Many terms can be used to describe Immunogen’s (IMGN) recent stock behavior, but it seems the word “schizophrenic” is the most suitable one. Immunogen gained almost 50%  in the three weeks prior to the ASCO annual meeting, just to give it all back in the 8 trading sessions following the conference, thus it is clear that the rollercoaster in the company’s share price had a lot to do with what was (or was not) presented at the conference. Immunogen is involved in multiple clinical programs, but for the past year the vast majority of the attention it has received was directed at T-DM1, which is being developed by Genentech (DNA) based on Immunogen’s technology. T-DM1 is garnering more attention than all the rest of Immoungen’s programs combined because it has all the necessary ingredients for the ultimate biotech story: Huge addressable market, a strong partner, impressive (yet preliminary) clinical activity and an opportunity to validate a disruptive technology. Accordingly, it is only reasonable to expect Immunogen to be traded in tandem with T-DM1’s development.

Wild swings in biotech stocks are commonly an outcome of clinical results publication, and indeed, the presented data at ASCO could be partially blamed for the violent market reaction. Nevertheless, in this particular case, Immunogen was affected from a lack of positive news rather than the release of negative news. Genentech had previously stated it would decide whether to advance T-DM1 into a registration trial during 2008, based on an ongoing phase II trial. This led many to believe that Genentech would use the ASCO platform to announce its intention to commence a phase III trial already this year. In the last day of the conference, when the market realized Genentech was not going to give the dramatic announcement nor was it going to release data from the ongoing phase II trial during the conference, the reaction was brutal.

T-DM1’s Clinical Data 

 

Genentech presented data from two phase I clinical trials with T-DM1, both showing encouraging signs of activity in advanced stage breast cancer patients. The two trials differed only in their treatment schedule, with the first trial evaluating T-DM1 given every three weeks and the second trial evaluating a weekly regimen of the same drug in the same patient population. In my previous article, I claimed that the weekly trial has a good chance of producing better results than the every three weeks trial because certain properties of T-DM1 make it more suitable for weekly administration. The main hypothesis was that weekly administration will result in higher cumulative doses as well as continuous presence of T-DM1 in circulation. The data from the weekly phase I trial did include some encouraging signs, but it was also a little bit ambiguous.

The two trials were dose escalation studies, where cohorts of 3 patients are accrued and evaluated for side effects and clinical activity. If no severe side effects are observed, additional cohorts of three patients are recruited and treated with higher doses, until there are signs of dose limiting toxicity, and the maximum tolerated dose (MTD) is established. At the MTD, the trial is expanded to additional patients in order to validate the lack of toxicities as well as the drug’s activity. In many clinical trials, activity, in the form of tumor shrinkage or disease control is measured at the MTD, as this is the dose that would be given in future trials. If the drug is active also in lower doses, patients from these cohorts may also be included in the overall assessment.  

In phase I trials, activity is typically measured by a decrease in tumor burden (tumors’ size), and patients can be stratified into 4 groups, based on their response to the drug. A complete response (CR) is defined as the disappearance of all lesions, which is very uncommon in phase I trials due to the advanced nature of the disease. A partial response (PR) is defined as a decrease of more than 30% in overall tumor burden.  Stable disease (SD) is defined as anything between a decrease below 30% and an increase below 20% in tumor size. Progressive disease (PD) is defined as an increase of more than 20% in tumor burden.  Two important rates in phase I trials are objective response rate ( ORR -percentage of patients who achieve CR or PR) and disease control rate ( DCR - percentage of patients who achieve CR+PR+SD).  Another trend investigators are always keen to see in a dose escalation trial is a dose dependent response, where higher doses lead to better clinical activity.

Before going over the data from the trials, it is important to understand that such analysis must be approached with a great deal of caution.  The trials include very few patients, which may lead to inaccurate results, as normal variability has a crucial effect on the trial’s outcome. For instance, in a 15 patient trial, each patient accounts for almost 7% of total data. In addition, both trials do not include a control arm, making the comparison of the data to historical figures even more problematic. Another issue is the fact that objective response is just an early surrogate for evaluating activity and is not necessarily correlated to other clinical end points such as overall survival and progression-free survival. Evidently, comparing results from two such trials is even more problematic, but for now, that is all we have.

The weekly trial started on the right foot, at a dose of 1.2 mg/kg per week, which is equivalent to the MTD in the every three weeks trials (3.6 mg/kg every 3 weeks). All three patients achieved a partial response, two of which were ongoing for nine months, as of Feb 29^th.  In the next cohort, out of three patients, only two had a PR and one patient did not respond to the drug. In addition, the two responses were maintained only for 4 months. The third cohort (2 mg/kg) had 2 PRs and one stable disease. The next dose, 2.4 mg/kg was defined as the MTD, due to the appearance of dose limiting toxicities at a higher dose, so accrual was extended at this dose level. At ASCO, Genentech reported data from 6 evaluable patients at this dose, 2 had a PR, 3 had SD and one had PD.

 

 

 

Based on the above data, not only did T-DM1 fail to show a dose dependent response, but there was an inverse dose-dependent response. There was, however, a clear activity at the MTD as well as at lower doses, and such fluctuations are very common in dose escalation trials. It is easy to envision a situation in which one of the patients from the lowest cohort was accrued to the MTD cohort, instead of the patient who did not respond (#12), to dramatically change the response rate at the MTD. This scenario, by itself, demonstrates the problematic nature of the data.

In addition, a comparison with the every three weeks trial raises some questions as well. In the every three weeks trial, five out of fifteen patients (33%) who were treated at the MTD achieved a PR, which is comparable to the ORR in the weekly trial. However, this comparison may be misleading. Because all doses in the weekly trial showed activity, the patients in those doses may also be included in final count. Adding these patients leads to an ORR of 60% (9 PRs out of 15). In the every three weeks trial, there was one additional patient with a PR at a lower dose than the MTD, which brings the ORR from all active doses to 37.5%, substantially lower than that of the weekly trial.  

While the data from the every three weeks trial is final, there are still additional patients in the weekly trial to be reported, not to mention the fact that some of the SDs in the weekly trial could turn into PRs.