Back

In Silico Investigations of Adaptive Therapy Using Two Cytotoxic or Two Cytostatic Drugs

Saha, D. K.; Anderson, A. R. A.; Cisneros, L.; Maley, C. C.

2023-05-14 bioinformatics
10.1101/2023.05.12.540626 bioRxiv
Show abstract

While the dose modulation (DM) protocols (DM Cocktail Tandem, DM Ping-Pong Alternate Every Cycle, DM Ping-Pong on Progression) involves adjusting drug dosages when the tumor burden changes, the fixed-dose (FD) protocols involves administering a specific, constant dosage of the drug only when the tumor is growing (Dose-Skipping) or when the absolute tumor burden is above the baseline level until it reduces to a certain percentage of the baseline (Intermittent). Moreover, two different drugs can be administered simultaneously (cocktail), or the drugs can be switched such that only one drug is applied at a given time (ping-pong), either every cycle (ping-pong alternate every cycle) or when the tumor grows (ping-pong on progression). The dose modulation protocols work well when treated with two cytotoxic drugs, however, the ping-pong protocols (DM Ping-Pong Alternate Every Cycle, DM Ping-Pong on Progression, FD Ping-Pong Intermittent) work well when treated with two cytostatic drugs. In general, adaptive therapy, using either two cytotoxic or two cytostatic drugs works best under conditions of high competition, such as high fitness cost, high replacement rates, and high turnover, although treatment using two cytostatic drugs works best under low turnover in many cases. Adaptive therapy works best when drug dosages are changed as soon as a change in tumor burden is detected, and it is best to pause treatment sooner rather than later when the tumor is shrinking. Adaptive therapy works best when an intermediate level of drug dosage is used, and both treatment with too little or too much drug leads to poor survival outcome.

Matching journals

The top 6 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.