In this week’s Journal of Clinical Oncology the final results of the phase III trial investigating chemotherapy with or without thalidomide in previously untreated NSCLC patients were published. Most people remember thalidomide for the horrifying birth defects (absence of arms and legs, etc) that resulted in the 1960s when pregnant women were given thalidomide for morning sickness. Since this debacle, thalidomide has been a highly controlled substance that has gotten a lot of understandably bad press, and this experience led to a number of necessary changes in the way the government investigates the safety of drugs before releasing them for widespread usage.

More recently thalidomide has become a standard treatment for multiple myeloma, a type of cancer of white blood cells. Much of the efficacy of thalidomide has been attributed to anti-angiogenic effects against vascular endothelial growth factor (VEGF) and fibroblast growth factor, both potentially important targets in lung cancer (we all know the story of Avastin (bevacizumab) in NSCLC, and anti-VEGF antibody). Thalidomide has shown some promise in preclinical models of NSCLC, and has also shown some indications of activity in small cell lung cancer (as Dr. West has covered in the past). It works well in multiple myeloma, but has some potentially worrisome side-effects even if you are not pregnant, principally increased risk of blood clots, constipation, and peripheral neuropathy.

In this most recent study, Dr. Lee and colleagues randomly assigned 722 patients in the United Kingdom with previously untreated NSCLC to either carboplatin and gemcitabine plus placebo or the same chemotherapy plus thalidomide, followed by up to 2 years of maintenance thalidomide (or placebo). The primary endpoint was overall survival. The thalidomide was started at 100mg daily, escalated to 150mg daily in the second month, and the maintenance dose was 200mg daily, a pretty standard dose.

The results were overall disappointing, essentially showing no hint of benefit from the addition of thalidomide in the overall group. The response rates were the same in the two groups (40% and 42% for thalidomide and control). The median overall survival of the thalidomide and control groups was 8.5 and 8.9 months.

PFS and Overall Survival Curves

In addition, the group decided to do a post-hoc (after the fact) subgroup analysis by tumor histology. We know that certain drugs, like pemetrexed, seem to have different effects on different types of NSCLC, so why not thalidomide? In fact, there did appear to be worse outcomes in the non-squamous patients treated with thalidomide, with only 8% of patients on thalidomide surviving at 2 years compared to 18% on placebo. In the squamous cell patients, however, it was just the opposite, with 20% of patients on thalidomide surviving 2 years compared to only 12% on placebo. It is hard to know how much weight to give this small but statistically significant difference.

OS by histology

One thing that was clear was that there was increased toxicity in the thalidomide arm compared to placebo. Specifically those on thalidomide were 74% more likely than those on placebo to develop a serious blood clot. There was also more skin rash and neuropathy in the thalidomide group.

Interestingly, it does not appear that a preceding phase II trial of this combination was ever done, which is usually done to show some level of efficacy before investing the enormous resources necessary for a phase III trial. There have been small studies testing thalidomide in combination with chemotherapy in NSCLC, but nothing that showed particular promise. There is also an American trial testing chemo plus thalidomide in locally advanced NSCLC still ongoing.

Based on this I think that thalidomide is probably not going to proceed any further in NSCLC. For one thing, the community is moving away from thalidomide and towards less toxic alternatives such as Revlimid (lenalidomide), which seems to be just as effective but easier to take. The possible differential effect based upon histology is interesting but unfortunately can only raise questions rather than answer them. We do need more therapies for patients with squamous cell carcinoma, so I could see another trial testing Revlimid in this population having some support.