Dual-Tracer PET-MRI-Derived Imaging Biomarkers for Prediction of Clinically Significant Prostate Cancer

Round 1

Reviewer 1 Report

The article is indeed interesting since advances in diagnostic tools for localized prostate cancers could avoid unnecessary surgical procedures.

I have some concerns regarding the use of two positron-emitting radiopharmaceuticals, both in terms of costs and practical advantages. The difference in terms of AUC between FMC and FMC+PSMA is low (however, not investigated by statistical analysis), but is it sufficient to justify their use? Moreover, PSMA alone, given the data from the latest studies in the biochemical recurrence setting, could probably be enough, thus sparing unnecessary radiation to the patients. Furthermore, is it practical for most of the referral centres to use this technology? These themes should be discussed in detail.

Some observations:

-          Inclusion ed exclusion criteria should be provided; a table with patients’ main clinical characteristics - including median basal PSA levels - should be shown in the article.

-          Statistical analysis: please specify which software did you use

-          Please specify the two acronyms: DCE and ADC

Author Response

I have some concerns regarding the use of two positron-emitting radiopharmaceuticals, both in terms of costs and practical advantages. The difference in terms of AUC between FMC and FMC+PSMA is low (however, not investigated by statistical analysis), but is it sufficient to justify their use? Moreover, PSMA alone, given the data from the latest studies in the biochemical recurrence setting, could probably be enough, thus sparing unnecessary radiation to the patients. Furthermore, is it practical for most of the referral centres to use this technology? These themes should be discussed in detail.

• thank you for this comment - as we have stated this study is not sufficient to use this in clinical standard practice. our cohort is retrospective, and this was a pilot study to see if there is any merit to this. also our cohort consist only of patients with cancer, so we cannot make any assumptions in a general population with elevated PSA only. Further investigation is needed to see if the added benefit warrants the additional resources needed.

Some observations:

        - Inclusion ed exclusion criteria should be provided; a table with patients’ main clinical characteristics - including median basal PSA levels - should be shown in the article.

• thank you. PSA levels and patient characteristics are shown in table 1, we have added them into the results section as well. inclusion and exclusion criteria are based on the initial prospective trial, we have added these into the methods section

-          Statistical analysis: please specify which software did you use

• Stata was used. has been added to methods

        - Please specify the two acronyms: DCE and ADC

• DCE: dynamic contrast enhanced
• ADC: apparent diffusion coefficient
• has been added

Reviewer 2 Report

The aim is to investigate if imaging biomarkers derived from 3T dual- 77 tracer (FMC and PSMA) PET/mpMRI can adequately predict csPC. The topic is interesting but some concerns about study design are to be emphasized.

-        First of all in the introduction please be more concise. A short framework and study aim are what is required. On the topic, I believe what follows should be included in the discussion “Introduction has to be A published meta-analysis including studies with suspected PC patients has shown that the positive predictive value (PPV) for csPC of PI-RADS score 3,4 and 5 were 12%, 48% and 72%, respectively, with a high heterogeneity among included studies”. 

-        Even if the database was maintained prospectively as part of a prospective trial, this is a retrospective study design. Please in materials be more accurate on the matter in order to avoid misinterpretation.

-        A big concern is the study target. Only patients with proven PC were included in the analysis, thus we cannot make any assumptions about the impact of dual-tracer PET/MRI in cancer detection, which would be valuable and most impactful. The editor should take it into account.

-        When talking about biomarkers in general. The tumor microenvironment should be considered. In this view, It is well known that autophagy plays a crucial role in cancer development in the dualism between significant and not significant PC. Recent findings have been conducted on molecular and biological pathways of autophagy please provide a brief description and include this recent paper (doi: 10.3390/ijms23073826; PMCID: PMC8999129; PMID: 35409187). I think this aspect should be discussed and included in your paper.

-        Correct typos.

Author Response

      - First of all in the introduction please be more concise. A short framework and study aim are what is required. On the topic, I believe what follows should be included in the discussion “Introduction has to be A published meta-analysis including studies with suspected PC patients has shown that the positive predictive value (PPV) for csPC of PI-RADS score 3,4 and 5 were 12%, 48% and 72%, respectively, with a high heterogeneity among included studies”.

• We thank you for this comment. We believe however that as this journal is not specifically intended for urologists/radiologists, but caters to a wider audience of health care providers treating malignant disease, that a slightly more extended introduction would be of benefit. the aforementioned sequence was included within the introduction, l to give the reader an overview of the current standard practice (namely mpMRI of the prostate) and the data that it has been deprived from.

      - Even if the database was maintained prospectively as part of a prospective trial, this is a retrospective study design. Please in materials be more accurate on the matter in order to avoid misinterpretation.

• While the word retrospective was clearly used, we understand the possibility of misinterpretation and have made this more clear in the first sentence of the methods section, thank you 

      - A big concern is the study target. Only patients with proven PC were included in the analysis, thus we cannot make any assumptions about the impact of dual-tracer PET/MRI in cancer detection, which would be valuable and most impactful. The editor should take it into account.

• Yes this is a very valid point. this study was a pilot investigation to see if there was any merit to using a dual tracer protocol (which has also been used in other situations such as LuPSMA studies). clearly we cannot make any assumptions on the performance in a general cohort. this, as well as the associated costs are of course big problems for this - yet this was again, not the aim of our study. we could show that within this cohort, some benefit was present, and further investigation will be needed.

      - When talking about biomarkers in general. The tumor microenvironment should be considered. In this view, It is well known that autophagy plays a crucial role in cancer development in the dualism between significant and not significant PC. Recent findings have been conducted on molecular and biological pathways of autophagy please provide a brief description and include this recent paper (doi: 10.3390/ijms23073826; PMCID: PMC8999129; PMID: 35409187). I think this aspect should be discussed and included in your paper.

• While the TME is obviously an important factor in tumor emergence and progression, this is a pilot study for imaging as a tool non invasive tumor grading in localised PCA. While the effects of TME are certainly relevant for FMC/PSMA expression (inflammation, androgen deprivation etc.), and their ligand, however this is out of scope for this paper. the citation you suggest is mostly concerned with changes within TME in metastatic tumors, and response to systemic treatment. we do not consider this relevant in our cohort of localised PCA, where the main distinction relevant at this stage of research is: significant vs insignificant PCA. markers of Autophagy as well as PET biomarkers are both currently not standard in diagnosis and/or treatment of localised PCA, once this becomes the case, surely an investigation into this topic is warranted.

      - Correct typos.

• manuscript has been corrected

Reviewer 3 Report

The article under review is all about prostate cancer patients who underwent 3T dual-tracer PET/mpMRI followed by radical prostatectomy (RP) between 2014 and 2017. The author performs performed a retrospective lesion-based analysis of all cancer foci and compared it to whole-mount histopathology of RP specimen. The author found that combined FMC+PSMA SUVmax were the only significant parameters for the prediction of prostate cancer patients. The study design is very good and effective for prostate cancer patients and it will reduce overtreatment in cancer patients.

Minor concerns:

1.       How does the combination of FMC+PSMA is going to distinguish the patients within grades or of Gleason score? The combination is only to distinguish between clinically significant and nonsignificant patients.

2.       Author mentioned in the article that n=77 was considered in the study however the patient’s details are missing. For instance, do those patients of the same age, what was the Gleason score, and whether the patients belong to BPH.

3.       How does the combination of FMC+ PSMA is more effective and accurate compared to the multimodal MRI-based radiomics nomogram method?

4.        Reference number 1 is not relevant to the cited text “Prostate cancer (PC) is the second most frequent cancer worldwide and it is responsible for 11% of all cancer-related deaths”.

5.       Why did the author the used DeLong test to compare the compare AUC´s?

6.       I agree with the author that a large cohort of patients is required to validate the efficacy in patients.  

Author Response

Dear Reviewer, we thank you for the constructive comments. Please find our responses below:

1. How does the combination of FMC+PSMA is going to distinguish the patients within grades or of Gleason score? The combination is only to distinguish between clinically significant and nonsignificant patients.

• We cannot distinguish between Gleason Scores/ISUP classification groups. however this is not the point of this study. clinically the most important question is currently "does the cancer need treatment or not", thus our primary focus was to distinguish between significant and insignificant PCA. This study was a pilot project to see if this approach had any validity, and a more thorough investigation is surely needed before applying this to practice.

2. Author mentioned in the article that n=77 was considered in the study however the patient’s details are missing. For instance, do those patients of the same age, what was the Gleason score, and whether the patients belong to BPH.

• Patient characteristics are given in Table 1. including age, PSA, clinical Stage and ISUP grade, we have included this information now additionally within the results section. ISUP grade is the newer version of Gleason Score and should be used in reporting, the classic Gleason Score is no longer relevant (e.g. Gleason 3+4 = ISUP 2). For BPH there is no effect that might be important to this study, that is why it is not reported. 

3: How does the combination of FMC+ PSMA is more effective and accurate compared to the multimodal MRI-based radiomics nomogram method?

• it was equal. however the multivariable model also included the FMC+PSMA values. FMX + PSMA and also the model (including these values) were both more accurate than any MRI parameter including PIRADS.

4: Reference number 1 is not relevant to the cited text “Prostate cancer (PC) is the second most frequent cancer worldwide and it is responsible for 11% of all cancer-related deaths”.

• We have changed this

5: Why did the author the used DeLong test to compare the compare AUC´s?

• Because we compared in at least two cases (FMC+PSMA and CHAID model) AUCs of models as opposed to binary markers.

6: I agree with the author that a large cohort of patients is required to validate the efficacy in patients

thank you

Round 2

Reviewer 1 Report

The authors have addressed my points.

Author Response

thank you very much for the comments

Reviewer 2 Report

Thank you for addressing the revised version of your paper. The big concern is still the study target, even if this study was a pilot investigation to see if there was any merit to using a dual tracer protocol. The editor should take it into account.

Author Response

Dear Reviewer,

We very much understand this point of criticism of our paper. We are aware of this and it is very clearly stated within the limitations (end of dicussion) that this is the case, and only patients with positive biopsy/proven PCA that were already scheduled for RP were included here.
We still think that this study has merit, as we know that PSMA is very good, yet not perfect. As this is the case, ways to improve upon our diagnostic options - to eventually achieve reliable non-invasive diagnosis if significant PCA - are constantly a field of research. The dual tracer approach has also been used as the combination of PSMA and FDG, top reduce PSMA false positives. The combination of PSMA and FMC is rather an attempt to reduce false negatives thus improve sensitivity. other options (as discussed in the paper) are additional genomic testing and so on. 

So we are aware - we have pointed it out in the manuscript very clearly - yet we think despite these limitations it is valuable insight.