Liver Cancer Fusion Predictor
A molecular blood test designed to flag signs associated with the presence of hepatocellular carcinoma, including small tumors. Learn what we report and review the research data.
What the test is designed to do
Detect a blood-based molecular indicator of HCC
The Liver Cancer Fusion Predictor is an HCC surveillance test developed to identify a pattern of cell-free cancer-associated fusion-transcript signals in blood. It is intended to help flag the possible presence of HCC, including small tumors, so that an abnormal result can be evaluated clinically.
The test is not a stand-alone diagnostic confirmation of HCC. It reports a surveillance finding that can inform follow-up rather than imaging-based tumor identification or pathologic diagnosis.
Proposed replacement role: The strategy described in MoleculeDx's FDA Breakthrough Device designation request would replace ultrasound plus AFP for HCC surveillance with the Predictor alone or with ultrasound. This proposed use remains subject to FDA review and authorization.
What we report
A clear Positive or Negative surveillance result
Molecular pattern associated with HCC detected
The result raises concern for the possible presence of HCC and merits prompt clinician-guided evaluation, which may include appropriate liver imaging and other tests. Positive is not the same as a confirmed HCC diagnosis.
HCC-associated signal not detected — HCC is less likely
A negative result is reassuring. The HCC-associated molecular pattern was not detected in this blood sample, suggesting a lower likelihood of HCC at the time of testing. However, a cancer can occasionally be missed.
For patients at high risk, periodic surveillance remains important. A clinician may recommend repeat Liver Cancer Fusion Predictor testing within an individualized monitoring plan. MoleculeDx is pursuing a proposed surveillance strategy in which the Predictor, alone or with ultrasound, would replace the usual ultrasound-plus-AFP approach. That replacement use has not been FDA authorized; patients should follow their current clinician-directed surveillance plan while it is being evaluated.
Illustrative interpretation only; this webpage is not a patient-specific laboratory report. The probability that a positive result truly represents HCC depends on disease prevalence, test specificity, and clinical context and cannot be calculated from the HCC-only subgroup data below.
How the technology works
MoleculeDx's Fusion-detect™ approach analyzes selected cell-free fusion-transcript RNA signals associated with cancer. Molecular assay results inform a proprietary multi-marker prediction process. This differs from measuring only a conventional protein marker such as alpha-fetoprotein (AFP).
What the test assesses: a molecular signal associated with HCC in the blood—not the physical location or size of a liver lesion. Tumor size and staging require clinical imaging and other assessment.
Research findings in confirmed HCC
Encouraging detection signals in smaller tumors
In the supplied research cohort, a fusion-based prediction model that included AFP showed higher observed sensitivity than an AFP-only rule in confirmed HCC cases, including tumors measuring 3.5 cm or less and those within Milan criteria. The AFP-only comparison used a 200 ng/mL threshold—a specific research cutoff, not a comparison against every guideline-recommended AFP strategy.
Small HCC subgroup findings
Sensitivity = proportion of people with confirmed HCC who had a positive result.
Full subgroup results from the supplied cohort
| Confirmed HCC subgroup | Fusion-based model* Positive / Total | Fusion sensitivity (95% CI) | AFP alone** Positive / Total | AFP sensitivity (95% CI) |
|---|---|---|---|---|
| Within Milan criteria | 21/22 | 95.5% (77.2–99.9%) | 8/21 | 38.1% (18.1–61.6%) |
| Outside Milan criteria | 8/8 | 100% (63.1–100%) | 5/8 | 62.5% (24.5–91.5%) |
| Tumors ≤3.5 cm | 9/9 | 100% (66.4–100%) | 3/9 | 33.3% (7.5–70.1%) |
| Tumors >3.5 cm | 20/21 | 95.2% (76.2–99.9%) | 10/21 | 47.6% (25.7–70.2%) |
Data source: User-supplied “Subtable for size and milan-testing cohort-website.docx,” Tables 1–4. * The source defines the “fusion prediction” rule as involving fusion markers plus AFP; proprietary thresholds are intentionally not published here. ** The AFP-only rule uses a cutoff of 200 ng/mL. The Milan-criteria row has 22 model-evaluable cases but 21 AFP-evaluable cases as supplied; this denominator difference should be reconciled in the underlying dataset before any formal comparative statistical claims. “Within Milan criteria” is not identical to “≤3.5 cm”; Milan criteria describe a tumor-burden category used in transplant selection. Confidence intervals are those provided with the source tables.
How to interpret the findings
These exploratory HCC-positive subsets demonstrate detection within the sampled confirmed cancers and support further evaluation of the model for smaller tumors. The nine-case ≤3.5 cm result has substantial uncertainty (95% CI 66.4–100%). The data shown do not measure specificity, positive predictive value, population screening performance, superiority to ultrasound plus AFP, or a reduction in mortality. They also do not by themselves establish the performance of every current commercial test configuration.
How this test fits HCC surveillance
Surveillance: The blood test identifies a potentially concerning molecular signal associated with HCC. Diagnosis: If the result is positive, a healthcare professional evaluates whether HCC is present using appropriate diagnostic imaging and, when indicated, further clinical or pathologic assessment.
Proposed FDA replacement indication: MoleculeDx is seeking evaluation through an FDA Breakthrough Device designation request for a surveillance strategy intended to replace ultrasound plus AFP with the Liver Cancer Fusion Predictor alone, or with the Liver Cancer Fusion Predictor plus ultrasound. This is the proposed intended use, not an FDA marketing authorization or an established claim of superiority over ultrasound plus AFP. The small-HCC data above compare the study's fusion-based model (including AFP) with AFP alone at 200 ng/mL; they do not establish superiority over the ultrasound-plus-AFP regimen.
Current clinical guidance: For eligible at-risk patients, AASLD recommends ultrasound plus AFP approximately every six months. Until an alternative surveillance strategy is adequately validated and authorized for the intended use, patients should follow their clinician's recommended surveillance plan. A positive molecular finding remains an indication for diagnostic evaluation, not a confirmed HCC diagnosis.
In appropriate patients, finding HCC at a stage that permits resection, ablation, or transplantation may expand treatment options; this study did not directly show that this molecular test reduces liver cancer deaths.
Published scientific foundation
A peer-reviewed study by Yu and colleagues in The American Journal of Pathology (2024) investigated circulating fusion transcripts and machine-learning assessment of HCC risk. The specific size/Milan subgroup figures above were supplied separately for this website and are not represented here as independently verified published results.
- Yu YP, Liu S, Geller D, Luo JH. Serum Fusion Transcripts to Assess the Risk of Hepatocellular Carcinoma and the Impact of Cancer Treatment through Machine Learning. Am J Pathol. 2024;194:1262–1271. DOI: 10.1016/j.ajpath.2024.02.017.
- Singal AG, et al. AASLD Practice Guidance on prevention, diagnosis, and treatment of HCC. Hepatology. 2023;78:1922–1965. DOI: 10.1097/HEP.0000000000000466.
Ordering and specimen collection
The current standard package listed in the secure booking portal is $160. Patients can review booking information and arrange mobile blood collection where available. Sample transportation, testing, and review determine the reporting schedule.
For patient-friendly explanations of the positive/negative report, see For Patients. For specimen protocols and clinical questions, contact MoleculeDx or discuss them with your ordering provider.
A blood-based signal for liver cancer surveillance
The Liver Cancer Fusion Predictor is designed to identify HCC-associated molecular signals for surveillance, with a proposed future role replacing ultrasound plus AFP, either alone or paired with ultrasound. That replacement use remains under FDA evaluation and is not yet authorized. A positive result warrants prompt clinical follow-up, including diagnostic imaging when appropriate; it does not independently confirm HCC.