Background: HER2/ERBB2 is a known therapeutic target in gastroesophageal adenocarcinoma (GEAC). HER3/ERBB3 is an emerging target in a variety of cancers but less defined in GEAC. Insights into the genomic and immunologic landscape of ERBB2/ERBB3 alterations (ERBB2/3-alt) are required to develop effective treatments.
Methods: We analyzed de-identified data from 2,050 GEAC patient tumor samples which underwent next generation sequencing with the Tempus xT assay (DNA-seq of 648 genes at 500x coverage, full transcriptome RNA-seq). Bivariate analyses were performed to compare demographics, immuno-oncology (IO) biomarkers, and co-mutations between ERBB2/3-alt groups. P-values comparing individual co-mutations between groups were adjusted for false discovery.
Results: Overall, 17.8% (365/2050) of tumors harbored an ERBB2 or ERBB3-alt, 3 of which had both and were excluded from analyses. Alterations were defined as pathogenic/likely pathogenic mutations, copy number amplifications (CN amp), or copy number loss (CN loss). Tumors were divided into ERBB2 CN amp (n=252, 12.3%), ERBB2 other (CN loss/mutation) (n=71, 3.5%), ERBB3 CN amp (n=17, 0.8%), ERBB3 other (n=22, 1%), and WT (n=1685, 82%) for both. No differences were observed between groups regarding baseline demographics (age of onset, race, ethnicity) aside from the percentage of females (p=0.002). Significant differences in IO markers and the co-occurring genetic alterations were seen. ERBB3 other (N = 22 mutations – No CN loss) had increased TMB, microsatellite instability high (MSI-H), and median neoantigen tumor burden.
Conclusions: ERBB2/ERBB3-alt are associated with significant changes in the tumor microenvironment in GEAC. Co-occurring genetic or immunologic alterations can be exploited to develop effective targeted or immune therapies.
Comparison between ERBB2 and ERBB3 groups regarding co-alterations and IO markers.
|
Overall N = 2,047 |
ERBB2 CN amp N = 252 |
ERBB2 other N = 71 |
ERBB3 CN amp N = 17 |
ERBB3 other N = 22 |
ERBB2/ERBB3 WT N = 1,685 |
p-value/ q-value |
TMB^ (N=1880) |
3.33 (1.92, 4.99) |
3.75 (2.69, 5.00) |
3.46 (2.31, 5.76) |
1.94 (0.82, 3.26) |
5.38 (2.78, 11.72) |
3.33 (1.92, 4.96) |
p = <0.001 |
TMB-H* (N=1880) |
75 (4.0%) |
5 (2.2%) |
4 (5.8%) |
0 (0%) |
6 (27%) |
60 (3.9%) |
p = <0.001 |
MSI-H |
63 (3.1%) |
1 (0.4%) |
2 (2.9%) |
0 (0%) |
5 (23%) |
55 (3.3%) |
p = <0.001 |
PDL1+ (N=1316) |
751 (57%) |
97 (54%) |
21 (54%) |
7 (64%) |
12 (67%) |
614 (57%) |
p = 0.8 |
Neoantigen Tumor Burden^ (N=1768) |
1.46 (0.97, 2.20) |
1.46 (0.98, 2.20) |
1.46 (0.98, 2.47) |
0.49 (0.49, 1.71) |
2.44 (1.22, 8.05) |
1.46 (0.97, 2.20) |
p = 0.008 |
KRAS |
380 (19%) |
22 (8.7%) |
6 (8.5%) |
4 (24%) |
4 (18%) |
344 (20%) |
q = <0.001 |
ALK |
316 (15%) |
28 (11%) |
11 (15%) |
2 (12%) |
8 (36%) |
267 (16%) |
q = 0.064 |
PIK3CA |
142 (7%) |
14 (5.6%) |
6 (8.5%) |
1 (5.9%) |
7 (32%) |
114 (6.8%) |
q = 0.023 |
TP53 |
1292 (63%) |
187 (74%) |
48 (68%) |
5 (29%) |
7 (32%) |
1,045 (62%) |
q = <0.001 |
TMB=Tumor Mutational Burden (mut/MB), * >=10 mut/MB, PDL1+=PD-L1 positive by internal IHC, ^ Median (Interquartile Range) |
VIEW THE PUBLICATION
VIEW THE POSTER