Comment from ACM Global Labs

I am interested in this regulation and posting comments that I hope to have more detailed discussions about in the next phase of the process. As a central laboratory, we have developed AI enabled solutions that help us analyze complex language of clinical trial testing and help extrapolate metadata into usable taxonomy and ontology that can enable us to operate more efficiently as a central laboratory. This is a critical intersection between the CRO or Sponsor, the Clinical Sites and Principal Investigators and the laboratory testing and outcomes data that will be needed when looking at any centralized and real-time monitoring, review and reporting. When looking at the proposed RFI, I look at this across the ecosystem of stakeholders that are relevant to the information. This should include transparency and access to both de-identified data as well as the ability to use a Unique Patient Code in identifying patients to those that need this information. For example, de-identified data may be preferred in large data sets, but the ability to identify patients to anomalous outcomes and safety concerns are of paramount importance. This is true for many aspects of the real-time solution that the FDA is interested in developing, including the tying of information both to and from EHR and other real-world data. Ultimately, the design of what is transparent to this pilot project and what remains blinded to the pilot are of paramount need in designing and structuring the interface with government regulators. This also needs to be thought through from users of the data at all access points, the site, transportation and tracking data, the laboratory and ensure that digital security measures are in place to control access and a single source of truth. Further recommendations would be to ensure that all parties that are currently working in concert or independently are willing to engage in the pilot process. Rightly so, no one wants to have real-time data be transferred to regulators if it is not accurate. Thus, having willing participants of Sonsors and CRO is still missing out on the ability to have real-time solutions be integrated into a centralized monitoring platform. This is where a Central Laboratory, like ACM Global Labs, can be a vital partner in the process. We have set up real-time monitoring solutions and API’s with our CRO and Sponsor partners currently and have worked through many of the challenges involved in this process. One must account for and ensure that the lab results are indeed QC checked and are aligned with the correct patient, site, sponsor and trial. This taxonomy can be used to help inform this at scale when looking at a central repository across the clinical trial landscape. Additional recommendation would be to look for a public-private partnership to help develop the criteria of this data structure for collection and transfer and allow this to be distributed to all laboratories that are involved in the trial itself. Thus, addressing the concern of what type of trials should be included in the pilot itself. The complexity and esoteric nature of the testing can cause anomalous data normalization and must be taken into account with any AI system that might be used to monitor. The last recommendation that for the pilot I would make is to look at selecting the right LLM to help ‘read’ free-text and use NLP that is accustomed to not only clinical trial lexicon, but also the potential esoteric reporting of specific testing results that are not simply a lab value. I look forward to the opportunity to discuss this more in future phases of the pilot and go into more detail around our current solutions and how we can partner to help bring down the cost and time it takes to bring new therapeutics and vaccines to market, benefiting the patient ultimately. Thank you, Thomas Mueller, BS, SM, MBA Chief Strategy Officer, ACM Global Labs, Rochester, NY