From the journal Digital Discovery Peer review history
Machine learning approaches to the prediction of powder flow behaviour of pharmaceutical materials from physical properties
Round 1
Manuscript submitted on 04 okt 2022![](https://www.rsc-cdn.org/pubs-core/2022.0.146/content/NewImages/icon-letter.png)
![](https://www.rsc-cdn.org/pubs-core/2022.0.146/content/NewImages/chevron-down.png)
25-Dec-2022
Dear Dr Florence:
Manuscript ID: DD-ART-10-2022-000106
TITLE: Machine Learning Approaches to the Prediction of Powder Flow Behaviour of Pharmaceutical Materials from Physical Properties
Thank you for your submission to Digital Discovery, published by the Royal Society of Chemistry. I sent your manuscript to reviewers and I have now received their reports which are copied below.
I have carefully evaluated your manuscript and the reviewers’ reports, and the reports indicate that major revisions are necessary. Especially the second reviewer feels that while you are aware of a paper recently published by industry in this space, it would be wise to take on board the implications of that work.
Please submit a revised manuscript which addresses all of the reviewers’ comments. Further peer review of your revised manuscript may be needed. When you submit your revised manuscript please include a point by point response to the reviewers’ comments and highlight the changes you have made. Full details of the files you need to submit are listed at the end of this email.
Digital Discovery strongly encourages authors of research articles to include an ‘Author contributions’ section in their manuscript, for publication in the final article. This should appear immediately above the ‘Conflict of interest’ and ‘Acknowledgement’ sections. I strongly recommend you use CRediT (the Contributor Roles Taxonomy, https://credit.niso.org/) for standardised contribution descriptions. All authors should have agreed to their individual contributions ahead of submission and these should accurately reflect contributions to the work. Please refer to our general author guidelines https://www.rsc.org/journals-books-databases/author-and-reviewer-hub/authors-information/responsibilities/ for more information.
Please submit your revised manuscript as soon as possible using this link:
*** PLEASE NOTE: This is a two-step process. After clicking on the link, you will be directed to a webpage to confirm. ***
https://mc.manuscriptcentral.com/dd?link_removed
(This link goes straight to your account, without the need to log on to the system. For your account security you should not share this link with others.)
Alternatively, you can login to your account (https://mc.manuscriptcentral.com/dd) where you will need your case-sensitive USER ID and password.
You should submit your revised manuscript as soon as possible; please note you will receive a series of automatic reminders. If your revisions will take a significant length of time, please contact me. If I do not hear from you, I may withdraw your manuscript from consideration and you will have to resubmit. Any resubmission will receive a new submission date.
The Royal Society of Chemistry requires all submitting authors to provide their ORCID iD when they submit a revised manuscript. This is quick and easy to do as part of the revised manuscript submission process. We will publish this information with the article, and you may choose to have your ORCID record updated automatically with details of the publication.
Please also encourage your co-authors to sign up for their own ORCID account and associate it with their account on our manuscript submission system. For further information see: https://www.rsc.org/journals-books-databases/journal-authors-reviewers/processes-policies/#attribution-id
I look forward to receiving your revised manuscript.
Yours sincerely,
Dr Kedar Hippalgaonkar
Associate Editor, Digital Discovery
Royal Society of Chemistry
************
![](https://www.rsc-cdn.org/pubs-core/2022.0.146/content/NewImages/icon-comments.png)
![](https://www.rsc-cdn.org/pubs-core/2022.0.146/content/NewImages/chevron-down.png)
Reviewer 1
Interesting research using a variety of data analysis techniques on a broad data set of standard powder flow properties. Perhaps 'data' and 'validation' CSV files called in the scripts may be included for easy code verification. All variables used in the scripts can be traced back to the XLSX database, except for variable '2Classes' whose meaning is unclear.
Reviewer 2
The authors present some interesting new information on the prediction of flow from particle size and shape information, which is an advancing field which has gained more prominence due to increasing utilisation of continuous direct compression as a mechanism for prosecuting dosage. The study is well designed and reported.
The following suggestions may help the future reader understand the work and its context more clearly:
1. The initial introduction seems to confuse and conflate wet granulation and roller compaction, which are both widely used techniques in the pharmaceutical industry. However only one of them involves heating and wetting, and has to be avoided for materials which cannot cope with these. Roller compaction is perfectly suitable for such materials, although it comes with its own liabilities. There is also emerging evidence that, in many cases, RC does not bind the majority of the drug in the granule. The intro does not distinguish between batch direct compression and continuous direct compression, which have different requirements for physical performance of the asset;
2. It is noted that PSD is thought to have an influence on particle size but it is difficult to relate that directly to flow in a model. Some of the papers quoted for that observation are now many years old and significant advances have been made since then. The authors should check whether Leane et al recommended 1000μm as a D90, as quoted.
3. The article notes the use of the Sympatec device as the mechanism for measuring particle size and, indeed, shape. There should be a discussion of the potential limitations of this device, particularly with regard to shape. When the particle size is small the detection CCD can measure the presence of a particle, and get an estimate of size, depending on how many pixels are occupied by the particle. However this limitation means that, for fine particles, the shape information is less sensitive than the size. If a particle takes up one pixel you can make an assumption about the size, but none about the shape. Some people have estimated that a particle must occupy a 10x10 matrix (100 pixels) to make a reliable measure of shape. It is possible (e.g. in figure 2 and with some models) that the system is overestimating how many of the particles are spherical based on the occupation of a small number of pixels, and assuming that the particles must be round, as it must with limited information. This shortcoming should be addressed. It might also be interesting to see if the system estimates the flow of larger particles much better than it does with small ones;
4. Table 5: the authors should confirm that the surface energy results are reasonable with respect to literature values. To this reviewer, who has used the technique, the numbers (e.g. 2.94 to 16.81mJ/m2) are not within the regular range of reported values. This technique is remarkably prone to artefacts, even in the hands of an experienced operator;
5. Notwithstanding the above the authors note that it was “unexpected” that surface energy and surface area did not contribute understanding to a model of flow. However they then quote another work which found exactly the same thing. The papers that proposed that there was a link between surface energy and flow were much older, before the widespread adoption of improved particle sizing techniques. The authors could propose a rationale for why, when improved particle size and shape information is available, surface area and energy appear to have a less useful influence on models.
6. The authors describe the particles as “tetrahedral” or “hexahedral”. The processing of organic solids mean that there is never a single shape for a particle, or group of particles, and each is generally unique, and that there are shape distributions (e.g. fine particles can be rounder due to being knocked off of larger particles, and larger particles can be rod shaped etc). A simple description of particles is often not helpful;
7. The conclusions could note that, although they use more sophisticated descriptors of particle size distribution, and a single number for shape (even though shape, as with size, is a distribution), even more sophisticated analysis (whole distribution of data analysis, for both size and shape) may allow a better understanding of the data and for more predictive models to be built;
8. The written language is generally good but occasionally the wrong term slips in. In places measurements are made in “triplicates”, which is generally not used, in others they are made in “triplicate”, which is less so;
9. I don’t know if this would be fixed by copy writers on publication but the authors do not standardise their Endnote references. The International Journal of Pharmaceutics, for instance, is designated in at least 5 different ways in various references, and other journals are similarly confused. This makes it annoying to check, at least for this reviewer.
Reviewer 3
This manuscript addresses a hot topic in pharmaceutical production with a novel, multidisciplinary approach to powder flow that can be used in a wide variety of materials. A good manuscript, well structured, clear in its content and with a remarkable effort behind it. However, it has to be emphasized that the focus on shear flow properties is limited to certain production processes (static flow) and is quite ineffective in the vast majority of dynamic processes such as die filling, convection processes, etc. Under static flow conditions compressive forces and shear stress acquire remarkable importance compared to dynamic flow conditions where cohesive phenomena acquire a different behaviour (e.g. attractive forces become more relevant). This is to be defined in the introduction where the application of this model in die filling where dynamic powder flow predominates is discussed.
![](https://www.rsc-cdn.org/pubs-core/2022.0.146/content/NewImages/icon-response.png)
![](https://www.rsc-cdn.org/pubs-core/2022.0.146/content/NewImages/chevron-down.png)
Reviewer 1: Interesting research using a variety of data analysis techniques on a broad data set of standard powder flow properties. Perhaps 'data' and 'validation' CSV files called in the scripts may be included for easy code verification. All variables used in the scripts can be traced back to the XLSX database, except for variable '2Classes' whose meaning is unclear.
Thank you for the comment. The “2 classes” variable refers to the dependent variable of the two-step model, dividing between free-flowing and non-free-flowing in the first step, and between cohesive and non-cohesive in the second step. The files called “data” and “validation” were put together into the file that was submitted with the manuscript.
Reviewer 2:
1. The initial introduction seems to confuse and conflate wet granulation and roller compaction, which are both widely used techniques in the pharmaceutical industry. However only one of them involves heating and wetting, and has to be avoided for materials which cannot cope with these. Roller compaction is perfectly suitable for such materials, although it comes with its own liabilities. There is also emerging evidence that, in many cases, RC does not bind the majority of the drug in the granule. The intro does not distinguish between batch direct compression and continuous direct compression, which have different requirements for physical performance of the asset;
Thank you very much for this comment. Changes have been made in the introduction to clarify the differences between WG and RC.
2. It is noted that PSD is thought to have an influence on particle size but it is difficult to relate that directly to flow in a model. Some of the papers quoted for that observation are now many years old and significant advances have been made since then. The authors should check whether Leane et al recommended 1000μm as a D90, as quoted.
The authors appreciate this comment. The papers in the first of the part of the paragraph are indeed many years old, but some other papers quoted towards the end of that same paragraph were have published more recently, i.e., Leane M, et al (2018). The PSD D90 value recommended by Leane et al. has been double checked, and therefore the text in the manuscript has been modified to clarify that the size proposed was PSD D90 smaller than 1000 µm.
3. The article notes the use of the Sympatec device as the mechanism for measuring particle size and, indeed, shape. There should be a discussion of the potential limitations of this device, particularly with regard to shape. When the particle size is small the detection CCD can measure the presence of a particle, and get an estimate of size, depending on how many pixels are occupied by the particle. However this limitation means that, for fine particles, the shape information is less sensitive than the size. If a particle takes up one pixel you can make an assumption about the size, but none about the shape. Some people have estimated that a particle must occupy a 10x10 matrix (100 pixels) to make a reliable measure of shape. It is possible (e.g. in figure 2 and with some models) that the system is overestimating how many of the particles are spherical based on the occupation of a small number of pixels, and assuming that the particles must be round, as it must with limited information. This shortcoming should be addressed. It might also be interesting to see if the system estimates the flow of larger particles much better than it does with small ones;
Thank you very much for this comment. The limitations of the Sympatec have been added to the methods section in the manuscript following the recommendations from this comment. Indeed, the differences in the accuracy of the prediction of flow depends on the size of the particles. In Fig 9 of the manuscript we present the combined confusion matrix of Step 1 and Step 2 of the two-step binary classification model. This confusion matrix reveals tha the model is better at predicting the powder flow of "free-flowing" materials than "cohesive" materials. We believe that this corresponds to the effect of the analytical error of the powder flow rheometer, i.e., the more free-flowing the material, the greater the error of the analytical measurement.
4. Table 5: the authors should confirm that the surface energy results are reasonable with respect to literature values. To this reviewer, who has used the technique, the numbers (e.g. 2.94 to 16.81mJ/m2) are not within the regular range of reported values. This technique is remarkably prone to artefacts, even in the hands of an experienced operator;
This comment is very much appreciated by the authors. The data presented in the manuscript was the specific surface energy (acid-base) experimentally obtained. The specific surface energy values are usually lower than the dispersive surface energy or the total surface energy. However, further investigation will be required to properly evaluate why the values obtained experimentally were not within the regular range of reported values.
5. Notwithstanding the above the authors note that it was “unexpected” that surface energy and surface area did not contribute understanding to a model of flow. However they then quote another work which found exactly the same thing. The papers that proposed that there was a link between surface energy and flow were much older, before the widespread adoption of improved particle sizing techniques. The authors could propose a rationale for why, when improved particle size and shape information is available, surface area and energy appear to have a less useful influence on models.
Thank you for this comment. The impact of surface area and surface energy on powder flow has been discussed in previous publications. However, Barjat et al. showed that the addition of surface area and surface energy data to their models did not improve the performance of the predictive models used to predict powder flow. Therefore, a similar result was expected in this work. The manuscript has been changed according to the recommendations made by the reviewer.
6. The authors describe the particles as “tetrahedral” or “hexahedral”. The processing of organic solids mean that there is never a single shape for a particle, or group of particles, and each is generally unique, and that there are shape distributions (e.g. fine particles can be rounder due to being knocked off of larger particles, and larger particles can be rod shaped etc). A simple description of particles is often not helpful;
Thank you for this comment. The manuscript has been changed according to the recommendations of the reviewer and hence, particle shape is no longer described as ”hexahedral” or ”tetrahedral”.
7. The conclusions could note that, although they use more sophisticated descriptors of particle size distribution, and a single number for shape (even though shape, as with size, is a distribution), even more sophisticated analysis (whole distribution of data analysis, for both size and shape) may allow a better understanding of the data and for more predictive models to be built;
Thank you for this comment. The conclusions have been expanded adding that the consideration of the whole particle size and particle shape distribution may allow a better understanding of the data and of the relationship between particle size and shape and powder flow, and hence, and improvement in the performance of the powder flow model would be expected.
8. The written language is generally good but occasionally the wrong term slips in. In places measurements are made in “triplicates”, which is generally not used, in others they are made in “triplicate”, which is less so;
Thank you for the comment. The use of these terms has been changed and now the same term is used consistently throughout the paper.
9. I don’t know if this would be fixed by copy writers on publication but the authors do not standardise their Endnote references. The International Journal of Pharmaceutics, for instance, is designated in at least 5 different ways in various references, and other journals are similarly confused. This makes it annoying to check, at least for this reviewer.
Thank you for the comment. The reference style has been updated according to the EndNote RSC Digital Discovery reference style.
Reviewer 3: This manuscript addresses a hot topic in pharmaceutical production with a novel, multidisciplinary approach to powder flow that can be used in a wide variety of materials. A good manuscript, well structured, clear in its content and with a remarkable effort behind it. However, it has to be emphasized that the focus on shear flow properties is limited to certain production processes (static flow) and is quite ineffective in the vast majority of dynamic processes such as die filling, convection processes, etc. Under static flow conditions compressive forces and shear stress acquire remarkable importance compared to dynamic flow conditions where cohesive phenomena acquire a different behaviour (e.g. attractive forces become more relevant). This is to be defined in the introduction where the application of this model in die filling where dynamic powder flow predominates is discussed.
Thank you very much for this comment. The intended application of this model for die filling has been added to the last paragraph of the introduction. Furthermore, some additional information has been added to the methods section “Powder Flow and bulk density – Powder Rheometer FT4” so highlight the value of this instrument in industrial settings due to its ability to assess dynamic flow, bulk and shear properties.
Round 2
Revised manuscript submitted on 25 feb 2023![](https://www.rsc-cdn.org/pubs-core/2022.0.146/content/NewImages/icon-letter.png)
![](https://www.rsc-cdn.org/pubs-core/2022.0.146/content/NewImages/chevron-down.png)
20-Mar-2023
Dear Dr Florence:
Manuscript ID: DD-ART-10-2022-000106.R1
TITLE: Machine Learning Approaches to the Prediction of Powder Flow Behaviour of Pharmaceutical Materials from Physical Properties
Thank you for submitting your revised manuscript to Digital Discovery. I am pleased to accept your manuscript for publication in its current form.
You will shortly receive a separate email from us requesting you to submit a licence to publish for your article, so that we can proceed with the preparation and publication of your manuscript.
You can highlight your article and the work of your group on the back cover of Digital Discovery. If you are interested in this opportunity please contact the editorial office for more information.
Promote your research, accelerate its impact – find out more about our article promotion services here: https://rsc.li/promoteyourresearch.
If you would like us to promote your article on our Twitter account @digital_rsc please fill out this form: https://form.jotform.com/213544038469056.
By publishing your article in Digital Discovery, you are supporting the Royal Society of Chemistry to help the chemical science community make the world a better place.
With best wishes,
Dr Kedar Hippalgaonkar
Associate Editor, Digital Discovery
Royal Society of Chemistry
******
******
Please contact the journal at digitaldiscovery@rsc.org
************************************
DISCLAIMER:
This communication is from The Royal Society of Chemistry, a company incorporated in England by Royal Charter (registered number RC000524) and a charity registered in England and Wales (charity number 207890). Registered office: Burlington House, Piccadilly, London W1J 0BA. Telephone: +44 (0) 20 7437 8656.
The content of this communication (including any attachments) is confidential, and may be privileged or contain copyright material. It may not be relied upon or disclosed to any person other than the intended recipient(s) without the consent of The Royal Society of Chemistry. If you are not the intended recipient(s), please (1) notify us immediately by replying to this email, (2) delete all copies from your system, and (3) note that disclosure, distribution, copying or use of this communication is strictly prohibited.
Any advice given by The Royal Society of Chemistry has been carefully formulated but is based on the information available to it. The Royal Society of Chemistry cannot be held responsible for accuracy or completeness of this communication or any attachment. Any views or opinions presented in this email are solely those of the author and do not represent those of The Royal Society of Chemistry. The views expressed in this communication are personal to the sender and unless specifically stated, this e-mail does not constitute any part of an offer or contract. The Royal Society of Chemistry shall not be liable for any resulting damage or loss as a result of the use of this email and/or attachments, or for the consequences of any actions taken on the basis of the information provided. The Royal Society of Chemistry does not warrant that its emails or attachments are Virus-free; The Royal Society of Chemistry has taken reasonable precautions to ensure that no viruses are contained in this email, but does not accept any responsibility once this email has been transmitted. Please rely on your own screening of electronic communication.
More information on The Royal Society of Chemistry can be found on our website: www.rsc.org
Transparent peer review
To support increased transparency, we offer authors the option to publish the peer review history alongside their article. Reviewers are anonymous unless they choose to sign their report.
We are currently unable to show comments or responses that were provided as attachments. If the peer review history indicates that attachments are available, or if you find there is review content missing, you can request the full review record from our Publishing customer services team at RSC1@rsc.org.