One of the few important lung cancer trials that was presented for the first time at the Korea World Conference on Lung Cancer was the Eli Lilly-sponsored trial "JMDB". Unlike most coded trial names, this isn't a cute acronym, but is rather one of Lilly's cryptic four letter codes that designates tumor type, drugs being studied, maybe geography...I don't know. Anyway, the study results were presented by Giorgio Scagliotti from Turin, Italy, who is a world leader in lung cancer and, befitting the name Giorgio, is very effusive and charming, often greeting friends with a hug and kisses on the cheeks. This includes me...but I digress.

At the Presidential Session, Giorgio described the design and highlights of the JMDB trial (abstract here), which is the largest trial ever conducted in the setting of advanced NSCLC, enrolling 1725 patients with previously untreated stage IIIB (with pleural effusion) and IV NSCLC. They were randomized to receive either cisplatin and gemcitabine, one of the most commonly used regimens in Europe and also somewhat in the rest of the world, or cisplatin and alimta. Both were given every three weeks for up to 6 cycles:

(Click to enlarge)

This was a trial enrolled from around the world. About half of the patients had adenocarcinoma, and 14% were never-smokers. Three quarters had stage IV disease, and about one quarter had stage IIIB NSCLC, and they excluded patients with uncontrolled malignant pleural effusions.

Looking first at toxicities, the patients on the cisplatin/gemcitabine arm experienced significantly greater drops in blood counts, as well as the rate of fevers with low white blood cell counts, which can be serious and generally requires us to bring patients in the hospital for IV antibiotics (1% vs 4%). Along with this, the freuency of transfusions and the need for "colony stimulating agents" like G-CSF were more common with cisplatin and gemcitabine. In terms of side effects not related to blood cells, significant hair loss was not very common with either arm but was seen more with cisplatin and gemcitabine (21% vs. 12%). The cisplatin/alimta arm had a higher rate of grade 3 or 4 nausea, but it was uncommon with either treatment (7% vs. 3%). Overall, both treatments were feasible. This shouldn't be surprisng, since both are extensively studied in the past and FDA-approved regimens in one setting or another, just not compared directly before this trial.

Both arms received a median of five cycles of chemotherapy, and the response rates were pretty comparable, 30.3% for cisplatin/alimta compared with 27.8% with cisplatin/gemcitabine. The median duration of response was also very similar for both chemo regimens, at about 4.5 - 5 months for each, and median progression free survival was also about 5 months on both arms. Overall survival was also the same, at 10.3 months for everyone.

So the summary sentence was that cisplatin/alimta had essentially the same efficacy as cisplatin/gemcitabine but with a side effect profile that was overall more favorable for cisplatin/alimta. While that's of some interest, it's not electrifying, since we already have many, many platinum-based doublets that have been shown to have comparable activity.

The real interest was in the breakdown of results by patient subgroups, including the pre-planned analysis by NSCLC histology. Patients with adenocarcinoma had a signficantly better overall survival with cisplatin/alimta (11.8 vs. 10.4 months, P = 0.03), as well as better response rate (31.9% vs. 24.5%). The same superior results with cisplatin/alimta in patients with large cell lung cancer (a small group, not statistically significant, but a sizable difference in the two approaches). I don't have a figure yet to show the magnitude of the separation of the survival curves with cis/alimta vs. cis/gemcitabine, but I was impressed by it.

On the other side of the coin, patients with squamous cell lung cancer had a significantly better survival with cisplatin/gemcitabine (9.4 vs. 8.5 months, p = 0.05), and a higher response rate with cis/gem as well (36.7% vs. 26.9%). Again, I don't have a figure for you yet, but the curves separated quite convincingly, to my eye.

In other subgroup analyses, we saw that never-smokers did significantly better than current or former smokers with either chemo, and they did not do clearly better with one treatment over another. In fact, they had a 74% longer survival overall on the trial. So the better results of never-smokers don't appear limited to EGFR inhibitors, unless this all came from later treatment with EGFR inhibitors after first-line chemo.

I'll add more later on the reasoning of why these differences may have occurred and what the results of the JMDB trial might mean in clinical practice.