In this study, we found that HER2-amplified PDAC has a prevalence of 2%, is detectable using contemporary genomic approaches, is associated with a clinical phenotype characterized by metastatic spread predominantly to the lungs and peritoneum with local recurrence, and can metastasize to the brain but tends to avoid the liver. It bears molecular similarities to HER2-amplified breast cancer, and is yet to be adequately assessed for potential responsiveness to anti-HER2 therapy. Multiple studies in large cohorts have shown that HER2-amplified breast cancers more commonly metastasize to the lung and the brain . In PDAC, clinical trial data indicate that the first site of distant metastases occurred most commonly in the liver (50%), and occurred in the lung in 9% of the cases . Autopsy studies have reported that 80% of distant metastases are to the liver, which occurred either alone or in combination with peritoneal and or lung metastases, while metastases sparing the liver made up the rest, and occurred in the peritoneum, lung, adrenal glands, and lymph nodes . Cerebral metastases were not found in these studies. In our cohort of 469 patients, the incidence of lung metastases without liver metastases was 8%, comparable to previous studies. We detected only one case with brain metastases (0.2%), although only this single patient was investigated specifically for brain metastases. If we include all 10 HER2-amplified cases with documented metastatic disease, none had evidence of liver metastases (P = 0.0028), and the rate of lung metastases was 50% (P = 0.0022). These data suggest that HER2-amplified PDAC may have a distinct clinical phenotype, and that liver metastases are not determined by physical factors such as portal blood flow, but by the pathophysiology of disease.
These findings have significant clinical implications. First, the detection of small lung nodules should not delay the diagnosis of metastatic disease originating in the pancreas or at relapse if the liver and other sites are clear, particularly in known HER2-amplified cases. Second, if an individual is known to have an HER2-amplified PDAC, then monitoring for disease progression in non-traditional sites such as the lung and the brain, with vigilance for neurological symptoms may be prudent. Finally, there is potential for anti-HER2 therapies in this subset of patients.
In situ hybridization studies in our reference laboratory identified HER2 amplification only in PDACs with high protein expression by IHC (score 2+/3+). Therefore, a reasonable and cost-effective approach to universal HER2 screening is to initially test all cases with IHC and then perform secondary ISH testing only on cases with 2+/3+ staining, as was initially performed for breast cancer. Using this approach, 8% of PDAC (2+ and 3+ cancers) will require HER2 ISH assessment, and of these one-quarter will be amplified.
In the current study, in-depth genomic analysis, apart from HER2 amplification, did not reveal any features that are atypical of PDAC, with mutations of KRAS and TP53 and loss of SMAD4 found, although the inherent heterogeneity of PDAC makes it difficult to draw conclusions about the other mutations detected. mRNA expression profiles clustered with HER2-amplified breast cancer, suggesting that HER2 may be an important driver of carcinogenesis in this subgroup of PDAC.
It is interesting to note that all three HER2-amplified cases with available genomic data harboured the KRASp.G12V mutation. This mutation is less common than the p.G12D mutation, and accounts for 32% of KRAS mutations in PDAC versus 40% for p.G12D . Given the small numbers of HER2-amplified cases, further studies of larger cohorts will be required before it can be determined if this association is sufficiently robust to be used diagnostically or targeted therapeutically.
Two clinical trials have assessed targeted trastuzumab therapy in PDAC [15, 16]. Both are single arm phase II trials used in combination with gemcitabine  and capecitabine . Although the latter performed HER2 FISH for 2+ expressing cases, the former did not, and neither verified the 3+ IHC cases by FISH. In addition, these were not standardized assays performed in reference laboratories, and resulted in a HER2 positive rate of over 10%. This likely overestimation underpowered the trials by over 80%, making a negative result uninterpretable.