Henry G Kaplan, Anna B Berry, Alexa Dowdell, Brian Piening
Background: Simultaneous ductal carcinoma in situ (DCIS) and invasive breast cancer (IBC) is very common. However, it is not clear how often these histologies are clonally related or the mechanism by which DCIS evolves into IBC. In this study we evaluated tissue samples of DCIS and IBC by DNA and RNA sequencing from 39 patients with concurrent disease.
Methods: Patients were identified from the Swedish Cancer Institute Breast Cancer Database starting sequentially with 2018 backwards to identify candidates with simultaneous DCIS and IBC. Individual pathologic review was performed from the cohort of patients who had had modified radical mastectomy to select 50 cases for analysis of normal breast, DCIS, and invasive disease. Sequencing was performed using the Tempus xT solid tumor assay (DNA-seq of 648 genes at 500x coverage, full transcriptome RNA-seq). The mutations identified by this assay include germline and/or somatic single nucleotide variants, insertions/deletions, and copy number variations. Results: Thirty-nine cases had adequate tissue for both DCIS and IBC for DNA and 31 for RNA analysis. DCIS and IBC were in the same quadrant of the breast for all cases. All IBC cases were ductal carcinoma. 38/39 IBC were estrogen receptor (ER) positive, 33/39 were progesterone receptor (PR) positive. 5 were HER2/neu amplified (three ER+/PR-, 1 ER+/PR+, 1 ER/PR-). 32/39 had Ki67 of 20% or less (IQR 14(9,19.5)). 35/39 were T1 or T2. Three were T3 and one T4. 21/39 were N0, 4/39 N1mic, 9/39 N1a, 3/33 N2a 2/39 N3a.For DCIS 24/39 had both genomic variants and copy number changes (V+C+), 7/39 were V+C-, 5/39 were V-C+ and 3/39 V-C-. For IBC 25/39 were V+C+, 7/39 were V+C-, 6/39 were V-C+, and 1/39 V-C-. In total, there were 36 unique genomic variants seen in DCIS but not the corresponding invasive sample. In contrast, there were 32 total unique variants seen in invasive but not the corresponding DCIS and 38 variants were seen in both samples. The one BRCA variant seen in both samples of a single patient was BRCA1 p.Q284, a known pathogenic variant but gBRCA normal. There were 86 copy number changes seen in DCIS but not IBC, 75 copy number changes seen in IBC but not DCIS and 47 copy number changes seen in both DCIS and IBC. 24/39 patients shared at least one common variant and 22/39 shared at least one copy number change. 4/39 had no copy changes in either sample but all of these had genomic variants. 5/39 had no variants in either sample but all of these had copy number changes. PIK3CA was seen in 2/39 IBC, 2/39 DCIS and 2/39 in both. TMB IQR was 1.6(1.1, 3.4) for IBC and 2.1 (1.1,3.2) for DCIS. All samples tested for MSI were stable. Only one patient had identical genomic variants and copy changes in both DCIS and IBC. RNA sequencing revealed 1178 genes with significant gene expression differences between paired DCIS and IBC (q<0.001). RNA-seq data was quantified using Kallisto and differential gene expression was assessed using DESeq2. Pathway enrichment was performed using rank GSEA against the MSigDB C2 collection (filtered to KEGG, Reactome and Biocarta pathways only). Upregulated pathways in DCIS involved more immunogenic pathways while IBC upregulated pathways involved proliferation and cell cycling.
Conclusions: The transition from DCIS to IBC is a dynamic process. How much of these differences are the result of tumor heterogeneity and temporal factors as opposed to truly separate clones will require more study. This could have significant implications for the treatment of DCIS.
View the full publication here.