Search

The potential of cell-free tumor DNA (ctDNA) for early tumor detection in asymptomatic patients is yet to be established. In the case of pregnancy associated breast cancer (BC), early detection is of special interest, since it is an entity of special aggressiveness due to a delay in diagnosis, along with the negative effect of mammary gland involution when BC is diagnosed during the postpartum period (PPBC). Indeed, PPBC has double metastatic risk and worst prognosis. With a potential applicability for cancer screening during breastfeeding, here we explored the presence of ctDNA in breast milk (BM) from women with BC associated to pregnancy. Matched samples from breast tumor, plasma and BM from a cohort of 14 women diagnosed during pregnancy or breastfeeding were analysed by droplet digital PCR and a targeted next generation sequencing panel (NGS). Thirteen patients had early-stage disease (11% Stage I, 61% Stage II and 28% Stage III) whilst one was diagnosed at advanced stage. BM harboured ctDNA, since mutations present in the tumor tissue were detected in 86% of the cases by ddPCR and in 71,4% by NGS (difference owing to technique sensitivity). Matched plasma samples had detectable ctDNA levels in only 8% of the cases. In one of the patients, a BM sample collected 18 months prior to BC diagnosis revealed the presence of a pathogenic PIK3CA mutation later detected in the surgically removed tumor. With the ultimate goal of applying the NGS in BM as a technique for early detection of BC in the postpartum period, we have collected samples from healthy volunteers and patients at high risk of developing BC (defined as women becoming pregnant at >40 years or carriers of germ-line pathogenic variants associated with BC -ie: BRCA1, BRCA2, PALB2, RAD51C/D). The application of NGS in BM as a technique for early detection of BC in the postpartum period, identified in a high-risk woman (criteria of enrolment was the age, 46yo) an AKT1 pathogenic mutation in the right-sided BM anticipating by 6 months the radiological diagnosis of a Luminal A tumor, stage pT1bN0M0. In summary, our data provides evidence that ctDNA in BM is highly prevalent even at initial tumor stages, and could be exploited for early breast cancer screening during breastfeeding. Citation Format: Cristina Saura, Carolina Ortiz, Enrique Javier Arenas, Judit Matito, Anna Suñol-Camas, Octavi Cordoba, Alex Martinez-Sabadell, Itziar Garcia-Ruiz, Ignacio Miranda, Clara Morales-Comas, Estela Carrasco Lopez, Cristina Viaplana, Vicente Peg, Paolo Nuciforo, Neus Bayo, Josep Maria Miquel, Marina Gomez-Rey, Guillermo Villacampa, Silvia Arevalo, Javier Carmona, Martín Espinosa-Bravo, Judith Balmaña, Rodrigo Dienstmann, Joaquin Arribas, Josep Tabernero, miriam sanso, Ana Vivancos. ctDNA IN BREAST MILK FOR EARLY DETECTION OF PREGNANCY ASSOCIATED BREAST CANCER [abstract]. In: Proceedings of the 2022 San Antonio Breast Cancer Symposium; 2022 Dec 6-10; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2023;83(5 Suppl):Abstract nr P1-05-14.

5547 Background: Copy number alterations (CNA) arise as a result of somatic changes to chromosome structure, resulting in gain or loss of genomic regions. A likely source of copy number variation is an incorrect repair of DNA damage that causes chromosome instability (CI). Hence, many tumors with a high degree of CNAs suffer CI. This is the case of high grade serous ovarian cancer (HGSOC), where deficiency in homologous recombination (HRD) is highly prevalent, and so are CI and CNAs. Clinically, HR status in HGSOC is a biomarker of iPARP response and required for proper patient management. Available HRD tests measure genomic loss of heterozygosity (LOH) features, closely related to CI. Such features are complex to acquire in samples with low tumor content; overall, 15% of HGSOC specimens fail HR testing. CNA profiling is technically more amenable to lower tumor cellularity and may be captured by a wider range of techniques and applications (gene panels, low pass WGS...). We present here the analysis and identification of CNA features in HGSOC as a novel biomarker of HR status. Methods: A cohort of 123 primary HGSOC tumors were analysed with a custom hybrid capture-based NGS panel (VHIO-300) that provides, along mutations in 450 genes, genome-wide CNA profiles. B-allele fractions, obtained from single nucleotide polymorphisms also in the panel, allowed HRD-LOH score calculation (Marquard et al., 2015). 41 CNA segments correlated to HRD-LOH scores (proportions test, FDR < 0.001) and selected as features to calculate a CNA-HGSOC score (range 0-100). Among them, 10 segments appeared altered almost exclusively in BRCA1/2 pathogenic/likely pathogenic mutant tumors. Results: The density plot of the CNA-HGSOC score showed a bimodal distribution with modes at 7 and 65. A cut-off value of 34 was selected based on the lowest density CNA-score value between the modes. Hence, high CNA-HGSOC was defined as tumors with scores ≥ 34. Our CNA-HGSOC score was strongly correlated as a continuous variable to HRD-LOH (Pearson correlation = 0.72; p < 0.0001) and ROC analysis (AUC = 0.84; CI 95% 0.77-0.91; p < 0.0001) demonstrated high predictability to classify tumors as if using an HRD-LOH test (HRD-LOH score; “very high”: 100-75, “high”: 74-50,”mid”: 49-25 and “low”: 24-0). BRCA mutation status was also accurately predicted using a subset of CNA features (AUC = 0.71; CI 95% 0.59-0.82; p < 0.0001). Conclusions: CNAs may provide a new powerful genomic resource to HRD determination. We identified 41 CNA features in tumors to inform HR status and a subset of 10 revealed mutation status of BRCA1/2 in HGSOC. Upon further validation, a CNA-HCSOC score would be easily implemented in routine analysis pipelines in clinical labs, allowing HR testing or even broaden its application to emerging fields, such as liquid biopsy.

e16302 Background: Approximately 80% of pancreatic ductal adenocarcinoma (PDAC) patients present with advanced disease at diagnosis, with a 5-year survival rate of less than 5%. According to molecular profiling, mutations in KRAS have been described in 90% of PDAC patients, and correlated with poor prognosis. Nevertheless, the invasiveness of tumor biopsy and the difficulties in obtaining quality samples in some patients remains a concern. Circulating free DNA (cfDNA) is considered a surrogate marker of tumor burden and may also act as a prognostic factor that can be used independently from tumor biopsy. In this study, we aim to explore molecular prognostic biomarkers using plasma samples. Methods: Patients with metastatic PDAC harboring KRAS mutation in tissue samples and treated with first-line chemotherapy were identified at Vall d’Hebron University Hospital. Clinical features and plasma samples were collected before starting any treatment. DNA mutations were analyzed in plasma and compared with tissue samples. In plasma, detection of KRAS mutation and total cfDNA concentration were evaluated for clinical significance. A cfDNA minimum 100 genome equivalent threshold was proposed based on results and utility. A Chi-square test was planned to prove the independence between cfDNA levels and KRAS detection in plasma. Survival analysis was performed using the Kaplan-Meier method. Results: We included 33 PDAC patients with a median age of 66.1 years old, 57% were women. All of them were treated with gemcitabine and Nab-paclitaxel. At the final analysis cut-off, 31 patients had died . KRAS mutation was detected in 25 patients’ blood samples (75.8%) and cfDNA-High levels were considered for 20 patients (60.6%). 19 of 20 (95%) patients with cfDNA-High correlated with KRAS blood detection, showing significant dependence (p-value 0.001). Median overall survival (OS) was 10.5 months in the KRAS detected group compared with 22.6 months in the KRAS non-detected group (HR for death 4.70; confidence interval (CI) 1.70-12.98, p-value 0.002). Better separation was seen using cfDNA where median OS was 9.8 months in the cfDNA-High group compared with 21.0 months in the cfDNA-Low group (HR for death 5.38; CI 2.17-13.31, p-value < 0.001). Conclusions: Liquid biopsy can provide quality prognostic information in metastatic PDAC patients. On account of this, the cfDNA level is proposed as a prognostic factor. However, prospective cohorts to validate this information are needed.[Table: see text]