Hutchison China MediTech — Update 9 December 2016

Investment summary

All eyes on TKIs: Future stars are aligning

Spotlight on NSCLC: Triple-pronged approach

Fruquintinib: Best-in-class VEGFR inhibition?

Savolitinib targets TKI resistance in NSCLC

Epitinib targets NSCLC with brain metastasis

Targeted therapies transform cancer treatment

Sensitivities

Valuation

Financials

Hutchison China MediTech (Chi-Med; HCM) is an innovative biopharma company focused on the highly lucrative global oncology and immunology markets. This note focuses on the Innovation Platform unit; we take a deeper dive into the group’s tyrosine kinase inhibitor (TKI) pipeline for non-small cell lung cancer (NSCLC), specifically savolitinib, fruquintinib and epitinib following the recent publication of positive top-line Phase Ib epitinib and full Phase II fruquintinib data at IASLC World Conference on Lung Cancer (WCLC). HCM has built a substantial pipeline of highly differentiated TKI drugs; we expect continued progress of the mid- to late-stage pipeline during the coming 12 months (including potential US and China regulatory filings) to catapult the company into the international spotlight. Additionally, HCM’s profitable Chinese healthcare business continues to benefit from the fast-growing domestic market; H116 results demonstrated robust double-digit growth despite currency headwinds. For a broader and more in-depth note on the group including the company’s full portfolio, see our note Stellar Evolution, published in May 2016.

We have increased our sum-of-the-parts (SOTP) valuation to $2.4bn (£32.2 share) from $2.3bn (£25.7/share). In US dollar terms, the overall value of the company has increased by $0.1bn, reflecting amendments to deal terms on savolitinib (we have increased R&D expenses and the royalty rate on sales) plus upgrades to our commercial platform (CP) business forecasts after a strong first half of the year. However, our per share valuation in sterling has increased materially related to the significant weakening of the pound versus the dollar (spot $1.25/£). We value the innovation platform (IP) at $1,789m; and placing CP’s 2016e share of net profit on a 22.5x rating gives $657m (867p/share). Adding in net cash of $80.6m at end June 2016 and netting out unallocated costs results in a value of $2.4bn (£32.2 share).

HCM is subject to the usual biotech and drug development risks, including clinical development delays or failures, regulatory risks, competitor successes, partnering setbacks, and financing and commercial risks. Expectations for the pipeline have increased and while our risk-adjusted NPV highlights the future sources of upside to the shares, the failure of one or more products to succeed would have a negative impact on the shares; savolitinib, fruquintinib and sulfatinib contribute ~65% to our valuation of HCM. CK Hutchison’s involvement in HCM materially reduces the myriad of risks associated with any direct investment in China; however, it does mean investors are minority shareholders. Additionally, the limited available free float reduces the shares’ liquidity.

HCM’s CP business posted robust H1 results, prompting us to upgrade 2016 sales and net income forecasts. Additionally, the amendment of the AstraZeneca (AZN) deal on savolitinib has increased our R&D expense expectations in the near term, offset in the longer term by a higher expected royalty rate on sales ex China for the drug across all indications (now 14-18% from 9-13%). The group remains well funded in the near term; we expect future cash resources to be bolstered by profits generated by the CP business in China as well as the anticipated property windfall relating to cash compensation on land from the Shanghai government.

HCM is approaching an inflection point; it is on the brink of the transformation of its Innovation Platform (IP) business into a late-stage portfolio story. A number of its next-generation TKI assets could reach the US (and ROW) and/or China market within the next few years (potential NDA submissions in the US and or China for savolitinib and fruquintinib, respectively, in 2017). We expect R&D driven newsflow, including multiple top-line clinical trial data plus initiation of pivotal registration trials over the coming 12 months, to focus investor attention on the pipeline potential; we summarise the newsflow by asset below. This note addresses HCM’s NSCLC portfolio, addressing savolitinib, fruquintinib and epitinib; collectively these three assets (across all oncology indications including NSCLC) contribute £18.4/share to our HCM valuation.

Savolitinib C-Met inhibitor for PRCC/RCC/NSCLC/gastric cancer

Fruquintinib VEGFR inhibitor for CRC/NSCLC/gastric cancer

Epitinib EGFR inhibitor for NSCLC with BM

Sulfatinib VEGFR/FGFR inhibitor for neuroendocrine tumours (NET)/thyroid cancer

Following the presentation of positive, proof of concept, top-line Phase Ib epitinib and full Phase II fruquintinib data at IASLC WCLC, 4-7 December 2016, we discuss the data in full within the context of HCM’s NSCLC portfolio. HCM has three tyrosine kinase inhibitors in clinical development for NSCLC; each drug has a specific positioning in the ever-evolving armament to treat the complex and constantly mutating disease that is cancer:

One of the clear strategic visions presented by HCM management is a focus on target selectivity, with the aim of having a potential first-in-class or best-in-class portfolio of TKI drugs within the targeted therapy space of cancer drugs. Multi-target kinase inhibitors, while efficacious, are hindered by side effect profiles (elicited from multiple target protein binding) and consequently lower clinical utility due to dose interruptions/reductions and or discontinuation of drug therapy. HCM is focused on designing and progressing candidates with superior kinase selectivity (eg savolitinib is a highly selective c-Met inhibitor); higher selectivity translates to lower dose reductions/interruptions and thus higher/more prolonged target coverage resulting in superior efficacy and better tolerability compared to first-generation TKIs. Below we discuss NSCLC pipeline in more detail. Exhibit 1 highlights the pivotal trials currently underway (not limited to the NSCLC indication): the seven shots at pivotal success as highlighted by the company.

Fruquintinib is an oral small molecule inhibitor to all three VEGF receptors. Fruquintinib is an oral small molecule that is a highly selective VEGFR1, VEGFR2 and VEGFR3 inhibitor, which in preclinical trials demonstrated fewer off-target toxicities, allowing higher drug exposure that translates to 24 hours a day VEGFR receptor inhibition. Fruquintinib’s most advanced indications are in NSCLC (third-line) and CRC (third-line) in China, and pivotal trials are currently underway in these indications; Exhibit 4 summarises its ongoing clinical trial programme. We assume success in third-line NSCLC and CRC will lead to development of fruquintinib in the second and first line settings.

The first-generation VEGF/R inhibitors (examples include the monoclonal antibody Avastin [Roche]) had combined reported sales of approximately $8.7bn in 2015 across multiple tumour types. This class of drug revolutionised the treatment of cancer by targeting the growth of blood vasculature that is essential for tumour growth (anti-angiogenesis). Avastin has become a gold standard for the first-line treatment of NSCLC globally. A Phase III study (BEYOND) evaluating Avastin plus platinum doublet chemotherapy for first-line treatment in advanced non-squamous NSCLC in a Chinese patient population led to median OS of 24.3 months. Oral, small molecule VEGFR inhibitors Nexavar (Bayer) and Sutent (Pfizer) are approved for use in some cancers but not NSCLC given lack of efficacy in Phase III trials, highlighting the unmet need for an oral anti-angiogenesis agent with a demonstrable impact on progression-free survival (PFS) in NSCLC.

In a Phase II, double-blind study, fruquintinib met the primary endpoint and significantly improved progression-free survival compared with placebo for third-line advanced NSCLC patients who had previously failed two lines of treatment. Data was presented at the recent IASLC World Conference on Lung Cancer (4-7 December 2016).

In the 91-patient study, patients were randomised to receive 5mg fruquintinib (n=61) once daily (three weeks on treatment followed by one week off) plus best supportive care (BSC) or placebo (n=30) plus BSC. The primary objective of PFS was met; as of August 2015, the fruquintinib group had a median PFS of 3.8 months versus 1.2 months on placebo (HR=0.27, p<0.001) (Exhibit 2). Of note was the disease control rate (DCR) of 70.5% on the fruquintinib arm compared to 16.7% on placebo (Exhibit 3). DCR or clinical benefit rate (CBR) is the percentage of patients who demonstrate a response to treatment, eg tumour shrinks or remains stable, ie it is the sum of complete, partial and stable disease rates. This data is highly encouraging given the third-line setting (Exhibit 2 and 3).

Comparison of overall survival between wild type and mutant demonstrated a marked benefit in EGFR mutant patients. Median overall survival in mutant was 8.44 months for fruquintinib (n=30) vs 5.49 months for placebo (n=15), (p=0.184). Median overall survival for wild type patients treated with fruquintinib was less then placebo (7.5 months vs 9.7 months, p=0.274). However, beyond 20 months, fruquintinib demonstrated a survival benefit. Due to the low power of this subgroup comparison and high p-values the significance of these results cannot be confirmed.

In terms of toxicity, 100% of patients on the fruquintinib arm had a treatment emergent adverse event vs 90% on placebo. The most common grade 1-2 events were palmar-plantar erythrodysaesthesia (PPE) syndrome (42.6%) and proteinuria (31.1%), and the most common grade 3-4 AE was hypertension (8.2%). Side effects of fruquintinib to date appear consistent with VEGFR inhibition. Hypertension, proteinuria and PPE are common side effects of many anti-cancer drugs, eg capecitabine, fludarabine and targeted therapies (including sorafenib and sunitinib).

FALUCA, the Phase III registration trial in China, was initiated in December 2015 as third-line treatment for non-squamous non-small cell lung cancer, following a positive Phase II proof-of-concept (POC) trial that reached the primary endpoint of PFS with no unexpected safety issues. FALUCA is a double-blind, placebo-controlled Phase III evaluating fruquintinib 5mg once-a-day plus best supportive care in four-week treatment cycles (three weeks on drug, one week off). 521 patients are being randomised to the fruquintinib group or the placebo group at a ratio of 2:1 across 45 centres in China. The primary endpoint is overall survival; secondary endpoints include progression free survival, objective response rate, disease control rate and duration of response; top-line data is expected early 2018; trial enrolment is estimated to complete mid 2017.

HCM expects to submit an NDA for fruquintinib in NSCLC to the China FDA in 2018, paving the way for a potential approval late 2018; we model 2019. The speed of any approval is less clear than with the US FDA, but we expect it to take between six and 12 months. We assume that the speed will depend on the quality of the data and a faster approval (nearer six months) could be achieved if the Phase III data is substantial. We expect further fruquintinib NSCLC trials to initiate in 2017, likely in combination with other targeted therapies. Notably the company has indicated a Phase Ib/II trial with Iressa in first-line NSCLC patients will start in H117.

Pivotal Phase III data for the treatment of third-line Chinese patients with colorectal cancer (CRC) will be HCM’s first Phase III readout and is a major inflection point for the company. We expect the data to be published in February/March with an NDA submitted to the Chinese FDA in mid-2017, with an approval six to 12 months later. The Phase III trial completed enrolment of 416 patients in H116; it is a randomised, double-blind trial comparing fruquintinib to placebo in third-line or above colorectal cancer patients. The primary endpoint is overall survival and patients who have had previous treatment with VEGF/R inhibitors are excluded.

HCM and partner Lilly are developing fruquintinib as a potential best-in-class drug due to its better side effect profile than the first-generation drugs, for it to compete in the global setting. Consequently, fruquintinib is being assessed in late-stage clinical trials across multiple indications in China. In October 2013 Lilly signed a deal to co-fund development of fruquintinib for the Chinese market, worth up to $86.5m in upfront fees and milestones, with tiered royalties (15-20%), with an option for global development. HCM has received a $6.5m upfront fee (2013), and as of December 2015 $19.2m in development milestones for the positive phase II POC clinical trials in CRC and NSCLC plus $13.9m in reimbursement for certain developmental costs. Furthermore if Lilly exercises its option to develop fruquintinib outside of China, HCM could receive up to an additional $300m of developmental, regulatory and commercial milestones payments. Importantly the partnership agreement with Lilly has enabled HCM to establish a manufacturing facility in Suzhou, China, which is currently producing Phase III clinical supplies and will on approval be used to produce fruquintinib for commercial supply.

In order to compete in the global anti-angiogenesis inhibition setting, clinical trials will focus on more proprietary combination studies (eg, fruquintinib plus savolitinib in clear cell renal cell carcinoma, fruquintinib plus Iressa in first-line NSCLC, fruquintinib plus Taxol in second-line gastric cancer). Fruquintinib’s lack of CYP450 inhibition/inducing is favourable for potential in combination treatment regimens, given that many drugs are metabolised through the cytochrome p450 enzyme pathway.

The relatively low penetration of Western developed drugs into China coupled with different patient demographics has generated a market that is open for opportunity. Many non-specific receptor tyrosine kinase inhibitors are marketed or in development that target combinations of VEGFR 1, 2 and 3. Due to their broad targeting they often have higher incidence of adverse events; as such more specific inhibitors of VEGFR 1, 2 and 3 are needed. In addition to fruquintinib there are four small molecule VEGFR inhibitors in development within China that we believe are comparable.

While at first glance it appears that the competition is intensifying in the China NSCLC targeted therapy market, we make the following observations. Fruquintinib’s lack of cytochrome-450 inhibition should enable high doses and a cleaner toxicity profile then competitors anlotinib, apatinib and famitinib. This enables it to be delivered in combination with chemotherapy agents that typically demonstrate increased toxicity when used with CYP450 inhibiting agents. Further improvements in toxicity could become evident as of the three VEGFR1, 2, 3 selective inhibitors (fruquintinib, anlotinib and lenvantinib) fruquintinib is the most selective. Fruquintinib’s reported grade 3+ AEs appear lower than the competition highlighting the rationale for combination studies. Finally, while pricing information is unknown at this time, we expect fruquintinib to be competitively priced. We expect HCM to capitalise on its knowledge of the Chinese market and to drive adoption through the utilisation of its existing sales structure. The market in China remains untapped, especially by Western companies; there is room for multiple agents with demonstrable efficacy in the first-, second- and third-line settings.

We forecast global peak sales for fruquintinib of $2.3bn across the potential CRC, NSCLC and gastric cancer indications. Exhibit 4 details fruquintinib’s full clinical trial programme, incident rates and penetration assumptions. We assume pricing of $5,000 per month in the US and ROW ex-China with a treatment course duration of 12 months, with China priced at a 50% discount. Our model assumes a tiered 15-20% royalty on China sales and 11% ROW sales payable to HCM from Lilly and up to $60m more in milestone payments. We have not included milestone payments on further sales after initial launch nor sales ex China, which would significantly enhance our valuation, given that if Lilly exercises its option to global development of fruquintinib, HCM could receive up to $300m in additional milestones.

Exhibit 5 details our peak sales forecasts and assumptions by indication. Note we forecast peak sales in China as seven to eight years from launch, and five years from launch for ROW.

Savolitinib (AZD6094), a highly selective first-in-class c-Met tyrosine kinase inhibitor, could be the first of its kind to market. Herein we discuss the clinical rationale for a c-Met inhibitor, the competitive market and finally our development and sales assumptions. Savolitinib is a novel, orally administered, small molecule tyrosine kinase inhibitor (TKI) that is being developed with partner AstraZeneca. The drug is a highly selective inhibitor of the c-Met signalling pathway and targets patients with resistant cancers whose tumour type tests positive for MET amplification or overexpression. Savolitinib is 1,000 times more selective for c-Met than the next kinase (PAK 3). MET activation is associated with poor prognosis in NSCLC and is also associated with EGFR TKI resistance. Pfizer’s c-Met targeting multi kinase inhibitor (MKI) Xalkori (currently only approved for ALK+ or ROS1+ NSCLC) has aimed to address this need, but there remains the clinical rationale for a highly specific c-Met inhibitor. A highly specific c-Met inhibitor should theoretically improve the efficacy and side effect profile of MKIs, enabling a wider range of combination treatment regimens for cancer.

Drug-related renal toxicity has been an issue with the first-generation selective c-Met TKIs that were under development and a factor in why none have reached the market to date. Savolitinib has been structurally designed by HCM’s scientists to address the renal toxicity issues that have, in part, prevented c-Met TKIs from gaining approval (the quinolone region of the compound has been replaced); clinical data so far shows it has not displayed any material renal toxicity (>370 patients to date).

Savolitinib is in 12 active clinical trials for renal cell carcinomas, NSCLC and gastric cancer. It is most advanced in its PRCC and NSCLC indications. Data from multiple Phase I/II studies support savolitinib’s early clinical benefit as a highly selective c-Met inhibitor in a number of cancer types. Given that savolitinib has demonstrated partial response in several solid tumours (Phase I overall response rate of 38% in PRCC indication compared to GSK’s foretinib 13.5% in Phase II), it is conceivable that savolitinib could be the first global, first-in-class c-Met inhibitor to reach the market. In Phase I studies the level of response to savolitinib by each patient correlated closely with the level of MET amplification. Savolitinib has additionally demonstrated synergistic effects with other cancer drugs in pre-clinical models; hence, its expansive clinical trial programme includes combination with other targeted therapies and immune therapies, eg the combination of savolitinib with AstraZeneca’s Tagrisso could shut down two resistant pathways accounting for 60-70% of all EGFRm+ TKI resistant NSCLC patients.

A global, pivotal Phase II/II trial is evaluating the combination of savolitinib and Tagrisso for the second-line treatment of patients with c-Met driven NSCLC (T790M±) as part of the TATTON study. Data is expected from the Phase IIb expansion part and, if positive, this could lead to the initiation of a global Phase III programme in H117. The initiation of the Phase II expansion triggered a $10m milestone payment from AstraZeneca to HCM in June 2016 (this expansion was initiated following encouraging early data from a number of patients enrolled in the TATTON study who received savolitinib in combination with Tagrisso). Importantly, overwhelmingly positive data could support a US NDA under breakthrough therapy designation for the NSCLC indication (second-line in combination with Tagrisso).

Additionally, positive data from the Phase II in PRCC in February/March 2017 (the study completed enrolment in October 2015) could enable a US NDA submission (under the breakthrough therapy designation) with potential US launch for the PRCC indication in late 2017 or early 2018.

The pivotal Phase III study in PRCC represents the first molecularly selected trial in renal cell carcinoma globally. Partner AstraZeneca has entered an agreement with US-based Foundation Medicine to develop a companion diagnostic test (tissue biopsy) to identify patients with c-Met driven cancers; the test is being developed in parallel with savolitinib’s clinical trial programme as part of a coordinated regulatory strategy and we anticipate it will form part of the regulatory filing with the FDA. We highlight that HCM comments that the PRCC Phase III companion diagnostic platform will be largely similar for other indications such as NSCLC and gastric cancer. Currently there is no standard method for testing for mutations; methods include gene sequencing, mass spectrometry and fluorescence in-situ hybridisation (FISH).

The NSCLC indication could be savolitinib’s largest opportunity given 1.7m new cases a year (8-10% MET amplification rate in NSCLC, source Frost & Sullivan). Exhibit 6 highlights where savolitinib could straddle the NSCLC treatment paradigm. The NSCLC clinical trial programme is assessing its utility in the first-, second- and third-line setting in certain identifiable patient populations. HCM estimates that the annual incidence of c-Met driven NSCLC in the US, European Union and Japan is around 40,000 to 50,000 patients across all treatment settings. Exhibit 6 shows the five opportunities for savolitinib within the current NSCLC treatment paradigm.

We believe the second-line NSCLC setting (#2 and #3 below) to be savolitinib’s largest commercial opportunity:

HCM see five opportunities for savolitinib in NSCLC patients as outlined below:

Renal (kidney) cell carcinoma (RCC) is the seventh most prevalent cancer in the Western world; RCC is diagnosed by histological subtype. Papillary renal cell carcinoma (PRCC) accounts for 14% of the 366,000 new renal cell carcinoma cases that occur worldwide, while clear cell renal carcinoma (CCRC) accounts for 74%. A number of targeted therapies (including Roche’s Avastin, Pfizer’s Sutent and Bayer’s Nexavar), immunotherapies and immune checkpoint inhibitors are approved to treat advanced renal cell carcinoma; however, there are no currently approved treatments for c-Met driven PRCC specifically. As MET amplification occurs in 40-70% of PRCC cases (100% of hereditary PRCC patients test positive for MET mutation), while c-Met over-expression occurs in 79% of CCRC (source Frost & Sullivan), it follows that a successful c-Met TKI inhibitor could transform renal cancer treatment. Savolitinib is currently in four Phase I/II trials for PRCC and CRCC (clear cell renal carcinoma).

Papillary renal cell carcinoma (PRCC) Phase II data is expected in February/March 2017, most likely at ASCO GI. This 109-patient, open-label Phase II study will report results of savolitinib (600mg qd) on its primary endpoint, ORR (objective response rate). While a single-arm, open-label study has its limitations; it is important that the data on survival from this study is in line with previous early stage data.

While HCM and AZN have agreed to proceed savolitinib into its Phase III global pivotal study for PRCC, overtly positive data from the open-label Phase II trial could lead to a US NDA application under breakthrough therapy designation for the PRCC indication, given the clear unmet need for c-Met driven PRCC. The final design of the pivotal trial is being agreed with international health authorities. Importantly this trial will be aligned with a companion diagnostic for c-Met driven PRCC; and this PRCC Phase III diagnostic will be similar for use in other indications.

In 2011, HCM granted AstraZeneca co-exclusive rights to develop, manufacture and commercialise savolitinib globally. HCM received an initial $20m non-refundable licence fee with up to a further $120m in clinical development and early sales milestones payable (as of June 2016 HCM had received $20m of those milestones) in addition to significant further milestone payments based on sales. This is in addition to a 30% royalty rate payable on China sales and originally tiered royalties of 9-13% of sales outside China. Under the terms of the 2011 deal AZN would pay 100% of the development costs ex-China and 75% of the costs for development in the China market (with HCM funding the remaining 25%). Under the 2016 amendment, HCM will contribute an additional $50m to accelerate the PRCC programme over a three-year period and in return will receive an additional 5% of royalty on sales on all indications ex-China, effectively taking the tiered royalty rate to 14-18%.

Importantly the collaboration with AstraZeneca has resulted in the addition of savolitinib to AstraZeneca’s Tagrisso and Iressa in two separate Phase Ib/II trials to address the opportunity for combination therapy as second-line and third-line treatment for NSCLC. Further combination studies are likely, eg savolitinib plus AZN’s PD-1 inhibitor durvalumab.

We forecast global peak sales for savolitinib of $3.4bn across the potential PRCC, CRCC, NSCLC and gastric cancer indications. Exhibit 7 details savolitinib’s peak sales potential by indication, incident rates and penetration assumptions. We assume pricing of $10,000 per month in the US and ROW ex-China with a treatment course duration of 12 months, with China priced at a 50% discount. We believe this is conservative given that AstraZeneca’s Tagrisso, a third-generation TKI, is priced at $12,750 per month, which is in line with the pricing being attached to most new lung cancer drugs, including ALK (anaplastic lymphoma kinase) inhibitors such as Pfizer's Xalkori and Novartis's Zykadia (source: Reuters). Furthermore, savolitinib could be moved into earlier lines of therapy as part of combination treatments, increasing the market opportunity, depending on the results of ongoing trials. Our model assumes a 30% royalty on China sales and 14-18% tiered royalty on ROW sales payable to HCM from AstraZeneca and up to $100m more in milestone payments. We have not included milestone payments on further sales after initial launch, which would significantly enhance our valuation. We note that under the terms of the agreement with AZN, the royalty rate is expected to step down to 10.5-15.5% upon reaching aggregate savolitinib sales of $5bn.

We have assumed higher overall penetration rates for PRCC given the 40-70% MET amplification and NSCLC 79% c-Met over expression rates. Our China penetration rates for NSCLC and gastric cancer err on conservatism. While we expect initial launch in 2017 for PRCC in China and the US, it follows that NSCLC and gastric cancer are larger opportunities given the patient populations. Note we forecast peak sales in China as seven to eight years from launch, and five years from launch in the rest of the world.

There remains a clear unmet need for a selective c-Met TKI inhibitor in the cancer armamentarium. The array of pathways that c-Met is involved in combined with the complex range of molecular alterations make it a difficult but extremely valuable target. It is increasingly apparent that the c-Met signalling network is complex, and combination therapy may be needed for optimal clinical efficacy. Exhibit 8 summarises the wider competitive landscape; this includes selective c-Met inhibitors, multi-kinase inhibitors and monoclonal antibodies. In our view, savolitinib’s closest competitor is capmatinib (INC280), a selective c-Met inhibitor that Novartis licensed from Incyte Corporation in 2009.

Per our analysis across the c-Met inhibitor space we view capmatinib as the closest compound to savolitinib in terms of target selectivity, clinical development programme and timeframe for NDA submission. Data to date shows promise; in preclinical work it demonstrated that it was 10,000-fold more selective for c-Met then 57 other human kinases. However, capmatinib unlike savolitinib retains a similar molecule structure (2-quinolinone molecule) to earlier drugs that failed due to renal toxicity; while no renal toxicity has been reported to date we do not expect Novartis to test capmatinib in patients with kidney cancer. As such in c-Met driven PRCC and clear cell renal carcinoma we would expect savolitinib to benefit from a lack of direct competition. Additionally through HCM’s collaboration with AZN, savolitinib benefits as the sole c-Met inhibitor in clinical development in combination therapy for NSCLC with the only EGFR/T790M inhibitor currently on the market (Tagrisso). As AZN look to position Tagrisso as a first-line therapy in NSCLC in developed markets, savolitinib could be well positioned as the only approved c-Met inhibitor that can be used in combination with Tagrisso. In emerging markets where pricing constraints are more common we expect an Iressa plussavolitinib combination protocol to dominate, partly due to the strength of Iressa plus savolitinib combination data in certain markets e.g. Chinese patients with NSCLC. HCM expect to be the first to launch a selective c-Met inhibitor in at least 50% of emerging markets. For the NSCLC indication, Novartis plans to file for capmatinib’s NDA in 2018. Capmatinib is being evaluated across multiple cancers; recent data presented at ASCO included data from a Phase I and Phase I/II study in NSCLC:

Capmatinib is currently in 16 active Phase I/II trials, eight of which are in NSCLC. Novartis plans to file an NDA in NSCLC in 2018, potentially one year after savolitinib. We believe there is scope for a number of selective c-Met inhibitors given the NSCLC market size and with oncology heavyweights AstraZeneca and Novartis in the space, we would anticipate a stronger awareness of the potential for the selective c-Met class of drugs among prescribers and payors.

In addition to c-Met specific inhibitors, a range of multi-target inhibitors that target c-Met are in development or marketed; these are summarised in Exhibit 8. One of the most clinically advanced is the marketed crizotinib (Xalkori). Crizotinib has been on the market since 2011 for treatment of NSCLC patients where it was approved for ALK (abnormal anaplastic lymphoma kinase gene)-positive patients and in March this year it was approved for ROS1 gene-positive NSCLC. It is currently in an array of Pfizer-sponsored and investigator-led trials for c-Met positive cancers.

A Phase I study in 13 c-Met amplified NSCLC patients treated with crizotinib were subdivided into three amplification groups (low [n=1], intermediate [n=6] and high [n=6]). Four partial responses were observed, one in the intermediate group and three in the high c-Met amplified group. Median duration of response was 35 weeks (95% CI). The most common adverse event (AE) was diarrhoea, which occurred in 50% of the patients: most AEs were grade 1 in severity. A Phase I trial in 17 (15 response evaluable) patients with MET exon 14 altered NSCLC treated with crizotinib observed that five patients (29%) had a confirmed partial response (an additional five patients had an unconfirmed partial response). The duration of treatment ranged from 0.5 to 9.1+ months. There were no deaths or progressive disease by the data cut-off date.

The correct identification of c-Met or MET status is needed to ensure efficacy of treatment. If savolitinib can demonstrate that it is effective in these patient subsets and is one of the first to market it could achieve significant success; however, the competitive environment still remains complex. The correct selection of patients in clinical trials could be fundamental in any c-Met inhibitor’s success. Theoretically savolitinib should provide an advantage over crizotinib as it provides full coverage of c-Met.

Epitinib is a selective EGFR tyrosine kinase inhibitor designed for optimal brain penetration, to target brain metastases (BM) associated with EGFR mutation positive solid tumours. Herein we discuss the clinical rationale for a next-generation EGFR inhibitor, the competitive market and finally our development and sales assumptions. Epitinib’s development rationale has been spurred on from preclinical studies where the drug demonstrated favourable drug exposures in the brain. TKIs as a class have revolutionised the treatment of EGFR positive cancers; however, few trials have been conducted to assess clinical utility in cancer patients with brain metastasis. In comparison to approved EGFR inhibitors (eg Iressa and Tarceva) preclinical data demonstrates that epitinib can cross the blood-brain barrier to reach more effective concentrations with potential higher efficacy in treating brain metastasis. Currently no targeted therapy has been approved that effectively can treat both NSCLC and brain metastases.

According to OncoLink, the exact incidence of brain metastases is not known. Studies suggest brain metastases occur in 10-30% of patients with cancer, including NSCLC. An estimated one-third of patients with EGFRm NSCLC experience disease progression due to development of brain metastasis. Furthermore, around 50% of NSCLC patients who eventually develop brain metastasis do not have effective treatments. Long-term prognosis is poor, with an estimated overall five-year survival rate less than 5% for patients with metastatic NSCLC.

The blood-brain barrier (BBB) is an anatomical structure that serves to control and limit what molecules are able to enter the brain through selective passage of substances. The BBB consists of brain capillaries – endothelial cells that are sealed together with tight junctions and are surrounded by astrocytes – that allow lipid soluble substances, as well as glucose, oxygen, and water through. Proteins, ions, and large molecules do not enter at all or enter very slowly. While the BBB functions to protect the brain from the entry of harmful pathogens and substances, it throws up challenges for drug development. The majority of orally delivered pharmaceutical agents used to treat CNS disorders are lipophilic small molecules that cross the BBB by transmembrane diffusion. Most oral chemotherapeutic agents are unable to cross the blood brain barrier, thus at present treatment for brain metastasis is mainly limited to surgical resection and radiotherapy (whole-brain radiotherapy and stereotactic radiotherapy).

Iressa and Tarceva, two approved EGFR inhibitors for NSCLC, have limited blood-brain barrier penetration and limited cerebral spinal fluid (CSF) concentrations of 2.5% for Iressa and 5% for Tarceva. Retrospective data analysis shows these TKIs have utility in treating EGFR mutated (NSCLC) in combination with whole brain radiation therapy (WBRT), however clinical trials have not directly addressed utility of these agents in treating NSCLC with BM. The impact of first-generation TKIs such as Tarceva and Crizotinib in CNS metastasis is modest, as limited CSF penetration has been seen with corresponding low rates of intracranial efficacy reported. HCM recognised the clinical need for a novel EGFR TKI with improved efficacy against brain metastasis, and has subsequently developed epitinib, which in preclinical studies demonstrated excellent brain exposure compared to gold standard EGFR TKI Tarceva (Exhibit 9).

Currently, a number of third-generation TKIs that have demonstrated better CNS penetration are at various stages of clinical development, including Tagrisso (AZN), ASP8273 (Astellas), HM61713 (Hanmi Pharmaceutical) and epitinib. Tagrisso is in Phase II for NSCLC with exon 19 deletion, T790 mutation. AZN is also developing AZD 3759; it has been specifically designed to penetrate the BBB but lacks activity against T790M.

HCM presented data from the ongoing Phase Ib first-line epitinib for EGFR+ve NSCLC with BM clinical trial at the recent IASLC World Conference on Lung Cancer. The primary objective of the trial was to evaluate the safety and tolerability of continuous dosing of epitinib in addition to determining the MTD or recommended Phase II dose (RP2D). The trial results demonstrated that epitinib was well tolerated with a safety profile consistent with approved EGFR-TKIs. MTD was not reached and 160mg once a day was determined to be the RP2D. In terms of efficacy, clinical utility was evident in EGFR +ve patients, and patients who were EGFR-treatment naïve demonstrated a strong response with an overall response rate of 61.9% and DCR of 90.5% (Exhibit 10). Key points were:

This Phase Ib trial has been expanded into a Phase Ib/II proof of concept trial, which is ongoing in China; mature PFS data is expected in 2017. HCM anticipates that a pivotal Phase III in first-line patients with EGFRm NSCLC with BM in China could be initiated in H117. We note that the China FDA has cleared the Phase II/III clinical trial protocol. This could pave the way for international studies and a potential application for US approval under the FDA’s breakthrough therapy designation.

We forecast epitinib peak sales of $905m for the NSCLC with brain metastasis indication, based on a 5% penetration of the NSCLC patient population with brain metastasis (20% of the 1,690,000 global new cases of NSCLC patients worldwide, 623,000 new cases in China), $5,000 per month ex-China (50% discount on price in China) and a 12-month duration of treatment. HCM currently retains all rights to the products worldwide. In our epitinib valuation, we assume a partnering deal worth $80m with a $20m upfront payment in 2019, and a further $60m in development milestones. We assume a 30% royalty rate on China sales, and 11% ROW.

Lung cancer is the leading cause of cancer-related deaths globally (19.4% of total cancer-related deaths); 1.8m new cases were reported worldwide in 2012. The US and China are two of the largest single markets, with 214,000 and 653,000 new cases, respectively, in 2012 (source: Globocan). NSCLC is the most common type of lung cancer, accounting for 85-90% of all cases. Five-year survival rates for NSCLC remain poor despite significant investment resources over the last 15 years; only 15% of patients diagnosed with lung cancer survive more than five years. The more recent availability of new treatment options as described below has improved outcomes and survival for patients, but there is still much need for more effective treatments across first-, second- and third-line settings.

Over the last 20 years, significant advancements have been made in the development of personalised medicine for cancer. The understanding that many cancers develop at a genetic level has translated in the clinical setting to “genotyping” an individual’s cancer, to identify the genetic abnormalities that are a causal factor in the development and progression of cancer. The discovery of oncogenes and tumour suppressor genes alongside the complete sequencing of the human genome were important advances in the understanding of the molecular mechanism of cancer (mutations that arise activate a specific gene, such that healthy cells become cancerous cells). Targeted therapies (for specific molecularly defined subsets of cancer patients) that focus on relevant gene expressions have demonstrated a marked improvement in overall survival in patients with amenable cancers compared to initial treatment with traditional chemotherapy. Such drugs work to close the abnormally activated molecular pathways in cancer cells leading to cancer cell death while minimally affecting healthy cells.

A wide variety of genetic mutations and potential targets continue to be identified in oncology. Some targets have become synonymous with a particular cancer (eg B-Raf and melanoma). However, genetic mutations can exist across a variety of cancer types; some lymphomas and lung cancers have an aberrant ALK (anaplastic lymphoma kinase) gene. EGFR and or VEGFR mutations are found in a wide variety of cancers including CRC and lung cancer. The MET oncogene has been observed to be highly expressed in a number of solid tumour types, across a range of carcinomas, including, but not limited to, lung (40%), kidney (70%), colon (78%) and gastric (>95%) cancers.

Historically NSCLC is histologically divided into adenocarcinoma (50%), squamous cell carcinoma (SCC) (30%) and large call carcinoma; it is now known that 60% of adenocarcinomas and 50-80% of SCC have identifiable ‘driver mutations’. NSCLC is not a single entity but a number of pathologies with different molecular abnormalities; subsets of NSCLC can be further defined at the molecular level by identification of ‘driver mutation’ that occurs across multiple oncogenes. Such mutations occur in the genome of the cancer cells within genes that encode for cell growth and survival and activation of mutant signalling proteins induce and sustain tumour growth. Lung adenocarcinomas are associated with KRAS (30%), EGFR (15%), ALK (5%) and MET (4%) mutations; see Exhibit 11. Common treatable oncogene mutations in NSCLC include the EGFR mutation and ALK translocation (see below, Current treatment modalities).

The epidermal growth factor receptor (EGFR) is involved in the pathogenesis and progression of many cancer types; approximately 15% of patients with NSCLC in the US and c45-60% in Asian patient populations have tumour associated EGFR mutations. Mutations in EGFR are seen more frequently in non-smokers. Genetic alterations such as amplification of the EGFR gene and mutations of the EGFR tyrosine kinase domain in cancer patients are correlated with a high probability of responsiveness to anti-EGFR drug therapy. The two most common activating mutations involve exon 19 and exon 21.

C-Met (mesenchymal-epithelial transition factor) overexpression, with or without gene amplification, has been reported in a variety of solid tumours, including lung cancers, and is correlated with poor prognosis. In the context of cancer, aberrant signalling through the c-Met receptor promotes pleiotropic effects that include growth, survival, invasion, migration, angiogenesis and metastasis. Additionally, up-regulation of MET can occur in patients after EGFR inhibitory therapy as a mechanism of resistance to this treatment; furthermore, as MET amplification occurs independently of EGFR (T790M) mutations it is a clinically relevant therapeutic target for some patients who acquire resistance to EGFR TKIs (Iressa and Tarceva).

One of the most prevalent MET-related mutations is exon 14 skipping, a result of defective messenger RNA (mRNA) splicing due to mutations or deletions at the acceptor or donor sites. This ‘skipping’ of exon 14 results in the loss of a key regulatory domain and promotes c-Met overexpression and tumour activation. Its prevalence in NSCLC has been shown in numerous academic studies and is recognised as a potential key indicator for efficacy for c-Met inhibitors. A 2014 study by the Cancer Genome Atlas Research Network of 230 resected lung adenocarcinomas found exon 14 skipping in MET mRNA in 4% of all cases.

While detection and inhibition of individual patient mutations is becoming increasingly relevant in cancer treatment, inhibition of more global pathways such as angiogenesis (eg through VEGF/R inhibition) remain important therapeutic strategies. Therapeutic intervention of tumour angiogenesis pathways have been identified in many cancer types including NSCLC. Angiogenesis is critical for the process of tumour growth, proliferation and metastasis. Vascular endothelial growth factor (VEGF) plays an important role in tumour angiogenesis, and inhibition of VEGF/R is a clinically proven target; monoclonal antibody Avastin (inhibits VEGF-A) remains the leading VEGF/R inhibitor on the market. To date small molecule inhibitors of angiogenesis have failed to produce meaningful improvements in overall survival.

Treatment of NSCLC varies and depends on the type and stage of the tumour as well as size and position in the lung; ~10% of lung cancers are surgically operated on, while the majority are treated with combinations of chemotherapy, radiotherapy and targeted drug therapies. Historically drug treatment decisions have been based on NSCLC tumour histology and platinum based chemotherapy has been the cornerstone of treatment; the main limitations of it are low survival rates alongside the troublesome side effects of these cytotoxic agents. More recently the treatment paradigm for the management of NSCLC has shifted with the availability of targeted therapies. Current guidelines from CAP (College of American Pathologists) and IASLC (International Association for the Study of Lung Cancer) recommend that patients with advanced NSCLC (adenocarcinoma) have the primary tumour or metastasis analysed for EGFR and ALK wherever feasible. Targeted therapies have improved progression-free survival in patients to 10-12 months versus six months on platinum doublet treatment in the clinical trial setting. In the clinical setting, treatment of NSCLC depends on stage diagnosed and molecular status of the tumour.

Historically chemotherapy has been the mainstay of treatment (before or/and after surgery); examples include cisplatin, carboplatin, paclitaxel, docetaxel and irinotecan. NSCLC patients often receive a combination of two drugs.

Targeted drug therapies are used for tumours with identifiable mutations or advanced tumours that progress despite chemotherapy (Exhibit 11):

VEGF/R targeting drugs eg Avastin (Roche) (VEGF-A) and Cyramza (Lilly) (VEGFR-2) are monoclonal antibodies that target VEGF/R in combination with chemotherapy. Avastin is approved in combination with carboplatin and paclitaxel (PC) for first-line NSCLC. In a pivotal 878-patient Phase III trial with PC in combination with Avastin, patients demonstrated a median overall survival (OS) of 12.3 months vs 10.3 months for PC alone.

More recent advances are in the field of immunotherapy, specifically immune checkpoint inhibitors (ICIs). ICIs block immune related receptors (eg programmed death PD-1) found on either T-cells or cancerous cells. Activation of such receptors enables cancer cells to hide from an individual’s immune system; by inhibiting these receptors, a patient’s immune system can find and destroy cancer cells. This field is currently dominated by PD-1/PD-L1 immune checkpoint inhibitors:

The NSCLC treatment paradigm will continue to evolve, as new resistance mechanisms and further driver mutations are identified. The treatment paradigm will move increasingly to a personalised level with genotype testing in all patients as soon as feasible. Furthermore, ongoing combinatorial drug studies are proving that combination therapy is more effective on progression-free survival rates and on challenging the resistance mechanisms; combinations such as but not limited to TKIs plus chemotherapy and TKIs plus immunotherapy (PD-L1) could become commonplace for identifiable patient subgroups. It is hypothesised that the combination of a PD-1 inhibitor with EGFR TKI could delay resistance and improve survival rates; such combinations have yet to be approved.

Targeted therapies already generate significant sales; in lung cancer, treatment was initially dominated by VEGF and EGFR inhibitors, eg Avastin and Tarceva, of which sales in Q316 were $1.7b and $252m, respectively, across all indications. As relevant new oncology targets have been identified, new and often highly effective drugs have been trialled and developed. Tagrisso, a novel drug for the treatment of EGFR T790M resistant NSCLC patients (launched in Q315) has demonstrated strong sales growth (45% increase in Q3 over Q2 [$133m from $92m]). Innovative new classes of drugs, eg immune checkpoint inhibitors, have posted significant sales to date; leading PD-L1 inhibitors Opdivo and Keytruda generated Q316 sales across all indications of $920m and $356m, respectively. EvaluatePharma’s consensus estimates predict the combined sales of the top 10 NSCLC drugs could reach $20bn by 2022.

HCM is subject to the usual biotech and drug development risks, including clinical development delays or failures, regulatory risks, competitor successes, partnering setbacks, and financing and commercial risks. The key sensitivities for HCM relate to crystallising value from the mid- to late-stage pipeline, in particular savolitinib, fruquintinib and sulfatinib: these three products contribute ~65% to our valuation. For the earlier-stage pipeline, both clinical development and partnering risks remain. We expect HCM to develop oncology assets epitinib and theliatinib and immunology assets HMPL-523 and HMPL-689 through to Phase II with initial proof-of-concept data before pairing. However, we have limited visibility beyond that on the terms and timing of any potential deal(s). CK Hutchison’s involvement in HCM materially reduces the myriad of risks associated with any direct investment in China; however, it does mean investors are minority shareholders. Additionally, the limited available free float reduces the shares’ liquidity. CK Hutchison is a HK-based multinational with a long and successful track record of prudent and prolonged investment and control of companies in China.

We have increased our SOTP valuation to $2.4bn (£32.2/share) from $2.3bn (£25.7/share), in the main due to the impact of currency (spot $1.25/£), amendments to deal terms on savolitinib (royalty rate 14-18% compared to previous 9-13% assumption) plus upgrades to our CP forecasts after a strong first half of the year. IP is valued at $1,789m; and placing CP’s 2016e share of net profit on a 22.5x rating gives $657m (867p/share). Adding in end June 2016 net cash and netting out unallocated costs results in a value of $2.4bn (£32.2/share).

Exhibit 12 details the breakdown of contribution from products by indication to our risk-adjusted NPV. However, we have included the non-risk adjusted NPV (in grey) to illustrate the potential value of the pipeline should all projects in our forecasts succeed. Projects in pre-clinical development or very early Phase I are not yet included in our valuation. For more detail on our valuation of the CP business, including a list comparable peers, see our note Stellar evolution, published in May 2016.

Interim results published in August 2016 highlighted robust growth within the commercial platform division (CP continues to be in the near term HCM’s primary profit and cash generative division), coupled with clinical progress within the IP portfolio. HCM has a healthy cash position with available cash resources of $197.5m (at 30 June) at the group level; this includes cash and cash equivalents ($75.9m), short-term investments ($46.6m) and unutilised bank borrowing facilities of $75m. We calculate net cash at 30 June of $80.6m and this will benefit in the second half of 2016 from a one-time gain of $38m related to property compensation (see below) and subsidies from the Shanghai government. Furthermore for the first time management has provided guidance across the group for 2016. Exhibit 13 highlights our forecasts vs guidance.

At the H116 results consolidated revenue at the group level grew by 27% from $82.5m to $104.5m, with the IP division contributing $22.3m (largely attributable from payments by partners) and consolidated CP sales contributing $82.3m (H115: $55.6m). We forecast IP revenues of $40m in 2016 and $54.5m in 2017 largely driven by developmental milestone payments from partners AstraZeneca and Lilly for progress of savolitinib and fruquintinib, respectively. We expect CP to continue posting double-digit top-line growth and forecast $156.7m in 2016 and $171.5m in 2017. These CP revenue numbers do not take into account revenues reported by the Shanghai Hutchison and Hutchison Baiyunshan joint ventures as these are accounted for using the equity method; thus only the net attributable profit of the JV contribution is reported as equity in earnings of equity investees, net of tax below the PBT line. In H116 this was reported as $21.3m and we forecast $65.8m in 2016 and $42.8m in 2017. A major component in the uplift is the share of profits from the land compensation, which we include in our model as $38.2m in 2016 and $8m in 2017. Importantly net attributable profit of the JV contribution also includes a negative contribution from the Nestlé JV (NSP), which is not recording revenues at present but is investing around $3m a year for the development of HMPL-004 (reformulation to enter Phase I trials in 2017).

Profit before tax at the group level reported a loss of $16.8m in H116 (versus loss of $0.1m in H115), largely due to the increase in R&D ($31.2m in H116 versus $21.2m in H115), S&M ($8.8m in H116 versus $3.8m in H115) and administrative expenses ($10.0m in H116 versus $7.5m in H115). Specifically, the increase in operating costs reflects a widening of clinical programmes, plus expansion of the small molecule manufacturing operations. We expect R&D expenses to increase to $80.0m and $110.2m in 2016 and 2017, reflecting the substantial need for investment in the burgeoning clinical trial programmes across the IP division, including the increased investment in savolitinib post the AZN deal amendment. We expect S&M and admin costs to increase marginally and, given higher cost assumptions, we forecast operating losses of $53.8m in 2016 and $56.8m in 2017 due to higher R&D costs in 2016 and 2017. In H116 net income from continuing operations, which includes the profit contribution from JVs (see above), reported at $2.8m and we forecast $6.6m in 2016 and loss of $22.1m in 2017. After stripping out minorities we forecast a net income attributable to the company of $3.8m in 2016 and a net loss of $26.6m in 2017.

HCM retains a healthy gross cash position. As of 30 June 2016, there was $122.5m in cash and cash equivalents with cash of $75.9m and short-term investments of $46.6m, offset by $41.9m in bank loans at the group level, translating to a net cash position of $80.6m. Reported cash benefits from the net proceeds of $95.9m raised from the Nasdaq listing in March 2016. At the JV level, not consolidated by the group, Hutchison Baiyunshan and Shanghai Hutchison had cash and equivalents of $78.3m (offset by a $26.5m loan) at the end of December 2015. We forecast a net cash position of $41m at end 2016 versus a net debt position of $18m at year end 2015.

Property windfalls aid additional cash injections; the appreciation in land values has benefited Shanghai Hutchison Pharmaceuticals Limited (SHPL) and Hutchison Whampoa Guangzhou Baiyunshan Chinese Medicine Company Limited (HBYS) old factory sites in Shanghai and Guangzhou, which are now in prime residential locations. The growth in production volumes and tighter regulations in China mean new, modern factories are needed. Around 98% of the ~$140m capex requirement has been invested (funded internally at the JV level) in two new factories in Shanghai and Anhui province, with the Shanghai factory having opened at the start of September 2016 and the Anhui factory is expected to come online early 2017. Fortunately, the compensation formulas (c 40% of land auction values) mean the total cash compensation is significant even after the JV capex requirements. The group is to receive $105m total cash compensation for the SHPL sites from the Shanghai government over 2015 to early 2017 with $31.1m received in H215. A further $73.9m from the Shanghai property compensation is expected at the JV level in 2016/early 2017; $59.5m was received in October 2016 with the balance expected early 2017. HCM’s share of this should feed in through special dividends over a number of years and we forecast an increase in dividends received from equity investees in 2016 to $32m from $6.4m in 2015. Further compensations are anticipated for the Guangzhou sites and we include a minor contribution only in our 2017 and 2018 forecasts, erring on the side of conservatism due to the uncertainty of the timing and outcome of any deal.