Advances and Challenges in Pharmacokinetic Modeling for PET Imaging: Compartment Models, Input Functions, and Quantitative Techniques
, Meltem Uyanik 1, Xue Li 1, Weijie Chen 1, Zhijin He 1, Caitlin Randell 1 and Alan McMillan 1,2
Round 1
Reviewer 1 Report
Comments and Suggestions for AuthorsJournal: Tomography
Manuscript Title: Advancements and Challenges in Pharmacokinetic Modeling for PET Imaging: Input Functions, Partial Volume Corrections, and Quantitative Techniques
Authors: J Wang, M Uyanik, X Li, W Chen, Z He, C Randell, A McMillan
Manuscript ID: tomography-4211864
This review manuscript on recent developments in pharmacokinetic (PK) modeling for PET imaging offers a solid overview of the field, including compartmental modeling, input functions, quantitative analysis, and emerging AI applications. It is relevant to a nuclear medicine audience; however, several areas require further attention prior to consideration for publication:
Although partial volume correction (PVC) is identified as important, the manuscript does not provide a systematic discussion of PVC methods, their assumptions, or their effect on kinetic parameter accuracy. The current treatment of PVC is too brief to support the emphasis placed on it in the title and abstract.
The discussion of one-, two-, and three-compartment models is useful, but the manuscript would benefit from practical guidance on how to choose among them based on tracer characteristics, tissue behavior, and clinical context. At present, the models are described more as abstractions than as tools for decision-making.
The manuscript is also heavily methodological, with limited discussion of clinical oncology applications. Adding more concrete examples would improve its translational relevance and make the review more useful to readers working in PET oncology.
Equations 9, 10, and 11 appear to be presented as separate expressions despite being identical in form. In addition, several multi-compartment equations appear incomplete or insufficiently numbered, which should be corrected for clarity and consistency.
The AI and machine learning section is too brief and would benefit from a more balanced discussion. In particular, it should address limitations such as generalizability, training data requirements, interpretability, and validation against gold-standard methods. The statement that AI removes the need for iterative calculations may also overstate the practical advantages, since computational trade-offs still remain.
A comparative table summarizing input function methods such as AIF, IDIF, PBIF, and MBIF would strengthen the manuscript substantially. Such a table should contrast invasiveness, accuracy, technical complexity, and clinical feasibility to help readers quickly compare the available approaches.
Finally, the manuscript requires thorough proofreading to correct typographical, grammatical, and formatting errors.
Overall, the topic is strong and timely, but the review would be significantly improved by clearer structure, deeper critical analysis, and more practical clinical guidance.
Author Response
• Although partial volume correction (PVC) is identified as important, the manuscript does not provide a systematic discussion of PVC methods, their assumptions, or their effect on kinetic parameter accuracy. The current treatment of PVC is too brief to support the emphasis placed on it in the title and abstract.
Response: Thank you for bringing this to our attention. To address this issue, the title was adjusted to “Advancements and Challenges in Pharmacokinetic Modeling for PET Imaging: Compartment Models, Input Functions, and Quantitative Techniques” to better reflect the content presented in this paper.
• The discussion of one-, two-, and three-compartment models is useful, but the manuscript would benefit from practical guidance on how to choose among them based on tracer characteristics, tissue behavior, and clinical context. At present, the models are described more as abstractions than as tools for decision-making.
Response: Thank you for this suggestion and we agree that this would be very helpful for the reader. So we added an additional paragraph at the end of the 2.1 classical compartment modeling section to help provide a bit more guidance via an example to help readers choose their model.
• The manuscript is also heavily methodological, with limited discussion of clinical oncology applications. Adding more concrete examples would improve its translational relevance and make the review more useful to readers working in PET oncology.
Response: We agree that this change would improve the impact and usefulness with those working in PET oncology. So, we added more examples throughout the paper.
• Equations 9, 10, and 11 appear to be presented as separate expressions despite being identical in form. In addition, several multi-compartment equations appear incomplete or insufficiently numbered, which should be corrected for clarity and consistency.
Response: Thank you for pointing this out, the equations 10 and 11 have been updated to the actual equation intended to be shown. In addition, the 1TCM equation was pulled out of “in-text” and listed out as a separate equation to help with clarity and consistency.
• The AI and machine learning section is too brief and would benefit from a more balanced discussion. In particular, it should address limitations such as generalizability, training data requirements, interpretability, and validation against gold-standard methods. The statement that AI removes the need for iterative calculations may also overstate the practical advantages, since computational trade-offs still remain.
Response: Thank you for bringing this to our attention. We adjusted the AI/ML section to incorporate more fundamentals and addressing limitations and requirements that standard AI/ML would demand and bringing these constraints within the contexts of PK modeling.
• A comparative table summarizing input function methods such as AIF, IDIF, PBIF, and MBIF would strengthen the manuscript substantially. Such a table should contrast invasiveness, accuracy, technical complexity, and clinical feasibility to help readers quickly compare the available approaches.
Response: We agree that the inclusion of this table will strengthen the manuscript. As such, we added a table 1 at the end of section 3 effectively serving as a summary and comparison of the input functions
• Finally, the manuscript requires thorough proofreading to correct typographical, grammatical, and formatting errors.
Response: Thank you for providing such a thorough review. Grammar and text have been modified throughout the text to correct for these errors.
Reviewer 2 Report
Comments and Suggestions for Authors1- The title is confusing, and I do not see a clear link between pharmacokinetics and OET imaging. Please revise
2- shorten the paragraph and provide a clear basis for the association between your topic of imaging and pharmacokinetic modeling
3- You need to add a paragraph in the introduction section linking the dots. How are pharmacokinetic modeling and PET imaging related? you need to simplify that for the reader
4- Section 2. Foundation of PK modeling in PET, which part of ADME is affected most? what are the applications? are there specific drugs or drug classes have more applications in PET imaging versus others? elaborate
5- In pharmacokinetic analysis, are there specific drug analyzing devices, such as HPLC are used? Or what other possible devices can be used in pharmaceutical analysis?
6- Can some of the concepts and formulas be simplified in figure or illustration?
Author Response
1- The title is confusing, and I do not see a clear link between pharmacokinetics and PET imaging. Please revise
Response: Thank you for pointing this out. The title was revised to more accurately reflect the content in the review paper: "Advancements and Challenges in Pharmacokinetic Modeling for PET Imaging: Compartment Models, Input Functions, and Quantitative Techniques"
2- shorten the paragraph and provide a clear basis for the association between your topic of imaging and pharmacokinetic modeling
Response: I agree that the association with PET imaging and PK modeling needs to be clearer. The second paragraph of the introduction was modified and shortened to be clearer to bridge these two topics
3- You need to add a paragraph in the introduction section linking the dots. How are pharmacokinetic modeling and PET imaging related? you need to simplify that for the reader
Response: I agree with this point, and so this was addressed in tandem with comment #2. The introduction was revised to be simpler and better “link the dots” between PK modeling and PET, emphasizing that analysis is driven by the need for tracking tracer kinetics.
4- Section 2. Foundation of PK modeling in PET, which part of ADME is affected most? what are the applications? are there specific drugs or drug classes have more applications in PET imaging versus others? Elaborate
Response: To clarify the content of this review paper, the wording at the beginning of section 2 and throughout the paper was modified to emphasize PK modeling with PET to be primarily driven by radiotracer kinetics. As such the drug class used is “radiopharmaceuticals” and the impact on ADME is elaborated upon in section 2.1.
5- In pharmacokinetic analysis, are there specific drug analyzing devices, such as HPLC are used? Or what other possible devices can be used in pharmaceutical analysis?
Response: Thank you for bringing this to our attention. We included a short section in 3.1 on a couple devices that could be used for pharmaceutical analysis – specifically for AIF determination.
6- Can some of the concepts and formulas be simplified in figure or illustration?
Response: Thank you for pointing this out. While equations 1-9 may appear to be complex, they are represented quite simply through the schematic illustration provided in Figure 2.
Round 2
Reviewer 2 Report
Comments and Suggestions for Authorsnone
