(W0930-06-35) A 3D Printing and Machine Vision Application for Quality Risk Management in Compounding Drug Products at the Point of Need

Purpose: The main objective of this study was to develop a patient centered, flexible and operator friendly solution to assure quality when preparing pharmaceutical products in Community and/or Hospital Pharmacies and thus mitigating the risks associated with healthcare-medication errors. The system designed is in line with the most recent regulatory directives incorporating Industry 4.0 key enabling technologies. A 3D printing process was employed to prepare personalized dosage forms when and where needed, while a combined machine vision and analytical methodology assured both quality and traceability during the whole production process (Figure 1).

Methods: Orodispersible films consisting of the Active Pharmaceutical Ingredient (API) Warfarin Sodium, the film former Hypromellose, Methocel^TM E4M, and water as solvent, were constructed with a semi-solids 3D printer ZMorph 2.0SX, where the API’s dose was accurately regulated by simply varying the film’s size. The automated in-process quality controls for the raw materials, bulk solution, and final films, were carried out by techniques such as UV/Vis, FT-IR, or Raman spectroscopy, coupled with Machine Vision for monitoring, identification and excluding defective films (Figure 2). The identification and quantification of the API was evaluated with one of the previously mentioned analytical techniques (Figure 3). A Deep Neural Network based Machine Vision system was finally developed to categorize final products into “conforming” and “not conforming”, with high accuracy.

Results: The compounding process and its final products were evaluated in near real time with in-process controls at each step (Figure 1). Initially, the bulk solution prepared was tested for assuring both the identity and API’s loading dose with an analytical method such as UV, FT-IR, or Raman spectroscopy, on-line or at line. Subsequently and if conformed to the specs, the liquid formulation was loaded to the 3D printer in order to prepare the personalized orodispersible films, which after drying and passing the final QC checks are ready to be used. The operator can control a series of the proposed system inputs such as the number of films to be printed, the size of the film, the printing speed, layer count or path width. The steps required for carrying out the whole production process and the relevant QC tests are presented in detail at the Master Manufacturing Formula, which can be thus easily followed by the trained Pharmacist and recorded. A cloud-based system effectively manages the above together with a series of supporting activities such as scheduling, maintenance, reporting/filing, cleaning SOPs etc., which can all be filed and remotely accessed, while maintaining data integrity. Therefore, the supply chain of a personalized drug product, from its raw materials and through its manufacturing process to the patient, is quality assured, recorded and fully traceable through a reliable systems thinking and acting approach.

Conclusion: The above design adequately addresses the risks associated with delivering the wrong API, and dose in a clinical setting through an automated Quality Control subsystem which simultaneously performs identification, weighing, assay, and integrity film checking in real time. This approach could be the basis for the production of small batches and fully customized dosage forms without the need of complex equipment and processes, at the point of need, complying with the quality guidelines and mitigating the patient risks associated with compounding activities.

References: Bouwman Y, Andersen LM, GMP and preparation in hospital pharmacies, European Journal of Hospital Pharmacy: Science and Practice 2012;19:469-473.
Council of Europe, Resolution CM/ResAP(2011)1 on quality and safety assurance requirements for medicinal products prepared in pharmacies for the special needs of patients. Adopted by the Committee of Ministers on 19 January 2011 at the 1103rd meeting of the Ministers’ Deputies, (accessed 15 April 2022).
Scott Gottlieb, M.D., FDA Commissioner, US Food & Drug Administration, Compounding Priorities Plan, January 2018, (accessed 15 April 2022).

Acknowledgements:The authors would like to thank the Special Account for Grants & Research, the National and Kapodistrian University of Athens and Kleon Tsetis foundation for providing the funding to attend the meeting.


Flow diagram of the compounding process and the relevant quality control tools coupled with the information storage and traceability


Examples of conforming and not conforming films as identified by machine vision.


Example of Raman spectrum of orodispersible films. Orange line: Warfarin sodium, green line: Hypromellose and blue line: orodispersible film.