Thermodox in HCC: The Science Points to Phase III HEAT Success

With all the confusion over the company deciding/not deciding to issue a PR for the occurrence of 190+ PFS events to trigger the interim analysis, I thought it would be important to re-focus investors' attention on what is, in fact, the most important question as it relates to Celsion and the HEAT trial:

Will Thermodox hit its primary endpoint (33% improvement in PFS-accelerated endpoint, eventually, overall-survival (OS) as well upon realization of 372 deaths) and prove successful in the HEAT study?

Whether this occurs at the interim (due to overwhelming efficacy, unlikely, but still on the table) or at the top-line read-out of 380 PFS events next year this time, I would like to present here my assessment of the scientific basis pointing towards a high probability of success in the HEAT trial. As you will see, much of the basis of my argument revolves around the well-documented patterns/factors associated with local recurrence in HCC, in addition to the proposed mechanism of action for Thermodox.

Rationale for the Trial

There exists a great unmet need in the treatment of liver lesions greater than 3 cm. Few would question the efficacy/safety of radiofrequency ablation for tumors smaller than 3cm. In such cases, local recurrence is not such a significant medical issue, although it still occurs. A consensus in the literature, however, without question, supports the notion that local recurrence is strongly associated with liver lesions greater than 3cm treated with RFA, the population of patients squarely targeted in Celsion's Phase III study. The extent to which the HEAT trial indeed is addressing an area of significant unmet need is no best exemplified by the National Cancer Institute's (NCI) Clinical Trials Planning Meeting (CTPM) designation of the HEAT trial as one of 8 priority trials in HCC in 2010. That, in and of itself, speaks volumes about the importance of this trial, and why the company has maintained that the results of the trial are highly anticipated by the medical community.

In an attempt to address some of these limitations, newer RFA devices/electrodes have been developed that can ablate larger areas of volume, but physician experience with these remains limited, and no large studies have corroborated their clinical utility. Similarly, RFA has been investigated in combination with transarterial chemoembolization (TACE) or bland embolization (TAE), with the rationale being to use TACE to shrink the tumor initially while cutting off blood supply to the lesion, and then using RFA after to ablate a smaller tumor. While this approach has had some success, again, no large-scale randomized trials have been published to definitively validate its clinical role.

Risk Factors for Local Tumor Progression

All of the following (not an exhaustive list) have been identified by a strong body of literature as being related to local tumor progression following RFA (Reference #1, specifically refer to Tables 1 and 2):

Reference #1- Annal Surg- Local Recurrence After Hepatic Radiofrequency Coagulation 2005.pdf

1. Lesion size >3cm
• Again, this is squarely the population targeted in the HEAT study, up to a maximum of 7cm.
  
  
2. Percutaneous vs Laparoscopic/Open-Surgical RFA
• Most patients in the HEAT trial have received percutaneous RFA, reflecting how RFA is done globally as part of today's SOC. Despite the higher chances of local progression with a percutaneous approach, clinicians likely favor it due to its convenience (can be done on outpatient basis) and its better treatment-related complication rates than a laparoscopic or open-surgical approach.
  
  
3. Lack of a 0.5/1.0 cm ablative safety margin
• Several papers, very recently, have identified that local tumor progression is associated with a lack of safety ablation margin, which essentially is removal of part of the healthy liver tissue surrounding the lesion via RFA. It is in these areas where micrometastases invisible to CT scans at 1 month follow-up can result in recurrence. Interestingly enough, the margins are precisely where Thermodox flexes its muscle. In fact, as I will show below, imaging results to date show the potential of Thermodox to add up to a 1-2 cm rim of heavily concentrated doxorubicin around the targeted lesion.
  
  
4. Proximity to major vessel
• This results in what is called the "heat sink effect", whereby the presence of a major vessel reduces the likelihood of achieving the clinically required RFA temperature for complete tissue necrosis. Depending on tumor location, heat sink can be a major barrier to overcome using RFA alone, which explains why it is indeed associated with recurrence. However, Thermodox activates at relatively low temperatures (39.5-42 degrees C) that are still attainable in these locations. So, in this instance, Thermodox would not only be working at the periphery, but would also have a greater effect inwards towards the center of the lesion. 

Having discussed the risk factors for local progression, it is important to highlight some nuances in the term "local progression", as the literature has often mixed up true local progression with intrahepatic distant progression. The former would be progression around the area of the ablated tumor, while the latter would be an entirely new lesion that surfaces in a remote location of the liver (see below for more on distant progression). Consistent with point #3 above regarding the use of an ablative safety margin, scores of literature confirm that local progression indeed often occurs right at the periphery of the ablated tumor. Hence, from this, we can postulate with a relatively high level of confidence that local control can be improved by expanding the zone of ablation.

Below, I would point the reader to two studies (Reference #2 and Reference #3) published in well-respected journals that speak to the site and patterns of recurrence, among other things discussed in the papers. Note that the overwhelming proportion of local recurrences are "contiguous" rather than "adjacent", or distant. Also, more importantly, notice that the most common first site of recurrence, is local as well. This is important given the PFS metric in the HEAT study, which captures any and all types of progressions, intra/extrahepatic, and all-cause death. As an aside, this paper once again shows the strong relationship between tumor size and progression, as well as safety margin and progression.

Reference #2- Annal Surg. Onc - Local Recurrence after Lap RFA 1032 tumors Jul 2008.pdf

 

Consistent with the study above, note in Table 2 of reference #3 that the most common type of local progression is "extrazonal, peripheral, nodular", irrespective of whether or not the tumor type was an HCC or metastases. This also has big implications for the colorectal liver mets study Celsion is planning to initiate. Also of importance, Table 1 shows that more than half of all progressions occur within the first 12 months, with a median of 12 months for HCC and just under 7 months for metastases (colorectal mets in particular are much more aggressive than HCC, more thoughts on this below), and do notice that this study looked at tumors much smaller than those being observed in the HEAT study. Keep this information handy as we discuss expectations for the HEAT trial control arm (RFA alone) below later in the article.

Reference #3- RJR- LTP After RFA of Liver Tumors- Analysis of Morph. Pattern and Site 2007.pdf
 

Data Collected to Date

At this point, it would seem logical to jump straight to the Thermodox Phase I data, but doing so should only be preceded by a review of studies conducted to date looking at non heat-sensitive liposomal doxorobucin (marketed by J&J as Doxil/Caelyx) in conjunction with RFA. This is clearly not what Doxil is indicated for, but again, referring back to the unmet need of successfully ablating large liver lesions, the medical community has recently examined how the combination of Doxil + RFA could enhance the area of ablation for RFA. This was actually a topic at the most recent World Conference of Interventional Oncology (WCIO) meeting (Reference #4). The reader is probably wondering how and why a non heat-sensitive liposome could possibly enhance the zone of ablation with RFA? The reasons for this synergistic effect boil down to the following:

1. Hyperthermia increases the porosity of tumors, enhancing liposome extravastion into tumor interstitium
   
   
2. Cellular stress through the production of oxidative nitrative lipid, DNA, and protein damage
   
   
3. Increased cell apoptosis

Refer to the following papers and presentations from my blog which support this synergistic potential. In reference #4, take special notice of the imaging results of RFA + non heat-sensitive liposomal doxorubicin. In reference #5, examine table 2 and look at the % increase in ablation volume with RFA + Doxil. As expected and seen in many prior Celsion company presentations, the ablation volume significantly decreases with RFA alone as time progresses.

Reference #6 is an interesting one, and I am compelled to include it here out of academic honesty. While focused on much smaller lesions, this small study did not show any impact on progression by adding Doxil to RFA, and more oddly, did not show any increase in ablation zone. It does stand out as somewhat of an anomaly in terms of the effect on ablation size with liposomal doxorobucin.

Reference #4- WCIO 2011 Presentation (Note the imaging results of RFA + non-heat sensitive liposomal doxorubucin)

Reference #5- AJR- RFA of Hepatic Tumors- Increased Tumor Dest, with Adj. Lip Dox 2002.pdf


Reference #6- Hepatol. Int. - RFA of Small HCC with intravenous peg lip dox 2011.pdf


So, the takeaway here is that normally, RFA alone results in a reduced size of ablation zone as time progresses (clearly noticeable via CT at 1 month scan), while the addition of non-heat sensitive liposomal doxorubicin generally leads to an enhanced area of ablation (with the exception of what was seen in Reference #6). In fact, so compelling is this information that experts in the academic community have recently suggested that a large randomized trial be conducted to evaluate RFA + non-heat sensitive for large liver tumors (Reference #7, Oddly enough, though the author was talking about RFA + Doxil rather than RFA + heat-sensitive liposome trials, he does not bother mentioning the late-stage HEAT study underway).

Reference #7- Radiology- Trials and RFA plus liposomal dox - 2010.pdf


One could hypothesize that RFA + Doxil would prove superior to RFA alone in a large-scale, randomized trial such as Celsion's HEAT study. Remember, we are talking about a liposome that was not even designed to be used in this manner, and one that does not have the unique characteristics of Thermodox that is designed to result in rapid intratumoral drug release in very high concentrations. Also, I don't believe there are any plans to initiate such a large study for Doxil + RFA, so the sheer level of evidence will clearly be in favor of Thermodox + RFA when the HEAT trial is complete.

Enter Thermodox: MOA and Results to Date
I won't go into much detail around what Thermodox actually is (I presume most are familiar with the basic technology), but as many of you know, it is a heat-sensitive, liposomal formulation of doxorobucin. The liposomes are technically much different than Doxil, even without the heat sensitivity aspects (Thermodox has a shorter half-life, among other things). Thermodox releases its payload of free doxorubicin only upon the application of locally applied heat, ideally, between 39.5-42 degrees Celsius. Thermodox is given as a simple intravenous (IV) solution 30 minutes before the application of heat, and as a function of the leaky vasculature of tumors, Thermodox nanoparticles (100nm in size) aggregate within the site of the tumor. This tumor liposome aggregation effect is especially pronounced in the liver. The following temperature gradient image taken from one of the references below shows one common type of RF electrode and the spectrum of temperature surrounding the centrally coagulated lesion area. In the 40-45 degree band below, that is where Thermodox would kick in. This is important, since we know from the papers above that local recurrence almost always occurs right at the periphery of the originally ablated tumor, and we also know that an ablative safety margin has been associated with less recurrence (still doesn't get rid of it entirely, since achieving an optimal margin under current imaging guidance is very challenging).

Cool-tip RFA Electrode Ablation Temperature Spectrum 

Before moving on to the actual data from the Phase I trial, to tie together exactly how Thermodox works, I would highly recommend the reader to view the following video from the inventor of the low temperature-sensitive liposome technology, Dr. David Needham from Duke University (lengthy, but a nice overview).

Thermodox Phase I Trial
The Phase I trial of 24 patients for Thermodox is not quite an apples to apples comparison to the HEAT trial. I will quickly point out some of the key differences:

1. Endpoint- Phase I trial endpoint was "treatment failure", operationally defined as disease progression (TTP) and/or initiation of a new therapy. The Phase III trial primary endpoint is progression-free survival (PFS), a broader/more sensitive metric that encompasses local/distant/extrahepatic progression, as well as all-cause death.
   
2. Patient Population- The Phase I was an "all comers" trial. A total of 9 patients had HCC, 15 had metastatic liver tumors from various sites. Generally, colorectal liver mets are the most aggressive from what I have seen (worse than HCC), while neuroendocrine mets are the least aggressive. (On a separate note, it is for this reason that I also perceive the fact that Celsion is initiating a randomized PII colorectal liver mets trial before having HEAT results in hand, as a nice sign of confidence.)

With all that said, the following article (Reference #8), recently published in Future Oncology by Celsion's Chief Medical Officer and one of the HEAT trial lead investigators (Dr. Ronnie Poon) presents a fantastic overview of the results. In my opinion, the most important point to take away from the trial is the strong dose response relationship seen up to the 50mg/m2 MTD. Overall, across all patients, TTF at the >=50mg/m2 dose was 374 days, while it was 80 days for those treated at <=50mg/m2 dose, again, suggestive of clinical activity and strong dose response. Of note, in the Phase 1 trial, there were 4 patients with tumors >5cm, 2 treated at <50mg/m2 (treatment failure at 25 and 93 days) and 2 >50mg/m2 (treatment failure at 261 and 374 days).  In terms of safety, the profile of Thermodox was consistent with that of regular doxorubicin.

Reference #8- Future Onc. - LTLD Adjuvant to Increase RFA Cure Rate- 2011.pdf
Imaging studies from the Phase I data (Reference #9) also tell a compelling story. After all, imaging studies (dynamic CT) will be telling the entire story in terms of identifying progression in the HEAT trial. Keep in mind the imaging studies I referenced above in regards to RFA + non-heat sensitive liposomal doxorubicin as you view this. Note that similar to that seen with RFA + non-heat sensitive liposomal doxorubicin, the area of ablation significantly expands following RFA + Thermodox, while again, the area of ablation for RFA alone clearly contracts within the first month. Take special note of this on slide 20, as well as slide 23, which shows a 1cm and 2cm drug "rim" at 28 days post-ablation

Reference #9- Imaging Features in RFA + Heat Nanoparticles.pdf


So, similar to the RFA + Doxil (non heat-sensitive liposomal doxorubicin), we see an enlarged ablation area with Thermodox, and this is precisely what we would expect given the synergistic effect many have hypothesized exists between RFA and liposomal dox in general. Of course, I did not highlight one very important difference between the two types of liposomal doxorubicin: With Thermodox, there is rapid, intratumoral release of doxorobucin at ~10x the dose as one would expect from regular systemic administration of regular dox (pay close attention to the video with Dr. Needham, above). Coupled with the well-documented high rates of progression seen in larger tumors, the relationship between treating the tumor margins and progression, and imaging results of liposomal dox + RFA (both Doxil and Thermodox), we can now piece together precisely why there is so much enthusiasm heading into the interim analysis of the 600 patient Phase III HEAT trial.

Summary and Key Factors to Be Aware of in the HEAT Trial

I think it is very safe to say that there is a lot of "activity" taking place within the treatment paradigm for HCC, particularly as it relates to intermediate and advanced disease. Ultimately, considering where the current standard of care is with respect to intermediate stage HCC and the documented unmet needs in treatment, I strongly believe a unique window of opportunity exists for Thermodox, with obvious clear competitive implications as well. Without question, there is a strong desire today within the interventional radiology community to overcome some of the very real limitations of RFA for larger sized liver lesions. Thermodox, given its minimal impact on the current standard of care (30 minute IV before RFA) and elegance with which it is expected to significantly extend the reach of RFA, could prove to be a major advancement in the treatment of HCC. Make no mistake, the HEAT study is squarely designed to address a significant unmet need.

Timing is quite important here for Celsion, and this relates back to the current standard of care undergoing a period of evolution, as I mentioned above. While this is not meant to be an exhaustive overview of competitive threats, here are some below that might be competitive with RFA + Thermodox in the future, or in other words, where the standard of care might evolve to in the future. Again, I posit that Celsion's Phase III HEAT study is being executed at just the right time, while there is still much ambiguity in the treatment paradigm, and with potential data coming out of the trial much earlier than many of the discussed treatment modalities below. Given the robustness of the HEAT study, a successful trial would in effect add significant clarity to a very ambiguous treatment paradigm for HCC.

• RFA + TACE- A fair amount of evidence shows significant promise for RFA + TACE, as well as more novel forms such as RFA + DEB-TACE (doxorobucin-eluting beads). However, no large studies such as the HEAT trial exist to firmly establish this treatment as a "gold standard" approach, particularly since there remains much debate over the timing of when TACE should be done relative to RFA, and the added morbidity presented by adding another inherently complicated procedure such as TACE. Also, I posed this question to management during a prior conference call, and Celsion CMO Dr. Borys made it quite clear that Thermodox as an RFA adjuvant could quite easily complement any combination approaches with TACE anyways.
  
  
• RFA + Sorafenib (Phase III STORM Trial)- I am unclear when data from this trial is expected to be released by Bayer/Onyxx, but I don't think it will be coming before top-line results from the HEAT study. Either way, given Sorafenib's established position in the advanced HCC market, this could prove to be a viable threat to Thermodox as an adjuvant to RFA. On the other hand, Sorafenib is very expensive, and thus, has really had limited success in the area of the world most afflicted with HCC, such as China. Also, similar to TACE, nothing precludes the use of Thermodox as part of triple therapy here.
  
  
• Newer RFA Devices (Multiple Probes)- RFA devices will continue to evolve on their own, as manufacturers make incremental improvements to expand an even greater area of ablation (particularly under percutaneous application). This is still very much evolving (see Reference #10 below). 5 years from now, such devices may be part of the "standard of care", but it certainly is not there yet. The same can be said about RFA + TACE for example. Without Thermodox, 5-8 years from now, the "standard of care" for intermediate HCC might have been RFA + TACE.  The fact is, we are not 5-8 years from now, and I once again highlight how opportune this period is for Celsion and Thermodox from a sheer timing perspective.  

Reference #10- Exp Rev Devices- RFA-Going Multipolar- 2011.pdf

Things to Watch Out For in the HEAT Study

1. Distant Intrahepatic Progression- I have discussed this before in another thread (http://celsion.blogspot.com/2011/08/perils-of-pfs-as-heat-study-primary.html) and it is worth pointing out once again, that as a function of the trial's PFS endpoint, distant intrahepatic progression will be captured and counted against both arms of the trial. One expects Thermodox to primarily flex its muscle locally, within the area of the ablated tumor(s), and while I have seen some reports illustrating the relationship between local progression and distant spread, it remains to be seen how Thermodox will impact such progression. It's ability to do so (i.e. prevent some intrahepatic distant spread), in my opinion, could mean the difference between overwhelming efficacy leading to an early NDA at the interim, versus having to wait for top-line data next year. Make no mistake, there will be distant intrahepatic spread in the trial, though one expects local spread to be the most common first site of recurrence.
   
   
2. Protocol Standardization Regarding an Ablative Safety Margin- I have called investor relations regarding whether or not the HEAT trial specifies a protocol for obtaining, for example, a 0.5cm or 1.0cm area of safety margin in the trial, or whether or not this decision is deferred to each of the trial sites. I am yet to hear an answer back from them. While there is a pretty clear relationship between such a safety margin and progression as I have discussed, this is still a fairly "hot" area of research in terms of the ideal margin. I would not be surprised if this decision was deferred to the participating trial sites, particularly since the choice of RFA device (with some broad guidance, of course) and type of procedure (open, laparoscopic, percutaneous) were left open-ended to the respective investigators. Not standardizing a 0.5cm or 1.0cm safety margin, in my opinion, is likely to favor Thermodox.
   
   
3. Control Arm Performance- There has been some debate over exactly what the control arm will look like in the HEAT study but I think it is fair to say there is no reason to think it will significantly deviate from what has been reported in the literature to date. Management has maintained an estimated 12 month median progression-free survival time for the control arm. I would caution the reader that few, if any papers, in the literature look at true progression-free survival as defined in the HEAT study, and thus, finding apples to apples comparisons has been challenging. Either way, I truly do concur with the approximate 12 month estimate management has publicly maintained, though I think the RFA-only arm in the HEAT study might be doing slightly better than that. Historically, percutaneous RFA is associated with higher rates of recurrence holding other factors constant, and we know that the overwhelming majority of patients in the HEAT study received RFA in this manner as well. The above-mentioned "safety margin" issue, for example, could impact how the control arm will look in the HEAT study. Lastly, some recent literature has pointed to newer multipolar RFA (discussed above) as being able to achieve a much larger ablation compared to older RFA probes. The data is still very preliminary, and I highly doubt investigators have enough experience using multipolar RFA such that it would be employed regularly in the HEAT study globally. Thus, I would expect recurrence, both local and distant, to be on the higher end of what is reported in recent literature.
   
   Having identified some of the factors that might impact the performance of the control arm in the HEAT study, below is an adapted image from Figure 1 of Reference #11 of what I believe the Kaplan-Meier progression-free survival curve will look like for the RFA-only arm. Keep in mind that PFS is a broad measure, capturing local progression, distant intrahepatic progression, extrahepatic progression, and death. With respect to Reference #11, disease-free survival (as defined by the authors) is an excellent surrogate for PFS, thus enhancing the utility of this reference. While the average lesion size is smaller than what we should expect in the HEAT study (in fact, this study did not find a difference in DFS by lesion size, interestingly), at 3.8cm, and laparoscopic RFA was employed, the study took place from 1997-2006. On the one hand, one could say that there have been some incremental technical improvements that might have been employed since the inception of the HEAT study in 2008. Yet, on the other hand, the study in Reference #11 used all laparoscopic RFA (in contrast to the HEAT study's predominant percutaneous focus), and the HEAT study is likely to have larger, more advanced HCC lesions in general. All in all, accounting for some of these differences, and after a brief email exchange with one of the authors of the study (Dr. Berber), my own assessment is that the original DFS curve from that study remains a solid reference from which one could estimate the outcomes in the control arm of the HEAT study. The flattening of the curve after roughly 2 years is squarely in line with what I have seen in the literature, particularly as it relates to local progression.
   
   
4. Sensitivity of Identifying Recurrence via CT Scans- Much of the variability in the literature regarding local progression rates following RFA, in my opinion, can be attributable to variations in how scans are interpreted. I cannot underscore this issue enough, as there exists today some debate about what actually comprises "progression." Obviously, this is standardized in the HEAT study, but the sensitivity of how the reviewers are judging progression will play a big role in PFS determination between both arms. I am just putting this somewhat random piece of information out there for you to consider. (Look no further than slides 4-9 from this very recent 2011 presentation made at WCIO from an earlier blog post to see what I am talking about: http://celsion.blogspot.com/2011/08/from-wcio-2011-highly-relevant.html)

Reference #11- HPB- Lap RFA of HCC long-term follow-up 2008.pdf

Estimated PFS Curve (Adapted from N. Ballem et al.)

Wrap-up

For as long as this article is, I likely omitted some other tangentially related topics. However, hopefully this article provides a solid overview of some of the salient issues to think about from a scientific perspective as it relates to the HEAT trial, while outlining some of the reasons why I remain highly confident in the success of Celsion's HEAT study. On the topic of what we should expect in the upcoming interim analysis, I will refrain from making any speculation, for example, by using the estimated control arm PFS curve from above and approximating the timing of the 190th event. As Dr. Borys has reiterated many times, the DMC will take into account the totality of the Thermodox risk-benefit profile in making their recommendation. With that said, a quick glance at the expected PFS curve for the control arm, coupled with the historic enrollment rates in the HEAT trial, gives me confidence that Thermodox is tracking well.

As I have also stressed many times, success in the HEAT study is, ultimately, validation for the entire LTSL platform. Given the significant near-term liver (HCC, CRLM) commercial opportunities and pipeline potential, much is at stake in terms of shareholder value. Of course, even more is at stake for potential future patients.

As always, I look forward to any feedback or comments you might have. Thank you.

Best,
Siavoche

*I would like to thank long-time cancer-cause advocate and LTSL-proponent Mitch Landgraf for editing this article.