Résumé
In 2015, the United States put forward the concept of precision medicine, which changed medical treatment from "one size fits all" to personalization, and paid more attention to personalization and drug customization. In the same year, Spritam®, the world's first 3D printed tablet, was in the market, marking the emerging pharmaceutical 3D printing technology was recognized by regulatory authorities, and it also provided a new way for drug customization. 3D printing technology has strong interdisciplinary and high flexibility, which puts forward higher requirements for pharmaceutical staffs. With the development of artificial intelligence (AI), modern society can perform various tasks, such as disease diagnosis and robotic surgery, with superhuman speed and intelligence. As a major AI technology, machine learning (ML) has been widely used in many aspects of 3D printing drug, accelerating the research and development, production, and clinical application, and promoting the new process of global personalized medicine and industry 4.0. This paper introduces the basic concepts and main classifications of 3D printing drug, non-AI drug optimization technology and ML. It focuses on the analysis of the research progress of ML in 3D printing drug, and elucidates how AI can empower the intelligent level of 3D printing drug in pre-processing, printing, and post-processing process. It provides a new idea for accelerating the development of 3D printed drug.
Résumé
A method to measure the antibody-dependent cell-mediated phagocytosis (ADCP) potency of anti-CD38 mAb was developed based on design of experiment (DoE) with a Jurkat/NFAT/CD32a-FcεRIγ transgenic cell line as the effector cell, the Daudi cell line as the target cells, and luciferase as the detection system. The DoE method was used for optimization of experimental parameters and methodological validation. The results show that anti-CD38 mAb exhibits a dose-response relationship with the following four-parameter equation: y = (A - D) / [1 + (x / C)B] + D. Several experimental parameters were optimized by statistical experimental design and determined as follows: the working concentration of anti-CD38 mAb was 800-20.81 ng·mL-1, the density of the target cells was 7.5×104 per well, and the density of effector cells was 2.5×104 per well, with an induction time of 6 h. The method showed good specificity. The recovery rate for samples from 5 different groups showed that the relative potencies of anti-CD38 mAb were (59.97 ± 4.74) %, (82.44 ± 5.15) %, (110.69 ± 11.71) %, (129.23 ± 5.22)% and (162.15 ± 3.66) %. The recoveries ranged from 103% to 120% and the RSDs of the above results were all less than 11%. The linear detection range was 50%-150%. Based on DoE design, this method for measuring ADCP potency of anti-CD38 mAb was optimized and validated with good specificity, repeatability and accuracy. This method can be used for evaluation of ADCP biological activity of anti-CD38 mAbs.
Résumé
This study aims to establish the design space of the key processes for drop-on-powder 3D printing based on design of experiment (DoE). By utilizing Minitab, an experimental scheme with three factors, two levels and three center points was designed to analyze the factors that significantly affected the tablet quality attributes. Furthermore, the factor interactions were analyzed using Minitab. subsequently, the computer aided drafting (CAD) software was used to adjust the model volume with fixed radius/height ratio (r/h = 1.25) and establish a linear regression equation between model volume and dose. As a result, the drug dose could be controlled in a flexible manner. The finally determined process parameters were: ink-jet level is 12, layer thickness is 150 μm, and the X-axis printing head speed of 635 mm·s-1. Regression equation between drug content (y) and model volume (x) was y = 0.062 x - 0.582 7 (R2 = 0.999 9) showing good linear relationship. This indicated that robust and feasible process parameters were obtained through DoE, and the preparation of personalized-dose tablets was realized with good reproducibility.