**5. Conclusions**

DTs are a crucial development of the close integration of manufacturing information and physical resources that raise much attention across industries. The critical parts of a fully developed DT include the physical and virtual components, and the interlinked data communication channels. Following the development of IoT technologies, there are many applications of DT in various industries, but the progress is lagging for pharmaceutical and biopharmaceutical manufacturing. This review paper summarizes the current state of DT in the two application scenarios, providing insights to stakeholders and highlighting possible challenges and solutions of implementing a fully integrated DT.

In pharmaceutical manufacturing, building blocks of a DT, including PAT methods, data managemen<sup>t</sup> systems, unit operations, and flowsheet models, system analyses methods, and integration approaches have all been developed in the last few years, but gaps in PAT accuracy, real-time model computation, model maintenance capabilities, real-time data communication, as well as concerns in data security and confidentiality, are preventing the full integration of all the components. To solve these challenges, several insights are provided. The development of new tools such as NIRS and in-line UV spectroscopy, iterative optimization technologies, and di fferent o ffline adaptive methodologies can help to resolve the existing issues in PAT methods. In order to reduce simulation time to achieve real-time computation, e fficient algorithms, and reduced order modeling approaches should be further studied for process models. In terms of model maintenance, adaptive modeling methods with online streaming data are to be investigated further. To have a fully integrated and automated DT, the information flow from the virtual component to the physical plant also needs to be established. The virtual plant should be able to change system settings and control the physical plant to help to achieve an optimized process within the design space. Ideally, all these components should be placed under appropriate physical and virtual security protocols.

In biopharmaceutical manufacturing, similar constituting components of DTs have been discussed, as well as the implementation challenges in each block. In terms of process monitoring, the development of NIR or Raman spectroscopy, material calibration, and chemometric methods can help to obtain an accurate predicting/measurement result. Advanced data integration and synchronization technology should be in place. For process simulation, there is no robust model that captures CPPs and CQAs for all the unit operations in the integrated process due to the computational complexity. Pre-analysis to screen the CPPs and CQAs is a promising approach to reduce the computational burden. Process models to capture the auxiliary equipment and process contamination need to be further investigated. To achieve a fully integrated DT, real-time data acquisition methods, data transferring systems, e ffective control and execution techniques, robust simulation methods, and anomaly detection are still in need, with other supporting functions.

It is noted that given the rapid development and publication rate in this area, and that this paper is merely a narrative literature review, the authors are not able to list and review all studies in these areas in detail. The papers selected and problems described in the manuscript are only a nonholistic subset used to represent the capabilities and drawbacks of a method or technology. Since the manuscript is organized using a conceptual and topical frame, the authors recommend interested readers to go through cited references to explore additional details. In addition to the summarized research opportunities, further research directions can include the development of a demonstrative case study of DT in pharmaceutical and biopharmaceutical manufacturing and a systematic review of the field.

**Author Contributions:** Conceptualization, Y.C. and M.I.; writing—original draft preparation, Y.C., O.Y., C.S., P.B., R.R., and M.I.; writing—review and editing, Y.C., O.Y., C.S., P.B., R.R., and M.I.; supervision, R.R. and M.I.; project administration, M.I.; funding acquisition, M.I. and R.R. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded by US Food and Drug Administration, gran<sup>t</sup> number DHHS-FDA-U01FD006487 and DHHS-FDA-R01FD006588.

**Conflicts of Interest:** The authors declare no conflict of interest.
