From the Editor: Electronic Alerts Would Help Oncologists Manage Therapy

It was not so long ago that designing an optimal treatment plan for an individual cancer patient required only a confirmed histologic/morphologic diagnosis of a particular tumor type and appropriate staging, which was surgically based in most solid tumors.

A number of additional pathologic factors such as tumor grade were recognized to be of reasonable prognostic significance, suggesting an individual patient would likely experience a relatively more or less favorable outcome, but with few exceptions (eg, estrogen receptor in breast cancer) these data were not of any real value in helping the clinician develop a successful management strategy. That is, a patient with a localized high-grade malignancy might have a greater risk for experiencing metastatic spread at some point in the future but this knowledge was not useful in suggesting approaches to reduce that risk.

During the past decade, however, there has been a virtual explosion in both the quantity and quality of clinically relevant data available to oncologists that can influence disease management, and it is almost certain this state of affairs will only accelerate in the future. From the presence of overexpression of HER2 in a breast cancer to the existence of mutations in the epidermal growth factor receptor (EGFR) in a non-small cell lung cancer, specific treatments of both common and uncommon malignancies will be increasingly based on the presence of particular molecular targets.

But wait. What we seem to “know for certain” today regarding the relevance of such testing may be challenged and changed tomorrow. Consider, for example, the status of an evaluation of KRAS mutations in patients with metastatic colon cancer. It has been well accepted that individuals whose cancers contain a KRAS mutation will not respond to a therapeutic antibody directed at the EGFR receptor.¹ However, a recent report has strongly suggested that while this conclusion is accurate for most KRAS mutations in colon cancer, such therapy may be quite beneficial in the presence of one specific mutation (codon 13).²

And, the list of highly specific data in unique clinical settings, knowledge of which may influence clinical decisions significantly, continues to grow rapidly. For example, a group of investigators recently reported that men with metastatic prostate cancer who had previously been castrated experienced a markedly increased clearance of docetaxel, a phenomenon that could substantially impact the activity of the agent in this population, compared with noncastrated patients.³ Routine measurement of troponin in women receiving trastuzumab has been suggested as a method of predicting the risk of cardiotoxicity associated with the agent.⁴ And, knowledge of specific genetic polymorphisms (pharmacogenetics) may be useful in predicting tamoxifen efficacy and irinotecan toxicity.⁵

What is a busy oncologist to do to keep track of these potentially highly clinically relevant associations between a particular molecular finding and specific positive or negative outcomes such as toxicity, biological activity, and survival? It is virtually certain that both the quantity of the information required to be processed in considering such data and the complexity associated with an appropriate interpretation in individual clinical settings are only going to increase in the future.

In considering a rational solution to this rapidly evolving dilemma, one can easily envision an important role for a highly interactive electronic medical record to assist in providing timely reminders for laboratory tests that might be ordered, or specific settings where attention needs to be focused on a particular issue.

For example, in a hospital/group practice pharmacy system an electronic “flag” could be placed when docetaxel is ordered for a patient with prostate cancer, reminding the physician of the association between prior castration and docetaxel clearance. An individual institution or clinical practice may elect to include routine orders for obtaining troponin levels in all patients receiving trastuzumab, or the physician might be automatically asked if he/she wanted to obtain the test since it may be useful in this setting.

Organizations may prospectively decide how they want to deal with the importance of knowledge of individual pharmacogenetic profiles when administering specific antineoplastic agents, and build that paradigm into the electronic order entry process. For example, it might be determined that it is essential to obtain an analysis of the germline variations associated with the metabolism/removal of certain drugs (eg, UGT1A1 with irinotecan or CYP2D6 with tamoxifen). Under these circumstances, the pharmacy could mandate that these data be obtained before the order is accepted. Alternatively, treating physicians may simply automatically be given the option (not required) to request the test when the particular anticancer agents are employed.

Finally, this process can be used not only for its gatekeeper function as a way to ensure optimal drug utilization and safety, but also as an excellent educational tool that provides current, clear, accurate, and comprehensive explanations regarding the relevance of particular tests and the implications of specific findings. Such information could be extremely helpful not only for members of the healthcare team but also for our patients and their families.