Kategori: Digitalization

Digitalization-tr 13 Eylül 2024

European Commission Highlights IRIS Technology Solutions for its Innovative Role in the AIDPATH Decentralized Advanced Therapy Production Project

IRIS Technology Solutions is proud to announce that we have been recognized as a “Key Innovator” by the European Commission’s Innovation Radar for our contributions to the AIDPATH project. This recognition highlights our innovative development of a monitoring strategy for optimal measurement of selected parameters in bioreactors, leveraging cutting-edge technologies such as Artificial Intelligence (AI), soft-sensors, and machine learning algorithms.

AIDPATH (Artificial Intelligence-driven, Decentralized Production for Advanced Therapies in the Hospital) is a high-impact EU-funded consortium dedicated to advancing the next generation of personalized medicine using gene-engineered immune cells at EU hospitals through AI technology. The project focuses on T cells modified to express a synthetic chimeric antigen receptor (CAR-T), a revolutionary treatment in hematology and oncology, with potential applications for infections and autoimmune diseases. AIDPATH aims to address the complexities of traditional CAR-T therapy, which is hindered by centralized manufacturing and inflexible clinical use schemes. By integrating patient-specific data and biomarkers, AIDPATH uses AI to enable flexible manufacturing and optimize CAR-T cell products, improving anti-tumor potency and reducing costs and resource utilization in hospitals.

Advanced Monitoring Strategy for Bioreactors

In this project, IRIS Technology Solutions played a pivotal role by developing a monitoring strategy that allows for the optimal measurement of selected parameters in bioreactors. Bioreactors are essential in bioprocessing as they provide the environment needed for biological reactions to occur, such as for the production of pharmaceuticals, biofuels and other bio-based products. In the case of AIDPATH, the bioreactor performs a so-called “perfusion” process which, starting from a small number of CAR T-cells taken from a patient, grows a much larger amount, which are then reintroduced into the patient for immunotherapy cancer treatment. However, maintaining optimal growth conditions within the bioreactor is a complex task that requires precise monitoring of variables such as pH, temperature, oxygen levels, and nutrient concentrations.

We tackled this challenge by designing a system capable of monitoring these parameters in real time, adapting dynamically to changes and ensuring that the bioreactor’s conditions remain optimal for the biological processes involved. One of the key achievements was our development of a control system that provides alerts based on the variations in these parameters. The system was programmed to follow specific control rules and thresholds to ensure expected behaviors during in-line monitoring, making it more efficient and reliable than traditional systems.

Innovation through AI-Powered Soft-Sensors

A standout innovation that contributed to this recognition was our integration of AI-powered “smart-sensors” within the monitoring system. Unlike the traditional approach that provides information based on a small number of hard sensors with a PID controller, our smart-sensors apply AI technologies (Fuzzy and Consensus Based algorithms) and advanced statistical techniques (Bollinger, sliding window), to aggregate the data from a larger and more varied set of hard sensors and convert it into actionable insights.

These smart-sensors have been designed to support decision-making by providing real-time insights into the bioreactor’s performance. For instance, instead of simply reporting temperature fluctuations, the soft-sensors analyze the behavior over time and relate it to the trends from other hard sensors. This approach allows for a more holistic feedback to the human expert operator of any significant changes within the bioreactor, ultimately leading to more efficient bioprocesses and higher yields.

Additionally, the soft sensors are integrated in the SCADA type interface (called COPE in AIDPATH) which displays the soft sensor outputs and alerts in real time on a dashboard which is easily interpretable by the human expert operator. Furthermore, the smart sensors are completely parameterizable (soft coded) by the end user, an advantage when calibrating the system for new use cases. Data analytics and machine learning can be used to further fine tune the control parameters and data processing rules.

Being recognized by the European Commission’s Innovation Radar underscores IRIS Technology Solutions’ leadership in advanced process monitoring and AI applications. This recognition not only acknowledges our efforts but also opens opportunities for future collaborations and partnerships with other innovators, businesses, and academic institutions. It highlights our commitment to developing cutting-edge technologies with real-world applications and impact.

Moreover, the Innovation Radar platform provides visibility to potential customers, investors, and partners seeking innovative solutions in the biotechnology and pharmaceutical sectors. As part of this platform, IRIS is poised to attract new interest and expand its network within the scientific and industrial communities.

Being named a “Key Innovator” by the European Commission’s Innovation Radar is a significant achievement that reflects our dedication to technological advancement and innovation, as well as the hard work and commitment of our team. We look forward to continuing to drive innovation in advanced process monitoring, artificial intelligence, and beyond.

IRIS Technology Solutions
Digitalization-tr, Industry-4-0-tr, Pharma-4-0-tr 3 Nisan 2024

Control of the coating process of granular forms by NIR spectroscopy

Control of the coating process of granular forms by NIR spectroscopy

In the pharmaceutical industry, there are many granular formulations that are coated to achieve a sustained or controlled release of the drug or active pharmaceutical ingredient (API) over time, a clear and well-known example being Omeprazole. In this paper we will discuss these extended release formulations and how it is possible to optimize the release time and potency analyses during the coating process using NIR spectroscopy.

Pelletisation process and traditional analysis

During the pelletisation process of modified release dosage forms, the correct application of the coating (e.g. an enteric release coating intended to prevent gastric digestion or degradation) will determine the subsequent efficacy of the drug and the mg/API release time of the drug and therefore controls are carried out throughout this process to ensure the quality and thus the expected pharmacological action.

 

Currently, this control is performed during the coating process with samples obtained from the coating equipment at different times and analysed in the laboratory using the analytical technique of HPLC or liquid chromatography and dissolution testing to demonstrate that the release of the active ingredient(s) is satisfactory. Both methods require sample preparation prior to analysis, require specialised personnel and consumables (materials), in addition to the duration (hours) of a dissolution test, whose main objective is to determine the bioavailability of the drug, meaning the relative amount of the drug that has entered the general circulation after administration, and the rate at which this access has occurred.

Therefore, the major problem with traditional analytics is that it is time-consuming to obtain the results and therefore does not allow for timely rectification of the coating process in case of failures or, in the frequent case of stopping the process for sampling, there is a risk that the quality of the semi-product will be altered.

 

An alternative and very effective tool that allows real-time monitoring of the coating process is NIR technology, since the spectral signature of each pellet can be related to its coating conditions, dosage and release times without the need to resort to traditional methods.

Development of an NIRS method for predicting release time and potency

In order to develop a predictive model for the real-time determination of release times and potency (mg API/g pellet) that is released at 1, 4 and 7 hours, we worked in coordination with a major Spanish pharmaceutical laboratory and the portable NIR spectroscopic analyser Visum Palm™ manufactured and marketed by IRIS Technology Solutions S.L. 

The data provided by the laboratory consists of the NIR spectra of several batches of two drugs based on, on the one hand, an antihistamine which, for confidentiality reasons, we will refer to as “DS”, and on the other hand, a form of vitamin B6 which, for the same reasons, we will refer to as “PH”.  In both cases, the active substance was part of the coating of the pellets constituting the vehicle. 

The spectra of the pellets were acquired at different times of the coating process, from both wet and dry samples and, in parallel, the respective sample was subjected to the usual analyses in these cases to determine the drug release at 1, 4 and 7 hours and the potency mg PI/g. 

The predictive models developed on the basis of the spectral data showed that it is not necessary to dry the samples for the acquisition of the spectra – so the control can be performed directly on the wet sample, saving time and handling – and that there is a clear relationship between the NIR spectra, the power and the release times of 1h, 4h and 7h, as we will see below.

PH compound - Coating process and NIR spectroscopy

Table 1: Quality parameters of the prediction models for the release at 1, 4, 7 hours and the potency in the samples with different stages of the PH coating process. The * symbol indicates that the model was built by using the average NIR spectra of the replicates of each sample.

Figure 1: Regression curves for PH a) All samples; b) Batches 1,3,4 y 7; c) Mean spectra of batches 1,3,4 y 7; d) Batch 7.

DS compound

Table 2 shows the quality parameters of the models for the analysis of wet DS samples. All samples were studied simultaneously: samples from batches 6, 8 and 10 together, and batch 6 separately. Batches 6, 8 and 10 were chosen for the study of a set of batches because they had the largest number of samples. In addition, batch 6 was chosen for individual analysis as it contained the most samples with the optimal release parameters for the case study.

Table 2: Quality parameters of the prediction models for the release at 1, 4, 7 hours and the potency in the samples with different stages of the DS coating process.

Figure 2 shows the regression curves resulting from the study for the active substance DS. The values of the quality parameters for the DS models show, in general, a good correlation. As an observation, it is noted that the error increases when data from different batches are used, probably because the process conditions of each batch are different due to the fact that the data come from the development and fine-tuning phase of the production process. The prediction of the release at 7 hours is worse than that of the other parameters, probably because the end of the release process has been reached in many cases before that time.

 

Figure 2: Regression curves for DS a) all samples; b) Batches 6, 8 y 10; c) Mean spectra of batches 6, 8 y 10; d) Batch 6.

Prediction of dry samples

Table 3: Quality parameters of the prediction models for the dry samples of DS batch 6 and PH batch 7.

The prediction models of the dry samples for individual batches of PH and DS show a good correlation. It should be noted that the prediction error is due to the few validation samples used.

 

Figure 3: Regression curves for Dry simples of a) DS batch 6 y b) PH batch 7.

Conclusions - Coating process and NIR spectroscopy

  • There is a clear correlation between NIR spectra with release times of 1h, 4h and 7h, as well as with potency, for both DS and PH, although it is slightly worse for PH.
  • In the case of the 7h release, the correlation seems a bit weaker, possibly because it is close to the maximum release (at the release plateau) or due to differences in the pH of the samples.
  • The different batch production conditions affect the robustness of this correlation, an inherent variability factor because the samples come from the development phase of the production process (fine-tuning phase) and not from the NIRS method.
  • Individual batch tests show a good correlation for both wet and dry samples. Since the results in both cases are similar, it can be concluded that drying is not necessary to correlate the studied parameters (release time and potency) with the NIR spectra.
  • Finally, from the analysis of the results analysed, it can be concluded that NIR spectroscopy can be used to optimise the control of the coating process of granular forms and that, from a technical point of view, it is a robust and evidence-based method. However, for all the cases evaluated in this document, definitive models have to be made once the production process has been fully developed.
IRIS Technology Solutions
Ai-tr, Digitalization-tr, Industry-4-0-tr, Innovation-tr, Pharma-4-0-tr 5 Eylül 2023

New Visum Palm™ AI-assisted handheld NIR analyser

handheld nir analyser

New Visum Palm™ AI-assisted handheld NIR analyser

IRIS Technology Solutions introduces the latest version of its Visum Palm™ portable NIR analyser to complement its Visum® range of real-time process analysers for industry.

The new Visum Palm™ is a fully portable NIR spectrophotometer that allows real-time analysis of different substances, products or mixtures, without the need for traditional laboratory and sampling techniques, allowing industry to obtain results on the spot to make decisions or correct production process parameters.

The new generation Visum Palm™ brings with it an innovative design and a radical change in the way users experience NIR technology, now assisted by AI with the Visum Master™ software, so that each manufacturer can automatically create their own predictive models or calibrations according to their control and analysis needs.

 

Design, autonomy and robustness

The Visum Palm™ handheld nir analyser offers an innovative and ergonomic design, as well as the possibility to perform analysis at any time and place without having to connect it to any external electronic device. This is possible because it incorporates an embedded touch screen and computer, which enable all the routine functionalities of the device.

The Visum Palm™ operates in the 900 to 1700 nm range, as this is the band that best combines availability of chemical information with price and technological maturity. It operates mainly in diffuse reflectance mode, for which it has specially designed and patented optics to extract as much information as possible from the sample. Specifically, it has a large illumination area (50 mm diameter) and a collection area of 10 mm. These features differentiate it from similar analysers in terms of its suitability for analysing heterogeneous samples, which is most often the case in real working conditions. In cases where heterogeneity is more evident, the device is configurable to calculate and report the average of a given number of repetitions.

The Visum Palm™ handheld NIR analyser is IP65 compliant, making it resistant to dust, moisture and water. It is also rugged enough to be carried and tested almost anywhere indoors or outdoors and even comes with a stand for desktop or tabletop use.

 

A new AI-assisted user experience

Unlike most common modelling and calibration software on the market, which requires the user to have some technical knowledge of chemometrics or entrust the task to a third party, Visum Master™ PC-based software makes NIR technology even more accessible by automating pre-processing, multivariate analysis algorithm selection and validation. This allows any user to generate models by simply inputting spectra and references (quantitative or qualitative) for routine real-time analysis to replace traditional analysis.

handheld nir

The new software also allows to extend and edit pre-existing models, synchronise with the portable analyser to import spectra, export models, download measurement results, automatically generate analytical method validation reports and audit reports for GMP environments, and to check the metrological performance of the device in a guided manner when needed.

 

For industry and GMP environments

While NIR technology has a myriad of applications in numerous industries such as plastics, food, chemical, agribusiness, wood, biofuels, to mention the most relevant but not the only ones; it is for the pharmaceutical industry and GMP environments where the new Visum Palm™ device introduces significant novelties at the level of usability and functionality. It is 21 CFR Part 11 compliant, allowing the generation and display of an automatic Audit Trail report, the record of all device activity, where comments and observations can be incorporated. It also allows the user to automatically generate the analytical method validations developed and perform metrological checks of the device when required and download the results at a later date.

“NIR technology today must be easy to use and understand, and at the same time it must give the user the freedom and autonomy to exploit it to the full and facilitate their day-to-day work. Technology must be an enabler. We will continue to take further steps in terms of automation and new functionalities because we are convinced that this is the right way forward and what the industry and the people in it need”, says Oonagh Mc Nerney, Director of IRIS Technology Solutions, S.L.

 

The new Visum Palm™ handheld NIR analyser is now available here, where you can also find technical information about the device, videos and contact IRIS Technology Solutions, S.L. for a demonstration or specific enquiry.

 

IRIS Technology Solutions