A few weeks ago, I received an email from a colleague at an affiliated hospital with the title, “Another one bites the dust.” Now I love Queen as much as the next guy, but I don’t like seeing this line in reference to new oncology drugs. In this case, the drug was picoplatin. Picoplatin is a novel platinum drug, somewhat like cisplatin or carboplatin. Like most chemotherapy, it works by poisoning DNA, causing rapidly dividing cells to die. As you can see, it shares the common platinum core with our old “friend” cisplatin, but the shape has been modified. The idea was that the drug might work in tumors that were resistant to cisplatin.

Picoplatin

Cisplatin

Picoplatin looked good in early studies, even earning attention on GRACE. Last year, Dr. West outlined the vital stats for the phase II of 77 patients for this new compound in 2nd line SCLC:

• Good side effect profile
• 10% response rate
• Median progression-free survival of 10 weeks
• Median survival of 27 weeks

Dr. West showed my favorite statistical graph, the waterfall plot, showing that about half of patients showed some degree of shrinkage:

(click on image to enlarge)

So the phase II data looked good. The purpose of a phase II trial is to provide a go/no go signal for whether a new compound is worthy of being tested in phase III—only the phase III trial can truly show the efficacy of a new drug and get it approved. So the company went to the FDA and got approval for the design of SPEAR (Study of Picoplatin Efficacy After Relapse) phase III study. This study enrolled 401 patients who were either refractory to a 1^st line platinum regimen or relapsed within 6 months of receiving it. Patients were randomized to either picoplatin or best supportive care. The study had a 90% power to detect a 33% reduction in the risk of death compared to best supportive care (goal of hazard ratio .67 with p<.05).

The trial was completed in March of 2009 and results were announced by press-release on November 16. The trial failed to meet its endpoint. With 320 evaluable events (euphemism for deaths) the hazard ratio was .82 with a p value of 0.089.

So is picoplatin active? A hazard ratio of .82 is meaningful to patient survival and the p value of .089 indicates that there was only an 8.9% chance that these results were found by chance alone. However, since the study failed to pass the magical p of .05, it must be called negative. However, the study certainly does suggest that picoplatin has activity in refractory SCLC.

Then why did picoplatin fail to pass the statistical bar? At the conference call, CEO Jerry McMahon, Ph.D stated, “The data indicates that more patients on the best supportive care arm received chemotherapy following progression than those on the picoplatin arm, and we believe that this may have been a significant factor contributing to the trial outcome, as picoplatin appeared to demonstrate a trend toward a survival advantage.”

Should we buy this argument? Should the FDA buy this argument? Was the choice of placebo control ethically appropriate? If it was, did it represent optimal clinical trial design? The patients treated in the phase III study had cancers that either progressed while on 1^st line platinum-containing chemotherapy or within 6 months of completing it. NCCN guidelines for this situation recommend: “Subsequent chemotherapy or Clinical trial or Palliative symptom management, including localized RT to symptomatic sites.” So NCCN waffles—it neither demands chemotherapy for this population nor endorses it. What agents are recommended when chemotherapy is chosen? NCCN says:

• Clinical trial preferred
• Relapse <2-3 mo, PS0-2 : ifosfamide, paclitaxel, docetaxel, gemcitabine, irinotecan, topotecan
• Relapse > 2-3 mo up to 6 mo: topotecan (category 1), irinotecan, cyclophosphamide/doxorubicin/vincristine (CAV), gemcitabine, paclitaxel, docetaxel, oral etoposide, vinorelbine

So, while the NCCN guidelines do not implicitly condone the placebo control, it also finds reasonable the use of a number of agents, with a nod towards topotecan for having category one evidence. We have already seen data comparing one of these agents, topotecan, to placebo that showed that patients lived longer on topotecan than on placebo:

So, perhaps while placebo control wasn’t necessarily unethical, a better choice of comparison might have been another drug known to be active, such as topotecan. That the CEO argues that this trial failed because more patients on the placebo group got subsequent chemo only strengthens this argument. My take on this data is not that picoplatin is inactive, but rather that we have not yet been shown evidence that it’s better than the stuff that we already have. If I worked for Poniard pharmaceuticals, the company that makes picoplatin, I’d consider a trial either in combination with another known active drug in the second line (perhaps irinotecan, topotecan, or amrubicin), or for third line therapy. But I don’t work for industry. Rather, I’m an oncologist with sick patients that I desperately want better therapies for. As such, I’m more excited about other new promising drugs for SCLC, particularly amrubicin (see http://cancergrace.org/lung/2008/06/23/amrubicin-at-asco-08/ and http://cancergrace.org/lung/2009/06/17/cancer-lifeline-relapsed-sclc-and-amrubicin/).

Hope springs eternal.