CF102: An Important Driver For Can-Fite

CanFite Biopharma Ltd. (CANF) is currently conducting a Phase 2 clinical study with CF102 for the treatment of hepatocellular carcinoma (HCC). The focus of the study is to test CF102, a selective A3 adenosine receptor (A3AR) agonist, as a second-line treatment of advanced HCC in subjects with Child-Pugh B cirrhosis. The 78-patient trial is currently recruiting patients in Israel and select countries in southeastern Europe (1). I’m expecting enrollment to complete by the end of the year, with top-line data expected in 2017.

The primary endpoint is overall survival (OS), so it is unlikely we will see primary endpoint data before the middle of 2017; however, there are several important secondary endpoints that may readout before that time, including time to progression (TTP), progression-free survival (PFS), objective response rate (ORR), and disease control (DC). Safety, hepatic dysfunction, viral hepatitis, and A3AR expression are other important secondary measures in the study.

I believe top-line data from this ongoing study is incredibly important to CanFite’s stock. The CF102 HCC program, along with the CF101 Phase 3 programs in rheumatoid arthritis and psoriasis, will drive meaningful value creation for shareholders over the next year. CanFite’s currently trades with a market capitalization of only $31 million, meaning the market has yet to factor in any upside from these programs. My article below is a look at the science behind the drug, a review of existing data, and the market opportunity with CF102 in HCC.

CF102 Mechanism Of Action

If you’ve ever looked at an investor presentation for CanFite Biopharma, you know that early in the slide presentation the company posts a question asking, “Why does cancer not metastasize to muscle?” It’s thought-provoking. After all, cancer metastasizes to other organs, bone, skin, lymphatic tissue, and blood. Rarely does cancer metastasize to muscle. Soft tissue sarcomas account for less than 1% of all malignancies, and only 5-10% of those is found in smooth muscle (2). That’s less than 0.005% of all cancers in the U.S.

Researchers have conducted studies to gain a better understanding of this phenomenon. Djaldetti et al., 1996, found that tumor cell lines co-incubated with newborn rat muscle cells showed marked morphological alterations. The authors also found that medium used to grow muscle cells inhibited proliferation of tumor growth (3). The conclusion: muscle cells produce a non-specific low molecular weight factor that can selectively inhibit the proliferation of tumor cells in vitro and in vivo.

Further research lead by CanFite’s current CEO, Pnina Fishman, PhD, in 1999 revealed that muscle cells release a number of anti-proliferative molecules, one of which is adenosine (4), a nucleoside involved in a number of cellular processes including energy production (through adenosine triphosphate) and signal transduction (through cyclic adenosine monophosphate). Fishman et al., 2000 later published work demonstrating adenosine exerts its antiproliferative effect mainly through binding to the A3 receptor (5), and that the cytostatic anticancer activity of adenosine, mediated through the A3AR, is a potential target for the development of anticancer therapies.

In 2007, now at CanFite, Fishman et al. found that A3AR is expressed in very low levels in most normal tissues, but overexpressed in a number of different tumor types including skin, colon, breast, renal, ovarian, and hepatic cancers (6).

Thus, the reason cancer almost never metastasizes to muscle is due to the A3AR agonists that muscle cells release (e.g., adenosine) that, in turn, activate the A3 adenosine receptor, which is found in large quantities on tumor cells. As such, targeting the A3AR through pharmaceutical intervention is a likely mechanism for inhibition of tumor cell growth. And, according to CanFite, the mechanism of action mediated via the A3AR includes modulation of key signaling proteins, such as PI3K, GSK-3β, PKA, PKB/Akt, IKK, and NF-кB, resulting in deregulation of the Wnt and the NF-кB pathways. Consequently, apoptosis of cancer and inflammatory cells takes place (7).

CF102 Preclinical & Phase 1 Data

Since the initial work on the activation of A3AR, Can-Fite has generated several additional publications showing that selective A3AR agonists (i.e. CF102) have anti-cancer activity in a wide range of tumor types, including melanoma (8, 9), prostate cancer (10), colon cancer (11), and importantly, HCC (12, 13). The data in HCC was the most responsive to treatment with CF102 among all the tumor types tested. In addition, a number of important conclusions were drawn from the two studies of CF102 in HCC, including:
– A3AR is highly expressed in liver cancer (A),
– CF102 inhibited the growth of liver cancer tumors in a rat orthotopic model (B), and
– CF102 treatment induced apoptosis in liver cancer cells (C).

The encouraging preclinical results noted above with CF102 in HCC were followed up by a Phase 1/2 trial in HCC patients (NCT00790218). According to the 2013 publication, median overall survival in the study population, 67% of whom had received prior Nexavar®, was 7.8 months, and for Child-Pugh B patients (28%) it was 8.1 months. Stable disease by RECIST was observed in four patients for at least 4 months. CF102 also helped maintain liver function over a 6-month period (14). The authors found a correlation between receptor over-expression levels at baseline and patients’ overall survival. One of the patients who presented with skin nodules that were biopsy-proven to be HCC metastases prior to the trial showed complete metastasis regression during three months of treatment with CF102. This would be an astonishing leap forward from the Phase 3 data in Child-Pugh B patients observed in the global program for Nexavar®.

Nexavar® Failures – A Wide Open Market

Nexavar® (sorafenib), a drug that has been shown to increase median overall survival by just shy of three months – 10.7 months in the sorafenib group vs. 7.9 months in the placebo group (hazard ratio = 0.69 [95% CI 0.55-0.87]; P<0.001) – in a randomized, double-blind Phase 3 study conducted at 121 centers in 21 countries around Europe, North America, South America, Australia, and Asia (15). However, 97% of the patients (581 of 602) in this Phase 3 trial were Child-Pugh A.

A second Phase 3 study with Nexavar® was conducted in 226 HCC patients at 23 centers in China, South Korea, and Taiwan. The median overall survival was 6.5 months in patients treated with sorafenib, compared with 4.2 months in those who received placebo (hazard ratio = 0.68 [95% CI 0.50-0.93]; p=0.014) (16). All 226 patients were Child-Pugh A.

A retrospective study conducted by DA Fonseca et al., 2015, found that Nexavar’s utility in Child-Pugh B patients is significantly lower than in the less advanced, Child-Pugh A, patients. For example, the median overall survival of Child-Pugh B patients from the two Phase 3 studies (n=20) was only 2.53 months [95% CI 0.33-5.92] vs. that of 9.71 months (95% CI: 6.22-13.04) for Child-Pugh A patients; the difference was statistically significant (P=0.002) (17). The results for the Child-Pugh A patients were very similar to what Cabibbo et al., 2012 published (9.8 months vs. 9.7 months), but the results in Child-Pugh B were far worse (6.1 months vs. 2.5 months). This seems to suggest that Nexavar® is simply an ineffective drug in HCC patients with Child-Pugh B cirrhosis.

The Phase 3 data on Nexavar® also show the drug to be poorly tolerated. For example, 80% of the patients reported adverse events (vs. 52% for the placebo group), with the biggest, statistically significant, issues being gastrointestinal side effects such as anorexia, weight loss, diarrhea, nausea, and vomiting. Other issues included alopecia, skin reactions, hypertension, abdominal pain, and bleeding. In contrast, in the CF102 Phase 1/2 trial, no serious drug-related adverse events or dose-limiting toxicities were observed. CF102 also showed no adverse effect on routine measures of liver function over a six-month period in 12 patients treated for at least that duration, suggesting that CF102 could be utilized safely in patients with cirrhosis or decreased liver function, like that of Child-Pugh B patients.

CF102 Can Be Big For Can-Fite

In September 2015, Can-Fite received U.S. FDA Fast Track designation for CF102 as a second-line treatment for hepatocellular carcinoma (HCC) (18). HCC is the most common form of primary liver cancer. This is an indication for which Can-Fite has also been granted Orphan Drug status given the high unmet medical need and lack of effective treatment options for patients failing first-line therapy (19).

American Cancer Society predicts there will be 39,230 new cases of primary liver cancer in the U.S. in 2016 (20). The ACS also predicts 27,170 deaths from the disease in 2016. That being said, liver cancer is much more common in countries in sub-Saharan Africa and Southeast Asia than in the U.S. In many of these countries, it is the most common type of cancer, likely due to the rampant infection rate of hepatitis B and C in these regions. The ACS predicts more than 700,000 people are diagnosed with this cancer each year throughout the world, of which nearly 400,000 or so live in China.

Nexavar® is indicated for the treatment of patients with unresectable hepatocellular carcinoma. The drug is also approved for advanced renal cell carcinoma (RCC). The drug did roughly $1 billion in sales in 2015, and management comments hinted that roughly 40-45% of the global sales is coming from the HCC indication (21). This suggests a penetrable global market for HCC at roughly $400 to $500 million. Nexavar® costs approximately $6,500 per month (source: GoodRx), or around $65,000 for an average 10 month treatment period. Keep in mind, patients will have likely failed transarterial chemoembolization (TACE), a form of intra-arterial catheter-based chemotherapy that selectively delivers high doses of cytotoxic drug to the tumor bed, before they progress to Nexavar®, so these are patients with limited options.

Based on the Barcelona Clinic Liver Cancer (BCLC) algorithm for the treatment of hepatocellular carcinoma, the estimated target market for Nexavar® is around 40% of the diagnosed population. This equates to roughly 14,000 patients in the U.S. It removes the 30-40% of patients that are cured through surgical resection or liver transplant and the 20% that respond to TACE. It also excluded the 10% Child-Pugh C patients that are terminally ill and likely to receive only supportive care.

Based on the available Nexavar® Phase 3 data stratified by liver performance status (22) the market opportunity for a drug like CF102 is still significant. The target patient population includes all Child-Pugh B patients and those that either do not respond to Nexavar® or are unlikely to respond to Nexavar® based on performance status. I estimate this is easily 10,000 to 12,000 patients in the U.S. each year. The course of treatment in these patients may be less than the approximate 10 month average for Nexavar®, but even if Can-Fite prices CF102 at a similar rate of $6,500 per month, at an average course of treatment for eight months that equates to $52,000 per patient.

With a minimum of 10,000 targetable patients in the U.S., the peak opportunity to Can-Fite looks like approximately $500 million. Several new competitors including Bristol-Myers with Opdivo® (nivolumab),  Exelixis (EXEL) with Cabometyx™ (cabozantinib), and potentially others including Medivation, Novartis, and AbbVie are expected to enter the market in the next few years. Nevertheless, if Can-Fite, through a commercial partnership similar to the one Onyx signed with Bayer/Amgen, can accomplish a similar level of penetration in HCC to sorafenib, the U.S. revenue opportunity for CF102 is $250 million. The opportunity in Europe is likely another $150 million in my view.

However, the interesting opportunity for Can-Fite is in countries in Southeast Asia where the rates of primary liver cancer are significantly higher due to epidemic-like infection rates of hepatitis B and C. The World Cancer Research Fund International estimates that 80% of the global primary liver cancer patients on earth are found in less developed countries outside the U.S. and EU. For example, the highest incidence rates can be found in Mongolia, Lao, Gambia, Vietnam, Korea, Thailand, Cambodia, and China. CF102 approval in these regions easily doubles my peak sales forecast.

I believe there is a meaningful opportunity for Can-Fite to sign licensing and distribution agreements in areas like China and Korea. Management has already out-licensed its other clinical-stage candidate, CF101, to Kwang Dong Pharmaceutical in Korea, so this management team has experience partnering with large Korean pharmaceutical companies. They also have experience negotiating term sheets with companies in China and Japan. Licensing deals that provide upfront cash to Can-Fite for CF102 in Asia will help support the U.S. and EU development plans. This cash could help extend the cash runway well into 2018.

CF102 For NASH – An Exciting New Opportunity

On November 23, 2015, Can-Fite announced the development of CF102 will be expanded into NASH based on compelling preclinical data. I wrote an article providing a brief analysis of this data and the opportunity in NASH back in November 2015, and I encourage investors to view that NASH article if they want to learn more.

NASH is believed to be the next big global pandemic given the soaring rates of obesity, diabetes, and metabolic syndrome all over the world. The mechanism of action for CF102 looks like a pan-hepatic improver of liver pathology, applicable to earlier-stage liver diseases such as NAFLD and NASH. I have yet to model sales of CF102 in NASH, but it is clear that this is an area that has attracted significant attention from some of the industry’s largest players.

The preclinical data on CF102 in NASH are impressive, but I believe that human proof-of-concept data must be generated before any monetary value to the asset should be ascribed. That being said, the value of CF102 in NASH could dwarf the value in HCC, or even what I model above for CF101 in RA and psoriasis. The NASH market is enormous and wide-open at this stage (23), so CF102 in NASH represents nice upside to investors in my view. A Phase 2 study is expected to start later in 2016.


Phase 2 data with CF102 in HCC should act as a major catalyst for the shares and create a significant valuation inflection for the company if the data are positive in 2017. CanFite stock has recently taken a hit due to the failure of CF101 in glaucoma. While any failure is a disappointment, I note that my model for CanFite included no contribution from CF101 sales in glaucoma and the failure of CF101 in this indication has no bearing on my expectations in rheumatoid arthritis or psoriasis, both which should start Phase 3 trials in the coming months. I continue to believe CanFite is worth $8-10 per share, and data with CF102 in HCC next year is a major part of that valuation.