Technical development, validation and industrialization has been concluded; clinical validation is sufficiently completed to support initial commercial development and is expected to continue in support of these commercial activities.
We finished the optimization of two versions of the HepatoPredict test (i) a kit version of HepatoPredict for use by pathology laboratories (primary version of HepatoPredict), a laboratory-developed test version, suitable for use in support of clinical research, quality control and back up for client laboratories. In both cases a cloud-based platform (MyOphiomics) allows access to the HepatoPredict.AI algorithm that generates a clinical report. Optimization procedures improved performance and usability of the original prototype, and included a detailed characterisation of the performance of the method that was the basis of the technical flies used for CE IVD marking. HepatoPredict achieved a precision of 94% (very high confidence mode), and a sensitivity of 130% (high confidence mode), using a non-trivial combination of variables and machine learning approaches, representing a significant improvement over existing clinical criteria and of the prototype at the beginning of the project.
We initiated the clinical validation procedures. In the retrospective setting we have completed the first cohort from CHULC (PT). We are currently validating on a cohort from another PT centre (CHUC). We have agreed terms and wait samples to arrive in July from H Henri Mondor (FR) and wait ethics approval for studies with with H de la Fe (ES) and Queen Elizabeth H (UK). The pandemic slowed us but the cohorts above serve to address all the validation needs raised by the hepatology/surgery community. In addition, we have discussed potential collaborations with several other hospitals in Europe, Asia and Americas, but these will be pursued in support of commercial development. Furthermore, we have, in collaboration with the CHULC (PT) designed, ethics-approved, registered and initiated a prospective clinical trial. 13 out of 40 patients were recruited and at least two were already transplanted.
We set up a manufacture laboratory and established a supply chain, detailed procedures compatible with the ISO13485 specifications (audit requested, pending), recruited and trained dedicated staff and can now support a business model based on a laboratory kit, with a manufacture capacity of 2000 kits/year and a roadmap for trebling this capacity. In addition, we set up a laboratory and formalised an extensive set of procedures that is now ready to support clinical studies, quality control activities and back up of client laboratories with an installed capacity for 5,000 samples/year, and a roadmap to doubling this capacity. This can also support an alternative business model of a laboratory-developed test.
