¹⁸F-FDG PET “Metabolic Response” to Neoadjuvant Systemic Therapy for Breast Cancer: Quo Vadis?

In their article in The Journal of Nuclear Medicine, Hennessy et al. (1) provide mature overall survival data on their neoadjuvant chemotherapy-free regimen of pertuzumab plus trastuzumab given to patients with localized human epidermal growth factor receptor type 2 (HER2)–positive breast cancer, monitored by ¹⁸F-FDG PET/CT, for which they previously reported a 34% pathologic response rate linked to an SUV (normalized by lean body mass) decline by at least 40% after 15 d of exposure to dual HER2 blockade.

The paper offers a unique opportunity to discuss the excitement of the oncology community about neoadjuvant trials, the importance of evaluating overall survival in addition to pathologic complete response (pCR), and the interest in this approach to explore treatment deescalation using molecular imaging.

Neoadjuvant systemic therapy has been initially applied as a downstaging procedure in patients presenting with surgically nonresectable breast tumors or tumors deemed unsuitable for breast conservation. The interest of the oncology community in this approach started to grow markedly when it became clear that the magnitude of response to neoadjuvant chemotherapy was associated with relapse-free survival (2) and when the seminal NSABP-B18 trial demonstrated similar outcomes for 4 cycles of doxorubicin-cyclophosphamide given before or after surgery (3).

At the turn of the century, many investigators embarked on the design and conduct of chemotherapy-based neoadjuvant trials with the choice of pCR as the primary endpoint for the comparison of newer, more sophisticated drug regimens. These efforts culminated in a pooled analysis of 12 trials involving 11,955 patients, all treated with preoperative cytotoxic drugs (few trials used targeted agents) (4). This analysis robustly confirmed a positive correlation, at the patient level, between pCR and event-free and overall survival. It also demonstrated that the best definition of pCR was the lack of invasive tumor in the breast and in the nodes. Disappointingly, however, pCR could not be proposed as a surrogate endpoint for event-free or overall survival, since at the trial level, there was only a marginal association between them. The dream of an accelerated new drug development for early breast cancer based on incremental gains in pCR, which can be achieved in a short time frame contrary to adjuvant trials, was not abandoned: trialists remained convinced that with powerful targeted drugs, such as the anti-HER2 monoclonal antibody trastuzumab, given on top of chemotherapy, larger differences in pCR would be predictive of better long-term outcomes at a population level. They embarked on a second pooled analysis, this time of 11 trials and 3,710 patients, all with HER2-positive breast cancer, but again, a weak association was found between pCR and long-term outcome at the trial level (5). An important finding, however, was that tumor burden (e.g., tumor size and nodal status) remains prognostic, even in patients reaching pCR, who can still experience relapse.

All this historical background has important implications for the new field of excitement in breast oncology: namely the identification of biomarkers able to predict, early on (e.g., after 2–6 wk of therapy), whether a patient will benefit from the selected drug regimen administered preoperatively: the hope, here, is a truly individualized approach, which will consist either of treatment escalation or of treatment deescalation based on observed biomarker changes.

¹⁸F-FDG PET is the most studied noninvasive and dynamic biomarker in this field, but so far, many trials such as Neoaltto (6) have suffered from the selection of a suboptimal endpoint, pCR, and from weak statistical power preventing robust correlations with long-term survival: the latter is what matters to patients, who will certainly welcome the news of complete tumor disappearance locally but who will continue to worry about their risk of distant relapse and death.

Two recent trials for HER2-positive breast cancer represent a step forward on the path toward chemotherapy deescalation based on early ¹⁸F-FDG PET metabolic responses.

The first is the single-arm U.S. trial TBCRC026 presented by Hennessy et al. (1). A strength of this trial is that it enrolled a relatively homogeneous population of 88 patients with HER2-positive/hormone receptor–negative tumors, known to be more addicted to the HER2 pathway than tumors coexpressing HER2 and hormone receptors and treated with dual HER2 blockade, namely trastuzumab plus pertuzumab, without chemotherapy and without endocrine therapy. With additional follow up, the study demonstrated that a maximum SUV (normalized by lean body mass) of no more than 3 at 2 wk was associated with improved relapse-free survival and overall survival, with the latter being statistically significant. Although this result is encouraging and underscores the potential role of molecular imaging in identifying a subgroup of patients who may safely forego chemotherapy, the trial has weaknesses: relapse-free survival and overall survival were only secondary endpoints, and postsurgical adjuvant therapy was left at the discretion of the oncologist; as a result, one third of patients who achieved pCR (6/18) nevertheless went on to receive adjuvant chemotherapy combined with HER2-directed therapy. This leaves some uncertainty regarding the robustness of the ¹⁸F-FDG PET response to predict which patients may indeed be spared chemotherapy.

The second trial (PHERGAIN), conducted in Europe, also chose ¹⁸F-FDG PET as an early, dynamic biomarker for the selection of patients suitable for dual HER2 blockade without chemotherapy (7). After 2 cycles of combined trastuzumab and pertuzumab given over 6 wk, 356 patients were randomized 1 to 4 to receive either standard chemotherapy with trastuzumab-pertuzumab or a biomarker response adaptive regimen in which a 40% or greater decline in SUV_max at 6 wk would allow for the continuation of targeted therapy (trastuzumab-pertuzumab) without chemotherapy until surgery: not only did these patients achieve an impressive pCR rate of almost 40% (86/227), but with longer follow up, only one local relapse was documented at a median follow-up of 3 y (8). Here, the strengths include a strict policy for adjuvant therapy, which did not authorize chemotherapy in pCR patients, and the selection of invasive disease-free survival as a coprimary endpoint together with pCR, but a particular feature is that the primary focus is the entire response-guided arm, which includes patients who went on to receive chemotherapy in view of their insufficient biomarker response. Nevertheless, the null hypothesis set by the investigators could be rejected given the excellent 95.4% invasive disease-free survival observed in the response-adapted arm (it was required to be ≥95%). Of note, the 3-y invasive disease-free survival was 98.3% in the biomarker-ignorant arm in which all patients received chemotherapy.

Can we derive lessons from these experiences and foresee new avenues of progress?

With the need to focus on truly important endpoints, namely event-free and overall survival, future trials will need to be large and therefore will require more extensive international collaboration. Standardization of PET imaging techniques among trial centers should not be a barrier in the 21st century, with the medical community claiming an ambition of delivering precision oncologic therapies.

Further progress could also emerge from the adjunct of circulating tumor DNA molecular responses to molecular imaging: indeed, a lack of circulating tumor DNA eradication in the neoadjuvant setting has been shown to predict a poor outcome (9). However, less than 50% of patients undergoing neoadjuvant chemotherapy have detectable circulating tumor DNA in their serum (10). Another potentially informative biomarker is represented by tumor-infiltrating lymphocytes, which have been reported to complement the value of metabolic response (11).

With some extra collaborative efforts, early response assessment with molecular imaging should soon leave the scene of clinical trials and enter routine clinical care.

• © 2023 by the Society of Nuclear Medicine and Molecular Imaging.

• Received for publication July 31, 2023.
• Revision received September 27, 2023.