Digital Immunophenotyping Predicts Disease Free and Overall Survival in Early Stage Melanoma Patients
, Francesca Galli 2,†, Romina Nassini 1, Filippo Ugolini 3, Sara Simi 3, Mara Cossa 4, Clelia Miracco 5, Andrea Gianatti 6, Vincenzo De Giorgi 7, Eliana Rulli 2, Antonio Cossu 8, Daniela Massi 3,*,‡ and Mario Mandalà 9,10,*,‡Round 1
Reviewer 1 Report
The authors suggest that study results may have a potential influence on treatment: "Since patients with absent/low intratumoral infiltration of CD8+ T-cell have a statistically significant shorter DFS and OS, they should deserve further treatments to reduce the recurrence rate and ultimately progression of melanoma." According to the study's limitations (mentioned by authors) and TILs/TAMs assessment methodology, it is a too strong conclusion. The implementation of image analysis for TILs/TAMs evaluation in routine diagnostics may be non-productive and time-consuming; on the contrary, counting TILs/TAMs cells per mm3 in "traditional" microscopy seems to be even more difficult and inconsistent. What are the cut-offs for high and low densities according to the main analysis? How was it counted? Could you please include that in methods. What are the practical tips of that study for pathologists?
Author Response
Reviewer 1
- “The authors suggest that study results may have a potential influence on treatment: Since patients with absent/low intratumoral infiltration of CD8+ T-cell have a statistically significant shorter DFS and OS, they should deserve further treatments to reduce the recurrence rate and ultimately progression of melanoma. According to the study's limitations (mentioned by authors) and TILs/TAMs assessment methodology, it is a too strong conclusion.
We thank the Reviewer for her/his comment. The statement has been mitigated in the Abstract as well as Discussion in light of the potential limitations of the study (page 2, line 3 from the bottom; page 11, line 8 from the bottom).
- The implementation of image analysis for TILs/TAMs evaluation in routine diagnostics may be non-productive and time-consuming; on the contrary, counting TILs/TAMs cells per mm3 in "traditional" microscopy seems to be even more difficult and inconsistent.”
We agree with the Reviewer that both image analysis and traditional microscopic assessment may have limitations. A new sentence on this regard has been added in the Discussion (page 11, line 12 from the bottom).
- “What are the cut-offs for high and low densities according to the main analysis? How was it counted? Could you please include that in methods.”
The cut-offs used to define high and low densities were the medians of the density of each biomarker calculated in the training cohort. As stated in the Statistical analysis paragraph, each immune cell biomarker was evaluated as a continuous variable and categorized as low or high according to its median and the same analysis carried out for the training cohort was adopted for the validation cohort using as biomarker cut-offs those calculated in the training cohort. In order to better specify this issue in Table 2, we modified the definition of “density according to main analysis median” with “density according to the median of the training cohort” (Table 2 from page 24).
- “What are the practical tips of that study for pathologists?”
While conventional H&E-based histopathological analysis remains the basic pathological standard approach, the message for pathologists is that computational pathology based on automated object detection and image registration using the HALO software can be applied for a more comprehensive spatial analysis of the tumor immune microenvironment. In the era of the digital transformation of the pathology workflow, computational pathology has the potential for discovering additional prognostic information in clinical research in the hope of supporting patients’ enrichment strategies and personalized treatments. Nevertheless, digital analysis cannot be considered a standard approach for the practicing pathologists and future prospective studies are needed to better understand the adding value of this technology compared to the standard evaluation of TME in early stages melanoma patients (page 11, line 12 from the bottom)
Reviewer 2 Report
The idea of paper is good, but the greatest back-draw of this paper is the obsolete training set (200-2015) and geolocation – Italy only; therefore the article should be resubmitted to Journal of Personalized Medicine (ISSN 2075-4426; CODEN: JPMOB3), which is more appropriate for this type of article (strictly addressed to Italians interindividual variations).
Abstract
DFS and OS: define them at their first appearance in text, respectively in the abstract
- INTRODUCTION
References 7-9 are quite old; can you search and add some newer references?
Please elaborate on the controversy from Ref. 7-9 and Ref. 10-11, add more references and data.
- MATERIALS AND METHODS
2.5. Image analysis
“Two experts pathologists...” – are they among the authors?
- Discussion
In the last paragraphs: the authors should emphasise the limitation of study at Italian patients and the importance for Personalised Medicine addressed to the Italian population.
Author Response
Reviewer 2
- “The idea of paper is good, but the greatest back-draw of this paper is the obsolete training set (200-2015) and geolocation – Italy only; therefore the article should be resubmitted to journal of personalized medicine(if 1.01) (issn 2075-4426; coden: jpmob3), which is more appropriate for this type of article (strictly addressed to Italians interindividual variations).”
We respectfully disagree with the Reviewer. The major scientific contributions to the study of TILs in melanoma derive from homogeneous cohorts of patients (e.g. Australia). In our opinion, the fact that our patients were observed and treated across Italy is a point of strength, as stated in the Discussion.
- “Abstract. DFS and OS: define them at their first appearance in text, respectively in the abstract”.
DFS and OS have been spelled out at their first appearance in the text (page 2, line 16 and 20).
- “Introduction. References 7-9 are quite old; can you search and add some newer references? please elaborate on the controversy from ref. 7-9 and ref. 10-11, add more references and data.
Following the Reviewer’s suggestion, more recent references have been added to the relevant old references already quoted in the text (new References: 10, 11, 12 and 13).
- “Materials and methods. 5. image analysis. “two experts pathologists...” – are they among the authors?
The two expert pathologists are included among the Authors and their initials are included in brackets in the text (page 6, line 11).
- “Discussion. in the last paragraphs: the authors should emphasise the limitation of study at italian patients and the importance for personalised medicine addressed to the Italian population”.
We respectfully disagree with the Reviewer. The fact that the data were collected from melanoma patients observed at Italian institutions is a point of strength, as stated in the Discussion (page 11, line 6).
Round 2
Reviewer 2 Report
Problem not solved:
“The idea of paper is good, but the greatest back-draw of this paper is the obsolete training set (200-2015) and geolocation – Italy only; therefore the article should be resubmitted to journal of personalized medicine(if 1.01) (issn 2075-4426; coden: jpmob3), which is more appropriate for this type of article (strictly addressed to Italians interindividual variations).”
-----------------------------
Problem not solved:
in the last paragraphs: the authors should emphasise the limitation of study at italian patients and the importance for personalised medicine addressed to the Italian population”.
