Ovarian ctDNA-MRD and Recurrence Detection

The clinical question this addresses

Epithelial ovarian cancer is notorious for its pattern of relapse. Many patients achieve a complete clinical response after cytoreductive surgery and platinum-based chemotherapy, yet a substantial fraction eventually recur — often from microscopic disease that was never detectable by imaging or serum biomarkers at the end of treatment. This gap between "no evidence of disease" and "cured" is precisely where the concept of minimal (or molecular) residual disease (MRD) lives.

Clinicians have long wanted a more sensitive way to answer two linked questions: (1) After primary therapy, does a patient still harbor residual tumor that has not yet declared itself radiographically? (2) During surveillance, can we detect recurrence earlier than CA-125 or imaging? Circulating tumor DNA (ctDNA) is being investigated as a potential blood-based answer. The important framing up front is that in ovarian cancer this remains investigational — it sits below the line of practice-standard testing [1][3].

What the assay is and how it works

ctDNA is the tumor-derived fraction of cell-free DNA (cfDNA) — short fragments of DNA shed into the bloodstream by tumor cells undergoing turnover. In a liquid biopsy, a routine blood draw is processed to isolate cfDNA from plasma, and assays then attempt to distinguish the tumor-derived fraction from the much larger background of normal cfDNA shed by healthy cells.

Two broad assay strategies are relevant here:

Tumor-informed assays, in which the patient's resected tumor is first sequenced to identify a personalized set of somatic mutations, and a bespoke panel is then built to track those specific variants in serial blood samples. This approach generally maximizes sensitivity for a given patient because it looks for known, patient-specific signals.
Tumor-naïve (tumor-agnostic) assays, which interrogate a fixed panel of commonly mutated regions or methylation signatures without requiring the tumor specimen first.

A key biological feature of high-grade serous ovarian cancer complicates both approaches: TP53 mutations are near-universal, but the overall mutational and shedding characteristics vary, and the amount of ctDNA shed can be low — particularly in the MRD setting where, by definition, residual disease is small [1].

What a ctDNA-MRD result can and cannot establish matters enormously for interpretation. A positive result — detectable ctDNA after curative-intent therapy — indicates that tumor-derived DNA is present in the circulation, which is evidence of residual or recurrent disease. A negative result is harder to interpret: it may mean no residual disease, or it may reflect disease that is present but shedding below the assay's limit of detection. In other words, ctDNA-MRD has an asymmetry — a positive finding is more informative about disease presence than a negative finding is about cure. The analyte reports on tumor DNA in blood; it does not localize the tumor, grade it, or by itself dictate any therapy.

The evidence

The available evidence in ovarian cancer is encouraging in direction but limited in scale, and it is explicitly framed as emerging rather than established.

Hou and colleagues examined ctDNA in epithelial ovarian cancer with a focus on recurrence [2]. Their work supports the broad observation that ctDNA dynamics track with disease behavior and that detection of ctDNA is associated with relapse — reinforcing the biological plausibility of using the analyte as a recurrence signal. The study contributes to the body of data suggesting that ctDNA changes can accompany or precede clinically evident recurrence, though the cohort and performance characteristics should be understood as part of early-phase, hypothesis-generating investigation rather than a validated surveillance tool [2].

Knisely's 2025 analysis of ctDNA-MRD in ovarian cancer is among the more recent contributions to this space [1]. It addresses the specific MRD question — whether ctDNA measured after primary therapy carries prognostic information — and situates ctDNA-MRD within the challenges particular to ovarian cancer: variable and sometimes low tumor shedding, the need to establish clinically meaningful detection thresholds, and the gap between demonstrating prognostic association and demonstrating that acting on a ctDNA result improves outcomes [1]. The evidence suggests ctDNA-MRD is prognostically meaningful in direction, but the data do not yet establish that intervening on ctDNA positivity changes the course of disease [1].

Taken together, these studies make a consistent case: ctDNA is detectable in a meaningful subset of ovarian cancer patients, its presence carries prognostic weight, and it can reflect disease burden over time. What they do not do is provide the kind of large, prospectively validated performance data — with locked assay thresholds, defined sensitivity and specificity in the MRD setting, and demonstrated clinical utility — that would support routine clinical use [1][2][3].

Where it stands: investigational, not standard

The honest status is that no ctDNA-MRD assay is a validated, practice-standard test for ovarian cancer MRD or recurrence monitoring. Expert consensus frameworks characterize ctDNA in ovarian cancer as experimental, reflecting that the available evidence has not yet met the bar for routine adoption [3]. This positioning is a statement about evidence maturity, not about biological implausibility.

Several concrete limitations explain why ctDNA-MRD sits below the practice-standard line:

Analytic sensitivity in the MRD setting. Residual disease, by definition, produces small amounts of ctDNA, and detection near the assay's limit challenges even sensitive platforms [1].
The negative-result problem. Because a negative test may reflect low shedding rather than true absence of disease, a "clean" ctDNA result cannot reliably reassure a patient that no residual disease exists [1].
Clinical utility versus prognostic association. Showing that ctDNA predicts outcome is necessary but not sufficient. The decisive question — does changing management based on a ctDNA result improve survival or quality of life? — has not been answered by adequately powered prospective trials in ovarian cancer [1][3].
Assay heterogeneity and standardization. Tumor-informed and tumor-naïve approaches differ in sensitivity and workflow, and the field lacks harmonized thresholds and validation cohorts of the size needed for clinical deployment [1].

For clinicians and patients, the practical implication is interpretive humility. If ctDNA-MRD testing is performed — most appropriately within a clinical trial or research protocol — a positive result may raise concern for residual or recurrent disease, and ctDNA detection can, in some contexts, be associated with eligibility considerations for therapy classes under investigation. But it should not be used to direct treatment decisions outside of a trial, and it does not replace established surveillance. Any linkage between a ctDNA result and a specific treatment is a matter for the treating team within the context of trial eligibility, not a conclusion a test result can make on its own.

Summary

ctDNA-MRD in ovarian cancer is a scientifically grounded and actively developing approach. The evidence suggests ctDNA detection after primary therapy carries prognostic information and reflects disease activity [1][2]. Yet the field has not demonstrated validated performance thresholds or clinical utility, and authoritative framing continues to classify it as experimental [3]. It is best understood today as a promising research tool that may, with larger prospective validation, eventually inform surveillance — but one that currently belongs below the practice-standard line.

References

Knisely A, et al. ctDNA-MRD in ovarian cancer. Clin Cancer Res. 2025. PMID 40748258.
Hou JY, et al. ctDNA and epithelial ovarian cancer recurrence. Gynecol Oncol. 2022. PMID 36117009.
Ovarian cancer expert consensus framework, 2026 notes — ctDNA experimental framing.

Ovarian ctDNA-MRD and Recurrence Detection: An Emerging but Investigational Tool