Inside the Atlanta Startup Rewiring How Cancer Gets Decided

There is a meeting that happens every week inside almost every cancer center in America, and if you have never had cancer you have probably never heard of it. It is called the tumor board. Around a table - or now, a video call - an oncologist, a surgeon, a radiologist, a pathologist and sometimes a geneticist argue over a single patient's scans, slides and lab work until they agree on a plan. It is the closest thing medicine has to a jury. And for decades it ran on projectors, printed charts and email chains.

OncoLens, a health-technology company based in Atlanta, built its entire business on the bet that this meeting deserved better software. Founded in 2017, the company sells a cloud platform that pulls a patient's data out of electronic medical records, labs and pathology systems, assembles it into a case packet, and uses artificial intelligence to flag who should be reviewed, tested for biomarkers, or matched to a clinical trial. Today it runs inside more than 225 cancer centers, touching roughly five million patients and 45,000 physicians.

01The problem is coordination, not science

---

Cancer care has a strange failure mode. The science is often available - the targeted therapy exists, the trial is open, the biomarker test is standard - but the right people never see the case at the right time. A patient can get one recommendation from a medical oncologist, another from a surgeon, and a third from an outside specialist, with none of them comparing notes. That gap is not a knowledge problem. It is a logistics problem.

OncoLens co-founder and CEO Anju Mathew learned this the hard way. Before starting the company she ran technology innovation and partnerships at McKesson and co-founded a cancer-imaging device company. Then her father was diagnosed with leukemia, and she watched him receive conflicting advice from physicians who never sat in the same room. Her co-founder, Dr. Lijo Simpson, is a board-certified hematologist/oncologist who knew the administrative side of that breakdown from the inside.

That pairing - a technologist who had felt the failure as a family member, and a clinician who had lived it as a practitioner - is the reason OncoLens reads less like a generic data platform and more like a tool built by people who have sat through a bad tumor board. The company's own name is a small tell: “onco” for cancer, “lens” for the single view of a patient it is trying to give a roomful of specialists who otherwise each see one slice.

02What the software actually does

---

The hard part is not the meeting. It is everything that has to happen before it. Someone has to find the patient, gather the scattered records, read the free-text notes, pull the pathology, and turn it all into a reviewable case. OncoLens automates that pipeline, and says it cuts tumor-board preparation time by more than 90 percent.

The AI layer is the part that turns a scheduling tool into something closer to a research engine. OncoLens says its natural-language processing can pull cancer staging out of unstructured clinical notes about 80 percent of the time, and identifies relevant patients with a reported 85 to 90 percent accuracy - no tidy structured field required. That matters because the most useful information about a cancer patient is usually buried in a paragraph a doctor typed at 9pm, not in a dropdown.

03The numbers centers report

---

OncoLens publishes a set of operational metrics from centers using the platform. Treat them as vendor-reported, but they point at where the value shows up - prep time, testing rates and data extraction.

Figures self-reported by OncoLens across customer deployments. Approximate.

04Who pays, and for what

---

OncoLens is a B2B SaaS business with two engines. The first is straightforward: cancer centers, academic medical centers and integrated delivery networks license the platform on subscription. Named customers include Houston Methodist, Ascension, Piedmont Healthcare, Emory Healthcare and Karmanos Cancer Institute.

The second engine is the OncoLens Research Network, which connects life-science and pharmaceutical companies to real-world data and clinical-trial matching across that network of centers. The company says roughly half of the top 20 pharma companies engage through it. In practice, the same data plumbing that preps a tumor board also makes a center legible to trial sponsors - which is a second product hiding inside the first.

This two-sided shape is what makes the model interesting. A center buys OncoLens to save its coordinators time and to stay compliant with accreditation bodies like the Commission on Cancer and NAPBC - the unglamorous features that actually close deals in health systems. Once the center is on the platform, its de-identified real-world data becomes valuable to sponsors hunting for trial sites and patient cohorts. The provider side funds the product; the research side compounds it. Each new cancer center makes the network worth more to pharma, and the pharma relationships help justify the platform to the next center.

05How it raised, round by round

---

OncoLens did not chase a splashy valuation. It stacked rounds as the customer base grew, mostly alongside Atlanta's BIP Capital. The seed came in 2019; the Series A grew to $7.25 million by 2021; and in October 2024 the company closed a $16 million Series B co-led by BIP Capital and Cross Border Impact Ventures, with Martin Ventures and SeedToB returning. Total raised sits near $27 million.

The Series B landed just after OncoLens ranked #1296 on the 2024 Inc. 5000, its second appearance on the list following a #316 debut in 2023. The company is SOC 2 certified and employs around 35 people. In 2025, Mathew was named a Georgia Titan 100 honoree.

06Where it sits in the market

---

The obvious rival is Roche, whose navify Tumor Board is the best-funded product in the category. Others in the mix include Xcures, MassiveBio and Genomet, plus the homegrown virtual tumor board tools many big centers build for themselves. OncoLens's wedge is breadth: instead of just running the meeting, it tries to own the whole data pipeline around it - intake, NLP, trial matching, registry reporting and the research network on top.

Capability	OncoLens	Typical rival
Virtual tumor board	Yes	Yes
NLP on unstructured notes	Yes	Varies
Automated trial matching	Yes	Sometimes
Registry & accreditation reporting	Yes	Rare
Life-science research network	Yes	Rare

There is also a structural reason a broad platform can beat a point tool in this market. Cancer centers are not short on software; they are short on integration. A radiology viewer, a genomics report, a registry system and an EMR rarely talk to each other, and every extra login is a reason a busy oncologist skips a step. By pulling those threads into one case view - and by doing the un-fun work of registry and CME reporting - OncoLens makes itself the layer a coordinator opens first, rather than one more tab.

The honest caveat: coordination software is easier to sell on prep-time savings than to prove on survival. OncoLens reports operational wins - faster boards, more biomarker testing, cleaner data - and those are real to the people doing the work. Whether that translates into better outcomes at population scale is the harder question the whole category still has to answer. The company's counter is a plausible one: more biomarker testing and more trial matches are exactly the inputs precision oncology needs, even if the outcome data takes years to catch up.

07The takeaway for builders

---

The most copyable thing about OncoLens is not its AI. It is the wedge. The company found an old, high-stakes ritual that ran on spreadsheets and email, sold software to automate the boring prep around it, and then quietly built a second business - the research network - on the data exhaust. Find the ritual nobody owns. Automate the paperwork. Sell to the people already in the room.