NEWS
Novel AI model works across multiple cancers
Researchers have developed a novel AI model that can analyse a routine whole histopathology image and simultaneously predict cancer subtype, specific genetic mutations, and survival outcomes across 32 different solid cancers rather than focusing on a single cancer type. The findings, from a study in The
American Journal of Pathology, highlight the potential of computational pathology to connect routine diagnostic imaging with molecular oncology. Currently, most deep learning-based
models are used for single-model concepts;
Proscia grows diagnostic options
Proscia has received a new 510(k) clearance from the US Food and Drug Administration (FDA) for Concentriq AP-Dx. The clearance includes compatibility with the Leica Aperio GT 450 DX slide scanner, cloud deployment, and a Predetermined Change Control Plan (PCCP). Concentriq AP-Dx is a digital pathology platform for primary diagnosis used by laboratories of all sizes. Through the PCCP, Proscia gains a more efficient pathway to expand Concentriq AP-Dx’s interoperability while maintaining regulatory oversight. The PCCP enables Proscia to validate and implement support for additional FDA-cleared scanners, image formats, and displays without requiring further 510(k) submissions. “Laboratories choose a digital pathology platform with the expectation that it will sit at the centre of their practice for years to come,” said Ian Cadieux, Head of Quality and Regulatory at Proscia. “The PCCP helps us to protect that investment, giving customers confidence that Concentriq AP-Dx will continue to be an interoperable foundation as their needs evolve.” Concentriq AP-Dx initially received
FDA 510(k) clearance for primary diagnosis with the Hamamatsu NanoZoomer S360MD Slide scanner in February 2024. The platform is also CE-marked under IVDR.
September 2026
WWW.PATHOLOGYINPRACTICE.COM 13
one model for one task. The researchers developed a single model that can generate seven outputs simultaneously from the whole histopathology image, including TP53 mutation status, TP53 RNA expression, tumour type, and survival- related outcomes at the slide level. The AI-based Vision Transformer model analysed routine haematoxylin and eosin (H&E) stained whole slide images of human solid tumours. The model was trained on a dataset that included more than 11,000 primary tumour cases retrieved from the Pan-Cancer Atlas,
with corresponding somatic mutation, RNA-sequencing, and clinical outcome data. The most significant result was that the model achieved a strong predictive accuracy score (AUROC of 0.766) for TP53 mutation detection across 32 solid tumour types in an independent validation set of 1,729 slides.
The American Journal of Pathology / Chaurasia et al.
Page 1 |
Page 2 |
Page 3 |
Page 4 |
Page 5 |
Page 6 |
Page 7 |
Page 8 |
Page 9 |
Page 10 |
Page 11 |
Page 12 |
Page 13 |
Page 14 |
Page 15 |
Page 16 |
Page 17 |
Page 18 |
Page 19 |
Page 20 |
Page 21 |
Page 22 |
Page 23 |
Page 24 |
Page 25 |
Page 26 |
Page 27 |
Page 28 |
Page 29 |
Page 30 |
Page 31 |
Page 32 |
Page 33 |
Page 34 |
Page 35 |
Page 36 |
Page 37 |
Page 38 |
Page 39 |
Page 40 |
Page 41 |
Page 42 |
Page 43 |
Page 44 |
Page 45 |
Page 46 |
Page 47 |
Page 48 |
Page 49 |
Page 50 |
Page 51 |
Page 52 |
Page 53 |
Page 54 |
Page 55 |
Page 56
