Screening for Prognostic Biomarkers in Metastatic Adrenocortical Carcinoma by Tissue Micro Arrays Analysis Identifies P53 as an Independent Prognostic Marker of Overall Survival

Round 1

Reviewer 1 Report

This paper by Hescot et al., aimed at analyzing expression of a large panel of proteins of the major deregulated pathways in adrenocortical carcinomas. For this, the authors conducted a retrospective study by immunohistochemistry on a TMA of 66 ACC patients (metastatic cancer) and evaluated correlations with prognosis including adjustements to the ENSAT and GRAS parameters. These analyses identified p53 and PDxK as prognostic of overall survival and showed that p53 was an independent marker of metastatic ACC after adjustment to mENSAT-GRAS parameters. Although these findings are not entirely novel, they rely on the concomitant assessment of 16 markers, a clearly defined cohort of 66 metastatic ACC and the novel GRAS scoring system. Therefore, they provide robust and interesting insight into prognostic relevance of key pathways in metastatic ACC. Before the paper can be published, I would like the authors to address the following comments.

 

 

Supplementary figures showing the different type of staining patterns for all markers are mandatory. In line with this, there should be an explanation of how the H-score was calculated and how “the majority of the tumour” was defined. Also, for staining of b-catenin, it is unclear which proportion of the tumour was taken into account to qualify for “nuclear expression” Same goes for LEF1, SF1, GsTP& etc... What was the threshold chosen to decide that a tumour was positive?

 

 

Addition of a color-coded table, like what is done for OMIC studies/Bioportal style, with all patients and all markers would help interpret the data. For example, it is impossible to tell from the current manuscript, if patients with abnormal b-catenin staining are also the patients with abnormal LEF1 accumulation. This table should also include hormonal secretion status, Ki67 and mENSAT-GRAS score.

 

Legend of table 2 should read “*defined as H-score>150 (high) or <150 (low)”

 

The large number of tumours showing P16 overexpression (50.8%) is interesting. Is there a link with other potential markers of senescence in these tumours? I would love to see staining for P21 and/or phospho-H2AX to evaluate this hypothesis. Is P16 staining linked with sex of patients or hormonal status?

 

Author Response

Please see the attachment

Author Response File: Author Response.docx

Reviewer 2 Report

The authors provide an interesting prognostic biomarker analysis in patients with metastatic ACC.

My main concern is that the article is centered around behaviour of the disease in the metastatic setting (primary endpoint is OS from time of metastatic diagnosis), whereas the tissue that was analyzed is mainly harvested from the primary tumor (55/66 cases), probably when the disease was not yet metastatic. Can the authors find a way of demonstrating that the conclusions are valid despite this discrepancy? Is it possible to display which proportion of patients with p53 overexpression belong to the group of primary tumours that were analyzed or did they all belong to the group who had metastatic tissue analyzed?

In the multivariable analysis, the authors adjust for mENSAT-GRAS parameters, however the performance of these variables are not shown in the table. It would be interesting to see the mutual ratio's of the complete model, so with the mENSAT-GRAS parameters included.

Furthermore, the article needs some editing by a native speaker (English). 

Author Response

Please see the attachment

Author Response File: Author Response.docx

Reviewer 3 Report

The authors analysed the expression of 16 proteins involved in known altered pathways in ACC (cell cycle, Wnt/ß-catenin, methylation) to identify and prioritize potential prognostic or predictive parameters in a cohort of 66 patients with metastatic ACC. Prioritisation was performed against modified ENSAT and GRAS parameters. They demonstrated that after adjustment to mENSAT and GRAS parameters, p53 and PDxK were prognostic of overall survival.

The topic is of some interest in the field of endocrine cancers, the study is well designed and well written. However, I have a number of concerns and requests that need to be addressed.

 

Major comments

1) Methods - Clinical and pathological characteristics as in Table 1: it is not entirely clear to me how the different variables have been calculated and considered. For instance, I understand that ENSAT tumour stage, hormonal secretion, and tumour size are from the time of diagnosis. Similarly, Weiss score, ki67 and resection status are from the primary tumour. Can the author confirm and better specify? Moreover, the initial mENSAT classification (at time of diagnosis) should be added in Table 1 and described. This is an essential part of the results but is not reported here.

Also, in Table 1 resection status other than R0 should be added (RX, R1, R2).

When describing details of therapy with mitotane, it is unclear if it is referred also as treatment of patients with advanced disease or includes also previous adjuvant treatment. This should also be clarified. Finally, response to treatment and disease control >12 months are referred to mitotane or platinum-based chemotherapy. Should be clearly specified.

2) Methods – immunohistochemistry: some details regarding the interpretation of the staining are missing. Were the slides analysed by only one investigator? This could potentially lead to some operator-dependent bias. How could the authors be confident that their results are reproducible for all the tested antibodies? I suggest to have at least two operators and calculate a correlation coefficient for the results, at least for less commonly used proteins.

3) Results - Clinical and pathological characteristics: how were the mENSAT-GRAS criteria calculated? I understand what are the variables included but it’s unclear how they were combined. Again, as this is an essential part of the analysis, it is important to have a detailed description.

Moreover, I understand that all characteristics included in the mENSAT-GRAS classification are those collected at baseline (time of diagnosis). Is this correct? Or have been some variables re-evaluated over the time (e.g. age, symptoms)? If so, this needs to be specified.

mENSAT-GRAS classification is then used as prognostic marker for the survival from the metastatic diagnosis. It is not entirely clear how the authors considered the impact of the time between the initial diagnosis and the occurrence of metastasis. Also, considering that all patients underwent pharmacological treatment to advanced disease (mitotane alone or combined with platinum-based chemotherapy), wouldn’t it be better to consider also the analysis for time-to-progression?

4) Results – expression profile: were there any differences between primary and metastatic tumour samples (55 vs 11)? Considering that this is a quite small and heterogeneous cohort, it is important to have this comparison.

5) Results – survival curves: it would be good to have also survival curves for the mENSAT classification and the GRAS parameters, individually and combined as then used for the multivariable analysis. This would be helpful to visualise differences.

6) Results – survival analysis: it would be good to have also survival curves of univariable analysis for the mENSAT classification and the GRAS parameters, individually and combined as then used for the multivariable analysis (Figure 3). This would be helpful to visualise differences.

Moreover, the authors say that mitotane duration and plasma levels were predictive of best response. Were these variables also associated with survival (overall or time-to-progression)? This should be specified. If yes, in my opinion, data regarding the mitotane treatment should be included in the multivariable analysis (Table 4).

7) Discussion: a more precise take-home message and potential suggested recommendation need to be added.

 

Minor points

1) the column regarding the “relevant pattern” in table 2 and table 3 is exactly the same (and therefore redundant).

2) in the description of the statistical analysis, tests used to evaluate the response to therapy  are missing.

3) Discussion: the authors mention in line 263-266 that two cases were available with matched primary-metsastasis samples. It would be good to have thid sinfromation reported in the results, with additional details.

Author Response

Please see the attachment

Author Response File: Author Response.docx

Round 2

Reviewer 1 Report

Hi thanks the authors for taking all of my comments into account. I feel the paper is now ready for publication.

Reviewer 3 Report

The authors replied to all my comments..

The revised version of the manuscript significantly improved and I have no further concerns.