Long-term results from one of oncology’s most closely watched breast cancer trials, published Wednesday in Nature Medicine, show that a subset of patients with HER2-positive early breast cancer can maintain five-year survival rates above 92% without ever receiving chemotherapy — but the field’s most rigorous experts are careful to draw a precise distinction: PHERGain demonstrates with striking clarity who might be spared chemotherapy, while the formal proof that omitting it is safe remains a question only a randomized trial can answer.

The five-year analysis of the PHERGain phase 2 trial found that patients whose tumors showed early metabolic response on PET imaging — and who subsequently achieved a pathological complete response (pCR, meaning no detectable residual cancer at surgery) — reached an invasive disease-free survival (iDFS) rate of 92.4% after five years, with just one distant recurrence recorded in that group over a median follow-up of 5.6 years. The paper also introduces long-term circulating tumor DNA (ctDNA) data from liquid biopsy, showing that molecular signals drawn from routine blood samples correlate with treatment response and long-term outcomes.

The study was led by José Manuel Pérez-García, Javier Cortés, Antonio Llombart-Cussac, and colleagues from MEDSIR, an international oncology research network based in Spain.

HER2-Positive Breast Cancer: Greatest Success Story and Its Next Frontier

HER2-positive breast cancer accounts for roughly 15–20% of all breast cancer diagnoses globally, translating to an estimated 47,000–63,000 new US cases each year. The introduction of trastuzumab (Herceptin) in the late 1990s fundamentally altered outcomes for patients whose tumors overexpress the HER2 receptor protein — a milestone that transformed a particularly aggressive subtype into one that responds well to targeted treatment. The subsequent approval of pertuzumab, which binds to a different domain of HER2 and works synergistically with trastuzumab by blocking a separate dimerization pathway, further improved pathological response rates.

Today, dual HER2 blockade with trastuzumab and pertuzumab combined with chemotherapy is the established neoadjuvant (pre-surgery) standard of care for most patients with stage II or higher disease. Yet despite these advances, the chemotherapy backbone carries real costs. Regimens incorporating taxane and platinum agents — the most common cytotoxic partners to HER2-directed therapy — can produce peripheral neuropathy, fatigue, hair loss, bone marrow suppression, and cardiac toxicity. For patients whose tumors respond powerfully to targeted therapy alone, those toxicities may be entirely avoidable.

That is the clinical premise PHERGain was designed to test — not whether de-escalation works in general, but whether a specific imaging-and-pathology framework can identify, at the individual patient level, which tumors are so exquisitely sensitive to HER2 blockade that chemotherapy adds nothing to their cure.

How PET Imaging Drives the Decision: The PHERGain Response-Adaptive Design

PHERGain (NCT03161353) was a multicenter, randomized, open-label, non-comparative phase 2 trial conducted across 45 hospitals in seven European countries. The study enrolled patients with centrally confirmed HER2-positive, operable stage I–IIIA breast cancer with tumors measuring at least 1.5 cm (0.6 in), and assigned 356 patients in a 1:4 ratio to a standard chemotherapy arm (Group A, 71 patients) or a response-adapted experimental arm (Group B, 285 patients).

Group B patients began treatment with trastuzumab and pertuzumab alone — no chemotherapy. After two cycles, each patient’s tumor was re-evaluated using 18F-fluorodeoxyglucose positron emission tomography (FDG-PET), a functional imaging technique that measures tumor metabolic activity by tracking the uptake of a glucose analog tracer. Cancer cells, which hypermetabolize glucose relative to normal tissue, accumulate the tracer at higher rates. The result is quantified as a standardized uptake value (SUV); PHERGain defined a metabolic response as a ≥40% reduction in peak SUV after just two treatment cycles.

Patients who met that threshold continued targeted therapy alone for six additional cycles. Those who did not were switched to chemotherapy combined with dual HER2 blockade — a built-in safety net ensuring that no patient with a poorly sensitive tumor remained undertreated. At surgery, patients who had achieved an early PET response but had residual disease at pathology received additional chemotherapy alongside continued targeted therapy. Only patients who achieved both an early metabolic response and pCR at surgery completed treatment without any chemotherapy exposure.

This architecture distinguishes PHERGain from earlier, cruder de-escalation attempts: the decision point is made prospectively, mid-treatment, on the basis of an objective biological signal — not on pre-treatment risk stratification alone.

Five-Year Outcomes: One Distant Recurrence in 86 Patients Over 5.6 Years

After a median follow-up of 5.6 years, the Group B results were striking in their durability. Among the 267 patients who underwent protocol-defined surgery, five-year iDFS was 89.5% (95% CI 85.7–93.4). Five-year distant disease-free survival — focused specifically on spread to distant organs — reached 92.7% (95% CI 89.5–96.0). Overall survival across the full randomized population was 98.1% in Group B versus 98.4% in Group A, a negligible numerical difference.

The most selective subgroup — the 86 patients who achieved both an early PET metabolic response and pCR at surgery, and who received no chemotherapy at any point — produced the headline finding. Five-year iDFS in this cohort reached 92.4% (95% CI 86.7–98.5). Distant disease-free survival was 98.7% (95% CI 96.3–100.0). No deaths were recorded in this entirely chemotherapy-free population during the five-year analysis. Crucially, just one distant recurrence was documented in this group over the entire follow-up period — a 1.2% distant relapse rate.

Approximately 30% of Group B patients were ultimately able to avoid chemotherapy entirely — suggesting that, in a real-world HER2-positive early breast cancer population treated at specialized centers, meaningful numbers of patients could potentially be spared cytotoxic treatment if the PHERGain selection model were applied.

Liquid Biopsy Adds a Molecular Layer: ctDNA Data and Their Limits

Beyond the imaging-and-pathology framework, the PHERGain five-year publication introduces a parallel biomarker story built around liquid biopsy. Investigators analyzed 932 plasma samples collected at baseline, after two treatment cycles, and immediately before surgery, using the Guardant Reveal assay — a tissue-free, methylation-based ctDNA test that detects tumor-derived DNA fragments in the bloodstream without requiring prior sequencing of the patient’s tumor.

Baseline landscape: ctDNA was detectable in 71.0% of evaluable patients before treatment. Detection tracked closely with disease burden — only 33.3% of stage I patients had detectable ctDNA at baseline, compared with 93.2% of stage III patients. Among node-positive patients, 87.8% had detectable baseline ctDNA, versus 55.0% of node-negative patients.

Early clearance and pCR: Among patients with detectable baseline ctDNA, 76.0% had cleared it after just two treatment cycles. Among those patients, the pCR rate was 48.4% in those who achieved early ctDNA clearance, versus 22.0% in those with persistent ctDNA (p=0.003). Before surgery, among 28 patients with still-detectable ctDNA, none achieved pCR — though this finding comes from a small subset and should be interpreted cautiously given the limited sample size.

Five-year prognostic correlation: Patients without detectable baseline ctDNA had five-year iDFS of 97.1%, compared with 86.7% among those with detectable ctDNA at baseline (HR 3.1; p=0.021). Detectable ctDNA after two cycles was associated with five-year iDFS of 77.5%, versus 94.4% in those who had cleared it (HR 4.3; p=0.0011).

The critical statistical caveat: These raw hazard ratios look compelling — but the investigators’ own analysis showed that after statistical adjustment for established prognostic factors (disease stage and nodal status), the association between early ctDNA clearance and iDFS was not statistically significant. This is the finding that separates a promising exploratory signal from a validated biomarker. ctDNA clearance strongly correlates with the same favorable tumor characteristics (early-stage disease, low tumor burden) that already predict good outcomes — making it difficult to determine whether ctDNA adds independent predictive information or is simply mirroring what stage and nodal status already tell us.

The PET scan and liquid biopsy also captured overlapping biology: 90.5% of PET responders had achieved early ctDNA clearance, compared with just 56.1% of PET non-responders, a concordance suggesting the two biomarkers track similar tumor sensitivity features.

As principal investigator Dr. Javier Cortés of the International Breast Cancer Center in Madrid and Barcelona described it: “In addition to the clinical outcomes, with nearly 90% of patients free from relapse five years after surgery, we see that the analysis of circulating tumour DNA in blood is emerging as a key tool to enable early identification of patients with a better prognosis and those who may benefit from more intensive treatment.”

What PHERGain Proves and What It Does Not

The distinction that matters most for clinical interpretation of PHERGain is this: the trial demonstrates with exceptional clarity who can be identified as an exceptional HER2-positive early responder. It does not formally demonstrate that withholding chemotherapy from those patients produces equivalent outcomes to giving it.

That distinction is not a matter of excessive caution. PHERGain was designed as a non-comparative phase 2 trial — which means it had neither the statistical power nor the trial architecture to formally test whether omitting chemotherapy is equivalent to including it. The trial was designed to estimate rates and demonstrate proof-of-concept, not to prove noninferiority. A randomized controlled trial, in which patients confirmed as early responders are randomly assigned to chemotherapy or no chemotherapy, would be needed to formally answer the equivalence question — and that trial has not yet been completed.

The favorable outcomes in the chemotherapy-free group reflect a genuine confound: patients who achieve pathological complete response are already biologically the best-responding tumors in any HER2-positive cohort. Their high sensitivity to HER2-directed therapy is itself a predictor of superior long-term prognosis, independent of whether chemotherapy was administered. Formally separating the de-escalation benefit from the underlying biology of highly responsive tumors requires the randomized design PHERGain does not have.

Similarly, the ctDNA analyses were explicitly exploratory. The study was not designed — and was not powered — to test whether modifying treatment based on ctDNA results improves outcomes; it demonstrated only that ctDNA clearance status was correlated with them. And as noted above, that correlation lost statistical significance after adjustment for known prognostic factors. Circulating tumor DNA standardization across laboratories and methods also remains a work in progress, limiting direct translation to clinical decision-making without validated thresholds.

There is also a practical clinical context issue: PHERGain enrolled patients between 2017 and 2019, before trastuzumab emtansine (T-DM1) became the standard adjuvant therapy for HER2-positive patients with residual disease after neoadjuvant treatment. The KATHERINE trial subsequently demonstrated that T-DM1 reduces the risk of recurrence by 46% in patients with residual disease (seven-year iDFS: 80.8% with T-DM1 versus 67.1% with trastuzumab alone). PHERGain’s residual-disease patients received less effective adjuvant treatment than today’s standard — which means the trial’s outcomes cannot be directly transposed onto the modern treatment landscape.

Despite these limitations, the five-year data are meaningfully reassuring in one critical respect: the chemotherapy-free strategy, as practiced in PHERGain, did not produce a pattern of late recurrences that might have undermined earlier three-year follow-up data. The one-distant-recurrence finding in 86 patients over 5.6 years is durable, not a short-term artifact.

A Practical Barrier: Not Every Cancer Center Has PET Capability

One constraint the PHERGain results do not resolve is infrastructure. The PHERGain response-adaptive model depends on access to FDG-PET imaging — a technology that requires a cyclotron to produce the radioactive tracer, specialized scanner hardware, and nuclear medicine expertise to interpret results quantitatively. These resources are concentrated at academic cancer centers and major community hospitals; they are not uniformly available across the cancer care ecosystem.

For patients receiving care at facilities without PET capability, the PHERGain approach as designed could not be implemented. As the field moves toward validating and potentially adopting response-adaptive de-escalation strategies, infrastructure equity will be a genuine access concern — particularly for patients in rural or underserved communities.

How Close Are We to Knowing Who Does Not Need Chemotherapy?

PHERGain’s publication arrives in a breast cancer research landscape that is rapidly converging on the same question from multiple directions. The companion PHERGain-2 trial, presented at ESMO Breast 2026 in May and simultaneously published in Annals of Oncology, enrolled 396 patients across 47 centers in six European countries and extended the de-escalation concept to even lower-risk HER2-positive patients — those with tumors between 5 mm (0.2 in) and 30 mm (1.2 in), no nodal involvement, and high HER2 protein expression (IHC 3+). PHERGain-2 observed a pathological complete response rate of approximately 60% in this selected low-risk population, with quality-of-life data supporting tolerability of the chemotherapy-free approach. Its primary efficacy endpoint is still pending.

Several other ongoing trials are pursuing related questions in this space. ATEMPT 2.0 is testing T-DM1 against paclitaxel plus trastuzumab in low-risk HER2-positive patients. ADEPT and CompassHER2-pCR are studying whether pCR response can formally guide adjuvant therapy de-escalation. The ADAPT trial’s HER2+/HR+ cohort is examining response-adapted therapy with endocrine approaches. None has yet reported formal noninferiority findings that would independently justify de-escalation in clinical practice.

Together, the PHERGain program represents one of the most mature and methodologically coherent efforts in oncology to answer the question of how much treatment is actually enough — and what kind of evidence it takes to know.

What PHERGain Means for Patients Today

For now, PHERGain’s data support an investigational posture, not a change to clinical practice. No major US or international breast cancer guidelines — from NCCN, ASCO, or ESMO, or the St. Gallen International Consensus — have endorsed routine chemotherapy omission for HER2-positive early breast cancer patients based on PHERGain data alone. Prospective, randomized trials with formal noninferiority designs remain the necessary evidentiary next step.

For patients who are identified as early exceptional responders — particularly those being treated at academic centers with PET imaging capability and access to clinical trial enrollment — PHERGain’s five-year data provide a meaningful evidentiary foundation for discussing chemotherapy-sparing approaches with their oncology team. Whether such a discussion is appropriate for any individual patient depends on their specific tumor characteristics, disease stage, response assessment, and clinical context.

The broader message from PHERGain’s five-year follow-up is precisely calibrated: the window for chemotherapy-free treatment in HER2-positive early breast cancer is biologically real, and it is durable over five years. What remains unknown is whether opening that window for patients outside carefully designed trial settings will produce outcomes equally good as the PHERGain results suggest — or whether the trial’s exceptional results reflect, in part, the exceptional selectivity of its patient identification process.

Frequently Asked Questions

Can HER2-positive breast cancer patients skip chemotherapy based on these results?

Not yet based on current guidelines. PHERGain’s five-year data are highly encouraging, but the trial was a non-comparative phase 2 study — it was designed to estimate outcomes in a selected group, not to formally prove that skipping chemotherapy is equivalent to receiving it. Major guidelines from NCCN, ASCO, and ESMO have not yet endorsed chemotherapy omission based on PHERGain data alone. Patients who are strong early responders and are being treated at centers with PET imaging capability may have the option to discuss this data with their oncologist in the context of clinical trial enrollment.

What is pathological complete response (pCR), and why does it matter for de-escalation?

Pathological complete response means no detectable residual cancer is found in the breast or lymph nodes when the tissue is examined after surgery. In HER2-positive breast cancer, achieving pCR is strongly associated with better long-term outcomes. In PHERGain, pCR status — combined with early PET metabolic response — was the gateway for patients to complete treatment without chemotherapy. The important caveat is that patients who achieve pCR are already biologically the most favorable responders, so their good outcomes may partly reflect their tumor biology rather than the de-escalation decision itself.

What is circulating tumor DNA (ctDNA), and is it ready to guide breast cancer treatment decisions?

Circulating tumor DNA is tumor-derived DNA found in the bloodstream. In PHERGain, a blood-based ctDNA assay (Guardant Reveal) was used to track molecular response alongside PET imaging. The results showed that patients who cleared ctDNA early had higher pCR rates and better five-year outcomes. However, after adjusting for known prognostic factors like stage and nodal status, the association between ctDNA clearance and survival was not statistically significant — meaning it may be mirroring what tumor stage already predicts. ctDNA liquid biopsy is not currently cleared by the FDA to guide treatment modification decisions in early breast cancer; it remains a research tool in this context.

How does PHERGain’s approach compare to standard treatment today?

Standard treatment for most HER2-positive early breast cancer patients currently includes dual HER2 blockade (trastuzumab + pertuzumab) combined with chemotherapy before surgery, followed by trastuzumab continuation after surgery. Patients with residual disease at surgery now typically receive trastuzumab emtansine (T-DM1) as adjuvant therapy — a practice that became standard after the KATHERINE trial demonstrated a 46% reduction in recurrence risk. PHERGain’s patients enrolled before T-DM1 was standard for residual disease, which means the trial’s residual-disease patients received less intensive follow-up treatment than patients would receive under today’s standard of care — a key reason the results cannot be directly transposed onto the current treatment landscape.