Pavurala Naresh, Xu Xiaoming, Krishnaiah Yellela S R
US Food & Drug Administration, Center for Drug Evaluation and Research, Office of Pharmaceutical Quality, 10903 New Hampshire Avenue, Silver Spring, MD 20993, USA.
US Food & Drug Administration, Center for Drug Evaluation and Research, Office of Pharmaceutical Quality, 10903 New Hampshire Avenue, Silver Spring, MD 20993, USA.
Int J Pharm. 2017 May 15;523(1):281-290. doi: 10.1016/j.ijpharm.2017.03.022. Epub 2017 Mar 19.
Hyperspectral imaging using near infrared spectroscopy (NIRS) integrates spectroscopy and conventional imaging to obtain both spectral and spatial information of materials. The non-invasive and rapid nature of hyperspectral imaging using NIRS makes it a valuable process analytical technology (PAT) tool for in-process monitoring and control of the manufacturing process for transdermal drug delivery systems (TDS). The focus of this investigation was to develop and validate the use of Near Infra-red (NIR) hyperspectral imaging to monitor coat thickness uniformity, a critical quality attribute (CQA) for TDS. Chemometric analysis was used to process the hyperspectral image and a partial least square (PLS) model was developed to predict the coat thickness of the TDS. The goodness of model fit and prediction were 0.9933 and 0.9933, respectively, indicating an excellent fit to the training data and also good predictability. The % Prediction Error (%PE) for internal and external validation samples was less than 5% confirming the accuracy of the PLS model developed in the present study. The feasibility of the hyperspectral imaging as a real-time process analytical tool for continuous processing was also investigated. When the PLS model was applied to detect deliberate variation in coating thickness, it was able to predict both the small and large variations as well as identify coating defects such as non-uniform regions and presence of air bubbles.
使用近红外光谱（NIRS）的高光谱成像将光谱学与传统成像相结合，以获取材料的光谱和空间信息。使用NIRS的高光谱成像具有非侵入性和快速性，这使其成为用于透皮给药系统（TDS）制造过程中过程监测和控制的有价值的过程分析技术（PAT）工具。本研究的重点是开发和验证使用近红外（NIR）高光谱成像来监测涂层厚度均匀性，这是TDS的关键质量属性（CQA）。化学计量学分析用于处理高光谱图像，并建立了偏最小二乘（PLS）模型来预测TDS的涂层厚度。模型拟合度和预测度分别为0.9933和0.9933，表明对训练数据拟合良好且具有良好的可预测性。内部和外部验证样品的预测误差百分比（%PE）小于5%，证实了本研究中开发的PLS模型的准确性。还研究了高光谱成像作为连续加工实时过程分析工具的可行性。当应用PLS模型检测涂层厚度的故意变化时，它能够预测小变化和大变化，并识别涂层缺陷，如不均匀区域和气泡的存在。
