One of the cardinal principles of modern medicine, as practiced in the United States, is active participation by patients—and often their families—in the process of clinical decision making.
Maurie Markman, MD
One of the cardinal principles of modern medicine, as practiced in the United States, is active participation by patients—and often their families—in the process of clinical decision making. Of course, there are rational and practical limits to this process. For example, patients generally will not be asked to decide what type of sutures to use in a surgical procedure or the specific antibiotic to be employed in managing an infection (except, of course, to inquire regarding the presence of prior allergic reactions).
However, when legitimate choices affect outcomes, including unique favorable or harmful characteristics of a particular option, the patient must be directly involved in the process and, after being provided with all available data, be permitted to make the final decision. In some situations, the clinical team may disagree with that choice, but—except in the most unusual of circumstances, such as mental incapacity—the patient’s desire needs to be respected.
With cancer and other serious medical conditions, available clinical data may indicate objectively striking differences for the probability of both benefit, such as a cure or long-term disease-free survival (DFS), and risk, including death, associated with 1 or more therapeutic options. In such circumstances, it is appropriate to argue that the decision can be made only by the patient, with the support and advice of the family and clinical team, based on an understanding of the available data and the patient’s beliefs, desires, and values.
Consider, for example, a recent report of the results of a randomized trial comparing 12 monthly infusions of cyclophosphamide with myeloablative autologous stem cell transplantation in the management of severe scleroderma.1 The trial, which involved 75 participants, revealed a statistically significant improvement (P = .01) in favor of an intensive myeloablative strategy employing a prospectively defined composite endpoint that included several highly clinically relevant events. A number of measures of clinical benefit were found to favor this approach at several time points following initiation of therapy. However, treatment-related mortality was 3% at 54 months and 6% at 72 months in the autologous transplantation group compared with no treatment-related deaths in the cyclophosphamide study arm.1
Therefore, the question is whether the unequivocal improvement in clinical outcome for the overall population associated with the aggressive therapy is justified, considering that there is a 1 in 20 chance of treatment-related mortality in this carefully selected and managed patient group. Of course, there is no single correct answer, just a decision that should be made by the individual patient.Turning to the realm of oncology, it is not difficult to find examples of study results in which individual patients may legitimately reach different conclusions regarding the interpretation of trial outcomes, including concerns about the risk of severe morbidity or treatment-related mortality versus the realistic potential for improved survival.
An excellent recent example: the reported phase II trial experience with the administration of everolimus to patients previously treated with cisplatin-based therapy for thymoma or thymic carcinoma, a notoriously treatment-refractory clinical setting.2 In a group of 51 patients, an objective response or disease stabilization was documented in 88% of the treated population, with a median progression-free survival (PFS) and median overall survival of 10.1 and 25.7 months, respectively. However, serious adverse events (AEs) were common, and 3 patients (6%) died of drug-induced pneumonitis.
Less serious but highly problematic toxicities may also characterize differences between therapeutic choices with different survival outcomes. For example, a recently reported phase III trial compared singleagent gefitinib versus vinorelbine plus cisplatin as adjuvant therapy for patients with stage II-IIIa EGFR-mutant non—small cell lung cancer.3
The study, which randomized 222 patients between the 2 treatment strategies, found a nearly 11-month median improvement in median DFS (28.7 versus 18.0 months; HR, 0.60; P = .0054) associated with the gefitinib treatment. However, by approximately 2.5 years after treatment initiation, the PFS curves for the 2 strategies were essentially superimposable. Also, although the incidence of serious AEs (grade 3-4) was substantially greater with chemotherapy, several low-grade but potentially problematic AEs, such as grade 1-2 skin rash (40%) and diarrhea (25%), were far more common with daily oral gefitinib therapy.
So, the question for patients in this clinical setting considering adjuvant therapy with either the targeted agent or cytotoxic chemotherapy is whether the observed benefit in delaying disease progression justifies the potential for symptomatic AEs and the impact on quality of life associated with gefitinib. Again, only the patient can make this very personal decision. A final example to be highlighted in this commentary is that of neoadjuvant chemotherapy versus an adjuvant strategy for early-stage breast cancer.4 The potential clinical utility of neoadjuvant therapy for a more conservative approach to breast conservation has been well established through a number of well-designed and well-conducted clinical trials. However, a recent analysis of the outcomes among 4756 women with earlystage breast cancer who participated in 10 randomized trials revealed that the risk of local recurrence may be somewhat greater (an absolute increase of approximately 5%) with the neoadjuvant approach compared with surgery followed by adjuvant chemotherapy (15-year risk, 21.45% vs 15.9%, respectively).
Therefore, the question is whether the cosmetic and quality-of-life benefits associated with neoadjuvant chemotherapy in well-defined clinical settings justify the small—but objectively measurable—increase in risk of local recurrence. Once again, the decision can be made only by the patient.