SPI Generative AI: Aignostics Looks to Build on Benchmarks in Pathology

Published: February 6, 2024

This Insight is part of Signify Research’s upcoming Signify Premium Insights – Generative AI service which will be fully launched in February. This content is only available to subscribers. To become a subscriber, or to register for a complimentary trial of the service, please follow this link.

Digital pathology AI vendor Aignostics has recently announced that the performance of its RudolfV histopathology foundation model on several well-known public benchmarks, including the PatchCamelyon dataset, the MHIST dataset and the CRC-100K dataset, has surpassed that of many of its larger peers.

The vendor claimed that its model was more accurate compared to published alternatives on several public data sets. This result is significant as the model was not trained on the highest number of slides in the comparison.

Given such a performance, is Aignostics generative AI technology ready for broader adoption, or is it still too early for such use cases.

The Signify View

In digital pathology, as in other areas of healthcare, generative AI is among the technologies that are drumming up the most excitement. Resultantly, many digital pathology vendors, whether they yet have a pragmatic use case, or a viable commercial product, are exploring generative AI and assessing how they can best take advantage of it. Essentially, generative AI offers such potential that many vendors feel they would be remiss not to at least explore the opportunity that it offers.