The PHERGain randomized phase 2 trial evaluated an 18F-FDG-PET-guided, pCR-adapted strategy to omit chemotherapy in operable HER2-positive early breast cancer. The study randomized patients with centrally confirmed HER2+ stage I–IIIA tumors of at least 1.5 cm to either standard chemotherapy plus trastuzumab and pertuzumab (group A) or trastuzumab and pertuzumab ± endocrine therapy with PET-guided adaptation (group B). After two cycles of dual HER2 blockade, the study team used PET response and subsequent surgical pathological response to decide whether to give chemotherapy. The authors previously reported a 3-year invasive disease-free survival (iDFS) of 94.3% in group B.
With a median follow-up of 5.6 years, investigators reported the prespecified secondary endpoint: 5-year iDFS for the adaptive arm (group B) was 89.5% (95% CI: 85.7 to 93.4) among 267 patients. The cohort included a post-randomization subgroup composed of patients who were PET responders after two cycles and who achieved pCR at surgery without having received chemotherapy; in that subgroup (n = 86) the 5-year iDFS was 92.4% (95% CI: 86.7 to 98.5). The study team noted one distant recurrence in this subgroup. Because that estimate is based on a subgroup defined after randomization by PET response and pCR, the authors framed it as descriptive and warned against interpreting it as causal evidence for safely omitting chemotherapy.
This set of outcomes provides longer-term context for the de-escalation approach. The overall 5-year iDFS in the adaptive arm declined modestly from the earlier 3-year estimate, while the PET-responder/pCR subgroup retained high event-free rates. Those findings speak to both the feasibility of selecting patients for chemotherapy omission and the need for cautious interpretation when post-randomization characteristics define subgroups.
PHERGain used an early imaging biomarker—18F-FDG-PET—to classify metabolic response after two cycles of trastuzumab plus pertuzumab ± endocrine therapy. Patients who did not show a PET response were scheduled to receive chemotherapy (docetaxel and carboplatin) before surgery. Patients who were PET responders continued a chemotherapy-free neoadjuvant course and were then assessed for pCR at surgery; PET responders who did not achieve pCR also received chemotherapy per the protocol. The first primary endpoint of pCR was previously reported: among PET responders who continued on the chemotherapy-free regimen, 37.9% achieved pCR at surgery. The trial therefore combined an early metabolic readout with pathological assessment to decide on escalation to chemotherapy.
This stepwise strategy shifts the decision point for chemotherapy from baseline risk features to on-treatment response. The investigators reported long-term iDFS data to evaluate whether that approach produced acceptable outcomes when chemotherapy was omitted for selected patients.
Investigators performed exploratory analyses of circulating tumor DNA (ctDNA) sampled at baseline and during treatment. Data indicated that detectable ctDNA either at baseline or during therapy was associated with inferior iDFS. By contrast, patients who cleared ctDNA during the treatment course showed higher rates of pCR and better long-term outcomes. The source did not report specific assay methods, thresholds for detection, or numerical hazard estimates for ctDNA associations in this summary; those methodological details were not provided in the article text excerpted here.
These exploratory observations suggest ctDNA could be an additional biomarker to refine selection for chemotherapy omission, but the findings remain hypothesis-generating. The authors did not present prespecified ctDNA-guided treatment rules in this report.
PHERGain was a randomized, non-comparative phase 2 trial; group B used adaptation by PET and pCR rather than direct randomization against a matched no-chemotherapy strategy. The reported 5-year iDFS estimate for the PET-responder/pCR subgroup is derived after randomization and is therefore descriptive. The authors explicitly stated this caveat, indicating the estimate should not be read as definitive evidence that omitting chemotherapy causes similar long-term outcomes. The source did not provide comparative hazard ratios against group A in the excerpt available here, nor did it furnish p-values or sample-size calculations for the 5-year endpoint in this text.
Readers should therefore view the long-term subgroup data as supportive of feasibility rather than conclusive proof of noninferiority. The adaptive framework produced encouraging event-free rates in patients selected by early PET response and pCR, but randomized comparative data with prespecified noninferiority margins would be required to establish equivalence of omission strategies.
The trial investigators noted the descriptive nature of the subgroup outcome and emphasized that the subgroup is defined post-randomization by on-treatment response. The source did not report detailed ctDNA assay parameters or provide comparative statistics versus the standard-chemotherapy arm in the excerpt available. Whether ctDNA can be reliably integrated into clinical decision algorithms remains an open question pending validation and standardized assay methodology. ClinicalTrials.gov identifier: NCT03161353.