Prospective longitudinal liquid biopsy study
Dynamic circulating tumour cell phenotypes and ctDNA evolution during PARP inhibitor therapy in HGSOC
Functional CTC heterogeneity and longitudinal molecular evolution in HGSOC
Can liquid biopsy reveal emerging treatment resistance before clinical relapse?
PARP inhibitor · schematic
HGSOC is a heterogeneous disease
High-grade serous ovarian cancer (HGSOC) contains diverse tumour subclones. Differences in DNA repair can shape how they respond to treatment.
PARP inhibition applies pressure
PARP inhibition exploits homologous recombination deficiency (HRD), making cells with impaired DNA repair more vulnerable to treatment.
Treatment-sensitive cells respond
As treatment-sensitive cells decline, imaging and CA-125 can reflect the clinical response.
A subpopulation may persist
Some tumour cells may survive treatment. Months can pass before clinical relapse, leaving their survival during therapy poorly characterised.
Objective
To longitudinally characterise functional CTC phenotypes and ctDNA molecular evolution during PARPi maintenance therapy.
Study design
Eleven patients, followed through treatment
A prospective longitudinal cohort of 11 patientswith advanced HGSOC receiving PARPi maintenance therapy, with peripheral blood collected at three points.
PARP inhibitor maintenance therapy
Cellular readout
Circulating tumour cells
Isolated, enumerated, and functionally characterised byPARP1 andγH2AX expression.
Molecular readout
Circulating tumour DNA
Tempus xF 105-gene cfDNA next-generation sequencing panel, in patients with sufficient plasma availability.
Sequential analysis
CTC burden, phenotypic profiles, variant allele frequency (VAF) and molecular evolution during treatment.
One blood sample. Two complementary windows into tumour evolution.
Per-patient CTC counts
View per-patient counts
| Patient | Outcome | Baseline | 6 weeks | 12 months |
|---|---|---|---|---|
| HSA-01 | Relapse | 0 | 0 | 16 |
| HVN-01 | No relapse | 7 | 1 | — |
| HVN-02 | Relapse | — | — | — |
| HVR-01 | Relapse | 1 | 0 | 6 |
| HVR-02 | Relapse | 9 | — | 4 |
| HT-01 | Relapse | 0 | 0 | — |
| HT-02 | No relapse | 8 | 3 | 0 |
| HT-03 | No relapse | 0 | 0 | 21 |
| HT-04 | No relapse | 4 | 4 | 0 |
| HT-05 | Relapse | 26 | 0 | — |
| HJ-01 | Relapse | 0 | 0 | 0 |
Baseline
60.0%
CTCs were detected in 60.0% of evaluable patients, with marked interpatient variability.
6 weeks
An early decrease in CTC counts was observed after 6 weeks of PARPi treatment.
12 months
CTCs persisted or re-emerged at 12 months in a subset of relapsing patients, suggesting dynamic biological adaptation.
PARP1⁺ / γH2AX⁻
PARP1 POSITIVE
The CTCs expressed PARP1: the enzyme the therapy is designed to inhibit was present in the cells still circulating.
γH2AX NEGATIVE
But γH2AX, the marker of DNA double-strand breaks, was largely absent. γH2AX-positive CTCs were rarely observed during follow-up: these cells were not registering the damage the treatment should be causing.
PARP1⁺ / γH2AX⁻ CTC PHENOTYPE
Patients with this CTC phenotype, by relapse outcome
80%
Patients with relapse
4/5 patients
25%
Patients without relapse
1/4 patients
This phenotype was detected more frequently in patients who developed relapse, potentially compatible with reduced DNA damage response activation.
Preliminary findings from a prospective longitudinal cohort.
TP53 · low-frequency variants, all below 1% VAF, tabulated in the text. The chart shows BRCA variants only.
View all reported variants
| Patient | Variant | Category | Baseline | 6 weeks | 12 months |
|---|---|---|---|---|---|
| HT-04 | BRCA1 p.Y655fs | Biologically relevant | 49.1 | 49.0 | 50.4 |
| HT-04 | TP53 p.M237I | Biologically relevant | 0.3 | 0.3 | — |
| HT-04 | TP53 p.D259Y | Biologically relevant | 0.2 | 0.4 | — |
| HT-04 | TP53 p.P151S | Biologically relevant | 0.1 | — | — |
| HT-04 | PBRM1 p.Y470C | Unknown significance | 49.8 | 48.0 | 51.0 |
| HT-04 | BRCA2 p.E1895K | Unknown significance | 48.4 | 50.4 | 51.4 |
| HT-04 | SDHA p.Q176H | Unknown significance | 0.5 | 0.3 | — |
| HT-05 | BRCA2 p.L2092fs | Biologically relevant | 49.7 | 49.9 | 47.3 |
| HT-05 | PMS2 p.P319S | Unknown significance | 46.2 | 47.5 | 46.5 |
Two variants, followed across a year
Sequential cfDNA sequencing tracked pathogenic BRCA alterations in patients HT-04 and HT-05 through all three timepoints.
Stable, and close to 50%
Persistent pathogenic BRCA alterations showed stable VAFs close to 50% across sequential samples, supporting their likely germline origin.
Following TP53
Three low-frequency TP53 variants (p.M237I, p.D259Y and p.P151S; VAF 0.1–0.4%) were detected in HT-04 at baseline, two at 6 weeks and none at 12 months.
| Variant | Baseline | 6 weeks | 12 months |
|---|---|---|---|
| p.M237I | 0.3 | 0.3 | — |
| p.D259Y | 0.2 | 0.4 | — |
| p.P151S | 0.1 | — | — |
TP53 is mutated in around 96% of HGSOC (TCGA 2011; Ahmed 2010), yet it was not detected in HT-05.
Non-detection is not clearance: near the assay's limit of detection (SNV sensitivity 93.8% at ≥0.25% VAF; Finkle 2021), calling depends on sample depth.
Below 1% VAF, tumour-derived ctDNA cannot be told from clonal haematopoiesis without matched leukocyte DNA.
Two windows onto the same evolving tumour
Function
CTCs
PARP1 / γH2AX
Genome
ctDNA
mutations / VAF
Tumour evolution
PARPi resistance?
The question mark is intentional. These are eleven patients, followed prospectively — enough to describe a pattern, not to establish a mechanism.
Longitudinal liquid biopsy enables real-time monitoring of functional CTC heterogeneity during PARPi treatment.
Persistence of PARP1⁺/γH2AX⁻ CTCs may reflect resistant tumour subpopulations associated with relapse.
Sequential ctDNA showed germline BRCA alterations and only low-frequency TP53 variants, whose origin remains unresolved.
Take-home message
Integrating functional CTC phenotyping with sequential cfDNA profiling may provide a minimally invasive strategy for treatment monitoring and identification of resistance mechanisms in HGSOC.