Application Note | July 17, 2026
Predicting Drug-Induced Organ Injury
Get the Resource
Application Note - Predicting DILI - Download
"*" indicates required fields
No safety failure ends more drug programs or forces more market withdrawals than drug-induced organ injury, and liver injury leads the list. Drug-induced liver injury (DILI) is the most common cause of acute liver failure in developed countries and the most frequent reason a drug is withdrawn from the market or carries a black-box warning. It ends compounds in preclinical development, halts them in the clinic, and forces label restrictions on molecules that are otherwise effective.
A safety failure caught in the clinic or after approval costs orders of magnitude more than one caught preclinical screens – which is why predicting DILI early, before it causes critical patient safety implications, should be the goal of any CRO seeking to be a drug development partner, rather than simply an assay vendor, to pharma and biotech teams.
Why Predicting DILI Early Matters – and What Standard Screening Modalities Miss
DILI-related failures persist because the tools used to screen for injury early do not translate to humans. Animal toxicology misses a large share of human hepatotoxicity, particularly the idiosyncratic injury that appears only in people. In vitro screens built on imaging, DNA, or gene expression report changes inside a cell or a model system, but they leave open the question that actually matters, which is whether the same change occurs in a patient and can be seen there. A signal that predicts injury in a dish is only useful if it also marks injury in a person, and most modalities cannot make that connection.
| Screening Modality | Limitation for Predicting Human Injury |
|---|---|
| Cell Painting / High-Content Imaging (IHC) | Predictive in vitro, but morphology cannot be measured in a patient, so an imaging hit cannot be confirmed in human blood. |
| DNA Sequencing | Describes what a cell could do, not what is actively happening, and is therefore not a real-time injury signal. |
| RNA Sequencing | Works in cultured cells, but because transcripts are confined to the intracellular space, confirming a signal in humans requires tissue that may not be readily obtainable. |
| Proteomics | Strong for tissue markers and mechanism, but not fast or cost-effective enough to screen thousands of compounds. |
Metabolites and lipids are the exception. They are the functional biochemistry of injury itself, changing within minutes to hours of an insult, and they cross cell membranes into the surrounding fluid and then into the blood. The same molecule that marks injury in a culture dish can therefore be measured in a patient’s plasma, which is precisely the property that transcripts and most proteins lack.
Use the form to download the Application Note for a deeper look into omics approaches for predicting DILI and other drug-induced organ injury.
Sapient's Approach: Screening + Human Translation, Not Just an In Vitro Assay
Using our proprietary ultra-throughput rapid LC-MS (rLC-MS) platform, Sapient’s approach to hepatotoxicity screening begins with running nontargeted metabolomics and lipidomics on both cells and spent culture media from compound-treated cultures. These rLC-MS systems measure more than 15,000 small molecule markers per run – including thousands of unannotated features that carry biochemical activity invisible to DNA, RNA, or protein-based assays.
Because metabolites and lipids are the downstream, functional output of a cell’s biochemistry, they report an active injury process rather than the potential for one. They also move freely across cell and organelle membranes, into the culture media, and into the circulation, which means an injury signature discovered in vitro can be carried intact to human confirmation.
The initial screen identifies which compounds injure cells, predicting DILI in vitro. The harder – and more commercially relevant – question is which of those signatures are truly specific to human injury.
That’s where DynamiQ™, Sapient’s proprietary molecular-clinical database, comes in: more than 67,000 plasma samples from over 15,000 individuals, profiled on the same rLC-MS-based metabolomic and lipidomic platform and linked to clinical outcomes across 60+ diseases, including hepatic, renal, cardiac, and neurological conditions.
An injury signature found in cell culture can be mapped directly onto the plasma signatures of patients in this database that have confirmed liver injury. Signatures that show up in vitro but never in injured patients get set aside as model artifacts; signatures that mark injury in both are the ones worth acting on.
Beyond Predicting DILI: One Platform for Drug-Induced Tissue Injury
While predicting liver injury is a priority because it drives the most attrition, Sapient’s platform can read out injury wherever it occurs – including in the heart, kidney, lung, or central nervous system. Because the readout is an untargeted signature rather than a fixed panel, a single screen can flag injury across multiple organ systems at once, including off-target toxicities a program may not have initially been looking for.
Start a Screening Study or Translational Safety Program
Historically, teams have screened for toxicity with assays that predict events in preclinical model systems and hoped the prediction would hold in people. Sapient reframes the question from, “what changes in a dish?” to “what change also marks injury in a patient?”, and answers it early – when it is still a chemistry or candidate selection decision rather than a clinical failure.
To discuss how your studies can benefit from predicting DILI and other organ injury earlier, schedule a time to meet with our team.
Get the Resource
Application Note - Predicting DILI - Download
"*" indicates required fields