphical options that would have increased dramatically the number of figures and the complexity of statistical description without modification of the results (e.g. using effect-sizes (ES) and 95% confidence intervals).

Curroncol-1653836.

On behalf of the authors, Youssef-El Ouadih, Jean-Jacques Lemaire

Clermont-Ferrand, Monday, April 11, 2022

Round 2

Reviewer 1 Report

Your paper is well written with proper revision along the lines reviewers' comments.