HER2 Protein Overexpression, mRNA Expression, and DNA Amplification Across Solid Tumors: Comparison of Next-Generation Sequencing–Based Assays With Immunohistochemistry

Abstract

PURPOSE

The tumor-agnostic approach has been increasingly adopted in precision oncology. Immunohistochemistry (IHC) is the standard biomarker testing for human epidermal growth factor receptor 2 (HER2)–targeted therapies, whereas next-generation sequencing (NGS) has been widely incorporated in routine clinical practices. Here, we investigated the concordance between NGS-based assays and IHC in HER2 testing. Interfering factors leading to discordant results between the assays were also studied.

MATERIALS AND METHODS

Over 78,000 solid tumors across various types from two independent cohorts in the United States and Japan were investigated for HER2 DNA copy number, mRNA expression, and protein overexpression by whole-exome sequencing (WES), whole-transcriptome sequencing (WTS), and IHC, respectively.

RESULTS

In the US cohort (n = 77,267), HER2 DNA amplification, mRNA overexpression, and IHC-positive (IHC-P) were detected in 4.9%, 10.1%, and 4.7% of the tumors, respectively. Positive results in at least one of the three assays were observed in 10.7% of tumors, while 3.9% were positive for all three assays. Using IHC as a comparator, WTS showed better sensitivity than WES but a lower positive predictive value. These results were consistent in the Japanese cohort (n = 1,225). Although the overall HER2 RNA expression level correlated well with IHC score and DNA copy number, the degree of correlation varied among tumor types. Heterogeneous distribution of IHC-P tumor cells was associated with discordant results between NGS-based assays and IHC.

CONCLUSION

HER2-positive status in protein, mRNA, and DNA showed concordance in general but varied among tumor types. NGS-based assays, especially WTS, could be a useful predictive tool for HER2 testing in tumor-agnostic settings. Intratumor heterogeneity in HER2 protein expression should be considered when bringing bulk sequencing tests into clinical settings.

INTRODUCTION

Trastuzumab, a humanized monoclonal antibody targeting human epidermal growth factor receptor 2 (HER2; gene name ERBB2), is one of the first US Food and Drug Administration (FDA)–approved molecularly targeted agents and has demonstrated substantial clinical success.¹ The FDA has approved trastuzumab for the first-line treatment of breast cancer (BC) as well as gastric and gastroesophageal junction cancers (GC). For HER2 biomarker testing, the gold standards are immunohistochemistry (IHC) and in situ hybridization (ISH) to assess protein overexpression and DNA amplification, respectively.

CONTEXT

• Key Objective

• Can next-generation sequencing (NGS) assays, whole-exome sequencing (WES), and whole-transcriptome sequencing (WTS) be an alternative to immunohistochemistry (IHC) to determine human epidermal growth factor receptor 2 (HER2) status in various solid tumors?

• Knowledge Generated

• With over 70,000 solid tumors, WTS assay demonstrated >90% sensitivity and specificity against HER2 IHC 3+. HER2 mRNA expression levels correlated well with IHC scores and DNA copy numbers. WES shows higher specificity and positive predictive value than WTS but lower sensitivity. Concordance between NGS assays and IHC is particularly high in tumors with a high (>70%) intratumor percentage of HER2 IHC-positive cells.

• Relevance

• NGS-based assays, especially the WTS assay, have considerable potential to play a role in HER2 testing in tumor-agnostic settings. Intratumor heterogeneity in HER2 protein expression should be addressed when bringing these assays into clinical settings.

Trastuzumab-deruxtecan (T-DXd) is an antibody-drug conjugate of trastuzumab that has been approved by the FDA for HER2-positive BC, GC, and HER2-mutated non–small cell lung cancer in the United States.² In BC, IHC 1+ or 2+/ISH-negative cases are also eligible for T-DXd.³ More recently, the FDA granted accelerated tumor-agnostic approval to T-DXd for unresectable or metastatic HER2-positive tumors, defined solely by an IHC score of 3+ without confirmation by ISH, based on three clinical trials.^4-6

As the demands of HER2 testing in tumor-agnostic settings have sharply increased, other modalities, particularly next-generation sequencing (NGS)–based assays, are being considered for alternative HER2 testing. Indeed, FoundationOne CDx has been approved by the FDA as a companion diagnostic for evaluating HER2 copy number alterations in BC. An ISH-based HER2 mRNA expression analysis has also been reported to correlate with protein expression in BC.⁷ However, reverse transcriptase polymerase chain reaction–based HER2 mRNA expression levels have demonstrated limited concordance with protein expression in BC.⁸ Furthermore, little is known about the association between mRNA and protein expression levels in cancer types other than BC.^9,10

Here, we analyzed the fidelity between results for whole-transcriptome sequencing (WTS)–based HER2 mRNA expression, IHC-based HER2 protein expression, and whole-exome sequencing (WES)–based copy number alteration status of HER2.¹¹ We also examined the possible utility of integrating WTS/WES into the HER2 testing pipeline and interfering factors influencing the concordance between NGS-based assay and IHC in tumor-agnostic settings.

MATERIALS AND METHODS

Study Population

This study examined two independent patient cohorts from Japan and the United States. The US cohort included patients who underwent biomarker testing at the Caris Life Sciences Molecular Pathology Laboratory (Phoenix, AZ). The Japanese cohort corresponded to patients registered in SCRUM-Japan MONSTAR-SCREEN-2, a nationwide multilayered molecular profiling study on advanced solid tumors (UMIN000043899), between May 2021 and June 2023.

For the Japanese cohort, the eligibility criteria in this study are described elsewhere.¹² The study protocol was approved by the Institutional Review Board (IRB) of the National Cancer Center, Japan (2020-496), and all patients provided written informed consent for biomarker analysis. The US cohort used retrospective, deidentified clinical data and was considered IRB exempt in keeping with 45 CFR 46.101(b) (4). Patient informed consent was waived, and exempt status was determined by Western Institutional Review Board Copernicus Group IRB. All procedures were performed in compliance with relevant laws and institutional guidelines such as the Declaration of Helsinki, the Belmont Report, and the US Common Rule, as required.

Next-Generation Sequencing Assays

NGS analysis of the tumor tissue was performed using Caris MI Profile (Caris Life Sciences, Phoenix, AZ) for both the Japanese and US cohorts, performed at a CAP/Clinical Laboratory Improvement Amendments–certified Caris laboratory. Hematoxylin and eosin–stained slides of the patient's tumor underwent review by a board-certified pathologist or trained pathologist assistant. Tumor enrichment was achieved by harvesting targeted tissue using manual microdissection techniques. WTS was performed on the NovaSeq 6000 sequencer (Illumina, San Diego, CA) and quantified the mRNA expression levels in tissues by sequencing all coding exons, and transcript levels were normalized to the transcript per million (TPM) values for each gene. With US cohort data, a receiver operating characteristic (ROC) curve was plotted for TPM against the IHC results to determine the optimal mRNA overexpression cutoff. Cases with a TPM equal to or above the cutoff were designated as WTS-positive (WTS-P), whereas those below the cutoff were designated as WTS-negative (WTS-N). WES of genomic DNA was performed by using a NovaSeq 6000 sequencer for evaluating DNA copy number. If all exons within the gene of interest had an average of ≥3 copies and the average copy number of the entire gene was ≥6 copies, the gene result was reported as amplified, and WES-positive (WES-P) was assigned. The copy number cutoff of six for amplification was determined internally at Caris as a standard.¹³ WES-negative (WES-N) was defined as a copy number of less than six.

Immunohistochemistry

IHC for HER2 was performed on 4-μm thick sections prepared from formalin-fixed paraffin-embedded tissue with an anti-HER2 antibody (clone 4B5; Roche Diagnostics, Tucson, AZ) using the VENTANA BenchMark ULTRA system (Roche Diagnostics) in a CAP-certified laboratory (Kashiwa, Japan) for the Japanese cohort and in the Caris laboratory for the US cohort. HER2 IHC was evaluated by board-certified pathologists in each laboratory according to the CAP-ASCP-ASCO guidelines for GC for all solid tumor types except BC, which was evaluated according to the ASCO-CAP guidelines for BC.^14-16 According to the recent FDA approval for T-DXd in tumor-agnostic conditions, an IHC score 3+ was considered positive (IHC-P), while patients with IHC scores of 0, 1+, and 2+ were designated as IHC negative (IHC-N).

In Situ Hybridization

ISH was performed for all the cases with IHC score 2+ with available residual tissue. For both the US and Japanese cohorts, VENTANA HER2 Dual ISH DNA Probe Cocktail was used on the BenchMark ULTRA platform (Roche Diagnostics), following the manufacturer's instructions. HER2/CEP17 ratio and average HER2 signal copy number per tumor nucleus were calculated and interpreted per ASCO-CAP guidelines.^15,16 When HER2 IHC score 3+ or IHC score 2+ and ISH was positive, the case was considered IHC + ISH-P.

Statistical Analysis

Welch t test and the Mann-Whitney U test were performed to compare the means and medians between groups. Progression-free survival (PFS) was calculated for each treatment after enrollment and was defined as the duration from the date of treatment initiation to the date of either disease progression or death from any cause. The median PFS of each group was estimated using the Kaplan-Meier method and compared using the log-rank test. All P values were two-sided, and statistical significance was set at P < .05. R Statistics software (version 4.4.1) (R Foundation for Statistical Computing, Vienna, Austria) was used for all statistical analyses.

RESULTS

HER2 Status in the US Cohort

In the US cohort, 77,267 cases for which all three assay results were available were included in the analysis. The most common cancer types were colorectal cancer (CRC), followed by BC and GC (Table 1). Overall, 4.7% and 4.9% were IHC-P and WES-P, respectively. The sensitivity and positive predictive values (PPVs) of WES for detecting IHC-P were 83.9% and 81.9%, respectively (Fig 1A).

TABLE 1.

Distribution of Cancer Types Across the Japanese and US Cohorts

Tumor Type	Japan, No. (%) (n = 1,225)	USA, No. (%) (n = 77,267)
Colorectal	305 (25)	26,874 (34)
Gastric/Esophagogastric	144 (12)	9,256 (12)
Breast	93 (8)	16,518 (21)
Others	683 (55)	24,619 (32)
Pancreas	102 (8)	257 (0.3)
Head and neck	84 (7)	88 (0.1)
Prostate	81 (7)	84 (0.1)
Urothelial	71 (6)	309 (0.4)
Biliary tract	64 (5)	4,665 (7)
Unknown	0 (0)	6,273 (8)
Miscellaneous	281 (22)	12,943 (16)

FIG 1.

Venn diagrams and sensitivity, specificity, and PPV of NGS assays comparing with IHC or IHC + ISH. The IHC + ISH category includes IHC score 3+ (IHC-P) and IHC score 2+/ISH-positive cases. (A) Venn diagram of the US cohort, illustrating the overlap between positive WTS, WES, and IHC. (B) Venn diagram of the US cohort, illustrating the overlap between positive WTS, WES, and IHC + ISH^a. (C) Venn diagram of the Japanese cohort, illustrating the overlap between positive WTS, WES, and IHC. (D) Venn diagram of the Japanese cohort, illustrating the overlap between positive WTS, WES, and IHC + ISH^a. ^aCases of IHC 2+ with no ISH data available were excluded from Figures 1B and 1D. Numbers indicate number of cases. The percentages in parentheses were calculated by taking the total number of cases shown inside the circles of the Venn diagram as 100%. IHC, immunohistochemistry; IHC-P, immunohistochemistry-positive; ISH, in situ hybridization; NGS, next-generation sequencing; PPV, positive predictive value; WES, whole-exome sequencing; WTS, whole-transcriptome sequencing.

With WTS data, the cutoff value of HER2 mRNA TPM against IHC-P was determined to be 83.7 by ROC analysis (Data Supplement, Fig S1). This cutoff grouped 10.1% of the cases as WTS-P. The sensitivity of WTS-P for IHC-P was 93.1%. Overall, 10.7% of patients were positive in at least one of the three assays, whereas only 3.9% tested positive in all three assays. When ISH was performed for IHC 2+ cases and incorporated into the analysis, specificity and PPV of WES and WTS increased, whereas sensitivity decreased (Fig 1B).

Validation of NGS-Based Prediction of HER Status in the Japanese Cohort

Using the Japanese cohort, we validated the cutoff value of HER2 mRNA expression and investigated the detailed relationship between clinicopathologic features and NGS-predicted HER2 status. Of 1,890 patients enrolled, 1,225 patients in whom all three assays were successfully performed were included in further analyses (Data Supplement, Fig S2). The most common cancer type was CRC, followed by GC and pancreatic adenocarcinoma (Table 1). The specimens comprised 41.9% biopsy specimens and 58.1% surgically resected specimens from 26 institutions. There were no statistically significant differences in the mean TPM between samples collected at National Cancer Center Hospital East and the other institutions (P = .34), or between biopsy and surgically resected samples (P = .66).

Among the 1,225 patients, 119 patients (9.7%) tested positive for at least one of the three assays, whereas only 3.4% were positive for all three assays. When IHC was used as a comparator, the sensitivities of WES and WTS were 61.4% and 80%, respectively. The PPVs of WES and WTS were 89.6% and 54.9%, respectively (Fig 1C). Consistent with the US cohort, specificity and PPV of WES and WTS against IHC + ISH-P increased at the expense of sensitivity (Fig 1D).

IHC-P was observed in 5.7% of the patients (Fig 1C), while 7.3%, 17%, and 70% were IHC 2+, 1+, and 0, respectively (Fig 2A). Among the tumor types in which more than 20 patients were examined, urothelial carcinoma (UC) showed the highest HER2 IHC-P rate, followed by cervical squamous cell carcinoma (CSCC), GC, and BC. Notably, very few or no IHC-P cases were observed in pancreatic adenocarcinoma, esophageal cancer (EC), prostate adenocarcinoma (PRAD), or neuroendocrine tumor/carcinoma (NET/NEC).

FIG 2.

(A) HER2 IHC score distribution across various cancer types (n ≥ 20), arranged from left to right in order of the cancer types with the highest to lowest IHC-P (score 3+) rates. Numbers indicate number of cases. (B) HER2 WES-P rates across various cancer types (n ≥ 20) in the same order as in (A). (C) HER2 WTS-P rates across various cancer types (n ≥ 20) in the same order as in (A). (D) Violin plot displaying the TPM distribution in various cancer types (n ≥ 20). The red line indicates the TPM cutoff for WTS-P. BC, breast cancer; CESC, cervical squamous cell carcinoma; CRC, colorectal cancer; EC, esophageal cancer; GC, gastric and gastroesophageal junction cancer; HER2, human epidermal growth factor receptor 2; HNC, head and neck carcinoma; IHC, immunohistochemistry; IHC-P, immunohistochemistry-positive; PDAC, pancreatic adenocarcinoma; PRAD, prostate adenocarcinoma; RCC, renal cell carcinoma; TPM, transcripts per million; UC, urothelial carcinoma; WES-N, whole-exome sequencing-negative; WES-P, whole-exome sequencing-positive; WTS-N, whole-transcriptome sequencing-negative; WTS-P, whole-transcriptome sequencing-positive.

WES-P was identified in 3.6% of the patients with a median gene copy number of 36. Among the tumor types in which more than 20 patients were examined, the highest frequency of WES-P was observed in GC, followed by head and neck cancer and BC (Fig 2B).

Using the cutoff value established in the US cohort, WTS-P was identified in 8.5% of the cases, with a mean TPM of 359.8. The positive rates for each cancer type varied, ranging from 0% (renal cell carcinoma, EC, NET/NEC and CSCC) to 30% (UC). UC showed the highest proportion in both IHC-P and WTS-P. However, the extent of correlation between the rate of WTS-P and IHC-P appeared to depend on cancer types (Figs 2A and 2C).

The TPM distribution was significantly different for each cancer type (Fig 2D). Among the tumor types for which more than 20 patients were examined, the highest mean TPM was 99.9 in biliary tract cancer (BTC), followed by BC and UC. NET/NEC showed the lowest mean TPM (8.5), which was significantly different from that in the UC (P = .004). Some cancer types (BTC, BC, UC, GC, CRC, and head and neck cancer) showed a wider range of expression levels compared with the others.

HER2 mRNA expression levels were significantly different across all IHC scores, with a median TPM of 21, 31.3, 45.7, and 248.4 for IHC scores of 0, 1+, 2+, and 3+, respectively (Fig 3A). This stepwise positive correlation was confirmed with the US cohort (Data Supplement, Fig S3). It is noteworthy that there were two patients who were IHC 2+, but with TPM of over 200 (1,406 and 259). Both cases showed only focal (5%) tumor cells with IHC score 3+ positivity. The one with TPM of 1,406 was a surgical specimen of GC with a high HER2 DNA copy number (96.5), though ISH could not be performed because of a lack of available residual tissue. ISH was positive in the other case.

FIG 3.

(A) HER2 IHC scores and mRNA expression. The violin plot illustrates the distribution of HER2 mRNA expression in TPM determined by WTS for each HER2 IHC score. The red line indicates the TPM cutoff for WTS-P. Significant differences were observed for an IHC 0 versus 1+ (P < .001), 1+ versus 2+ (P < .05), and 2+ versus 3+ (P < .001). (B) Scatter plot illustrating the distribution of HER2 gene copy numbers determined by WES versus HER2 mRNA expression in TPM. Only WES-P cases in the Japanese cohort are shown. The HER2 gene copy number and mRNA expression level are positively correlated (r = 0.70, P < .001, Spearman rank correlation). HER2, human epidermal growth factor receptor 2; IHC, immunohistochemistry; TPM, transcripts per million; WES-P, whole-exome sequencing-positive; WTS, whole-transcriptome sequencing; WTS-P, whole-transcriptome sequencing-positive.

The HER2 gene copy number determined by WES was also positively correlated with the mRNA expression levels in patients with HER2 DNA amplification (r = 0.700, P < .0001, Spearman rank correlation; Fig 3B).

Factors Associated With Discordance of HER2 Status Across Assays

To identify the factors causing discordant results between NGS-based and IHC assays, we focused on heterogeneity in HER2 expression by examining the proportion of tumor cells exhibiting strong membranous HER2 positivity, designated as HER2-positive tumor cells (HPTC). We divided the IHC-P patients into three groups based on the proportion of HPTCs: HPTC-low (HPTC-L) as tumors with <30% HPTCs, HPTC-intermediate (HPTC-int) as those with between 30% and 70% HPTCs, and HPTC-high (HPTC-H) as those with >70% HPTCs. All HPTC-H patients were WTS-P, and more than 90% were WES-P, whereas over 50% of the HPTC-L patients were WES-N/WTS-N, and only 13% were WES-P/WTS-P (Fig 4A). These findings suggest that patients with HPTC-L are likely to exhibit false-negative results in NGS-based assays, despite IHC-P.

FIG 4.

(A) Stacked bar chart showing the distribution of different WES/WTS results across the three groups: HPTC-H (H), HPTC-int (int), and HPTC-L (L). (B) Box plot showing the distribution of mRNA expression (TPM) in IHC-P cases across the three groups: HPTC-H (H), HPTC-int (int), and HPTC-L (L). (C) Box plot showing the ERBB2 DNA copy number across the HPTC-H (H), HPTC-int (int), and HPTC-L (L) groups. (D) Box plot comparing mRNA expression in TPM between two groups: false positives (WTS-P but IHC-N) and true positives (WTS-P and IHC-P). HER2, human epidermal growth factor receptor 2; HPTC-H, HER2-positive tumor cells-high; HPTC-int, HER2-positive tumor cells-intermediate; HPTC-L, HER2-positive tumor cells-low; IHC-N, immunohistochemistry-negative; IHC-P, immunohistochemistry-positive; NS, not significant; TPM, transcripts per million; WES, whole-exome sequencing; WTS, whole-transcriptome sequencing; WTS-P, whole-transcriptome sequencing-positive.

Consistently, HER2 mRNA TPM was significantly higher in the HPTC-H group (median, 516.9) than in the HPTC-int group (median, 204.5; P = .002) and the HPTC-L group (median, 64.7; P < .0001; Fig 4B). The HPTC-int group had a significantly higher TPM than the HPTC-L group (P = .003).

HER2 DNA copy numbers also correlated with the proportion of HPTCs. Among the WES-P cases, the HER2 DNA copy number was significantly higher in the HPTC-H group (median, 36.4) than in the HPTC-int (median, 15.2; P = .018) or HPTC-L (median, 10.8; P = .008) groups (Fig 4C).

The proportion of HPTCs varied depending on the cancer types. For example, most of the IHC-P cases in UC were HPTC-L, whereas most of the IHC-P in BC were HPTC-H (Data Supplement, Fig S4).

In contrast to the high sensitivity of WTS to predict IHC-P, PPV was relatively low in both the United States and Japanese cohorts. Among the WTS-P group in the Japanese cohort, HER2 mRNA expression levels were significantly higher in the IHC-P group than in the IHC-N group (P < .0001; Fig 4D), which may suggest that using a higher TPM cutoff value specifically optimized for tumor type instead of a global one may reduce the WTS-P/IHC-N rate and improve the PPV of WTS. So, we explored a lineage-specific TPM threshold specifically determined for BC in the US cohort. Replacing the global cutoff (83.7) with the BC-derived cutoff (100.9) increased the PPV of WTS from 53.8% to 68.4% (Data Supplement, Fig S5).

Among the patients with GC who received HER2-targeted therapy, either trastuzumab or T-DXd without immune checkpoint inhibitors, HPTC-L patients had poorer responses to HER2-targeted therapy than those who were HPTC-H (P = .024; Fig 5). All patients were HER2 IHC 3+, except two—one in each group—who were IHC 2+.

FIG 5.

Kaplan-Meier survival plot comparing the PFS of CRC and GC cases between two groups: patients with tumors containing >70% IHC 3+ tumor cells (HPTC-H; in blue) and those with <30% (HPTC-L; in yellow). CRC, colorectal cancer; GC, gastric and gastroesophageal junction cancer; HER2, human epidermal growth factor receptor 2; HPTC-H, HER2-positive tumor cells-high; HPTC-L, HER2-positive tumor cells-low; IHC, immunohistochemistry; PFS, progression-free survival.

DISCUSSION

Although IHC offers several advantages, including widespread accessibility and relatively low cost, there are limitations including limited quantitative precision, variability because of fixation conditions, and the requirement for interpretation by trained pathologists. Additionally, HER2 IHC testing has a narrow dynamic range that does not accurately reflect the true protein expression.¹⁷ In recent years, there has been a growing need for comprehensive pan-cancer HER2 testing, particularly with the recent tumor-agnostic approval of T-DXd. To overcome the drawbacks of IHC, NGS-based assays are the most anticipated testing modalities for various cancers.^9,10,18-21

Intratumor heterogeneity of HER2 protein expression significantly affects the accuracy of NGS-based HER2 status determination. Among 42 Japanese cases in which all three assays returned positive results, 69% were HPTC-H tumors, while only 12% were HPTC-L tumors. By contrast, among the NGS false-negative cases, 92% in the WTS-N/IHC-P cases and 78% in the WES-N/IHC-P cases were HPTC-L. These findings indicated that a low proportion of HPTCs compromises the concordance between results by WES/WTS and IHC. Moreover, the proportion of HPTCs showed organ-specific trends (Data Supplement, Fig S4). The cancer types such as CRC, GC, and BC could be referred to as HPTC-H–type tumors and may be suitable for NGS-based HER2 testing. Cancer types with more heterogeneous HER2 protein expression, such as UC and BTC, could be classified as HPTC-L type may be more challenging to incorporate the assay into clinical settings.

Intratumoral molecular heterogeneity is known to contribute to poor prognosis in BC.^22-24 In GC, we and others have reported that heterogeneity in protein expression has a negative impact on the response to HER2-targeted therapy.^25-27 In this study, we demonstrated that HPTC-H GC had prolonged PFS compared with those with HPTC-L tumors treated with HER2-targeted therapy (Fig 5). This suggests that patients with HPTC-H tumors may derive greater benefit from HER2-targeted therapy than those with HPTC-L tumors. Since mRNA expression levels correlated well with the proportion of HPTCs (Fig 4B), the WTS-based assay has the potential to serve as a valuable tool for efficiently identifying HPTC-H cases that are likely to benefit from HER2-targeted therapies, as previously suggested.⁹ Indeed, 12 of 13 WTS-N/IHC-P and all 29 WTS-N/IHC + ISH-P cases in the Japanese cohort were HTPC-L tumors.

Our study indicated a relatively low PPV of the WTS assay to predict IHC. Previously, Shayeb et al²⁸ analyzed the concordance between HER2 mRNA overexpression, DNA amplification, and protein overexpression with a relatively small number of samples (n = 723). They found a higher frequency of mRNA overexpression than protein overexpression, which is consistent with our results. Although they employed polymerase chain reaction to evaluate mRNA levels and used a different cutoff for IHC-P (IHC 2+, 30%) compared with our study, the findings are still comparable.

Analysis of 44 false-positive (WTS-P/IHC-N) cases in the Japanese cohort revealed that approximately 75% of these cases were BC, UC, PRAD, and BTC: These cancer types exhibited the highest mRNA expression levels irrespective of the protein expression levels (Fig 2D). Inherently high mRNA expression levels may produce a higher rate of false positives. To reduce the incidence of false-positive results, it may be beneficial to identify and apply tailored TPM cutoff values for specific patient subgroups. We explored lineage-specific TPM thresholds for BC in the US cohort; raising the cutoff from 83.7 to 100.9 improved the PPV of WTS from 53.8% to 68.4% (Data Supplement, Fig S5). Using lineage-specific TPM cutoffs is likely to improve diagnostic performance and clinical utility.

Several points should be considered when using NGS-based assays, including their high cost and longer turnaround time compared with IHC. One of the key strengths of the NGS-based approach is its ability to simultaneously detect HER2 gene-alternation and expression of other potentially targetable genes.^29,30 As such, with proper quality control and an optimal cutoff, WTS can especially serve as a powerful tool for accurately identifying patients eligible for HER2-targeted therapy.

In conclusion, we demonstrated concordance between HER2 DNA copy numbers, mRNA, and protein expression determined by WES, WTS, and IHC across various solid tumors. Positive rates of each assay and distribution of mRNA expression varied among cancer types. NGS-based assays, especially WTS, have potential clinical utility for predicting HER2 IHC-P status especially in HPTC-H tumors. HPTC types as well as background and HER2 mRNA expression levels among cancer types may be the factors that need to be addressed to bring NGS-based assays into clinical settings.

DATA SHARING STATEMENT

A data sharing statement provided by the authors is available with this article at DOI https://doi.org/10.1200/PO-25-004413. The data will be available in response to any reasonable request made to the corresponding author.

AUTHOR CONTRIBUTIONS

Conception and design: Michiko Nagamine, Takao Fujisawa, Yoshiaki Nakamura, Hiroji Iwata, Kenjiro Namikawa, Kanako C. Hatanaka, Kenichi Taguchi, Milan Radovich, Takayuki Yoshino, Takeshi Kuwata

Provision of study materials or patients: Naoya Sakamoto, Yoshito Komatsu, Eiji Oki, Akitaka Makiyama, Chigusa Morizane, Milan Radovich

Collection and assembly of data: Michiko Nagamine, Takao Fujisawa, Naoya Sakamoto, Shigenori Kadowaki, Makoto Ueno, Shogen Boku, Yoshito Komatsu, Eiji Oki, Chigusa Morizane, Hidemichi Watari, Kenjiro Namikawa, Kanako C. Hatanaka, Kenichi Taguchi, Daniel Magee, Takayuki Yoshino, Matthew Oberley, Takeshi Kuwata

Data analysis and interpretation: Michiko Nagamine, Takao Fujisawa, Makoto Ueno, Akitaka Makiyama, Norio Nonomura, Susumu Okano, Kenjiro Namikawa, Yutaka Hatanaka, Kanako C. Hatanaka, Kenichi Taguchi, David Spetzler, Milan Radovich, Daniel Magee, Takayuki Yoshino, Matthew Oberley, Takeshi Kuwata

Manuscript writing: All authors

Final approval of manuscript: All authors

Accountable for all aspects of the work: All authors

AUTHORS' DISCLOSURES OF POTENTIAL CONFLICTS OF INTEREST

The following represents disclosure information provided by authors of this manuscript. All relationships are considered compensated unless otherwise noted. Relationships are self-held unless noted. I = Immediate Family Member, Inst = My Institution. Relationships may not relate to the subject matter of this manuscript. For more information about ASCO's conflict of interest policy, please refer to www.asco.org/rwc or ascopubs.org/po/author-center.

Open Payments is a public database containing information reported by companies about payments made to US-licensed physicians (Open Payments).