Back

Predicting Tissue of Origin from Bulk Tumor Gene Expression using a Pre-trained Transformer Model

Mellors, T.; Spitmann, M.

2024-12-05 bioinformatics
10.1101/2024.12.01.626105 bioRxiv
Show abstract

Identifying the tissue of origin for cancers is essential for enhancing diagnostic precision, selecting effective treatments, and guiding clinical decision-making. In this study, we developed a predictive model to classify the tissue of origin across various cancer types. Using a dataset with 10,300 samples from 32 unique tissue types, the model achieved an overall accuracy of 88% in distinguishing among all 32 classes, with an average accuracy of 99.2% within each class. When tested on metastatic skin tumors, it reached an accuracy of 87%, underscoring its potential in addressing challenging metastatic cases. These results demonstrate the models reliability in oncology applications, offering a promising tool for improving diagnostic accuracy and supporting personalized cancer treatment strategies.

Matching journals

The top 8 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.