Triple-negative breast cancer (TNBC), characterized by the absence of estrogen, progesterone and human epidermal growth factor receptors, has the poorest prognosis amongst breast cancer subtypes and makes up about a third of breast cancer recurrences. Patients complete neoadjuvant chemotherapy and immunotherapy, undergo surgery, and receive further adjuvant treatment; they then spend the years under surveillance, hoping that routine mammograms and physical exams will catch any recurrence early enough to successfully intervene. A new study in BMJ Open from researchers at Value Analytics Labs, Abbott’s cancer diagnostics business, and Harvard Medical School asks an important question: how effective is that surveillance system? Dr. Heather Parsons, who now heads the Breast Oncology Program at Fred Hutch Cancer Center, is senior author on the study.
The team created a microsimulation model, meaning they tracked individual virtual patients (rather than modeling a population as a single aggregate group), month by month, over their entire lifetime. They simulated the journey of a newly diagnosed stage II or III TNBC patient: neoadjuvant chemotherapy with immunotherapy, surgery, adjuvant treatment, and then years of follow-up care as laid out in the National Comprehensive Cancer Network (NCCN) guidelines. The NCCN guidelines recommend a screening mammogram every 12 months after surgery followed by a diagnostic mammogram and core needle biopsy in case of a suspicious finding.
The model was built using real clinical trial data (primarily the pivotal KEYNOTE-522 trial that tested the addition of PD-1 inhibitor pembrolizumab to neoadjuvant chemotherapy). The model simulated whether and when each virtual patient experienced a recurrence, based on patterns observed in clinical trial data. The model was rigorously checked against independent real-world cancer registry data to confirm it was producing realistic predictions. This type of model lets researchers incorporate variables such as a patient's specific history that are hard to represent with population-level models.
As lead author Dr. Sumeyye Samur explains, this kind of tool didn't previously exist: "To our knowledge, this is the first clinically validated microsimulation model to integrate the natural history of operable Stage II and III TNBC with current treatment and post-treatment surveillance. Our study quantifies the performance and resource use of current TNBC surveillance and provides a validated modeling platform for evaluating whether emerging approaches, including biomarker-based strategies such as circulating tumor DNA, could improve recurrence detection and patient outcomes."
What the model revealed was a noticeable gap between recurrence and detection of that recurrence for a significant number of patients. Over five years, the simulation projected that 7.4% of TNBC patients would experience a recurrence, but current surveillance methods would catch only 64% of recurrences for stage II cancers and 51% for stage III cancers. About a third of locoregional (breast or nearby lymph nodes) recurrences and 44% of distant recurrences remained undetected within the 5-year period. Current guidelines don't recommend routine imaging for distant recurrence unless a patient develops symptoms, which means a substantial share of metastatic cancer can go unnoticed.
The model also revealed the discrepancy between how resource intensive the current surveillance approach is and what it catches. Detecting a single locoregional recurrence required, on average, 399 mammograms and five follow-up biopsies. Catching one distant recurrence took 279 imaging tests and 39 biopsies. That's a massive amount of testing, healthcare cost and patient anxiety for a system that, based on the model projections, still misses roughly 40% of recurrences overall within 5 years in the simulation. Sensitivity analysis showed that the model’s detection rate depended heavily on patient adherence to mammography schedules, highlighting one potentially modifiable way to improve outcomes.
On the bright side, a new generation of detection tools is emerging. Circulating tumor DNA (ctDNA) blood testing, which can detect molecular residual disease before a recurrence becomes clinically or radiographically visible, has shown promise in other cancers for catching relapse earlier than imaging alone. Through ctDNA testing was not used for detection in this model, the model incorporated ctDNA status as marker of each patient’s underlying recurrence risk. This lays the groundwork for testing whether such a biomarker-based strategy could close the detection gap.
That's precisely where Samur says the team is headed next: "Our findings raise an important question: Can emerging surveillance approaches, such as ctDNA-based minimal residual disease testing, close the recurrence-detection gap identified in our study? We plan to use this validated model to evaluate their potential clinical and health-economic impact relative to the current care."
Dr. Gebra Cuyun Carter, Health Economics & Outcomes Research Director for Abbott’s cancer diagnostics business, weighs in: “This study helps fill that gap by establishing a benchmark for current surveillance. By better understanding where today’s approach performs well and where gaps remain, we can more clearly evaluate whether emerging tools could help detect recurrence earlier or more effectively. Future research can explore whether more sensitive, risk-adapted strategies could identify signs of recurrence sooner and, importantly, whether earlier detection can ultimately create meaningful opportunities to improve patient care.”
Fred Hutch/University of Washington/Seattle Children’s Cancer Consortium Member Dr. Heather Parsons contributed to this research.
This spotlighted research was supported by Exact Sciences Corporation.
Samur S, Hoban O, Gu NY, Carter GC, Palomares M, Arrick B, Gursel E, Yildiz Y, Ayer T, Chhatwal J, Parsons HA. 2026. Recurrence detection in patients with triple-negative breast cancer following the current standard of care: a microsimulation model. BMJ Open. doi: 10.1136/bmjopen-2025-115240.
Science Spotlight writer Kelly Mitchell is a postdoctoral fellow in the Paddison Lab at Fred Hutch Cancer Center. She utilizes live cell reporters and CRISPR screening to study how glioblastoma cancer cells resist chemotherapy and radiation treatment. She obtained her PhD in cellular biology from Albert Einstein College of Medicine.