Efforts Continue to Uncover Early Indicators of Clinical Effect

NEW ORLEANS-Thecontinuing search for surrogate markersfor testing new anticancer drugsreceived a critical update at a scientificsymposium at ASCO on "Early Indicatorsof Clinical Effect. Do TargetedTherapies Hit Their Targets?"New concepts in assessing inhibitorsof epidermal growth factor receptor(EGFR) signal transduction infailureclude pharmacodynamics, pharma-cogenomics, and pharmacodiagnostics,according to Manuel Hidalgo,MD, of Johns Hopkins Medical Centerin Baltimore. "But inhibition of amolecular target is not the same as anantitumor effect," he cautioned.Effects on target markers are especiallyimportant if early studies do notshow objective tumor responses. Thefailureto inhibit tumor growth mightbe a sign of inadequate dosing, he said.Failure to inhibit the target indicatesthat a new drug is needed. Failure toinhibit tumor growth despite inhibitingthe target marker means that a newtarget marker or a combination of dif-ferent agents is needed.Dr. Hidalgo emphasized the criticalneed for validated surrogate tissuemarkers for pharmacodynamic studiesthat could be used in large clinicaltrials. "We can't do the studies neededin phase II trials using tumor tissuebecause it is not feasible to do pairedtumor biopsies in hundreds of patients.We need validated surrogatesto use early after treatment to predictwhich patients will do well," he said."The traditional divergence betweendiagnostic and therapeuticgroups must be transcended in orderfor biomarker research to be moresupported and better integrated intotherapeutics development," notedJohn W. Park, MD, of the Universityof California, San Francisco. He emphasizedthat biomarkers should bedefined in advance and appliedthroughout the development process.Pathway as ParadigmAlex A. Adjei, MD, PhD, of theMayo Clinic, Rochester, Minnesota,used the ras-MAP-kinase pathway as aparadigm for discussing aspects of targetedtherapy in non-small-cell lungcancer (NSCLC). Distinguishing betweenbiomarkers and surrogate markersis important, he noted. Biomark-ers are objectively measured and usedto evaluate biological processes. Surrogatemarkers are accurate measuresof drug effects but may not measurethe effect on the drug target. "Surrogatemarkers can be biomarkers, butbiomarkers are not necessarily surrogatemarkers," Dr. Adjei said.Major uses of biomarkers includepharmacodynamic markers, responseprediction, and patient selection.Pharmacodynamic markers show thatthe drug inhibits the target, and, theyare useful in phase I trials. They can beused with surrogate tissues, althoughtumor tissue is preferable. "In mostcases, surrogate tissue is inappropriatefor phase II trials unless you canshow that the effect is the same as whatwould happen in the tumor. The targetmust be critical for the activity ofthe drug," he added.Markers for response predictionmight include negative predictors toexclude patients who will not respondto the therapy.For markers used for patient selection,such as HER2/neu or EGFR, tumorsamples are key, according to Dr.Adjei. "The target must be critical forthe action of the drug, and the assaymust be validated," he stressed. "Don'tforget the pharmacology: what doesthe host do to the drug? Markers mustbe validated. The take-home messageis that you must do your homeworkbefore you start a clinical trial."Discovery of DifferencesDavid Carbone, MD, PhD, ofVanderbilt University Medical Center,Nashville, Tennessee, discussedtumor proteomics in targeted therapy,particularly with regard to lungcancer. "Lung cancer survival is reallypoor, and we don't have very goodtreatments," he said. "The number ofcases each year is almost the same asthe number of deaths."The search for molecular signaturesis currently a hot area in lung cancerresearch. "We all believe that there aresome properties of cancers that determinewhy one patient responds andanother does not; why one patient iscured and another isn't; why one patienthas metastases and another doesnot," Dr. Carbone said.The discovery of differences relatedto mutations in the EGFR is "themost important finding in the geneticsof lung cancer," he said. However,some responders and patients whodevelop stable disease in response toEGFR inhibitors do not have the identifiedEGFR mutations, and some whodo are resistant to EGFR inhibitorssuch as gefitinib (Iressa). "Our hypothesisis that this is a complex phenotypethat can be predicted by analysisof complex gene/protein patterns,"he said.Researchers have already found thatmetastatic potential is encoded in thebulk of the primary tumor rather thanin subclones within the tumor and thatmany solid tumors have gene/proteinpatterns similar across tumor types."The name of the game is clinicalbenefit. It would be nice to have prepredictorsof clinical benefit, not just ofsuper-responders," Dr. Carbone said.Proteomic ProfilingProteomics is important becausemost nucleic acid sequences exert theireffect on translation. "Activity comesfrom the protein, not the gene," Dr.Carbone said.One method that can be used forproteomic profiling and requires onlya few cells is matrix-assisted laser desorption/ionizing mass spectrometry(MALDI). Dr. Carbone said that MALDIanalysis has already identified proteinprofiles associated with improvedsurvival in lung cancer."MALDI uses frozen samples oftumor and requires no processing orextraction. It can distinguish primaryfrom metastatic tumor and can detectmediastinal lymph node involvement,which has major implications for managementof lung cancer. This couldpotentially improve lung cancer outcomes.We have found a 15-proteinpattern that predicts prognosis in lungcancer and might be useful for identifyingpatients for further therapy."Dr. Carbone also described the useof MALDI in nonsurgical samples. Themethod was 99% specific at detectingbronchial preneoplastic changes in preliminarystudies. "Each MALDI testcosts only a couple of dollars to perform,once you buy the million-dollarmachine," he addedThe downside is that MALDI canbe performed only on small proteins,and formalin-fixed tissues cannot beused.