Just Enough Palliation: Radiation Dose and Outcome in Patients With Non–Small-Cell Lung Cancer

Lung cancer is a common disease, increasing in worldwide incidence, for which current strategies of prevention, detection, and therapy remain unsatisfactory. In the United States, 161,840 deaths from lung cancer are estimated to occur in 2008. Recurrence is common even among patients with apparently localized disease amenable to definitive local therapies. Most patients present with locally advanced or metastatic disease and are not appropriate candidates for resection. Although death resulting from lung cancer may come predominantly from distant disease, much suffering comes from uncontrolled local disease. Studying palliative therapies for patients with lung cancer is necessary, complex, and difficult. Patients vary in their prior therapy, such as previously untreated stage IIIA or IIIB and not a candidate for curative intent treatment, newly diagnosed stage IV, locally recurrent after surgery, relapsing after a response to initial chemotherapy, and refractory to initial chemotherapy. In addition, they vary in the extent of known disease, including intrathoracic only and intrathoracic and extrathoracic (with or without intracranial disease), and in age and comorbidities. The concept of locally advanced “incurable” disease is culturally as well as medically defined. Symptoms that may warrant palliation range from those clearly caused by intrathoracic tumor (such as airway and vascular obstruction, pain resulting from chest wall invasion, cough, hemoptysis), through those likely resulting from a combination of cancer and underlying benign pulmonary disease (such as dyspnea), to systemic symptoms such as fatigue, anorexia, and malaise that may have only a tenuous relation to the status of intrathoracic disease. The response of these symptoms may be assessed by the patient, the physician, or both. The study design may select any of a variety of primary end points including frequency and duration of symptom palliation, survival, or quality of life. Much of the rise in lung cancer incidence is occurring in parts of the world where the availability of radiation therapy is quite limited. It is estimated that in 2020, approximately 70% of the world cancer incidence will occur in countries of moderate or low income, and that these will typically be diagnosed at advanced stage with most patients in need of palliative treatment. The proportion of these cancers related to tobacco will also be high, in the range of 10%. Thus, there is an increasing need for better determination of the appropriate dose and fractionation requirements for effective and resource-efficient palliative radiation therapy for patients with incurable non–small-cell lung cancer (NSCLC). In those parts of the world with well-developed and distributed health care systems, we debate the convenience of one or two fractions compared with five or 10 fractions. In other areas, where radiation therapy is available for only a small portion of cancer patients, we may be talking about explicit rationing of care: who gets treated and who does not. Although palliative radiotherapy is cost effective by North American and European metrics, such arguments hold little weight when the required treatment facilities and the expertise to run them safely and effectively are simply not available. The study by Fairchild et al in this issue of Journal of Clinical Oncology reports the results of a systematic review of 13 trials involving 3,475 patients in which patients receiving palliative radiation therapy for NSCLC were randomly assigned among two or more radiation dose/fractionation schedules. Most of these trials compared very short (oneor two-fraction) with moderately short (fiveup to 15to 20-fraction) schedules. Their results show a modest but statistically significant benefit in improvement of total symptom score (but not individual symptoms) and survival for the higher doses up to doses of 35 Gy. One-year survival was 26.5% for patients in high-dose arms compared with 21.7% for patients in low-dose arms (P .002). Along with these benefits came an increase in physician-assessed dysphagia of modest (5.6%) but significant (P .01) degree. These results are not too surprising, and I suspect confirm the clinical impressions of many, if not most, radiation oncologists. The authors concluded that short schedules were satisfactory for most patients but that longer schedules and higher doses should be considered for patients of better performance status and realistic potential for longer term survival. This is not the first overview of these randomized trials of palliative fractionation for patients with NSCLC. The Cochrane Collaborative has published an analysis of a similar data set of 13 trials through 2002 and updated in 2006. Because of considerable heterogeneity in dose regimens, patient characteristics, and the methods of evaluating and reporting outcomes, a formal meta-analysis was not felt to be appropriate and a narrative synthesis and interpretation was made. They found that short and longer palliative schedules had similar response rates. There was no clear evidence for an overall survival difference, but several trials suggested a benefit for longer-course, higher-dose regimens in patients with good performance status (Eastern Cooperative Oncology Group performance status of 0 to 1). The Fairchild study differs methodologically from the Cochrane analysis in converting the doses in each treatment regimen to Biologically Equivalent Dose10 values and in comparing doseoutcome relations among all the trials despite the clinical heterogeneity of the patients studied.