Logo: to the web site of Uppsala University

uu.sePublications from Uppsala University
Change search
CiteExportLink to record
Permanent link

Direct link
Cite
Citation style
  • apa
  • ieee
  • modern-language-association
  • vancouver
  • Other style
More styles
Language
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Other locale
More languages
Output format
  • html
  • text
  • asciidoc
  • rtf
Image analysis with a machine learning model (MLM) of semisolid extrusion (SSE) oral dosage forms on a tapering schedule
Uppsala University, Disciplinary Domain of Science and Technology, Technology, Department of Materials Science and Engineering, Nanotechnology and Functional Materials.
Uppsala University, Disciplinary Domain of Medicine and Pharmacy, Faculty of Pharmacy, Department of Pharmacy. Uppsala University, Disciplinary Domain of Medicine and Pharmacy, Faculty of Medicine, Department of Women's and Children's Health.ORCID iD: 0000-0002-2514-6618
Uppsala University, Disciplinary Domain of Science and Technology, Technology, Department of Materials Science and Engineering, Nanotechnology and Functional Materials.ORCID iD: 0000-0002-5496-9664
Division of Pharmaceutics, Utrecht Institute for Pharmaceutical Sciences, Utrecht University.
Show others and affiliations
(English)Manuscript (preprint) (Other academic)
Abstract [en]

Dose-adjustment is a very common practice in pharmacies, with accuracy particularly relevant for medications on a tapering schedule. Semisolid extrusion (SSE) has shown potential to provide a more accurate form of dose-adjustment than traditional methods (i.e. tablet splitting and liquid solutions). These methods, however, usually lack validation and quality assurance, as no validation is typically conducted post dose-adjustment. Machine learning (ML) based image analysis can provide this necessary validation. Fluoxetine-containing SSE tablets of three tapering steps were created to investigate with image analysis using a MLM. The tablets had a high degree of uniformity in mass and dimensional measurements. The average drug content of the SSE tablets was within ±3.5% of the desired dose for all tapering steps, showing higher accuracy than doses made via tablet splitting (±28.2% dose accuracy) and liquid solutions (±17.0% dose accuracy) produced by a licensed pharmacist. Utilizing a MLM for image analysis and images of tablets taken from the top and bottom of each tablet, the ability to categorize tablets based on the view of the tablet in the image (top or bottom of the tablet) and identification of the dose in the image showed over 99.87% confidence. The MLM also had the ability to identify mass outlier and non-outlier tablets well, with 90% correct identification of test images for tablets with a tablet height consistent with the average. Here, the need for more than one tablet viewpoint (i.e. greater variety of data), was more evident than for determining the drug dose, where the test tablets were all still identified correctly with over 99.65% confidence in all cases. This study identifies the feasibility of using a MLM to identify tablets as a means of validation based on images. 

National Category
Nanotechnology
Identifiers
URN: urn:nbn:se:uu:diva-552330OAI: oai:DiVA.org:uu-552330DiVA, id: diva2:1944285
Available from: 2025-03-13 Created: 2025-03-13 Last updated: 2026-04-22
In thesis
1. Semisolid Extrusion and Selective Laser Sintering in Pharmaceutics: From Clinical Application to Mass Customization
Open this publication in new window or tab >>Semisolid Extrusion and Selective Laser Sintering in Pharmaceutics: From Clinical Application to Mass Customization
2025 (English)Doctoral thesis, comprehensive summary (Other academic)
Abstract [en]

Oral medications are not readily available for a subset of the population, including pediatric patients, patients with comorbidities, and patients on a tapering schedule. Additive manufacturing (AM) provides a viable solution to this shortcoming in current medication standards, allowing for tailored oral dosage forms. Two methods, semisolid extrusion (SSE) and selective laser sintering (SLS) show particular promise for this application and are used in this thesis. SSE has particular applicability for clinical applications, while SLS has shown an aptitude for creating an array of doses of medication at a larger scale. 

Investigation into printing oral dosage forms using SSE in a hospital setting showed promising results. Oral dosage forms were created with consistency from batch-to-batch in mass and drug content. The drug content achieved, additionally, aligned closely to the desired drug content. Interviews with pharmaceutical professionals yielded information leading to a suggested workflow overhaul, utilizing SSE, for implementation in hospitals. 

Further investigation into SSE led to the development and printing of oral dosage forms on a tapering schedule. These oral dosage forms were examined with image analysis using a machine learning model (MLM). This added a layer of validation, where there is typically no validation to dose-adjusted and unlicensed medication. MLMs were able to determine differences in the images and classify images with a high-degree of success, particularly with the use of more than one viewpoint of the oral dosage forms.

The impact of geometry on the fundamental properties of SLS oral dosage forms was next studied. It was found that geometry, with shape and surface-area-to-volume ratio (SA/V) have an impact on not just the dissolution profile of the oral dosage forms, but also the resultant print quality of the oral dosage forms in terms of mass and volume compared to the theoretical values. 

Analysis into different polymers and polymer grades was performed for SLS oral dosage forms. The findings indicated that the type of polymer and polymer grade impact the resultant oral dosage forms. A general trend of slower laser sintering and higher printing temperatures, within an appropriate printing window for the material, yielded oral dosage forms that best adhered to European Pharmacopoeia guidelines.

Overall, SSE and SLS have shown distinct advantages for pharmaceutical application. The outcomes demonstrated in this work indicate the viability of SSE, particularly with image analysis validation, in a clinical setting. This work has also shown that properties such as geometry and polymer choice have a large impact on SLS printing. 

Place, publisher, year, edition, pages
Uppsala: Acta Universitatis Upsaliensis, 2025. p. 89
Series
Digital Comprehensive Summaries of Uppsala Dissertations from the Faculty of Science and Technology, ISSN 1651-6214 ; 2511
Keywords
additive manufacturing, pharmaceutics, semisolid extrusion, selective laser sintering, materials.
National Category
Nanotechnology
Research subject
Engineering Science with specialization in Nanotechnology and Functional Materials
Identifiers
urn:nbn:se:uu:diva-552332 (URN)978-91-513-2416-6 (ISBN)
Public defence
2025-05-08, Heinz-Otto Kreiss, Ångström, Lägerhyddsvägen 1, Uppsala, 09:15 (English)
Opponent
Supervisors
Available from: 2025-04-14 Created: 2025-03-13 Last updated: 2025-07-16

Open Access in DiVA

No full text in DiVA

Other links

Pre-print in full-text

Search in DiVA

By author/editor
Paulsson, MattiasStrömme, Maria
By organisation
Nanotechnology and Functional MaterialsDepartment of PharmacyDepartment of Women's and Children's Health
Nanotechnology

Search outside of DiVA

GoogleGoogle Scholar

urn-nbn

Altmetric score

urn-nbn
Total: 95 hits
CiteExportLink to record
Permanent link

Direct link
Cite
Citation style
  • apa
  • ieee
  • modern-language-association
  • vancouver
  • Other style
More styles
Language
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Other locale
More languages
Output format
  • html
  • text
  • asciidoc
  • rtf