English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Deep-learning models based on whole-slide digital pathology images (WSIs)become increasingly popular for predicting molecular biomarkers. Instance-basedmodels has been the mainstream strategy for predicting genetic alterationsusing WSIs although bag-based models along with self-attention mechanism-basedalgorithms have been proposed for other digital pathology applications. In thispaper, we proposed a novel Attention-based Multiple Instance Mutation Learning(AMIML) model for predicting gene mutations. AMIML was comprised of successive1-D convolutional layers, a decoder, and a residual weight connection tofacilitate further integration of a lightweight attention mechanism to detectthe most predictive image patches. Using data for 24 clinically relevant genesfrom four cancer cohorts in The Cancer Genome Atlas (TCGA) studies (UCEC, BRCA,GBM and KIRC), we compared AMIML with one popular instance-based model and fourrecently published bag-based models (e.g., CHOWDER, HE2RNA, etc.). AMIMLdemonstrated excellent robustness, not only outperforming all the five baselinealgorithms in the vast majority of the tested genes (17 out of 24), but alsoproviding near-best-performance for the other seven genes. Conversely, theperformance of the baseline published algorithms varied across differentcancers/genes. In addition, compared to the published models for geneticalterations, AMIML provided a significant improvement for predicting a widerange of genes (e.g., KMT2C, TP53, and SETD2 for KIRC; ERBB2, BRCA1, and BRCA2for BRCA; JAK1, POLE, and MTOR for UCEC) as well as produced outstandingpredictive models for other clinically relevant gene mutations, which have notbeen reported in the current literature. Furthermore, with the flexible andinterpretable attention-based MIL pooling mechanism, AMIML could furtherzero-in and detect predictive image patches.

A robust and lightweight deep attention multiple instance learning algorithm for predicting genetic alterations