CF102

Namodenoson is Can-Fite BioPharma's oral twice-daily small molecule that activates the A3 adenosine receptor, a signaling protein found at unusually high levels on liver cancer cells and at much lower levels on healthy tissue. It is in Phase 3 (NCT05201404, trial name LIVERATION) for advanced hepatocellular carcinoma (HCC, primary liver cancer) in patients with Child-Pugh class B7 cirrhosis [1][2]. The Child-Pugh score is a 5 to 15 point scale used to grade how badly cirrhosis has damaged liver function: class A (5-6) is well-compensated, class B (7-9) is moderate, class C (10-15) is severe. B7 sits at the lighter end of the B band, B8 and B9 are sicker, and this distinction matters because all of the currently approved HCC immunotherapy combinations are restricted to Child-Pugh A. The Phase 3 trial randomizes 471 patients 2:1 with overall survival as the primary endpoint. The commercial thesis is narrow but real: there is no systemic therapy approved specifically for CPB7 HCC, and Can-Fite reported a survival signal in this exact population from its Phase 2 trial that did not hit on the broader CPB cohort [3]. Can-Fite is a small-cap clinical-stage company built around the A3 adenosine receptor mechanism, so the readout is largely binary for the equity (NYSE American: CANF).

Namodenoson is a novel compound, not approved for any indication anywhere, and an oral twice-daily small molecule dosed at 25 mg BID in the Phase 3. It has FDA orphan drug and fast track designations for HCC, plus EMA orphan medicinal product designation, and an additional FDA orphan drug designation in pancreatic cancer. The Phase 3 LIVERATION trial (NCT05201404) opened enrollment in March 2023 with an originally estimated primary completion of February 2026, but ClinicalTrials.gov still shows recruiting status as of mid-2026, meaning the trial is materially behind its original timeline [2]. A realistic readout window is now late 2027 to 2028 unless Can-Fite publicly updates the timeline. Can-Fite has not publicly committed to a specific readout date. Beyond HCC, namodenoson is in Phase 2 for non-alcoholic steatohepatitis (NASH/MASLD, NCT04697810, n=114), with topline expected per company communications in 2026 but no firm month disclosed [4], on the back of a prior Phase 2 published in Aliment Pharmacol Ther 2021 that met its primary ALT endpoint at the 25 mg dose and showed reduction in liver fat [5]. A small Phase 2 in advanced pancreatic cancer (NCT06387342, n=20) is essentially a safety and signal-finding study [6]. The drug carries ChEMBL ID CHEMBL431733 and chemical synonym Chloro-IB-MECA; the sponsor code CF102 still appears in legacy trial records. The asset has been in clinical development since 2008 (NCT00790218 Phase 1/2) [7], so long that RxNorm has assigned a CUI (2622280) despite no approval yet, which is unusual.

Adenosine is a small molecule the body makes constantly from ATP breakdown and uses as a local signal telling nearby cells things like 'slow down' or 'damp the inflammation.' Cells sense it through four receptors (A1, A2A, A2B, A3). The A3 adenosine receptor (ADORA3) is the target here. In healthy liver, A3 receptor levels are low. In hepatocellular carcinoma cells the receptor is overexpressed, and namodenoson exploits that difference [8]. When it binds A3 on a tumor cell, it triggers a signaling cascade that downregulates two of the main 'keep growing' pathways tumor cells lean on, NF-kB and Wnt/beta-catenin, and pushes the cell toward apoptosis (programmed cell death) [8][9]. NF-kB is essentially a master switch for cell survival and inflammation: when it is overactive in a tumor cell, it tells the cell to keep dividing and resist death signals, so shutting it down strips a tumor of one of its core defenses. Wnt/beta-catenin is a growth-signaling relay used heavily in embryonic development that many cancers, HCC in particular, hijack to drive proliferation, so blunting it removes another major proliferation engine. A3 activation in non-cancerous hepatocytes is the opposite story, protective, dampening inflammation and oxidative damage, which is the rationale for the NASH program [9]. The biology is plausible and supported by preclinical and clinical pharmacodynamic data, but it is not genetically validated. No human geneticist points to ADORA3 loss-of-function variants protecting against HCC, and no other A3AR agonist has reached approval in oncology. The strongest piece of human validation is the A3-high-expression-in-HCC observation paired with the Phase 2 CPB7 survival signal [3].

NCT05201404 (LIVERATION) is a randomized, double-blind, placebo-controlled Phase 3 in 471 patients with advanced HCC and Child-Pugh class B7 cirrhosis specifically (B8 and B9 excluded). Patients must have progressed on or be ineligible for first-line systemic therapy. Patients are randomized 2:1 to oral namodenoson 25 mg twice daily versus matching placebo, both on top of best supportive care, in continuous 28-day cycles. The primary endpoint is overall survival, with progression-free survival and objective response rate as secondaries [2]. Placebo is a defensible control because no systemic therapy is approved specifically for CPB7 HCC: atezolizumab plus bevacizumab excludes CPB outright due to variceal bleeding risk, durvalumab plus tremelimumab was tested only in CPA in HIMALAYA, and dose-reduced sorafenib is sometimes given but tolerated poorly in this group. In practice, CPB7 patients today typically receive symptom management and palliative supportive care, with reduced-dose sorafenib used off-label in a minority of cases despite poor tolerability and limited evidence; major guidelines (NCCN, EASL, AASLD) do not recommend systemic therapy for Child-Pugh B patients as a class. So the placebo arm here approximates real-world standard of care, not a no-treatment straw man. The 471-patient target is large for a CPB7 trial because median survival in this population is short and the analysis needs roughly 350 events for adequate power. Pace of recruitment is the unspoken risk: the trial opened in March 2023 with a Feb 2026 estimated completion and remained recruiting through mid-2026, suggesting slow accrual against tight inclusion criteria. Sites are global with concentration in Eastern Europe and Israel, where Can-Fite has historically run trials.

The model puts this drug's approval odds at 10%. That starts from the historical rate for Phase 3 drugs in this area (about 48%), then adjusts based on ten facts about the trial and sponsor. The estimate falls well below that baseline mainly because of heavier-than-usual blinding, the sponsor's thin or weak approval record, weak earlier-phase results, and a randomized design. The remaining factors are near average for this stage and do not shift the number much either way.

Efficacy is the dominant risk. The Phase 2 missed primary OS in the full CPB cohort, and the CPB7 subgroup driving Phase 3 was a stratified analysis [3]. If that signal was a statistical artifact, Phase 3 fails. Mechanism risk compounds it. ADORA3 has not been genetically validated in HCC and there is no second oncology asset hitting this target to compare against, so a failed readout would not just kill the program, it would call the underlying biology into question. Safety risk is low based on cumulative human exposure. Namodenoson has been in clinic since 2008 across HCC and NASH, and the published clinical data review shows benign tolerability, mostly mild gastrointestinal events, with no on-target cardiac signal reported despite A3AR expression in the heart [9]. Execution and financing risk is elevated. Can-Fite (NYSE American: CANF) reported approximately $5.5 million in cash plus $3.0 million in short-term deposits as of December 31, 2025 (about $8.5 million liquid), with management guiding to roughly 12 months of runway, and raised approximately $4.3 million al warrant exercises in March 2026, bringing the liquid position to roughly $12.8 million. Against ongoing Phase 3 burn and an expected 2027-2028 readout, Can-Fite will almost certainly need to raise again before data, and has historically funded operations with dilutive equity raises. Investors should re-check the most recent 6-K for the latest cash figure and an updated runway estimate before sizing any position. Commercial risk depends on label width. If approved narrowly for CPB7 HCC, the addressable US population is in the low five figures annually, and pricing will need to be positioned carefully against a sponsor with no commercial infrastructure. Can-Fite would likely need a partner or acquirer for launch, which is a feature for the equity but a constraint on uptake speed.

Worth watching but small-cap binary. The data point that converts this from watchlist to signal is OS in the CPB7 stratum with a hazard ratio at or below 0.75 and a p-value clearly below 0.05; that would justify accelerated regulatory discussion and would force a partnering or M&A conversation given Can-Fite's size. Check back at the next AASLD (typically November) or ASCO GI (typically January) meeting where Can-Fite has historically presented updates, and watch ClinicalTrials.gov for the move from 'Recruiting' to 'Active, not recruiting' as the enrollment-complete tell [2]. A second checkpoint: any 6-K or 8-K from Can-Fite announcing financing terms before readout, that tells you whether the company can hold the equity through the data event or whether shareholders are about to take dilution. The NASH Phase 2 readout (NCT04697810, n=114), expected per company communications in 2026 but without a firm month [4], is a useful tell on the broader A3AR thesis, but the population and endpoints are different enough that a NASH hit would not directly de-risk the HCC trial. Bottom line: a low-probability, high-orphan-value Phase 3 from a sponsor that has been chasing this single mechanism for over a decade. The trial is well-designed for the question and the question is real, but the target biology has never been validated outside Can-Fite's own program, and the financing clock is the second hand to watch alongside enrollment.