Back

Extending the Boundaries of Cancer Therapeutic Complexity with Literature Data Mining

Niezni, D.; Taub-Tabib, H.; Harris, Y.; Sason-Bauer, H.; Amrusi, Y.; Azagury, D.; Avrashami, M.; Launer-Wachs, S.; Borchard, J.; Kusold, M.; Tiktinsky, A.; Hope, T.; Goldberg, Y.; Shamay, Y.

2022-10-30 bioinformatics
10.1101/2022.05.03.490286 bioRxiv
Show abstract

Drug combination therapy is a main pillar of cancer therapy but the formation of an effective combinatorial standard of care (SOC) can take many years and its length of development is increasing with complexity of treatment. In this paper, we develop a path to extend the boundaries of complexity in combinatorial cancer treatments using text data mining (TDM). We first use TDM to characterize the current boundaries of cancer treatment complexity and find that the current complexity limit for clinical trials is 6 drugs per plan and for pre-clinical research is 10. We then present a TDM based assistive technology, cancer plan builder (CPB), which we make publicly available and allows experts to create literature-anchored high complexity combination treatment (HCCT) plans of significantly larger size. We develop metrics to evaluate HCCT plans and show that experts using CPB are able to create HCCT plans at much greater speed and quality, compared to experts without CPB. We hope that by releasing CPB we enable more researchers to engage with HCCT planning and demonstrate its clinical efficacy.

Matching journals

The top 12 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.