HER2 expression and extensive molecular characterization of resected brain metastases from colorectal cancer: The HEROES study.

2025 Background: Brain metastases (BM) from primary colorectal cancer (prCRC) are rare (1-3%); few is known about CRC BM predictive factors, prognosticators and molecular pathways. High rate of HER2 amplification ( HER2+) in CRC with BM was previously described, being HER2+ overall rare in CRC (<5%); however, HER2+ impact on prognosis of CRC with BM is uncertain. Enrichment in high microsatellite instability (MSI-H) and BRAFV600E mutations (mut) was also documented in BM tissue compared to matched prCRC. We designed this study to describe molecular landscape and clinical characteristics of CRC with BM, with special focus on HER2. Methods: HEROES was a retrospective-prospective, observational study in which patients (pts) with resected BM from CRC and treated at our institution from 1 January 2010 until 31 December 2021 were enrolled to perform extensive molecular analysis of matched prCRC and BM tissue. Molecular characterization and TMB were obtained with Next Generation Sequencing (NGS, FoundationOne CDx). RAS/BRAF and MSI were respectively assessed with MassArray (Myriapod Colon Status kit) and immunohistochemistry (IHC) to validate NGS results. HER2 status was assessed with IHC/in situ hybridization (ISH). Tumor-infiltrating lymphocytes (TILs) were counted on hematoxylin-eosin-stained tissue. TMB was defined high if ≥ 5.02 mut/mB; TILs if ≥ 1.6 (cutoff set with ROC curve using R software v 4.1.2). Primary objective was to describe molecular landscape of paired BM and prCRC, with special focus on HER2. Secondary objectives were to search for new prognosticators of PFS after BM resection (BM-PFS), intracranial-only PFS (BM-iPFS) after BM resection, and OS in pts with resected BM from CRC. Results: Out of 100 pts with BM from CRC, 22 underwent BM resection and were included in the analysis. 18 out of 22 pts were aged ≥ 70 (82%); 19 (86%) had left colon or rectal origin. 17 (70%) had concomitant lung metastases. HER2+ was found on 4 (18%) BM, 3 (14%) of which had also HER2+ in matched prCRC; 3 (14%) BM carried BRAFV600Emut; 2 (9%) BM had MSI-H. Acquired HER2+ and BRAFV600Emut on BM were reported in two different pts. KRASmut were consistent between BM and prCRC. Factors positively influencing BM-iPFS were low TMB (HR 0.36; 95%CI 0.12-1.10; p = 0.0275) and absence of HER2 ( HER2neg) on BM (HR 0.20; 95%CI 0.03-1,52; p = 0.0013). HER2neg BM were also related to longer BM-PFS (HR 0.35; 95%CI 0.07-1.76; p=0.0402), as well as KRASmut BM (HR 0.35; 95%CI 0.11-1.06; p=0.0096). Longer OS was found in pts with ECOG PS 0-1 (p<0.0001), with ≤ 1 metastatic site (HR 0.30; 95%CI 0.09-0.94; p = 0.0076) or with high TILs in prCRC (HR 0.32; 95%CI 0.10-1.03; p = 0.0368). Conclusions: Even with the limitation of small sample size, this study supports HER2+ enrichment in both prCRC and BM from CRC. HER2neg, low TMB or KRASmut BM conferred better prognosis. ECOG PS 0-1, ≤ 1 metastatic site and TILs-enriched prCRC were related to better OS.