Abstract 1509: Estimation of lesion-specific mutation shedding in cfDNA from rapid autopsy study

Sampling cfDNA using liquid biopsies offers clinically important benefits for monitoring of cancer progression. A single cfDNA sample represents a mixture of shed tumor DNA from all known and unknown lesions within the patient. However, the amount of DNA shed by any one specific lesion is not well characterized. We design an LSM (Lesion Shedding Model) that is able to order the lesions from the strongest to the poorest shedding for a given patient. Our framework intrinsically models for missing/hidden lesions and operates on blood and lesion cfDNA assays to estimate the potential relative shedding levels of lesions into the blood. We verify the accuracy of the model under controlled conditions using a simulation approach to create a synthetic blood cfDNA sample per patient, where specific lesions are assigned predefined shedding levels. The simulated data mirrors real data lesion genomic similarities. The LSM correctly obtains a partial order of the lesions, i.e. accurately stratifies the lesions by their assigned shedding levels, in 83% (10/12) simulated data sets. We then applied LSM to a cohort of 12 cancer patients with available cfDNA blood samples and multiple biopsied lesions and found that indeed there were lesions that were consistently shedding more than other lesions into the patients9 blood. From this analysis we are then able to probe whether there are detectable tissue-specific differences in the shedding behavior of lesions. In most patients a top shedding lesion(s) was identified. Trends emerged across lesions whereby pancreatic and gallbladder lesions tended to be higher shedders than other lesions and abdominal cavity lesions tended to be low shedding lesions. The LSM represents one of the first attempts to characterize the contributions of distinct lesions in cfDNA. With an accurate understanding of lesion cfDNA contribution, we can begin to look for patterns in shedding behavior that associate with clinical features of the lesions, such as response to treatment, tissue of origin, age, growth, etc. As sample size increases, the LSM will also be able to address the question of whether a cfDNA biopsy would provide sufficient coverage of the entire space of lesions present in the patient, or, if biopsies of individual lesions would be necessary. Such insights would further increase the impact and applicability of liquid biopsies in the clinic. Citation Format: Filippo Utro, Kahn Rhrissorrakrai, Chaya Levovitz, Ignaty Leshchiner, Liz Martin, Christopher Chen, Christopher Pinto, Brian Danysh, Kara Slowik, Lipika Goyal, Read Allen, Maida Williams Broudo, Dejan Juric, Gad Getz, Laxmi Parida. Estimation of lesion-specific mutation shedding in cfDNA from rapid autopsy study [abstract]. In: Proceedings of the Annual Meeting of the American Association for Cancer Research 2020; 2020 Apr 27-28 and Jun 22-24. Philadelphia (PA): AACR; Cancer Res 2020;80(16 Suppl):Abstract nr 1509.