Pharma Strategy Blog

Interesting news arrived in my email box this morning:

"Roche (SIX: RO, ROG; OTCQX: RHHBY) today announced that the U.S. Food and Drug Administration (FDA) issued a Refuse to File letter for accelerated approval for the company’s trastuzumab-DM1 (T-DM1) Biologics License Application (BLA). As planned Roche will continue with its ongoing Phase III EMILIA registration study. Roche will continue to work with the FDA and expects a global regulatory submission of T-DM1 mid 2012.

The BLA submitted in July 2010 requested accelerated approval for T-DM1 based on the results of a single-arm Phase II study, which showed T-DM1 shrank tumors in one-third of women with advanced HER2 positive breast cancer, who had received on average seven prior medicines, including two HER2 targeted agents."

There was a lot of excitement around the initial phase II results (one third of the women experienced tumour shrinkage, for example).  Yet, to much surprise, the application was rejected:

"In their review of the BLA, FDA stated the T-DM1 trials did not meet the standard for accelerated approval because all available treatment choices approved for metastatic breast cancer, regardless of HER2 status, had not been exhausted in the study population."

At first reading of the press release, my thoughts centred around:

1. How could the FDA possibly reject such an application?
2. Was an SPA agreed up front?

The first question is easy to be indignant about, but what criteria were used for the trial and what does the data actually show?

Looking at the available clinical trials for T-DM1, this one looks most likely as 100 patients were required and the inclusion criteria stated:

• HER2-positive disease
• Metastatic breast cancer
• Disease progression on the last chemotherapy regimen received in the metastatic setting
• Prior treatment with an anthracycline, trastuzumab, a taxane, lapatinib, and capecitabine in the neoadjuvant, adjuvant, locally advanced, or metastatic setting and prior treatment with at least two lines of therapy (a line of therapy can be a combination of two agents or single-agent chemotherapy) in the metastatic setting
• At least two lines of anti-HER2 therapy must have been given in the metastatic setting as monotherapy or combined with chemotherapy or hormonal therapy. The HER2-targeted agent can include trastuzumab, lapatinib, or an investigational agent with HER2-inhibitory activity.

There are only two approved treatments for HER2-positive breast cancer, trastuzumab and lapatinib, plus others in clinical trials, eg pertuzumab, which was also allowed as one of the prior therapies.  All patients appear to have been refractory to at least two of these drugs, most likely trastuzumab and lapatinib.

The prior chemotherapies included anthracyclines, taxanes and capecitabine, which is quite heavy pre-treatment and includes all of the considered standards of care for several lines of therapy.  Indeed, the results of the trial presented at the San Antonio Breast Cancer Symposium last December showed that the average number of prior treatments in the metastatic setting was 7.

The others that could be used in the treatment of breast cancer include nab-paclitaxel (Abraxane), which I'm assuming would be covered in the taxane group and ixabepilone (Ixempra), an epothilone approved in by the FDA for metastatic breast cancer following progression on a taxane and anthracyclines, with or without capecitabine.

Ixabepilone is not a taxane or an anthracycline and therefore technically not covered in the inclusion criteria as a prior therapy, although some of the women would likely have received it, but not all.  It is, however, a taxane-like compound in that it targets the microtubules, as described in the PI:

"Ixabepilone is a semi-synthetic analog of epothilone B. Ixabepilone binds directly to B-tubulin subunits on microtubules, leading to suppression of microtubule dynamics. Ixabepilone suppresses the dynamic instability of aB-II and aB-III microtubules."

If that was the criteria for rejecting the trial as not truly refractory in a very heavily pre-treated setting, I would be surprised.  It's a bit of a technicality and nit-picking. Ixabepilone is not a commonly preferred treatment at all and BMS reported very low sales in 2009.  Most US physicians appear to prefer the well established taxanes such as paclitaxel, docetaxel and nab-paclitaxel rather than a synthetic taxane-like agent and probably consider it as a last resort. It is used as a standalone therapy in the refractory/salvage setting.

However, we're talking about a potential indication for a targeted agent that inhibits HER2 dimerization not a chemotherapy, so I would have thought that most of the important bases were covered by including the most common chemotherapies, trastuzumab and lapatinib in the inclusion criteria.

With regards to the latter question, Twitter chat this morning suggested that no, an SPA wasn't formally agreed upon, but Roche held discussions with the FDA that led them to file for accelerated approval. Thanks to Adam Feuerstein of The Street for being the first to answer that question.

My general opinion is that if you have an agreed Special Protocol Assessment (SPA) and meet the defined targets, it's much easier to move forward and gain approval. If you don't, things may turn into a crapshoot and it can go either way.  And that seems to be what has happened here.

The other factor that comes into play in this discussion is the ongoing discussion of the accelerated approval for Avastin in metastatic breast cancer and the overwhelming negative ODAC opinion last month.  That will not have helped, although one would like to think it shouldn't influence any decision making at the FDA.  We would probably be naive to think otherwise given the full protocol included PFS rather than OS as the endpoint.

To be fair, Roche appear to be addressing this issue, in their announcement this morning:

"Roche will amend the Phase III randomized EMILIA study in order to rigorously evaluate overall survival in addition to progression-free survival and will submit data from EMILIA to support a global regulatory submission in mid 2012."

Overall, I think it's a disappointing decision by the FDA given the heavily pre-treated population and lack of options for women who are refractory to both trastuzumab and lapatinib. Chemotherapy has not been shown to be particularly effective in HER2 disease, and by then the women generally have a poor prognosis. When you look again at the SABCS data, you find an objective response rate (ORR) of approximately 30% and a clinical benefit rate of around 40%, which is quite startling in a heavily pre-treated group.

Meanwhile, for now we'll have to wait another two years to see what the survival data looks like.  That's a rather long time for women who have failed Herceptin and Tykerb to wait for a drug that appears to have significant activity.

{UPDATE: I posted this analysis in a hurry, uncaffeinated, and in a rush to head out to a meeting.  Of course, one remembers later that SPA's apply to phase III not phase II trials, so a formal agreement wouldn't apply here. Thanks also to the two people who gently reminded me of this. Usually though, there are some discussions with FDA around the trial design for accelerated review. Essentially, in layman terms, it means patients are refractory to existing treatments for the disease concerned. Note that Roche, in their press release defined it as, "Consideration by the FDA for accelerated approval requires recognition of a defined patient population of unmet need (a life-threatening disease with limited treatment choices)." It looks like the FDA are applying it more strictly to ALL therapies, although the number of women who might have taken ixabepilone AND be HER2-positive will likely be miniscule.  I would still maintain that refractory to all available HER2 therapies and most chemotherapies apply for the majority in this case.}