Volpara reaches break-even early as Lunit integration advances

Volpara Health, a breast cancer screening software company headquartered in New Zealand, reached break-even more than a year ahead of plan and is pressing its advantage after being acquired by Korea’s Lunit last year.

The pairing brings together two companies known for rigorous clinical validation in AI imaging. Volpara’s commercial reach in the U.S. and its strengths in density analytics and risk modeling now sit alongside Lunit’s cancer detection algorithms, creating what executives describe as a uniquely comprehensive software ecosystem for breast screening.

In an interview with Korea Biomedical Review in mid-August, Volpara CEO Craig Hadfield outlined how tighter execution delivered the earlier-than-expected break-even and why the company’s next phase hinges on Prognosia’s image-based risk model and a steady march toward autonomous AI.

Earlier in June, Lunit acquired Prognosia's "core technology asset," according to Lunit's public filing on Aug. 13. Prognosia is a risk prediction algorithm developer spun out from Washington University School of Medicine. 

Early break-even, market expansion, and integration after acquisition

Hadfield said Volpara reached break-even more than a year ahead of plan by leaning on Volpara’s installed base to begin selling Lunit’s software in the U.S.

“Last year we achieved 28 percent revenue growth, and in the first half of this year we grew another 22 percent,” he said. “We grew our costs, but less than the top line, which allowed us to hit break-even early.”

Much of the growth came from Volpara’s existing products, with an increasing contribution from its role as Lunit’s commercial partner in the U.S.

Since taking the helm six months ago, Hadfield has concentrated on four priorities that anchor the post-deal playbook. 

Hadfield explained that his immediate focus has been on strengthening Lunit’s presence in the U.S. by tapping into Volpara’s established network. 

“Lunit was slightly later to market, but Volpara has the largest footprint among all competitors in the U.S., which helps us open doors quickly,” he said.

At the same time, the company is pursuing international expansion with renewed intensity, particularly in Latin America.

 “We have doubled down on the region,” he said. “We added headcount, put systems and processes in place, and ported Volpara’s more mature commercial discipline to the market.”

The company is expanding its North American footprint with a dedicated sales unit in Canada.

Product development remains another priority. 

Volpara’s established tools -- such as breast density measurement and the questionnaire-based Risk Pathways model -- have continued to deliver growth, but Hadfield highlighted Prognosia’s image-based risk technology as the next driver.

The acquisition will allow Volpara to integrate the software seamlessly into its existing cloud infrastructure, enabling hospitals to activate it instantly.

Integration with Lunit is the final strand of his agenda. Bringing together two mid-sized companies from different cultural and geographic backgrounds is a complex undertaking, he acknowledged. “We are not massive, but we are not small either. Add the differences between Korea, New Zealand, and the U.S., and it will take years,” he said. “There are many positives, but integration requires constant effort.”

Competitive positioning is central to the growth plan. 

Hologic dominates mammography hardware in the U.S., but Hadfield argues that software is a different contest. 

“Volpara and Lunit together have about 20 percent of the software market,” he said. “Our software is vendor-agnostic, so hospitals running Hologic, GE, Fuji, or Siemens can use our suite.”

The company is software-first and science-first, which is why customers increasingly prefer buying multiple products from Volpara rather than stitching together point solutions, he added.

Another important factor was the company’s deployment model, which eliminates on-site hardware. 

Data are anonymized inside the hospital network before transmission, then processed in Microsoft Azure, and systems are audited under ISO 27001 with routine penetration tests. 

“Hospitals want virtual, cloud-deployed software that is centrally managed,” he said. “We can switch on additional products later with a simple configuration change, which makes adoption easier and faster.”

The business model remains flexible. Volpara’s traditional approach is hospital-pay per patient across volume. 

Lunit has also demonstrated a patient-pay model for certain AI services. As reimbursement evolves in the U.S., hospitals could see additional paths to adopt new tools without adding cost. 

“Where reimbursement is available, hospitals can move more quickly, and in some cases create a new revenue stream,” Hadfield said.

Prognosia acquisition and the road to autonomous AI

Volpara’s acquisition of Prognosia was not just about adding another product to its portfolio.

For CEO Craig Hadfield, the deal was a strategic step to accelerate the company’s long-term vision of autonomous AI in breast screening. Lunit and Volpara had already built complementary strengths -- detection on one side and risk modeling on the other -- but the missing piece was an image-based tool that could bridge the gap between early detection and personalized risk stratification.

“Traditional risk models combine patient information, family history, and breast density. Image-based risk analyzes the mammogram itself to calculate five-year risk,” Hadfield said. “It is not one versus the other, it is one plus one equals three.” 

By layering Prognosia’s technology onto Volpara’s infrastructure, the combined company is now better positioned to expand its market and move closer to the future of AI-driven screening.

The next challenge, however, lies in translating this scientific advantage into clinical practice, where physicians need clear and trusted thresholds to act on risk scores.

Hadfield pointed to the familiar twenty percent lifetime-risk cut-off in guidelines like NCCN for supplemental screening, and he expects analogous five-year image-based thresholds to be established through ongoing work with Prognosia’s founders and U.S. bodies as the product moves through the FDA.

“For traditional models, twenty percent lifetime risk is the line for additional screening like ultrasound or MRI,” he said. “For image-based models, it may be three to five percent over five years.”

There is no formal guideline yet, which is why the company is doing the work now so that by the time it launches commercially next year, the framework is in place or close, he added. 

Prognosia’s model already holds FDA Breakthrough Device Designation, a strong signal that the approach addresses an important unmet need and a pathway that can enable faster coverage decisions. Volpara plans to submit for 510(k) clearance later this year. 

“We have a long history working with the FDA, with multiple 510(k) clearances across Volpara and Lunit,” Hadfield said. “There is now precedent for image-based risk in the U.S., so we feel confident about the regulatory path.”

Demand dynamics also favor image-based risk, he added.

Risk Pathways has grown between twenty and thirty percent annually for five years despite the inherent friction of questionnaires and data collection.

By contrast, image-based risk runs on the image already captured in routine care and returns a score without extra steps. 

“The mammogram is taken, uploaded, our algorithm runs, and the result is delivered back to the radiologist’s workstation,” Hadfield said. “When you combine ease with strong science, you get adoption.”

For market launch geographically, the U.S. remains the primary focus because of market size and cloud readiness, with Europe close behind given significant interest from ongoing image-based risk studies. 

Latin America is opening as a commercial opportunity as hiring and processes take root. 

Asia is on the roadmap as cloud adoption matures. 

“Ultimately we intend Prognosia to be global,” Hadfield said.

Beyond risk stratification, the acquisition advances a longer arc toward autonomous AI. 

Hadfield defines this as an integrated end-to-end workflow in which AI confidently handles the large volume of low-risk exams, and radiologists focus on the thirty to fifty percent of cases where their expertise is most valuable.

“We need to bring our products together so there are no extra clicks and no redundancy,” he said. “We know the features that are missing, we are engaging the FDA on the evidence required, and we are building toward that.”

The company is probably three to five years away from full autonomy, but compared to others we are further along the pathway, he added. 

The company is also mindful of patient communication as some people worry that risk scores may create anxiety or lead to unnecessary interventions. 

Hadfield said the evidence shows transparency improves outcomes.

“Women would rather know than not know,” he said. “Catching breast cancer at stage one leads to very high survival, while stage four outcomes are much worse. Information matters.”