hile HER2 overexpression has been observed in various solid tumors, HER2-targeted therapies have been largely limited to use in patients with HER2-positive (HER2+) breast and gastric cancers. Emerging clinical data have demonstrated that HER2-targeted therapies, like T-DXd, zanidatamab, and tucatinib, may offer clinical benefit for patients with any HER2-overexpessing cancer. As the use of HER2-targeted therapies expands into other solid tumors, in which tissue samples may be more limited and HER2 expression may be more heterogenous, more sensitive and less invasive diagnostic methods are needed. Liquid biopsies and advances in ctDNA analysis offer a wide range of potential clinical applications, complementing standard clinical, imaging, and tissue assessments in order to optimize tumor-agnostic cancer diagnosis and patient treatment management.

The Current Status for HER2 Testing

There is a large portfolio of available human epidermal growth factor receptor 2 (HER2)-targeted therapies that have an associated companion diagnostic (CDx) in breast and gastric cancers which confirms ERBB2 amplification status (i.e., HER2 gene amplification) and/or HER2 protein overexpression on the surface of tumor cells. These include drugs such as trastuzumab, pertuzumab, ado-trastuzumab emtansine (T-DM1), and fam-trastuzumab deruxtecan (T-DXd). While the indications and mechanisms of action may differ between each of these therapeutics, patient selection for treatment with HER2-targeted drugs is well established. However, it’s limited to a handful of similar, commercially available immunohistochemistry (IHC) tests for detection of increased expression of HER2 protein, and in situ hybridization (ISH) for confirming ERBB2 amplification status. Both methodologies utilize tumor tissue samples requiring invasive procedures.

The Benefits and Disadvantages of Tissue-Based Diagnostics

The FDA-approved tissue-based IHC methods follow ASCO/CAP guidelines to help ensure uniformity for HER2 scoring with algorithms specific to breast and gastric cancers (Current CAP Guidelines | College of American Pathologists). An IHC score of either 0 or 1+ is defined as negative or low, while HER2-overexpressing status of 3+ is unequivocally amplified and a score of 2+ is equivocal. The latter usually requires additional testing (and tissue) using an ISH method to confirm ERBB2 amplification status at the chromosomal level.

Although widely used and accepted, there are drawbacks to using IHC diagnostics where biomarker detection is dependent on selectivity and specificity of antibody staining and antigen retrieval methods in a fixed tumor specimen. Subsequent interpretation of the HER2 results requires trained pathologists to review. The subjective nature and manual counting of the scoring can lead to varied interpretations of results, and pathologists’ reads may not be reproducible, especially true in borderline cases with HER2 IHC 2+ equivocal score. In addition, the heterogenous expression of HER2 that is often observed in non-breast cancer tumors can make the scoring of IHC and ISH tests difficult. For instance, can a serially sectioned 4-micron FFPE (formalin-fixed, paraffin-embedded) slide of a tumor specimen accurately convey the full picture of a metastatic disease, or is it only representative of a small section or subclone of the tumor? In the future, the advent of AI-driven digital pathology tools could improve reproducibility as well as speed up the turnaround time and tumor coverage of such tests, ultimately benefiting patients and reducing costs.

Acquiring the tissue for IHC and ISH requires an invasive surgical procedure and places the patient at a significantly higher risk for procedure-related complications. Additionally, biopsies may be infeasible in some cancer subtypes at the early cancer stage due to small tumor size, inaccessibility, or in later-stage patients where invasive surgery is not recommended. Later-stage patients already selected using IHC who then relapse on a HER2-targeted therapy may have transiently downregulated levels of HER2 and/or other mechanisms of resistance to HER2. For these patients, new biopsies may be required to be eligible for alternative treatments. This could temper their enthusiasm or restrict recruitment for novel HER2-targeting compounds, capable of overcoming such resistance.

Liquid Biopsy: A safe alternative, and a valuable supplement to invasive tumor tissue testing?

Alternatives to detecting tumor antigens by IHC and ISH, for the diagnosis of HER2 overexpression, include liquid biopsy tests which detect CTC (circulating tumor cells), soluble tumor markers (e.g., CEA, CA19-9, sHER2-ECD), and fragments of shed tumor DNA present in the blood of cancer patients. Liquid biopsy tests, like next-generation sequencing (NGS) of plasma circulating tumor DNA (ctDNA), offer several advantages over assessing tumor status by IHC or ISH methods. NGS testing of plasma ctDNA is minimally invasive, allowing for repeated sampling and real-time monitoring of the tumor dynamics without the need for tissue biopsies. This method also provides a comprehensive genomic landscape, detecting not only gene amplification but also other relevant mutations and alterations and offers a more detailed understanding of tumor biology. In doing so, it may provide a new diagnostic platform to support novel therapeutics and tumor-agnostic indications.

Several NGS tumor tissue and liquid biopsy tests, like Guardant360 and FoundationOne CDx, have recently been approved for non-small cell lung cancer (NSCLC) and with tumor-agnostic indication to guide targeted therapy treatments. These companion diagnostics enable healthcare providers to tailor treatments to the molecular characteristics of each patient’s cancer, improving outcomes and providing more personalized care options.

These methods continue to evolve with the increasing numbers of genes captured and with improved sensitivity and specificity. Techniques have extended beyond single-nucleotide variants (SNVs) to amplifications (notably, ERBB2) and now to epigenetic changes indicative of aberrant expression of tumor drivers. However, results of samples collected from the same patient still show discordance between tissue-based NGS and these blood-based assays.

Heterogeneity and tumor evolution may explain these discordant results, but in some cases the explanation may be insufficient specificity and sensitivity or non-shedding tumors. The latter have been described in early-stage tumors, which sometimes lack sufficient vascularization to enable ctDNA levels to rise above the limit of detection. Although tumor-derived nucleic acids or ctDNA can be readily detected in appropriately collected and processed blood samples, estimates of the stability of ctDNA without preservatives range from minutes to several hours. In addition, normal cell lysis and apoptosis can introduce cell-free DNA (cfDNA) from normal cells which can interfere with the detection of ctDNA from the tumor. Thus, sample handling and timing of collections is of paramount importance when dealing with ctDNA.

A major benefit of utilizing ctDNA includes the ease of serial collections at appropriate timepoints, which would require repeat biopsies to monitor using IHC/ISH or tumor tissue NGS. The serial monitoring of ctDNA has proven effective and informative in addressing dynamic changes associated with tumor growth or response to treatment. Publications have drawn a clear correlation between elevated levels of ctDNA before and after treatment as a poor prognostic indicator, perhaps an indication that a tumor is continuing to grow and metastasize and is unresponsive to treatment. Conversely, when levels of ctDNA drop after treatment this can be an indication of response; again it might be hypothesized that this signifies the growth of the tumor has been slowed or reversed.

To address lack of concordance between tumor and blood samples attributable to sensitivity or a non-shedding tumor, several companies (e.g., Guardant and FMI) are now examining the epigenetic markers found on ctDNA fragments. These can be utilized in discriminating between tumor-derived and normal cell DNA, increasing the level of confidence of a positive result. In addition, epigenetics is now being used to identify expression-level changes, with heterochromatin on inactive promoters being a mark of repressed expression. Additional analysis that can identify cell-type-specific markers may shed light on the tumor’s clonal evolution.

The advances in ctDNA analysis are allowing researchers to better understand the dimensional and heterogeneous data which would otherwise require multiple samples from a tumor to replicate. This complex multidimensional data is being analyzed using statistical and machine learning algorithms, with the end goal of producing robust and accurate classifiers for diagnosis, prediction of response to treatment, and prognosis. Importantly, this can be done with minimal inconvenience to the patient in a longitudinal manner heralding a new albeit still relatively costly analysis technique. Arguably, the minimal inconvenience to the patient and the ability to correctly prescribe a drug with efficacy could more than offset this cost.

One other potential application of liquid biopsies will need to address additional questions. The most fundamental of these is in very early diagnosis where the correlation between presence of ctDNA and disease progression or outcome will require additional and extensive studies, potentially over many years. However, in more advanced cancers post-conventional and established diagnostic methods, like IHC/ISH, liquid biopsies appear capable of significant contribution to the patient journey. Noninvasive and longitudinal sampling can reveal either sustained response, methods of resistance, or other aspects of tumor evolution and in the future may allow treatment(s) to be tailored appropriately.

As with all methods and tools in the fight against cancer, there is unlikely to be a one-size-fits-all method. Patients whose tumors do not shed ctDNA may not benefit using current techniques. However, further advances in sensitivity through the inclusion of other markers in addition to DNA alterations (e.g., expression and/or heterochromatin) may eventually reveal non-shedders to be low shedders just below current levels of detection.

Summary

Conventional tumor biopsies are likely to remain the standard of care for some time yet. However, the less invasive nature of liquid biopsies, the additional layers of prognostic information, and advantages of real-time monitoring will likely be highly beneficial to most if not all patients agnostic of tumor type.

Table 1: HER2-targeted therapies, by indication and associated companion diagnostic (CDx)