One of the most pressing issues in lung cancer research is in identifying patients who could benefit from a particular drug, both to increase their chances of having a good outcome and to spare everyone else from an ineffective drug with unnecessary toxicity. There have been some exciting advances in this field, but before I elaborate I want to give some (simplified) background on how drugs are traditionally developed. Classically, potential cancer drugs are tested on cancer cell lines in a Petri dish, and if the drug appears to kill the cells, it is then tested in animals. If the animals don’t die the drug may eventually be tested on small numbers of humans with advanced cancer as part of a phase I trial. The drugs are normally given to patients with no regard as to how likely the drug is to work on that type of cancer, and the results are generally predictable. If the patients are able to tolerate the drug, and some of them have shrinkage of their tumors, the drug can then be tested on larger groups of patients with the same kind of cancer as the patients from the phase I trial who seemed to benefit. However, if none of the patients in the phase I trial respond the drug is usually abandoned. Rarely, a drug will jump through all these hoops and lead to an approved chemotherapy drug, usually over a period of a decade or more. However, even “effective” drugs typically work only in a minority of patients. Drug companies are typically more interested in developing drugs that work on ever larger numbers of patients rather than on identifying why the minority of patients who responded did so. When the effectiveness of this process plateaued with the current stable of chemotherapy drugs, companies began to focus more on “targeted drugs” like Gleevec, Herceptin, and Tarceva. These drugs were designed to affect only a single (or small number of) proteins such as the platelet-derived growth factor receptor (PDGFR), Her2, or the epidermal growth factor receptor (EGFR). The best example of this is Gleevec, which turns off the single protein that drives chronic myelogenous leukemia (CML) cells (bcr-abl), and has revolutionized the treatment of CML. In lung cancer, the best examples are the EGFR inhibitors Tarceva and Iressa, which can cause dramatic tumor responses and prolong survival in about 10% of Western NSCLC patients. It has been shown that the tumors in these patients are “addicted” to signaling through mutated EGFR, and shutting off this signal kills the cells. In the other 90%, however, there is either no benefit at all from Tarceva or some stabilization of disease, probably because EGFR is important to these tumors but not to the same extent as in the EGFR mutants.

The critical thing to remember with these drugs is that a “targeted” drug is only useful if the tumor relies on that target to survive! Most drug companies still don’t get it. Even today most clinical trials are combining targeted drugs with each other and/or with chemotherapy in order to get more and more unselected patients to benefit, instead of figuring out the right patients for a particular drug. It makes financial sense if you think about it, since a drug that is approved for all NSCLC patients would be used in hundreds of thousands of patients each year, while a drug that is useful in 5% of NSCLC patients would have a much smaller market. Despite this resistance, some researchers are looking for more of these targets in lung cancer. EGFR was the first, and likely represents that largest slice of the NCSLC pie for a single targeted drug (10%). However, there is reason to believe that smaller numbers of patients might have cancer that relies on specific targets in much the same way that EGFR-mutants rely on EGFR signaling. Michael Comb, the founder of Cell Signaling Technologies, Inc, is leading an effort to identify these targets using a technology called PhosphoScan, which can identify proteins called tyrosine kinases which have been turned on in NSCLC cell lines and tumor samples. By looking at over 3700 proteins in 150 tumor NSCLC samples, he was able to identify four additional proteins that may together be behind another 10% of NSCLC cases. These proteins are Met (1%), PDGFR (4%), Ros (1%), and Alk (4%). (Click to enlarge) Abnormal signaling through these tyrosine kinases drives lung cancer cells in some cell lines and in patient samples, and drugs which target these proteins are effective at killing these cells, at least in a culture dish. The next step will be to validate the importance of these pathways by testing drugs that inhibit the proteins. Already a clinical trial using an Alk kinase inhibitor is underway in NSCLC patients who are positive for the abnormal Alk protein (present in 4-5% of patients), and PDGFR inhibitors already exist in drugs like Gleevec, Sorafenib, and Sutent. Met inhibitors are also in clinical development, although I am not sure if a Ros inhibitor has yet been developed. Together, we may now be able to explain the underlying cause of 20% (1 in 5) of newly diagnosed NSCLC patients, and drugs are already being developed that may be effective in these patients. What to do in the other 80% is still up in the air, but many of those patients will have perturbations of a limited number of pathways, and combinations of targeted drugs may be effective in those patients. For example, some tumors rely on signaling through both Met and EGFR, such that inhibiting one or the other is ineffective while combining EGFR and Met inhibitors may work. The message here is that it is at least as important to identify the important targets in lung cancer as it is to create and test new drugs to hit the targets. If we invested half as much money and effort into studies like this as we do to gain FDA approval for another drug that works in only 30% of patients, I truly believe we could develop effective treatments for the majority of lung cancer patients in a short period of time. Hopefully pharmaceutical companies will come to understand that there probably aren’t any more blockbuster lung cancer drugs waiting in the wings, and that the future will be in carving out more tiny slices of the pie. As everyone knows around Thanksgiving, it doesn’t take that many slices of the pie before it’s all gone.