Reply to “Critical Evaluation of HER2 PET Imaging in Metastatic Breast Cancer: Addressing Key Challenges and Future Directions” and “The Role of HER2 PET in Treatment of Breast Cancer”

REPLY: We want to thank Drs. Song and Ye and Drs. Bading and Mortimer for sharing their insightful perspective and commentary in their Letters to the Editor, addressing the key limitations of our study (1), and highlighting the important knowledge gaps of human epidermal growth factor receptor 2 (HER2)–targeted PET radiotracers in patients with metastatic breast cancer (mBC). We fully agree there are several major limitations in our study analysis and, on a broader level, our collective knowledge as a scientific community of HER2-targeted PET imaging, mechanisms of action in vivo, and potential clinical applicability in patients with mBC. Further research is certainly necessary to explore the mechanisms behind HER2 PET uptake in certain mBC lesions and, equally importantly, the lack of uptake in other lesions, correlation with HER2 status on pathology, and the reasons behind discordant findings between HER2 PET and pathology.

Regarding false-positives on HER2 PET discussed by the letter by Song et al., our main goal for undertaking this retrospective reanalysis of our prospective clinical trials of ⁸⁹Zr-labeled HER2-targeted monoclonal antibodies (⁸⁹Zr-HER2-mAbs) was to attempt to explain the high false-positive rates for HER2-positive malignancy in our prior HER2 PET trials, with most biopsied HER2 PET–avid lesions resulting in HER2-negative pathology (2–4). We were discouraged by these confusing results and speculated during this time that this discordance may be due to several factors such as biopsy issues (i.e., sampling error and bone pathology decalcification), nonspecific binding of free ⁸⁹Zr, or differences of HER2 affinity (2). However, these 2 trials were conducted in 2017 and 2020, before the landmark DESTINY-Breast04 study in 2022 that showed unprecedented efficacy of the HER2-targeted antibody–drug conjugate (ADC) trastuzumab deruxtecan (T-DXd) in patients with HER2-low mBC and thereby established HER2-low status as an important and separate classification beyond the traditional binary classification of HER2-negative and HER2-positive used for many years (5). During our first-in-human trial of our newest ⁸⁹Zr-HER2-mAbs, ⁸⁹Zr-site-specific (ss)-pertuzumab, in 2021–2022 (6), we noticed a similar pattern of discordances when correlating HER2 PET–positive lesions with pathology and hypothesized that ⁸⁹Zr-HER2-mAbs may actually detect a wide spectrum of HER2 expression, without the ability to differentiate between HER2-positive and HER2-negative as defined by a pathology classification on immunohistochemistry/fluorescence in situ hybridization, and HER2-low lesions may explain the high false-positive rates in our prior trials. Published in the midst of our trial, the dramatic results of DESTINY-Breast04 opened our eyes to HER2-low as a plausible explanation. Thus, we went back and reanalyzed paired biopsy samples from our prior trials and current ongoing ⁸⁹Zr-ss-pertuzumab trial to see if prior HER2 PET false-positives were in fact HER2-low lesions. Our results suggest that ⁸⁹Zr-HER2-mAbs detects HER2-expressing lesions (both HER2-positve and HER2-low lesions) and may even detect HER2-zero lesions, with all prior false-positives explained by HER2-low (1). We agree with Song et al. that a thorough comparison of the sensitivity and specificity of ⁸⁹Zr-HER2-mAbs in detecting HER2-low lesions would be ideal and important for clinical application; however, recruiting patients to undergo multiple biopsies of different metastases is difficult and not feasible, especially for an imaging study.

Yet, we acknowledge that our results from this study certainly do not explain the full picture of ⁸⁹Zr-HER2-mAbs nor the pharmacodynamics or complex biologic interactions of these radiotracers with cancer cells, including differences in mAb binding sites on the HER2 receptor (trastuzumab, subdomain IV; pertuzumab, subdomain II), rates of mAb–HER2 internalization and intracellular trafficking, and interactions with the surrounding tumor microenvironment, as mentioned by Song et al. In addition, we did not explore variations in drug delivery and pharmacokinetics of different radiolabeled mAbs, as mentioned by Bading et al., and associated factors such as the enhanced permeability and retention effect; however, we would not expect any differences in enhanced permeability and retention effect given similar molecular sizes of trastuzumab and pertuzumab (∼148 kDa). We have previously reported on the pharmacokinetics, biodistribution, and dosimetry of the 3 ⁸⁹Zr-HER2-mAbs, and there are certainly differences in whole-body and serum biologic clearance, with ⁸⁹Zr-DFO-pertuzumab and ⁸⁹Zr-ss-pertuzumab demonstrating more similar pharmacokinetics, biodistribution, and dosimetry compared with ⁸⁹Zr-trastuzumab (6–8).

We postulate that these differences would likely not significantly influence our primary results that prior HER2 PET false-positives were due to tracer uptake in HER2-low lesions; however, all of the aforementioned factors could affect our surprising findings of HER2-low lesions exhibiting SUV_max higher than that of HER2-positive lesions. In addition, these counterintuitive findings may be explained by SUV differences of different ⁸⁹Zr-HER2-mAbs, organ or site of biopsied lesion, total tumor burden, and line of therapies between the trials and enrolled patients. The longer half-life of ⁸⁹Zr (78.4 h) and the uptake period (5–8 d) of our HER2 radiotracers may also explain why our results conflict with those of Zhao et al. using ^99mTc-NM-02 nanobody SPECT/CT, in which ^99mTc-NM-02 uptake was positively correlated with HER2 expression (9). With the longer time period, ⁸⁹Zr-HER2-mAbs may be localizing to both extracellular membranous HER2 and intracellular HER2 within endosomes (or even degraded by lysosomes into free ⁸⁹Zr in the cytoplasm), the latter which would not be captured by immunohistochemistry analysis. HER2 internalization and intracellular trafficking are essential for effective cytotoxic payload delivery into cancer cells for HER2-targeted ADCs (10), such as T-DXd, and if, in fact, ⁸⁹Zr-HER2-mAbs are partly imaging and quantifying HER2 internalization and not just membranous HER2 expression, then there is rationale and potential for their use as a PET imaging biomarker to improve prediction of HER2-targeted ADC response, in conjunction with pathology HER2 status used in clinical practice. However, this is merely our attempt to postulate potential mechanisms behind HER2 PET uptake and intensity, and further research is needed to advance our understanding of “what exactly we are imaging” with HER2 PET through reverse translational research via preclinical in vitro cell and in vivo animal model studies, which is currently ongoing in the field. In parallel, further exploration of the HER2 internalization, trafficking, and recycling pathway and how this pathway is affected by different HER2-targeted mAbs or cells with varying levels of HER2 expression (i.e., HER2-low vs. HER2-positive) is critical to better designing HER2-targeted drugs and companion imaging agents.

Lastly, we fully concur with Drs. Barding and Mortimer in their statement that our study findings of higher SUV_max in HER2-low compared with HER2-positive lesions may limit the ability of ⁸⁹Zr-HER2-mAbs to assess intertumoral lesion-to-lesion HER2 heterogeneity and predict response to T-DXd and other HER2-targeted therapies. Certainly, their innovative trial demonstrating ⁶⁴Cu-DOTA-trastuzumab HER2 PET predicts response to the HER2-targeted ADC, trastuzumab emtansine, in patients with HER2-positive mBC (11), and similarly, the ZEPHIR trial with ⁸⁹Zr-trastuzumab HER2 PET (12) served as the paradigm and inspiration for our ongoing clinical trial, funded by the SNMMI Mars Shot Research Fund (13), of ⁸⁹Zr-ss-pertuzumab HER2 PET in predicting T-DXd response in patients with HER2-low mBC. We agree that pertuzumab may not as accurately model the biodistribution and drug delivery of T-DXd as trastuzumab would; however, in our experience with 3 ⁸⁹Zr-HER2-mAbs, including ⁸⁹Zr-trastuzumab, the image quality and tumor-to-background signal was best with ⁸⁹Zr-ss-pertuzumab and appears to detect more lesions, with higher lesion conspicuity and intensity (6). In our recently completed ⁸⁹Zr-ss-pertuzumab pan-cancer trial in patients with HER2-positive solid tumors (unpublished), we noticed a promising trend of higher HER2 PET/FDG PET concordance in responders versus nonresponders to T-DXd in the subset of patients who subsequently received this ADC, giving us cautious optimism that this tracer will work in patients with HER2-low mBC.

DISCLOSURE

No potential conflict of interest relevant to this article was reported.

Randy Yeh*

Icahn School of School of Medicine at Mount Sinai, New York, New York

E-mail: randy.yeh@mountsinai.org