Precision genomics is moving from specialty service to enterprise infrastructure
Precision genomics is no longer confined to a molecular pathology lab or a high-volume oncology center. It is becoming an operating model for health systems that want genetic information available at the bedside, in the chart, and in the treatment plan without delay.
An HHS analysis cited in the article gives the scale of the shift: Medicare Part B paid for 18 million genetic tests in 2024, a 160% increase from 2018, and genetic tests now account for 43% of Medicare’s clinical laboratory spending. Those figures reflect more than growing demand. They reflect a health care system in which genomic data is increasingly tied to real prescribing decisions, surveillance strategies, and treatment selection.
The pressure point is no longer scientific validity alone. The bigger issue is whether health systems can build the infrastructure to use genomic information consistently across service lines.
Oncology has become the proving ground for genomic medicine
Cancer care is where precision genomics has already changed day-to-day practice. Hereditary cancer testing, comprehensive genomic profiling, and molecular residual disease testing now sit inside many cancer pathways, not as experiments but as tools that help determine who needs what therapy, and when.
Wells Messersmith, MD, chief medical officer of oncology services at UCHealth, framed the clinical imperative in blunt terms: “You cannot care for a cancer patient in 2026 without precision medicine.” That statement reflects the way oncology has evolved. Tumor biology now shapes treatment selection as much as the organ site or stage at diagnosis.
Saketh Guntupalli, MD, chair of obstetrics and gynecology at the University of Colorado School of Medicine and a gynecologic oncologist at UCHealth, pointed to the change in women’s cancer care, where cell-free DNA and MRD testing have altered both surveillance and response assessment. The clinical logic is straightforward. If tumor fragments can be detected in blood before a scan turns positive, recurrence can be recognized at a molecular stage, when disease burden is still low and therapeutic choices may matter more.
One stark example comes from the CIRCULATE-Japan GALAXY study, published in Nature Medicine, which found that Natera’s Signatera test detected recurrence a median of nearly six months before imaging in a 2,240-patient analysis. A six-month lead time is not a cosmetic gain. In cancer care, it can alter the timing of imaging, escalation of therapy, and patient counseling.
MRD testing changes how clinicians think about recurrence and response
MRD testing is built around a simple biologic premise: tiny fragments of tumor DNA circulate in blood, and their presence can signal residual disease long before a lesion becomes obvious radiographically. In practical terms, this turns liquid biopsy into a surveillance tool that can complement scan-based follow-up.
The report also points to work by Bratman and colleagues showing that Signatera could identify exceptional responders and was strongly prognostic for recurrence risk after just two cycles in the context of certain treatments, including immunotherapy. Early response assessment matters because treatment exposure without benefit can carry toxicity, cost, and lost time. When a biomarker can identify nonresponse early, it gives clinicians a more precise basis for changing course.
A single blood test cannot replace the rest of oncologic assessment. It can, however, sharpen it. In a field where recurrence often emerges after a long period of uncertainty, molecular data offer a more immediate readout of disease behavior than waiting for a lesion to cross a radiographic threshold.
The bottleneck is workflow, not science
The article makes a clear distinction between the scientific promise of genomics and the operational challenge of using it at scale. The challenge is not whether the assays exist. It is whether results can be ordered, delivered, interpreted, and acted on without friction.
Messersmith described the scale problem in operational terms: moving from 10 patients to 100, 1,000, and then 100,000 requires the electronic health record and industry partnerships to do the heavy lifting. That is the real inflection point for health systems. Genomic testing fails when results sit outside the chart, when follow-up depends on manual tracking, or when different service lines use incompatible testing pathways.
James Stover, PhD, senior vice president and general manager of health systems at Natera, said the major opportunity lies in workflow improvement. The science, in his view, is already established. The value comes from simplification: easier ordering, easier access to results, and clinical decision support that helps clinicians and patients act on the data.
A fragmented testing environment makes that harder. Stover noted that some health systems work with as many as 10 genomics partners. Multiple vendors can create inconsistent ordering patterns, duplicated infrastructure, and uneven interpretation across departments. In genomic medicine, fragmentation is not just inefficient; it can distort care pathways.
EHR integration is becoming the new standard of care infrastructure
The future model described in the article is not a standalone lab relationship. It is an EHR-integrated enterprise structure in which tissue and blood testing are triggered automatically, genomic results are uploaded into the shared record, and the treating clinician sees a usable molecular profile at the point of care.
Guntupalli described a near-future workflow in which a patient is diagnosed, automatically triaged to tissue and blood testing, and the important mutations are already visible in the chart when the oncologist opens it. That workflow compresses the time between diagnosis and biologically informed treatment. It also reduces the chance that a key result will be buried in a separate portal or missed during a busy clinic day.
Messersmith said the ideal system would ingest all information into the EHR, deliver it to both physician and patient, and close the loop so results are actually seen. Closed-loop processes matter because genomic testing has little value if the result never reaches the next decision-maker.
Genomics is moving beyond oncology, but other specialties are still catching up
Oncology has built the clearest evidence base, but the article argues that precision genomics should not stop there. HonorHealth has made this a strategic priority, extending precision genomics beyond cancer into cardiovascular and neurocognitive care. Laura Goetz, MD, who was director of precision medicine at HonorHealth Research Institute when interviewed, described that expansion as a five-year goal at scale.
The biology supports the broader ambition. Inherited mutations, tumor mutations, and pharmacogenomic differences can affect diagnosis, prognosis, and treatment response across specialties. If a patient metabolizes a drug poorly, or carries a mutation that changes risk, that information belongs in clinical decision-making regardless of organ system.
Still, non-oncology fields face a familiarity gap. Goetz said other disciplines often have confusion about what precision medicine is and how to apply it. That hesitation slows adoption even when the clinical rationale is strong.
Partnership now determines how quickly genomics can scale
The article draws a direct line between enterprise adoption and strategic partnership. A vendor sells a test. A partner helps build the workflow around it, integrate it with the EHR, standardize ordering, and connect results to outcomes.
Messersmith said UCHealth has increasingly turned to partners rather than homegrown labs because the field is moving too quickly for isolated systems to keep pace with new assays and treatment options. That is especially true in a field where innovation touches not only the test itself but also the reporting format, the downstream workflow, and the data architecture.
The most ambitious vision is a continuously learning system that aggregates de-identified genomic and clinical data and uses those signals to refine treatment in real time. Messersmith described the hope that AI-enabled systems and shared datasets could reveal patterns clinicians would otherwise miss. If that model works, genomic testing will do more than guide one patient at a time. It will feed a feedback loop that improves care design across the enterprise.
The remaining challenge is not whether genomic medicine has utility. The challenge is whether health systems can build durable, interoperable infrastructure fast enough to keep the results in front of clinicians when decisions are made.