If you or someone you love has been told that genomic testing found “nothing actionable,” or that the cancer has stopped responding to treatments matched to DNA mutations, you are not necessarily out of options. A newer test called DarwinOncoTarget reads your tumor’s biology more deeply than standard genomic platforms and may uncover drug targets that DNA-based testing cannot detect. Developed at Columbia University and recognized as one of Fast Company’s Top 10 Most Innovative Biotech Companies of 2025, this technology is still early in its clinical journey, but it represents a genuinely different approach that deserves attention, particularly for patients facing treatment-resistant disease.
This article explains what DarwinOncoTarget is, how it differs from the genomic tests you may already know (FoundationOne, Guardant360, Caris, Tempus, and BostonGene), and when it may make sense to ask your oncologist about it.
The Problem: Why Genomic Testing Sometimes Hits a Wall
Most cancer patients today are familiar with some form of genomic testing. Your oncologist may have sent a tissue sample or blood draw to a company like Foundation Medicine, Guardant Health, Caris Life Sciences, Tempus, or BostonGene. These tests examine your tumor’s DNA, looking for specific mutations or genetic alterations that can be matched to targeted drugs. When they find a clear match, the results can be life-changing.
But here is the hard truth that patients with advanced or treatment-resistant cancer often discover: genomic testing based on DNA mutations alone helps only a fraction of patients. Published estimates suggest that only about 10% of patients with treatment-resistant cancer experience meaningful clinical benefit from therapies matched solely to DNA mutations. The largest precision oncology trial ever conducted, NCI-MATCH, screened roughly 6,000 patients and found that while about 38% had potentially actionable mutations, only 18% could actually be assigned to a treatment, and only a fraction of those responded.
Why does this happen? Because DNA mutations tell only part of the story. A mutation is like a typo in an instruction manual. Sometimes that typo changes everything about how the final product works. Other times, the cell compensates, routes around the error, or the typo sits in a section of the manual that is not actively being read. The same exact mutation in the same exact gene can behave very differently depending on which tissue the cancer arose from, what other molecular machinery is active, and how the tumor has adapted to previous treatments.
This is where treatment-resistant cancer becomes especially frustrating. A tumor that initially responded to a drug, as determined by genomic testing, may eventually develop resistance, not necessarily through new DNA mutations but through shifts in protein activity, changes in the regulatory networks that drive cell behavior, and adaptations in the tumor’s internal operating system. Standard genomic tests, which read the DNA “blueprint,” may see the same mutations as before and offer no new direction.
What DarwinOncoTarget Does Differently
DarwinOncoTarget does not look at DNA mutations. Instead, it reads your tumor’s RNA, which is the active, working copy of your genes. Think of it this way: if DNA is the master blueprint locked in a filing cabinet, RNA is the set of instructions the cell is actually following right now. By analyzing the complete RNA profile of your tumor (a process called whole transcriptome sequencing), DarwinOncoTarget can determine which proteins are actively driving the cancer at this moment, not which ones might theoretically be involved based on a DNA mutation.
The core technology is an algorithm called VIPER (Virtual Inference of Protein-activity by Enriched Regulon analysis), published in Nature Genetics in 2016. In plain terms, VIPER looks at thousands of genes that are turned up or down in your tumor and works backward to determine which upstream proteins are responsible for those patterns. It is reading the tumor’s behavior, not just its genetic code.
An important clarification: DarwinOncoTarget is a computational test. It does not grow your tumor cells in a dish and physically expose them to drugs (that is a different category called functional drug sensitivity testing). Instead, it uses algorithms and regulatory network models to predict which drugs are most likely to be effective against the specific protein drivers in your tumor. Currently, the test evaluates 193 druggable proteins and matches them against 94 FDA-approved cancer drugs and 239 investigational drugs in advanced clinical trials. For a typical patient, it identifies somewhere between 5 and 20 potential drug targets.
DarwinHealth also offers a companion test called DarwinOncoTreat, which goes a step further. OncoTreat identifies tightly connected networks of “master regulator” proteins that work together to sustain the cancer state. Rather than targeting one protein at a time, OncoTreat seeks drugs that can collapse these networks, potentially addressing the adaptive resistance that makes treatment-resistant cancers so difficult to control.
A Closer Look at the Standard Genomic Platforms
To appreciate what DarwinOncoTarget adds, it helps to understand what the major genomic tests do well and where their limitations lie:
- FoundationOne CDx (Foundation Medicine/Roche) is probably the most widely used tissue-based genomic test. It sequences 324 cancer-related genes to detect DNA mutations and has more than 30 FDA-approved companion diagnostics. It is considered a gold standard for comprehensive genomic profiling of solid tumors. Its limitation: it reads only DNA, so if your tumor’s driver is a protein activity problem rather than a DNA mutation, FoundationOne may return no actionable findings.
- Guardant360 CDx (Guardant Health) is a liquid biopsy, meaning it requires only a blood draw rather than a tissue biopsy. It analyzes circulating tumor DNA shed by the cancer into the bloodstream. This makes it especially convenient for monitoring and for patients with tumors that are difficult to biopsy. Results typically come back in under seven days. Its limitation: it analyzes a smaller set of genes (74), and not all tumors shed enough DNA into the blood to be detected reliably.
- Caris Molecular Intelligence (Caris Life Sciences) takes the broadest approach among conventional platforms. Caris combines DNA sequencing, RNA sequencing, and protein analysis, giving oncologists multiple layers of information. Of all the standard platforms, Caris comes closest to a functional picture of the tumor because it includes some protein-level data. It also draws on a database of over 740,000 matched patient records to power AI-driven treatment insights.
- Tempus xT offers one of the largest gene panels, with 648 genes, and combines DNA and RNA sequencing with access to one of the world’s largest clinical datasets. By adding RNA data, Tempus reports it can match roughly 43% more patients to targeted therapies than DNA-only panels. This is a meaningful step forward, though the analysis still centers primarily on identifying specific genetic alterations.
- BostonGene’s Tumor Portrait uniquely integrates DNA, RNA, and tumor microenvironment classification. It categorizes tumors by their immune landscape, which helps predict how likely they are to respond to immunotherapy. BostonGene reports a 98% clinical actionability rate, though the platform is primarily oriented toward research and biopharma partnerships.
All of these tests are valuable, and many patients benefit from them. They should generally be the first step in any precision oncology workup. But they all share a fundamental limitation: they identify molecular features in the tumor’s genetic code, then cross-reference databases to predict which drugs might work. The keyword is “might.” They are making educated guesses based on what the DNA (and sometimes RNA) looks like. None of them directly assesses the functional protein machinery that is keeping the cancer alive at this moment in this patient.
That gap is exactly what DarwinOncoTarget was designed to fill.
Why This Matters Most for Treatment-Resistant Cancer
If your cancer responded initially to treatment and then progressed, or if multiple lines of therapy have failed, DarwinOncoTarget may deserve special consideration. Here is why.
Treatment-resistant cancers are, by definition, cancers that have outsmarted the drugs thrown at them. In many cases, the tumor has not simply acquired a new DNA mutation (though that can happen). Instead, the cancer has rewired its internal regulatory networks, activating alternative protein pathways to survive. Standard genomic tests may see the same DNA profile as before, leaving your oncologist with few new directions to explore.
DarwinOncoTarget reads the tumor at the level where this rewiring actually occurs. By identifying which proteins are abnormally active right now, it can pinpoint drug targets that would be invisible to a test focused solely on DNA mutations.
This is not just a theory. A 2023 study published in Cancer Discovery enrolled approximately 130 patients with diverse treatment-resistant malignancies and used the DarwinHealth platform to identify drug matches. When the predicted drugs were tested in patient-derived tumor models (human tumors grown in mice), OncoTreat-predicted drugs achieved a 91% disease control rate, OncoTarget-predicted drugs achieved 68%, and drugs not predicted by the platform achieved 0%. Those are striking numbers, though it is important to note that the drugs were tested in mouse models, not in the enrolled patients. The researchers themselves noted that not all predicted drugs would necessarily work in human patients. Still, the contrast between predicted and unpredicted drugs is noteworthy and suggests the platform is identifying real biological vulnerabilities.
Additional evidence supports the approach. A study in Nature Cancer demonstrated that VIPER-based protein activity scores could predict which metastatic breast cancer patients would respond to the drug ricolinostat. A separate study published in the New England Journal of Medicine showed that a VIPER-derived marker accurately distinguished responders from non-responders to selinexor in patients with triple-class refractory multiple myeloma, one of the most treatment-resistant blood cancers.
The bottom line for patients: when genomic testing has been done and either found nothing actionable, or the matched treatments have stopped working, DarwinOncoTarget offers a scientifically credible additional investigation that reads the tumor’s biology through a different lens.
How to Get the Test
DarwinOncoTarget is ordered through the Laboratory of Personalized Genomic Medicine at Columbia University Irving Medical Center. Any oncologist in the United States can place the order. The patient does not need to be seen at Columbia. Tissue specimens are shipped to Columbia’s laboratory for analysis.
The test requires a tumor tissue sample, typically formalin-fixed paraffin-embedded (FFPE) tissue from a previous biopsy or surgery. The sample must contain at least 50% tumor cells. For leukemia patients, whole blood or bone marrow can be used instead. If your oncologist has tissue from a recent biopsy stored at a pathology lab, it may be possible to use that specimen without requiring a new procedure.
Results are typically available within about 21 days. The report identifies which proteins are abnormally active in the tumor, which FDA-approved drugs target them, and which investigational drugs in clinical trials may also be relevant. Your oncologist can then use this information to consider new treatment strategies, ideally during a tumor board discussion.
DarwinOncoTarget works across all human cancers, both solid tumors and blood cancers. The companion OncoTreat test is currently validated for a more limited set of cancer types, including breast cancer, glioblastoma, meningioma, neuroendocrine tumors, and sarcomas, with additional types being added as the research progresses.
What to Know About Cost and Coverage
Insurance coverage for DarwinOncoTarget is not yet standardized. The test has a billing code that allows it to be submitted to Medicare and private insurers, but there is no formal national coverage determination. Some patients may face out-of-pocket costs. One published estimate put the test cost at approximately $1,600, though this may vary.
Perhaps the bigger coverage question involves the drugs themselves. Because DarwinOncoTarget may recommend FDA-approved drugs for a cancer type they were not originally approved for (called off-label use), obtaining insurance authorization for those medications can be a hurdle. This is a challenge that a feasibility trial published in PLOS One in June 2025 specifically identified as a key barrier. Your oncologist’s office or a patient advocate may be able to help navigate prior authorization or appeal processes.
An Honest Look at What Has and Has Not Been Proven
The scientific foundation behind DarwinOncoTarget is substantial. The underlying research has been published in some of the most respected journals in medicine and cancer biology, including Cell, Nature Genetics, Nature Cancer, Cancer Discovery, and the New England Journal of Medicine. Few diagnostic companies can point to a publication record of this caliber.
However, it is important to be forthright about what has not yet been demonstrated. No randomized controlled trial has compared outcomes in patients treated according to DarwinOncoTarget recommendations versus those treated with standard approaches. No published data yet show that using the test improves progression-free survival or overall survival in human patients. The preclinical validation, particularly the 91% disease control rate in mouse models, is encouraging, but the leap from animal models to proven human benefit has not yet been documented in published form.
A feasibility trial at Good Samaritan University Hospital in New York, published in PLOS One in 2025, represents one of the first community oncology studies designed to assess whether the approach can work in everyday clinical practice and to lay the groundwork for a future randomized trial. Columbia University is reportedly conducting a larger 3,000-patient study, and additional trials are underway in metastatic pancreatic cancer, HER2-positive breast cancer, and other tumor types.
The test is not yet incorporated into NCCN or ASCO treatment guidelines, so your oncologist may or may not be familiar with it and may have varying levels of comfort with its recommendations.
When to Consider DarwinOncoTarget: A Practical Framework
Based on the available evidence and practical realities, DarwinOncoTarget is most worth discussing with your oncologist if several of the following apply to your situation:
Your cancer has progressed through one or more lines of standard therapy, and previous genomic testing (FoundationOne, Caris, Guardant360, Tempus, or BostonGene) found no actionable mutations, or the drugs matched to identified mutations are no longer working. Adequate tumor tissue from a recent biopsy or surgery is available, or you are able to undergo a biopsy. Your overall health and energy levels are sufficient to tolerate additional treatment. You have the time, meaning your cancer is not progressing so rapidly that a 21-day turnaround for results would be clinically untenable. And you and your oncologist are open to exploring off-label drug use or enrolling in a clinical trial based on the results.
DarwinOncoTarget should not be seen as a replacement for standard genomic testing. It is best understood as a complementary, deeper-level investigation, an additional tool in the precision oncology toolkit that reads a layer of tumor biology the DNA-focused tests do not reach. For patients who have already traveled the conventional genomic testing road and found themselves at a dead end, this test offers a scientifically serious reason to keep looking.
Looking Ahead
DarwinHealth co-founder Andrea Califano now serves as President of the Chan Zuckerberg BioHub in New York, bringing the same AI and network biology algorithms to broader applications in cancer and other diseases. The company’s pharmaceutical partnerships with Bristol Myers Squibb, Daiichi Sankyo, and others suggest that the underlying science is being taken seriously by major players in oncology drug development. The technology is also being explored for pediatric cancers through the GAIN Consortium initiative led by Dana-Farber Cancer Institute.
The coming years should bring the definitive clinical trial data that will determine whether DarwinOncoTarget’s impressive preclinical track record translates into proven survival benefits for patients. In the meantime, for those facing treatment-resistant cancer with limited options, it is a conversation worth having with your oncology team.
If you have questions about whether DarwinOncoTarget might be appropriate for your situation, ask your oncologist or contact the Laboratory of Personalized Genomic Medicine at Columbia University Irving Medical Center.
References
- Alvarez MJ, Shen Y, Giorgi FM, Lachmann A, Ding BB, Ye BH, Califano A. Functional characterization of somatic mutations in cancer using network-based inference of protein activity. Nat Genet. 2016;48(8):838-847.
- Alvarez MJ, Subramaniam PS, Tang LH, Grunn A, Aburi M, Rieckhof G, Komissarova EV, Hagan EA, Bodei L, Clemons PA, Dela Cruz FS, Dhall D, Diolaiti D, Fraker DA, Ghavami A, Kaemmerer D, Karan C, Kidd M, Kim KM, Kim HC, Kunju LP, Langel U, Li Z, Lee J, Li H, LiVolsi V, Pfragner R, Rainey AR, Realubit RB, Remotti H, Roses R, Rustgi A, Sepulveda AR, Serra S, Shi C, Yuan X, Barberis M, Bergamaschi R, Chinnaiyan AM, Detre T, Ezzat S, Frilling A, Hommann M, Jaeger D, Kim MK, Knudsen BS, Kung AL, Leahy E, Metz DC, Milsom JW, Park YS, Reidy-Lagunes D, Schreiber S, Washington K, Wiedenmann B, Modlin I, Califano A. A precision oncology approach to the pharmacological targeting of mechanistic dependencies in neuroendocrine tumors. Nat Genet. 2018;50(7):979-989.
- Chari A, Vogl DT, Gavriatopoulou M, Shain KH, Masszi T, Beksac M, Spicka I, Hungria V, Meletios A, Berdeja JG, Oriol A, Gothelf A, Dimopoulos MA, Holmberg L, Bringhen S, Weisel K, van der Holt B, Kim K, Maisnar V, Moreau P, Raab MS, Bahlis NJ, Facon T, Pour L, Jie S, Qiu L, Mateos MV, Chen Y, Grosicki S, Jagannath S, Leleu X, Ludwig H, Hajek R, Delforge M, Agha ME, Lonial S, Touzeau C, Laubach J, Kumar SK, Richardson PG, Gatt ME, van de Donk NWCJ, Cavo M, Anderson KC. Oral selinexor-dexamethasone for triple-class refractory multiple myeloma. N Engl J Med. 2019;381(8):727-738.
- Mundi PS, Dela Cruz FS, Grunn A, Diolaiti D, Mauguen A, Rainey AR, Guillan K, Siddiquee A, You D, Realubit RB, Karan C, Ortiz MV, Douglass EF, Accordino M, Mistretta S, Brogan F, Bruce JN, Caescu CI, Carvajal RD, Crew KD, Decastro G, Heaney M, Henick BS, Hershman DL, Hou JY, Jones PA, Kalinsky K, Manji GA, Pentsova EI, Shu CA, Su JA, Iyer G, Boire A, Gavathiotis E, Bhatt RS, Alvarez MJ, Kung AL, Califano A. A transcriptome-based precision oncology platform for patient-therapy alignment in a diverse set of treatment-resistant malignancies. Cancer Discov. 2023;13(6):1386-1407.
- Radigan R, Sangal A, Zhang D, Weber J, Schmalzle M, Finkelstein A, Agarwal M, Califano A, Alvarez MJ, Rosenzweig KE. Feasibility trial of Darwin OncoTreat and OncoTarget precision medicine testing to improve outcomes for patients with limited metastatic disease that failed first-line systemic therapy. PLoS One. 2025;20(6):e0325769.
- Zeleke TZ, Pan Q, Chiuzan C, Onishi M, Li Y, Tan H, Alvarez MJ, Honan E, Yang M, Chia PL, Mukhopadhyay P, Kelly S, Wu R, Fenn K, Trivedi MS, Accordino M, Crew KD, Hershman DL, Maurer M, Jones S, High A, Peng J, Califano A, Kalinsky K, Yu J, Silva J. Network-based assessment of HDAC6 activity predicts preclinical and clinical responses to the HDAC6 inhibitor ricolinostat in breast cancer. Nat Cancer. 2023;4(2):257-275.
