How do you know if two cancer drugs can be combined safely?
Two drugs each work, but will combining them stack toxicities? How do I tell promising pairings from intolerable ones?
Two effective drugs don't make a safe combination if they share a toxicity, and most databases only list single-drug side effects, so combination risk is invisible. We used Hydra to extract efficacy by subtype and adverse-event profiles across the landscape, then score pairings: in TNBC it flagged a complementary pair that real clinical data confirms.
Who this is for: Translational, clinical-pharmacology, and safety scientists assessing whether two agents can be combined without stacking toxicities.
- every pairing scored on efficacy × tolerability
- 2 axes
- efficacy normalised (TNBC, PD-L1+, and more)
- by subtype
- side-effect profiles extracted, not just response rates
- AE-aware
every pairing scored on efficacy × tolerability
efficacy normalised (TNBC, PD-L1+, and more)
side-effect profiles extracted, not just response rates
The problem with judging combination safety
Two effective drugs don't automatically make an effective combination. If both carry an overlapping toxicity, combining them stacks that risk and the regimen becomes intolerable long before it becomes more effective, which is exactly why safety testing often restricts which combinations ever reach efficacy studies.
There's a data gap that makes this hard: most side-effect databases only contain adverse events for single drugs, not for combinations, so the overlapping-toxicity risk is invisible until the clinic. Reading every relevant trial for both response rates and adverse-event profiles, then reasoning about which toxicities compound, is the slow manual synthesis that gates combination strategy.
What teams in this space search for
- How do I know if two cancer drugs can be combined safely?
- How do you predict overlapping toxicity in drug combinations?
- Which drug combinations stack side effects?
How we solved it with Hydra
“Across the competitive landscape for this indication, extract each agent's efficacy by subtype and its adverse-event profile from the trial literature. Score which pairings combine well without stacking overlapping toxicities, and flag the ones that would. While you're at it, surface patient subpopulations with poor response so we can feed them back into target discovery.”
What Hydra ran
The Combination-Safety Scoring module normalised efficacy across the landscape, which agents work best in which subtypes (TNBC, PD-L1-positive, and so on), and, crucially, also extracted side-effect profiles from the same trial sources, closing the gap left by single-drug databases.
It then scored each candidate pairing on two axes together: combined efficacy and whether the two agents share an overlapping toxicity that would compound.
What it found
Worked example in triple-negative breast cancer: one agent had a high response rate but notable pneumonitis; a complementary agent was also effective with low pneumonitis. Because they don't share that toxicity, the module flagged the pairing as promising and scored it high, and real clinical results bear that out.
By contrast, two agents from a class with additive side effects scored poorly and were not recommended. The same read also surfaced subpopulations with low response to current options.
What we learned
Efficacy is only half the question, scoring has to read adverse events, not just response rates, and it has to reason about combinations even when the databases only describe single drugs.
Safety screening is also a discovery input: the low-response subpopulations it surfaces are fed straight back into target discovery, pointing toward genuinely unmet need rather than dead-ending as a 'no'.
What you get
- Promising, low-overlap combinations flagged and scored, confirmed by real clinical results
- Additive-toxicity pairings ruled out before they reach the clinic
- Efficacy and safety weighed together, across the whole landscape
- Low-response subpopulations routed back into target discovery
Data sources used
- ClinicalTrials.gov (trial design & populations)
- Published trial outcomes (efficacy & adverse events)
- cBioPortal / TCGA (subtype context)
Figures reflect analyses PharosBio ran on public datasets and public benchmarks. Named competitors, collaborators, and logos are withheld at this stage; the methods and results shown are real and repointable to your own target.
Sources & methods
- Trial efficacy & adverse-event data: ClinicalTrials.gov; published trial outcomes
- Subtype context, cBioPortal / TCGA
Frequently asked questions
Why can't I just use a side-effect database?
Most adverse-event databases only list side effects for single drugs, not combinations, so the overlapping-toxicity risk that actually sinks a regimen isn't captured. The scoring has to extract per-agent profiles and reason about which toxicities compound.
Has the scoring ever matched real outcomes?
Yes. In the TNBC example, a high-response agent with notable pneumonitis paired with a low-pneumonitis agent scored as promising, and that pairing is supported by real clinical results, while an additive-toxicity class pairing scored poorly.
Run this analysis on your question
Hydra plans, executes, and validates, so you reach a defensible answer in hours, not weeks.