Plasma HER2ECD a promising test for patient prognosis and prediction of response in HER2 positive breast cancer: results of a randomized study - SAKK 22/99

Epidermal growth factor receptor-2 (HER2) is an oncogene of key importance in breast cancer (BC). Its amplification occurs in 15–25% of primary BC patients and is age-dependent [1, 2]. HER2 amplification identifies an intrinsic subtype [3] of particularly aggressive BC correlating with higher levels of proteases [4]. HER2 represents the target of monoclonal antibodies, such as trastuzumab (T), pertuzumab and T-DM1, effective in the cure of patients with HER2 positive BC [5, 6] as well as of tyrosine kinase inhibitors (TKIs) [7, 8]. This protein located in the cytosolic membrane is composed of three domains: the internal tyrosine kinase reactive part, the transmembrane lipophilic linker and the extracellular carboxyl-terminal tail called extracellular domain (ECD) [9]. Metalloproteases can cleave the ECD from the cell surface. The HER2_ECD with a length of p105 represents the shed ECD product, which is circulating in blood of patients with BC cells overexpressing this oncogene and high levels of proteases. An enzyme-linked immunosorbent assay has been developed to dose HER2_ECD blood levels and proposed as peripheral marker for monitoring disease progression and predict therapy response [10‐12]. This is of clinical interest since treatment of advanced BC is historically based on the HER2 status of the primary tumor and several studies have suggested a possible clinically significant discordance of up to 42% between the HER2 status at primary and later disease stages [13, 14]. The discrepant HER2 status during the course of the disease is however reported to be minimal (2%) when based on in situ hybridization detections and performed with exactly the same methods in a single center [15].

The HER2_ECD is detectable in serum and plasma, but the majority of studies analyze serum samples. The HER2_ECD test is Food & Drug Administration (FDA) approved and numerous clinical studies have demonstrated that both a HER2_ECD serum level > 15 ng/ml and a reduction of ≥20% between 2 successive blood draws is predictive for significant trastuzumab response [16].

The present study was designed as translational research investigation of the SAKK22/99 clinical study [17] to investigate associations of plasma baseline HER2_ECD values and longitudinal time points’ variations with treatment efficacy to possibly improve identification of patients requiring upfront combination of trastuzumab with chemotherapy (TChemo).

Our translational research study aimed to address the medical need to identify a robust non-invasive marker able to select patients with particular profit from a sequential anti-HER2 chemotherapy treatment strategy. An approach which, according to our results of the randomized Phase III trial SAKK22/99, might help deferring chemotherapy and its toxicity fitting in a de-escalating intention in the treatment of HER2-positive disease [17].

Our results indeed indicate that patients with low baseline HER2_ECD levels may profit of single agent trastuzumab, while no difference between arms is observed when the baseline HER2_ECD levels are higher (Fig. 3c/d). Moreover, since the HER2_ECD levels under trastuzumab monotherapy remain fairly constant over time and start increasing only shortly before disease progression, monitoring HER2_ECD levels may help identifying the optimal moment for introducing chemotherapy.

To our best knowledge, this is the first study investigating plasma HER2_ECD or any other blood marker, including circulating tumor cells (CTCs) or free DNA, to address specifically the optimal time point for the introduction of chemotherapy after anti-HER2 monotherapy in HER2+ metastatic BC.

Several studies have investigated the value of serum HER2_ECD in patients with BC. Our results are in line with a recent review summarizing 27 studies (10 in metastatic BC, 13 in early BC and 4 in patients with early or metastatic BC) [18]. Even though there are discrepancies between studies, including the definition of the optimal threshold, the majority suggest that serum HER2_ECD might be particularly useful in metastatic BC as an indicator of cancer progression and predictor for anti-HER2 therapies efficacy.

The mechanistic reasons why higher levels of HER2_ECD could correlate with poorer prognosis and with less anti-HER2 therapeutic effectiveness are several. First of all, the truncated form that remains in the cancer cell membrane and cytoplasm after cleavage of the “shedding” ECD has a higher rate of constitutive tyrosine kinase activity and is therefore more oncogenic than the intact HER2 [19]. Second, HER2 is truncated and released into blood circulation by a proteolytic mechanism, which has been attributed to various zinc-containing metalloproteases including members of the matrix metalloproteinase and the ADAM family [20]. Subsequently, the presence of free HER2_ECD is also an indicator of presence in the cancer cells of proteins degrading the extracellular matrix and known per se as markers of invasiveness and aggressiveness. Third, if the external domain is missing the anti-HER2 therapy might bind the free HER2_ECD instead of the HER2 external domain lacking the targeted receptor blockage [18].

These mechanistic models along with our longitudinal monitoring of plasma HER2_ECD in the two treatment-arms help explaining our findings. Indeed, we observed longer TTP for patients with plasma HER2_ECD below 15 ng/ml only when treated with sequential treatment. Trastuzumab alone in these patients binds to the full-length HER2 domain and is effective. In this situation, there is no interaction between the free HER2_ECD and trastuzumab and the free HER2_ECD levels remain constant until the cancer progresses. At this point, the HER2_ECD levels increase and sequential HER2_ECD monitoring would allow identifying the exact moment for chemotherapy administration. According to protocol plasma collection timelines (baseline, after 3 weeks, at first response evaluation (8–9 weeks) and at PD) it was not possible to know how long before PD the HER2_ECD levels started to raise. Prospective studies should integrate monthly detections in order to estimate the possible anticipation of chemotherapy administration.

In contrast, we observed that HER2_ECD levels in patients treated with upfront TChemo, as reported in several other studies [21], continuously diminished, up to < 20% of the baseline level, when chemotherapy is effective. Consequently, patients with high levels of free ECD should be treated with upfront combination therapy.

Another recent meta-analysis based on 15 prospective and 8 retrospective studies, investigating the prognostic value of HER2_ECD with the FDA approved threshold of 15 ng/ml, concluded that higher levels are correlated with poorer OS with a hazard ratio (HR) of 2.3 (CI:2.0–2.6) [23]. Our results are very much in line with this meta-analysis despite our threshold, with respect to OS for patients treated with upfront TChemo, had to be set at a higher level (35 ng/ml).

One possible explanation for our higher threshold and overall expression values of the plasma HER2_ECD could be that the plasma test also detects the extracellular domain of HER2 present on CTCs. If this hypothesis is correct, plasma HER2_ECD could be more sensitive and representative of the aggressiveness of the disease and subsequently more clinically useful than the serum test, indicating the concomitant presence of cleaving proteases and CTCs facilitating metastatic processes.

Further, in this translational study, we confirmed that monitoring of HER2_ECD values represent a clinically relevant complementary assessment in order to compensate inter-laboratory and inter-observer discrepancies in HER2 overexpression, particularly when detected in small biopsies of metastatic disease or in very old tissue specimens.

Another relevant finding is the high correlation of plasma levels at baseline and within 24 h after trastuzumab infusion in both arms. This observation is clinically important since it supplies evidence for avoiding unnecessary double blood assessments and suggests that trastuzumab does not bind immediately to the free soluble HER2_ECD.

This study was well conceived at time of trial start almost 20 years ago. Unfortunately, the trial recruited too slowly and did not achieve the planned number of patients. Therefore, the overall analysis is underpowered and the current treatment of choice pertuzumab/trastuzumab is investigated in a more recent study (SAKK22/10).

In conclusion, our data suggest the detection of plasma HER2_ECD may help identifying patients with HER2+ BC who would profit from a sequential treatment and deciding when to introduce chemotherapy. According to our study and to the available literature serum or plasma HER2_ECD levels should enter routine clinical practice for monitoring metastatic HER2+ BC. Even though fine-tuning on the threshold is necessary, serum and plasma HER2_ECD might help to personalize anti-HER2 and chemotherapy regimens.