Cancer Network speaks with Joel Neal, MD, PhD, assistant professor in the department of medicine at Stanford University. Dr. Neal was a discussant in the non–small-cell lung cancer session at this year's annual meeting of the American Society of Clinical Oncology (ASCO) and is an expert on non–small-cell lung cancer.
—Interviewed by Rachel Warren
CancerNetwork: I know there was a lot of news about non–small-cell lung cancer at ASCO this year, and that this is your area of research—can you give me a wrap-up of what you think the most important findings were?
Dr. Neal: Well, I think what was most important depends on who is listening. For the community oncologists, the biggest finding is probably the overall survival analysis of the PARAMOUNT trial. This trial, presented by Dr. Paz-Ares, showed that maintenance pemetrexed, after induction cisplatin/pemetrexed had an overall survival benefit, in addition to the progression-free survival response that we saw earlier. The overall survival showed that patients who got maintenance pemetrexed following four cycles of cisplatin/pemetrexed had an overall survival of 13.9 months. If you count from the time of induction it was actually 16.9 months. The placebo group was 11 months, or 14 months from the time of induction. This was a significant result with a hazard ratio of 0.7, so there is some evidence to say that continuation maintenance of pemetrexed, in addition to switch maintenance pemetrexed, which is already FDA-approved, may be useful.
Some of the other data that we saw was from Dr. Lilenbaum, showing us phase III data with a couple of hundred patients showing that carboplatin and pemetrexed in performance status 2 patients vs pemetrexed alone was superior in terms of progression-free and overall survival. This tells us that patients—even with decreased performance status—should get carboplatin if they can tolerate it, and we’d seen similar data 2 years ago from the French group.
Some of other results were toward specific molecular subtypes of lung cancer. We saw some data in RET, we saw some data in ALK-resistant, we saw some promising data in KRAS mutants in non–small-cell lung cancer with a drug called selumetinib, which is a MEK inhibitor, and some data that showed that erlotinib doesn’t look as good as docetaxel in the TAILOR trial in EGFR wild-type patients.
We also saw that, in EGFR mutants (in the LUX Lung-3 trial) afatinib, which is a next-generation EGFR inhibitor which is irreversible and may have some activity against T790M, was superior frontline to cisplatinum/pemetrexed, with a progression-free survival of 11 months vs 7 months, but we haven’t seen the overall survival results so that goes along with previous trials that have told us that for EGFR mutants, first-line tyrosine kinase inhibitor therapy is important.
One of the most exciting results I think we saw was from Julie Brahmer, about the programmed death 1 (PD-1) drug. PD-1 is an inhibitory receptor expressed on T cells and a lot of tumors produce something called PD-1 ligand; it’s expressed on many solid tumors. In a phase I trial including about 100 non–small-cell lung cancer, they saw a response rate of about 18% in heavily pretreated patients, and some of those responses lasted for more than 24 weeks. The drug had fairly low side effects, just a little rash, diarrhea, fatigue, and itching because it’s an immunotherapy, so really this shows us that moderating the immune system may be a potential target in non–small-cell lung cancer.
And finally, another molecular target was ROS1. ROS1 is very much like ALK, which is treatable with crizotinib, and we saw from Dr. Shaw at Mass General that ROS1 patients, similarly to ALK patients, had a 57% response rate on crizotinib, which is FDA approved for ALK.
So that’s a wrap-up of the big things in the oral sessions for non–small-cell lung cancer, at least in the metastatic session.
CancerNetwork: Give us your thoughts on the study that showed FDG-PET scanning has substantially lower sensitivity and specificity than was previously thought?
Dr. Neal: This data was from Dr. Grogan and this was an early-stage clinical trial; it was clinical stage I non–small-cell lung cancer and it enrolled almost 1,000 patients. The trial was trying to do serum proteomic analysis of resected non–small-cell lung cancer to demonstrate whether or not you could use proteomics to diagnose non–small-cell lung cancer. For a secondary analysis they used FDG-PET scans and compared whether or not the PET scans plus the CTs were reliable for predicting whether someone had non–small-cell lung cancer. The question at its essence was, "If someone has a small lung nodule and that was diagnosed by CT only, can you use a PET to tell whether or not it might be cancer and therefore whether or not it might need to be resected?" What they found was that PETs were 82% sensitive, which is not terrible, but they were very nonspecific—with only a 31% specificity. There were a lot of false positives—a lot of these false positives were granulomas, suggesting prior or current fungal infection. The accuracy of FDG-PET improved with the size of the lesion, presumably because granulomas would generally be small.
This doesn’t mean that for stage IV, metastatic non–small-cell lung cancer, PETs are insensitive at all, they’re a very valuable tool for staging patients and they can show distant metastatic disease, but what it means is that for the majority of lesions that are PET avid, if you do a PET scan on people as a screening tool, it will not be a great discriminator of whether something is lung cancer or not so they probably have to be paired with another test.
CancerNetwork: Finally, can you talk a little about where we are as far as biomarkers for lung cancers—both subtypes and targets?
Dr. Neal: For non–small-cell lung cancer, we saw a lot of evidence at the meeting for exciting new targets, as well as more data about the targets we already knew. Within adenocarcinoma, which is the one that’s best molecularly characterized, the most common are KRAS mutations, which are somewhere between 20% and 30% in frequency. We saw that selumetinib plus docetaxel is a potentially effective therapy for that group, although that’s still in clinical trials and has yet to be validated. EGFR-mutant non–small-cell lung cancer is the next biggest share, and for this group we saw that in the LUX Lung-3 trial, afatinib vs cisplatinum/ pemetrexed, that afatinib looks like a great drug in first-line treatment of EGFR-mutant lung cancer.
We also saw that ALK mutants, which are around 5% of patients, fall into a few subcategories, in a series of patients from the University of Colorado. In some of those mechanisms, resistance may be overcome. And finally in ROS1, which is about 1% frequency, and RET translocations, which are about 1% frequency, these may each be targetable with drugs that are already available. So even though they’re very low frequency, we might be able to find effective therapies. We have yet to see those effective therapies work, but with a response rate of 57% in ROS1 to crizotinib, and with crizotinib already FDA-approved, and with the identification of RET translocations, which are a different subgroup, that may respond to sorafenib or sunitinib, I think we have hope that we may not have to start reinventing drugs and starting from ground zero, and that we might be able to use already approved things.
We also saw in squamous non–small-cell lung cancer, which up until a couple of years ago had no identified driver alterations, some data from Paul Paik at Memorial Sloan-Kettering. They’ve used some squamous multiplex testing and have identified FGFR1 amplification, by FISH analysis, as a recurrent alteration in about 25% of squamous-cell carcinomas, and this could be targeted potentially by FGFR inhibitors that are starting to enter clinical trials. They also reported that PI3 kinase alterations and P10 alterations are recurrent and common. Dr. Govindan from Washington University presented a larger series in the earlier oral session of 180 patients, and also identified FGFR1 and the P10 and PI3 kinase alterations as being recurrent and common in squamous non–small-cell lung cancers. So it looks like squamous might have some targetable alterations as well—which is something relatively new since the last ASCO.