Can the Mystery of Late Recurrence in Breast Cancer Be Solved with the Oncotype DX Test?

TAILORx and RxPONDER Are No Longer Just About Chemotherapy Decisions: They Are Becoming a Way to Unravel the Molecular Mystery of Late Breast Cancer Recurrence

The Greatest Mystery of HR+/HER2-Negative Breast Cancer: Late Recurrence

Hormone receptor-positive, HER2-negative breast cancer is generally regarded as a slower-growing disease that is sensitive to endocrine therapy. Although true, this is an incomplete view. In this subtype, recurrence risk is not concentrated only in the first 5 years; some patients may develop distant metastatic recurrence even after 10, 15, or 20 years. Thus, HR+/HER2-negative breast cancer may appear quieter early on, but display biologically persistent behavior during long-term follow-up.

Today, treatment of early-stage breast cancer has become more personalized through options such as adjuvant chemotherapy, endocrine therapy, and CDK4/6 inhibitors. However, a fundamental question remains incompletely answered: Why do some patients experience recurrence years later despite appropriate treatment, while other patients with similar clinical characteristics never recur?

EA1241 focuses precisely on this unknown. By examining tissue samples from prospective, well-characterized clinical studies with long follow-up, such as TAILORx and RxPONDER, using modern molecular analysis and artificial intelligence tools, the researchers will try to identify the biological signature of late recurrence.

They Answered the Question of Who May Not Need Chemotherapy

In early-stage HR+/HER2-negative breast cancer, TAILORx and RxPONDER are two of the most important studies of the de-escalation era. De-escalation means reducing unnecessary treatment burden without compromising treatment success. One of the clearest examples in breast cancer is identifying patients who will not actually benefit from chemotherapy and safely managing them with endocrine therapy alone.

TAILORx showed how the 21-gene recurrence score can guide chemotherapy decisions in lymph node-negative, HR+/HER2-negative early breast cancer. Overall, it demonstrated that chemotherapy can safely be omitted in many patients with a recurrence score of 0–25 and endocrine therapy alone may suffice; risk was interpreted more cautiously in some younger/premenopausal subgroups.

RxPONDER extended this question to lymph node-positive patients. In HR+/HER2-negative patients with 1–3 positive axillary lymph nodes and recurrence scores of 0–25, menopausal status was decisive: adding chemotherapy showed no clear benefit in postmenopausal women, whereas a signal of chemotherapy benefit was identified in premenopausal patients.

Primary Tumor and Recurrence Tissue Will Be Compared in the Same Patient

EA1241 is a translational study aiming to match primary tumor samples obtained from TAILORx and RxPONDER participants with recurrence samples to be collected from patients who later developed recurrence. Enrollment of 600 participants is planned. These participants will either have experienced breast cancer recurrence or will be considered a group of special interest because of high genomic risk.

Participants will consent to their recurrence tissues being included in the biobank system and will provide annual clinical information for 5 years. Molecular analyses will therefore not remain solely laboratory data, but will be matched with the patient's actual clinical course.

New Layers of Data Beyond Oncotype DX

Oncotype DX is a very powerful and clinically validated tool based on the expression of a specific set of 21 genes. Breast cancer biology, however, is not limited to 21 genes. A much deeper risk map can be created by jointly assessing genomic mutations, copy number alterations, RNA expression profiles, the tumor microenvironment, histopathological image patterns, and clonal evolution developing under treatment pressure.

In EA1241, Caris Life Sciences will perform comprehensive sequencing and molecular profiling using the MI Cancer Seek platform. According to the ECOG-ACRIN announcement, these analyses will be evaluated within a broader research framework combining whole-exome analysis, whole-transcriptome analysis, whole-slide imaging, and advanced machine learning approaches.

A Clinical Study No Longer Produces Only Results; It Produces New Science

TAILORx and RxPONDER were clinical studies testing treatment decisions in the traditional sense. Their value, however, did not end when enrollment was completed. Long follow-up, stored tissue samples, and detailed clinical information now give these studies a new scientific function.

EA1241 should therefore be viewed not as a treatment trial, but as a discovery platform . The aim is not to obtain approval for a drug directly today, but to generate data that will develop future risk models, biomarkers, and recurrence prevention strategies.

Extending Endocrine Therapy Is Not Suitable for Every Patient

In HR+/HER2-negative early breast cancer, endocrine therapy reduces recurrence risk. As treatment duration increases, however, hot flashes, joint pain, sexual health problems, bone loss, cardiometabolic effects, and the burden on quality of life may increase. An approach of “longer endocrine therapy for every patient” is therefore not ideal.

Today, clinicians decide on extended endocrine therapy by assessing clinical risk, pathological features, lymph node status, tumor size, grade, age, menopausal status, genomic scores, and patient tolerance together. However, this is still not a perfect selection process. A better understanding of the biology that truly determines late recurrence may clarify who needs longer treatment, closer monitoring, or new targeted prevention strategies.

Combining Pathology Images, Clinical Data, and Molecular Profiles

Artificial intelligence is not a magic solution on its own in cancer research. However, it can be a powerful tool for combining multilayered data. In studies such as TAILORx and RxPONDER, clinical outcomes were followed prospectively, tumor tissues were biobanked, and treatment decisions were made within a standardized framework. Such datasets are highly valuable for AI models to learn multilayered risk relationships rather than images alone or genes alone.

Early analyses previously presented by Caris and ECOG-ACRIN using TAILORx data suggested that AI models combining clinical, molecular, and histopathological data could provide more powerful information for predicting recurrence risk than current methods. EA1241 has the potential to take this approach a step further by also including recurrence tissues in the analysis.

It Will Influence Tomorrow's Risk Map, Rather Than Today's Patient

This initiative does not immediately offer patients a new treatment. In the long term, however, it may mean better risk classification, more accurate treatment selection, and more personalized follow-up strategies for patients diagnosed with early-stage breast cancer.

One of the hardest questions for patients in particular is: “My treatment is over, but will this disease return years later?” Today, we answer this question using statistical risks. Studies such as EA1241 may allow us to provide more molecular, more personal, and more predictive answers in the future.

Studies That Changed Practice Are Now Seeking Answers to Questions That Will Reshape Our Understanding of Biology

A New Research Agenda to Prevent Late Recurrence

From Studies That Reduced Chemotherapy to Science That Prevents Late Recurrence

TAILORx and RxPONDER made chemotherapy decisions more informed in early-stage HR+/HER2-negative breast cancer and spared many patients the burden of unnecessary chemotherapy. The same studies are now taking on a new scientific role through stored tumor tissues, long-term clinical follow-up, and modern molecular technologies.

EA1241 will seek the molecular causes of late recurrence that can develop 10–20 years later by comparing primary tumor and recurrence tissue in the same patient. This may reshape not only initial treatment in early-stage breast cancer, but also long-term recurrence prevention strategies.

The future of breast cancer research is moving beyond the question “Should this patient receive chemotherapy?” toward “Can the biological seeds that will create a recurrence risk in this patient 15 years later be seen today?” The transformation of TAILORx and RxPONDER into a discovery platform can, in this respect, be read as the second major phase of personalized treatment in early breast cancer.