Cell-free DNA (cfDNA) sequencing allows non-invasive detection and monitoring of cancer signals in human bio-fluids. It can be considered as a liquid biopsy. The majority of cfDNA sequencing approaches aim to detect mutations reflecting the presence of cancer cells. The fraction of cfDNA fragments exhibiting tumour-derived mutations is low. The majority of cfDNA molecules being released by cancer cells have no mutations, and thus are not detectable by most liquid biopsy approaches (PMID:36380865). Moreover, such approaches are blind to detect the majority of hallmarks that make cancer.
The NAB team aims to develop experimental and computational methods to non-invasively access cancer biology and improve our understanding of cancer. This encompasses tools to enrich and analyse non-invasively the methylome of cancer patients using our custom T7-MBD-seq approach, together with computational tools to leverage the multiple layers of features within this data.
T7-MBD-seq was used to profile DNA methylation in >1300 samples, including >600 cfDNA samples, across several studies including SCLC models and cfDNA from patients with SCLC. Together with the BBS team, we have developed a sensitive tumour/normal prediction classifier for disease monitoring and utilised differences in methylation patterns between the predominant molecular subtypes, based on NE transcription factors ASCL1,NEUROD1 and a double negative, to derive a molecular subtype classifier. These blood tests are now being validated in clinical cohorts, and other malignancies.
Key Collaborators
• BBS team and Preclinical team, NBC
• Prof. Charles Rudin, Memorial Sloane Kettering Cancer Centre, New York
DNA methylation patterns are cell-specific, and this specificity can inform on the source of cfDNA in the blood. Using T7-MBD-seq data, the NAB and BBS teams have developed a tissue-of-origin (TOO) classifier called CUPiD (Cancer of Unknown Primary ID) to pinpoint the origin of tumour-derived cfDNA. This classifier aims to support treatment decisions in cancer of unknown primary (CUP). CUP describes a metastatic cancer cohort, with unknown primary tumour, making selection of beneficial treatment challenging.
CUPiD allows accurate TOO predictions across 29 tumour classes. CUPiD was tested on 143 cfDNA samples from patients with 13 cancer types, alongside 27 non-cancer controls, with overall sensitivity of 84.6% and TOO accuracy of 96.8%. In an additional cohort of 41 patients with CUP, CUPiD predictions were made in 32/41 (78.0%) cases, with 88.5% of the predictions clinically consistent with a subsequent or suspected primary tumour diagnosis, when available (23/26 patients). Combining CUPiD with cfDNA mutation data demonstrated potential diagnosis reclassification and/or treatment change in this hard-to-treat cancer group. CUPiD is now being validated in larger cohorts, and more challenging scenarios. Beyond CUP, improving the resolution of TOO classifier to deconvolute the cellular and immune origin of cfDNA has an untapped potential in immuno-oncology and toxicology.
Key collaborators:
• BBS team, NBC
• Dr Rebecca Lee, DCR team, NBC
• Dr Natalie Cook, Christie NHS Foundation Trust
A rich landscape of biological signatures can be observed in the blood of cancer beyond mutations and methylation patterns. The structural properties of cfDNA and its fragmentation are not random in cancer. This fragmentomic signal can be harnessed from a range of sequencing data to boost the potential of existing liquid biopsy methods, or directly as a new class of biomarker.
From the whole-genome sequencing, we retrieved the cfDNA fragment-end composition using a new software (FrEIA [fragment end integrated analysis]), as well as the cfDNA size and tumour fraction in three independent cohorts (n=925 cancer from >10 types and 321 control samples). At 95% specificity, we detected 72% cancer samples using at least one cfDNA measure, including 64% early-stage cancer (n=220). cfDNA detection correlated with a shorter overall (p=0.0086) and recurrence-free (p=0.017) survival in patients with resectable oesophageal adenocarcinoma. Integrating cfDNA measures with machine learning in an independent test set (n=396 cancer, 90 controls) achieved a detection accuracy of 82% and area under the receiver operating characteristic curve of 0.96. In conclusion, harnessing the biological features of cfDNA can improve, at no extra cost, the diagnostic performance of liquid biopsies.
Such signal can be complemented by additional features (e.g. fragmentomic signal, mitochondrial DNA (mtDNA), extracellular vesicles (EV)), and is agnostic to the sequencing platform used. For example, using Nanopore sequencing we analysed the genomic and fragmentomic profile of plasma samples from lung cancer patients, and urine from bladder cancer patients, within 24 hours from sampling.
Key collaborators:
• Prof Sarah-Jane Dawson, Peter McCallum Centre, Melbourne, Australia
The NAB team has developed a portfolio of GCP-compliant liquid biopsy tests and applied these to the CAcTUS, DETECTION and DyNAMIc melanoma trials. The CAcTUS trial (CirculAting Tumour DNA gUided therapy Switch) used a ddPCR ctDNA assay to measure mutated BRAF levels in patient blood to inform treatment switch from targeted to immunotherapy for advanced cutaneous melanoma: 44 patients were screened and validated ctDNA data returned to clinic within 7 days. The DETECTION trial (Circulating tumour DNA guidEd Therapy for stage IIB/C mElanoma after sugiCal resecTION) opened to recruitment in late 2021. DETECTION involved ddPCR assays to monitor tumour activity and burden (TAB) levels for the upcoming DyNAMIc trial (Circulating tumour DNA guided Adaptive BRAF and MEK Inhibitor therapy), which uses BRAF V600 ddPCR assays in plasma cfDNA to monitor TAB to inform adaptive BRAF-MEK inhibitor therapy stage III unresectable/IV cutaneous melanoma. DyNAMIc opened in 2024.
Key collaborators:
• QA team, NBC
• Prof Paul Lorigan and Rebecca Lee (UoM/CFT)
