Evaluating the Correlation Between FISH and NGS in Assessing ERBB2 Amplification in Breast Cancer: Insights and Implications
Authors
Viktor Smirnov, Polina Turova, Vladimir Kushnarev, Gleb Khegai, Sheila T. Yong, Anna Butusova, Nikita Kotlov, Konstantin Chernyshov
- BostonGene, Corp. Waltham, MA, USA
Abstract
Introduction
Fluorescence in situ hybridization (FISH) is the gold standard for evaluating ERBB2 amplification in breast cancer (BC) when HER2 status is equivocal by immunohistochemistry, per ASCO guidelines (15–20% of cases). Next-generation sequencing (NGS) can confirm the HER2 status of BC samples and, in turn, help reduce the incidence of equivocal samples. Here, we examined the correlation between FISH and NGS in annotating HER2 status of BC samples and evaluated the clinical utility of NGS as a support tool for improving analysis outcomes.
Methods
The BostonGene internal breast cancer (total n=358, FISH annotation=124) and TCGA-BRCA (total n=972, FISH annotation n=285) cohorts were assessed together for ERBB2 expression via RNA sequencing and ERBB2 mutation and amplification via DNA sequencing. High ERBB2 gene expression was set at >8.5 log2(TPM+1)1. For NGS, Gain was defined as an increase of one gene copy relative to sample ploidy, while Amplification referred to an increase of two or more gene copies compared to sample ploidy. Samples were assessed for FISH positivity (+) per ASCO/CAP guidelines.
Results
There was substantial agreement between FISH and NGS analysis for both cohorts combined (Cohen’s kappa=0.57). All FISH+ samples had significantly higher ERBB2 expression (n=71, median (med)=8.8) than all FISH- samples (n=338, med=6.6, U-statistics (stats), p<0.001). Among the FISH+ samples, 53% (n=37) had high ERBB2 expression. Samples with amplification as shown by NGS had significantly higher ERBB2 expression (n=35, med=9.6) than non-amplified samples (n=335, med=6.6, U-stats, p<0.001). High ERBB2 expression was observed in 31% (n=23 of 74) of samples with ERBB2 gain or amplification. FISH+ samples with ERBB2 gain had significantly higher ERBB2 expression (n=14, med=10.5) than FISH- samples (n=13, med=7.9, U-stats, p<0.001).
NGS also detected 23 unique ERBB2 mutations, among which 18 were classified as missense, 4 as frameshift, and 9 as gain-of-function. Interestingly, 20 non-amplified samples with ERBB2 mutation exhibited significantly higher ERBB2 expression (med=7.4) than non-amplified samples with wildtype ERBB2 (n=999, med=6.9, U-stats, p=0.004).
Conclusion
Our study demonstrated substantial agreement between FISH and NGS (Cohen’s kappa=0.57) in evaluating HER2 status in BC, highlighting NGS as a viable support tool for assessing HER2 status, particularly in equivocal cases. First, NGS reliably detected significantly higher ERBB2 expression in FISH+ samples than in FISH- samples. Next, it detected ERBB2 mutations in non-amplified samples that were not covered by FISH. Since these mutations may be associated with increased ERBB2 expression, they are clinically relevant and warrant further study. Last but not least, NGS detected focal ERBB2 amplifications and gains at a higher resolution than FISH. While associated with varied FISH results, these events consistently concurred with higher ERBB2 expression, reflecting the heterogeneous nature of HER2 amplification in BC that FISH does not take into account. The ability of NGS to identify these nuanced genomic alterations yields valuable insights into the tumor biology and potential resistance mechanisms of BC that are crucial for designing personalized treatment strategies and improving patient outcomes.
Fluorescence in situ hybridization (FISH) is the gold standard for evaluating ERBB2 amplification in breast cancer (BC) when HER2 status is equivocal by immunohistochemistry, per ASCO guidelines (15–20% of cases). Next-generation sequencing (NGS) can confirm the HER2 status of BC samples and, in turn, help reduce the incidence of equivocal samples. Here, we examined the correlation between FISH and NGS in annotating HER2 status of BC samples and evaluated the clinical utility of NGS as a support tool for improving analysis outcomes.
Methods
The BostonGene internal breast cancer (total n=358, FISH annotation=124) and TCGA-BRCA (total n=972, FISH annotation n=285) cohorts were assessed together for ERBB2 expression via RNA sequencing and ERBB2 mutation and amplification via DNA sequencing. High ERBB2 gene expression was set at >8.5 log2(TPM+1)1. For NGS, Gain was defined as an increase of one gene copy relative to sample ploidy, while Amplification referred to an increase of two or more gene copies compared to sample ploidy. Samples were assessed for FISH positivity (+) per ASCO/CAP guidelines.
Results
There was substantial agreement between FISH and NGS analysis for both cohorts combined (Cohen’s kappa=0.57). All FISH+ samples had significantly higher ERBB2 expression (n=71, median (med)=8.8) than all FISH- samples (n=338, med=6.6, U-statistics (stats), p<0.001). Among the FISH+ samples, 53% (n=37) had high ERBB2 expression. Samples with amplification as shown by NGS had significantly higher ERBB2 expression (n=35, med=9.6) than non-amplified samples (n=335, med=6.6, U-stats, p<0.001). High ERBB2 expression was observed in 31% (n=23 of 74) of samples with ERBB2 gain or amplification. FISH+ samples with ERBB2 gain had significantly higher ERBB2 expression (n=14, med=10.5) than FISH- samples (n=13, med=7.9, U-stats, p<0.001).
NGS also detected 23 unique ERBB2 mutations, among which 18 were classified as missense, 4 as frameshift, and 9 as gain-of-function. Interestingly, 20 non-amplified samples with ERBB2 mutation exhibited significantly higher ERBB2 expression (med=7.4) than non-amplified samples with wildtype ERBB2 (n=999, med=6.9, U-stats, p=0.004).
Conclusion
Our study demonstrated substantial agreement between FISH and NGS (Cohen’s kappa=0.57) in evaluating HER2 status in BC, highlighting NGS as a viable support tool for assessing HER2 status, particularly in equivocal cases. First, NGS reliably detected significantly higher ERBB2 expression in FISH+ samples than in FISH- samples. Next, it detected ERBB2 mutations in non-amplified samples that were not covered by FISH. Since these mutations may be associated with increased ERBB2 expression, they are clinically relevant and warrant further study. Last but not least, NGS detected focal ERBB2 amplifications and gains at a higher resolution than FISH. While associated with varied FISH results, these events consistently concurred with higher ERBB2 expression, reflecting the heterogeneous nature of HER2 amplification in BC that FISH does not take into account. The ability of NGS to identify these nuanced genomic alterations yields valuable insights into the tumor biology and potential resistance mechanisms of BC that are crucial for designing personalized treatment strategies and improving patient outcomes.
Latest publications
