Single-cell circulating tumor cell genomics of KRAS-independent oncogenic sub-populations and longitudinal clonal evolution in metastatic pancreatic adenocarcinoma.

4216 Background: KRAS mutations are a dominant oncogenic driver in metastatic pancreatic ductal adenocarcinoma (mPDAC) and are routinely detected using ctDNA. However, ctDNA reflects pooled signals from dominant high-shedding clones and may incompletely capture intra-patient heterogeneity or tumor evolution. Circulating tumor cells (CTCs) enable single-cell genomic interrogation and longitudinal tracking of evolving sub-populations. We compared genomic alterations, tumor heterogeneity, and pathway dynamics using ctDNA and single-cell CTC DNA in mPDAC. Methods: Under an MCW IRB approved protocol (PREDICT-MCW; NCT05802069), 16 patients with mPDAC receiving standard of care were enrolled. Peripheral blood was collected at baseline and longitudinally at 90-day intervals. Live single CTCs were isolated using the OncoIncytes platform with anti-EpCAM–based glass bead capture and gentle release. Genomic DNA from individual CTCs underwent linear amplification and targeted sequencing using a 1080-gene panel on the Illumina NextSeq 2000. Matched ctDNA was analyzed using the same panel. Variant calling used the iCore pipeline. Alterations were compared at gene and patient levels. Tumor heterogeneity was assessed using mutant-allele tumor heterogeneity (MATH). Results: CTC counts were highest at baseline (mean 6.5 CTCs per patient; range 2–11) and varied longitudinally with marked inter-patient heterogeneity. At first on-treatment assessment, 50% of evaluable patients showed reduced CTC counts, while others exhibited stable or increased levels. Across all timepoints, 49 altered genes were identified. Single-cell CTC profiling revealed broader genomic representation, with 32 genes (65%) detected exclusively in CTC-DNA, 14 (29%) shared with ctDNA, and 3 (6%) detected exclusively by ctDNA. At the patient level, CTC-DNA identified genomic alterations in all patients (16/16), whereas ctDNA detected alterations in 75% (12/16). At baseline, 61% of CTCs achieved ≥11× coverage across KRAS exon 2 and codon 12. KRAS alterations were detected in only 6% (1/16) of patients by CTC-DNA compared with 44% (7/16) by ctDNA, persisting longitudinally (13–19% vs 59–76%, CTC-DNA vs ctDNA). KRAS-negative CTCs harbored recurrent alterations in MAPK, DNA damage response, receptor tyrosine kinase, and tumor suppressor pathways. CTCs demonstrated higher intra-patient heterogeneity than ctDNA (MATH 50.5 vs 25, p < 0.05). Conclusions: Single-cell CTC profiling reveals KRAS-independent tumor sub-populations present from baseline and dynamic clonal evolution not captured by ctDNA alone, revealing tumor heterogeneity and KRAS-independent disease biology. Paired CTC and ctDNA analysis provides complementary resolution of persistent atypical oncogenic and tumor suppressor sub-clones not evident from bulk liquid biopsy.