Development of a Semi-Mechanistic Modeling Framework for Wet Bead Milling of Pharmaceutical Nanosuspensions

Round 1

Reviewer 1 Report

Objective of the current research work is novel. 

Reference cited are not form latest articles, so need to revise those.

Introduction part needs to be revise, currently flow of the introduction part is not reader friendly. Please revise it in concise form.

Various models belonging to the mechanistic–phenomenological–empirical (statistical) model para needs to revise.  

Describe details with technical justification for current research work like selected models. Needs to revise last para by incorporating technical rationale or aim of study, scope of the work, test performed and expected outcome of the study. 

Please verify 8 different drug products or 6?

Bead loading for DV50 is 75, 85 or 75-85? Verify.

Mention the difference between equipment selected to manufacture drug product. What was the batch size for drug products?

Please mention various milling times and also mention time of samples taken form mill outlet.

All tables needs to verify for column and raw data. 

Minor editing required

Author Response

Reviewer 1

Objective of the current research work is novel.

Response:  The authors thank the reviewer for the feedback.   The manuscript has been updated based on the reviewer’s comments and the updated sections are highlighted in yellow for the ease of review. 

1. Reference cited are not form latest articles, so need to revise those.

Response: The manuscript has been updated to include additional references to better reflect the scientific literature on this topic.   

2. Introduction part needs to be revise, currently flow of the introduction part is not reader friendly. Please revise it in concise form.

Response: The introduction section of the manuscript has been revised to make it more concise, enhance the flow, and facilitate the readability.  

3. Various models belonging to the mechanistic–phenomenological–empirical (statistical) model para needs to revise.

Response: This relevant paragraph in the Introduction section has been revised accordingly.  

4. Describe details with technical justification for current research work like selected models. Needs to revise last para by incorporating technical rationale or aim of study, scope of the work, test performed and expected outcome of the study.

Response:  A paragraph has been added in the Introduction section to clearly state the aim of the study, the scope, and the targeted outcome. 

5. Please verify 8 different drug products or 6?

Response: In the manuscript, we used eight different drug products. In addition to the generic drugs griseofulvin (GF) and fenofibrate (FNB), there are six innovator drug products included in this study, referred as DP numbers from one to six.

6. Bead loading for DV50 is 75, 85 or 75-85? Verify.

Response:  We want to clarify that two levels of bead loadings were used ( 75% and 85%).   Therefore, in Table 1, we have summarized them as 75% and 85%.

7. Mention the difference between equipment selected to manufacture drug product. What was the batch size for drug products?

Response: We added a row in Table 1 that provides the ranges of batch volumes used for each manufacturing scale.

8. Please mention various milling times and also mention time of samples taken form mill outlet.

Response:  We added a row in Table 1 that provides the ranges of milling times used for each scale. We also added a sentence in the methods section about sampling intervals, indicating that it varied for each experiment, depending on the total milling time.

9. All tables needs to verify for column and raw data.

Response:  The authors thank the reviewer for this feedback.  All raw data have been reviewed and verified.

Reviewer 2 Report

The manuscript entitled "Development of a Semi-Mechanistic Modeling Framework for Wet Bead Milling of Pharmaceutical Nanosuspensions" evaluates the development of a practical semi-mechanistic modeling framework to predict particle size evolution during wet bead milling of pharmaceutical nanosuspensions over a wide range of process conditions and milling scales. Three models were developed (A, B and C), out of which, Model C was the simplest of all, showing good prediction performance.

Overall, the authors presented here a nice piece of work. The applied experimental method was appropriate, and the obtained results were of high quality. Few minor comments before accepting the manuscript for publications are given below:

1. Please reduce the size of the introduction.

2. Give more details on the milling beads used (i.e. type of material etc.).

3. Table 3 – 1^st column is not clear was it means (section?)

4. Figure 3-7 and 8-10, please improve resolution.

Author Response

Reviewer 2

The manuscript entitled "Development of a Semi-Mechanistic Modeling Framework for Wet Bead Milling of Pharmaceutical Nanosuspensions" evaluates the development of a practical semi-mechanistic modeling framework to predict particle size evolution during wet bead milling of pharmaceutical nanosuspensions over a wide range of process conditions and milling scales. Three models were developed (A, B and C), out of which, Model C was the simplest of all, showing good prediction performance.

Overall, the authors presented here a nice piece of work. The applied experimental method was appropriate, and the obtained results were of high quality. Few minor comments before accepting the manuscript for publications are given below:

Response:  The authors thank the reviewer for the kind words and the feedback.   The manuscript has been updated based on the reviewer’s comments and the updated sections are highlighted in yellow for the ease of review. 

1. Please reduce the size of the introduction.

Response:  The introduction section of the manuscript has been revised to make it more concise, enhance the flow, and facilitate the readability.   We have edited the content further to reduce the size of this section.

2. Give more details on the milling beads used (i.e. type of material etc.).

Response:  The details for the milling beads have been covered in section 4.1.3 of the manuscript.  In addition, the Materials section has been updated to include the details of the beads used in this study.

3. Table 3 – 1st column is not clear was it means (section?)

Response:  The 1^st column in the Table 3 of the original manuscript was intended to cross-refer to the corresponding sections or figures of the relevant studies. To better convey the information, we have merged the 1^st and 2^nd columns in the revised manuscript for the each of cross-reference and readability.

4. Figure 3-7 and 8-10, please improve resolution.

Response:   We thank the reviewer for the feedback.  We have verified the quality and the resolution of all Figures in the manuscript and can confirm that the current versions of the figures appropriately reflect the desired image quality in the pdf format for the manuscript.  

Reviewer 3 Report

It is an interesting study to investigate the process understanding in wet bead milling for process scale up and operation, in particular, the development of practical modeling framework for its QbD application in pharmaceutical industry. My comments below:

(1) Overall, the structure of the manuscript is well written. Possible improvement would be to add some brief and key definitions in Appendix A, B, C to the place it was referenced in the manuscript. For example, I was very interested in N2 to N4 parameters, especially N2 as it was related to the milling rate or relative milling rate on page 13, while its detail definition was provided in Appendix C at the end of the manuscript. 

(2) Additionally, in the nomenclature, the symbol A was used multiple times, only A was listed for the slope term. There are Aj*, Aj**, Aj; same for N, Nj was not listed. The list would be a very quick guide while reading this long manuscript full of variables and equations. Could you improve the list?

(3) On page 11, line 453, could the authors also identify the scale of each milling equipment in Table 1, lab, pilot, or commercial?

(4) On page 17, line 662, if Ostwald ripening is one of the key factors, I am curious why only the large particle is strongly impacted? The fine particle could dissolve and decrease in size. And how long of the milling process and how fast the sample was analyzed for PSD?

(5) Last, but not the least, the fitting of the model tells one story, while the model validation/application would be another? Could the authors provide some comments or examples of the model C validation/performance for unseen experimental data?    

Author Response

It is an interesting study to investigate the process understanding in wet bead milling for process scale up and operation, in particular, the development of practical modeling framework for its QbD application in pharmaceutical industry. My comments below:

Response:  The authors thank the reviewer for the feedback.   The manuscript has been updated based on the reviewer’s comments and the updated sections are highlighted in yellow for the ease of review. 

(1) Overall, the structure of the manuscript is well written. Possible improvement would be to add some brief and key definitions in Appendix A, B, C to the place it was referenced in the manuscript. For example, I was very interested in N2 to N4 parameters, especially N2 as it was related to the milling rate or relative milling rate on page 13, while its detail definition was provided in Appendix C at the end of the manuscript.

Response:   The authors thank the reviewer for the suggested improvements.  Additional cross-references have been incorporated in the revised manuscript accordingly.

(2) Additionally, in the nomenclature, the symbol A was used multiple times, only A was listed for the slope term. There are Aj*, Aj**, Aj; same for N, Nj was not listed. The list would be a very quick guide while reading this long manuscript full of variables and equations. Could you improve the list?

Response:   The authors thank the reviewer for the feedback.  The authors would like to clarify that j has been defined in the “Indices” section of the Nomenclature as the index for particle size quantile (10, 50, 90). * and ** have been defined in the “Indices” section as intermediately derived constants.  For the convenience of the readers, we have updated the “Symbols used” section with A, B, N, x symbols with index j.  We also added Aj* and Aj** in the list as intermediately derived slope terms of Yt function.

(3) On page 11, line 453, could the authors also identify the scale of each milling equipment in Table 1, lab, pilot, or commercial?

Response:   Typical batch volumes and processing times were added to Table 1 to give the reader a sense of scale for each mill.    

(4) On page 17, line 662, if Ostwald ripening is one of the key factors, I am curious why only the large particle is strongly impacted? The fine particle could dissolve and decrease in size. And how long of the milling process and how fast the sample was analyzed for PSD?

Response:   The authors thank the reviewer for the insightful feedback and are aligned with the reviewer’s comment.  From a product performance perspective, the largest size fractions have the greatest anticipated impact on differentiated dissolution behavior, and these fractions tend to preferentially grow.  Therefore, that aspect has been of our primary consideration during product design.   However, we acknowledge that Ostwald ripening is a product specific phenomenon dependent on material attributes.  No update was made to this section of the manuscript to avoid confusion. 

PSD analysis of all samples was performed immediately after sampling from the mill.      

(5) Last, but not the least, the fitting of the model tells one story, while the model validation/application would be another? Could the authors provide some comments or examples of the model C validation/performance for unseen experimental data?   

Response:

The authors thank the reviewer for the insightful feedback.  The authors confirm that prospective model use / validation has been performed successfully three times.  In the interest of the brevity of presentation, we did not cover this topic in the original manuscript.   The model was mostly built on the first three drug products (DP1-3), and then successfully used for a priori prediction of scaleup for drug products DP4-6.   To specifically address this point, the manuscript has been updated to include a new section 5 to provide an illustrative example for DP6 showing prior usage of DV50 or DV150 data to predict the outcome of DV300 scaleup batch.