What Do We Mean by Targeted Therapy?

Abstract

Oncology, like all other avenues of life, is subject to fashion. Every few years, a new theory (eg, the dose-intensity hypothesis) or new term (eg, “paradigm”) will sweep down on the field, rather like hula hoops or the Macarena dance craze. The current fashion in oncology is “targeted therapy.” Sometimes the term is expanded to “molecularly targeted therapy.” One wonders what one might target other than molecules? Virtually every week I receive a mailing for a conference devoted to targeted therapy in this or that cancer; some of them, I must confess, I have spoken at. Each one of these conferences seems to equate targeted therapy with whatever is new, regardless of whether it actually is targeted. Every chemotherapy agent introduced in the past decade is now claimed to be targeted, at least by those marketing the new agent. Because I believe targeted therapy to be both real and of enduring importance, the misuse of the term concerns me. I therefore humbly propose some criteria for calling a therapy targeted. I do not suggest that every agent must meet every one of the proposed criteria, but rather that the criteria proposed represent the goals towards which targeted therapy wanna-bes should strive. These include: 1. The drug has a specific molecular target. This is the most trivial definition of targeted therapy. By definition, every agent has at least some molecular target, whether we recognize it or not. Otherwise, how could it work? In recent years, the target has been identified before the therapy has been developed, though some bona fide targeted therapies (eg, ovariectomy for metastatic breast cancer) were developed before the target (the estrogen receptor) was discovered. The more promiscuous the agent in its molecular targets, the less we should consider it a targeted therapy. 2. The target is biologically relevant (ie, is a part of the malignant phenotype) for specific cancers. The estrogen receptor is biologically relevant to the growth of specific human breast cancers, as is HER2. Replicating DNA is relevant to the malignant phenotype in a global sense, but is not a very specific target (rendering cyclophosphamide a poor exemplar of targeted therapy). 3. The target is reproducibly measurable in individual patients. Tamoxifen is useful because we can measure estrogen receptors. Trastuzumab is useful because we can measure HER2 with some reliability; try to imagine the development of trastuzumab had we been unable to assay (however imperfectly) HER2. The more reproducible the assay, the more able we are to determine clinical benefit. 4. The presence of the measurable target correlates with clinical benefit when the therapy is employed. If targeted therapies are used for every patient with a given disease, one has to wonder to what extent they are targeted. A few years ago, tamoxifen was proposed, on the basis of a somewhat dubious and tortured understanding of tumor biology, as a therapy for all breast cancer patients in the adjuvant setting. It is now clear that it only benefits patients who are truly estrogen receptor positive. Similarly, it is now reasonably clear that trastuzumab’s real benefits are confined to patients with fluorescence in situ hybridization (FISH)-positive cancers. These criteria may be criticized on individual grounds. Rituximab, for instance, has as a target an antigen not currently known to be crucial to the malignant phenotype of lymphomas, though it is relatively tumor specific. We currently lack any truly useful way of targeting individual patients with anti-epidermal growth factor receptor (EGFR) agents, yet EGFR is clearly a specific biologic target. Whether these definitional failures represent current lapses in knowledge or technique, as opposed to something more basic from a biologic standpoint, is uncertain. Nevertheless, the ability to define a concept is what gives a concept meaning. I suggest that the closer an agent is to meeting the 4 criteria outlined above, the more targeted (and the less scattershot) it becomes. Shotguns are effective, but not particularly subtle. As we develop the new generation of agents, we owe it to our patients to make these agents as precise as possible. Otherwise, these agents are just chemotherapy masquerading under another name.