Radiation Therapy at End of Life in Children

Introduction

Pressures related to the substantial resources devoted to patient treatments with advanced incurable conditions has led to an emerging literature in the adult population describing use of radiotherapy (RT) at the end of life (EOL). For purposes of this manuscript we define EOL as the 30 days preceding death. EOL RT carries with it minimal palliative benefit in most cases, and nonexistent survival benefit in adults.^1,2 Dual pressures of decreasing cost of care while simultaneously improving access to care for more patients make it critical that a national discussion continue regarding which aspects of RT at EOL in adults may justifiably be considered futile care.

The landscape of the discussion for children is far different. Unlike adults, for example, meningeal spread of cancer in children is not uniformly fatal, ³ and significant palliation and cure may be accomplished even in the direst cases. ⁴ Thus we anticipate that RT would be less frequently delivered at EOL for pediatric cancer patients.

Intense selection is involved in treating children with RT, and in particular with proton therapy (PrT). The relative scarcity of this technology^5,6 requires that geographic, financial, and availability variables all coincide to allow PrT access to highly selected patients. Thus, we anticipate that the population of pediatric PrT patients would be even less likely to be treated at EOL than standard pediatric RT patients. We sought to quantify the practice of providing RT at EOL to children and to assess the impact of PrT as a scarce technology.

Methods

After approval by institutional review boards, the records of the Indiana University Health Proton Therapy Center (IUHPTC) and University of Miami Radiation Oncology Department (UM) were reviewed for patients receiving radiation therapy at age ≤21 years between June 1, 2008 and June 1, 2013. Of these patients, specific analysis was made of patients not completing prescribed courses of RT and of patients dying within 30 days of receiving their final fraction of RT. Data were collected beginning in September 2013 in both sites. This allowed for 30 days to elapse from the children treated in June 2013: the ending date for both sites. Further analysis was made of differences between children (defined as ages ≤12) and adolescents/young adults (AYA; ages 13 to 21). Finally, comparison data for PrT and conventional RT was performed. Descriptive statistics were sufficient for this analysis.

Results

On January 1, 2010, IUHPTC became accredited by the American Association for Accreditation of Ambulatory Surgical Facilities. As part of that process, the center is required to submit cases of patients dying within 30 days of the end of their treatment, so processes were instituted to rigorously collect those data. These data were analyzed in October 2013. Only 2 children died of 272 treated (0.7%) (patients #8, 9 in Table 1). The first experienced aspiration at home that required hospitalization and subsequent discharge to hospice, and one developed progressive disease on treatment. Neither child received RT on the day they died.

At UM, corresponding data between June 2000 and June 2013 revealed 9 truncated courses of 464 pediatric conventional RT treatments (2%). Five of these children survived 30 days after RT and four died within 30 days. Of 425 patients for whom full follow-up data are available, another 3 patients were found to have died with 30 days of RT for a total of 7 children dying either on treatment or within 30 days afterwards (1.6%). Four patients received RT on the day they died.

The 30-day death rates for PrT and conventional RT were not significantly different (Fisher's exact test, p=0.21). These data were no different for children versus AYA patients because of limited events.

Discussion

We were able to find no other publications describing prevalence of RT at EOL in children. In 2012 we published RT at EOL results of a targeted review of adult patients presented at a departmental morbidity and mortality (M&M) conference. ¹ In this select population, over half of patients receiving RT within 30 days of death were being treated at time of death, and most of them were less than halfway through the prescribed RT course. Six patients received RT on the day they died. A subsequent multi-institution review ² included an entire year's patients; in that cohort, 6.3% died within 30 days of receiving their last treatment, and 2.3% received RT within their last week of life. For patients treated at EOL, the median time until death from completion of therapy was only 12.5 days. Similarly, Kapadia and associates ⁷ analyzed a population of nonsmall cell lung cancer patient outcomes. The authors reported 10% of patients who died had received RT at EOL, and nearly half did not complete the prescribed course. Guadagnolo and colleagues ⁸ described congruent results in patients older than 65 years from SEER-Medicare data. In this series, 7.6% received RT in the last month of life, and of these, only 17.8% received more than 10 treatments. It appears that RT at EOL in children is less frequent than in adults by about an order of magnitude (1.6% with photons, 0.7% with protons).

There are several important clinical differences between the adult and pediatric populations in this arena. First: many adult patients have comorbid conditions that might impact survival even if their cancer is controlled; this is rarely the case with children. Next, adults often have less capacity to tolerate treatment toxicities. Finally, adults, especially elderly adults, process risks and benefits differently from children, AYA, or their caretakers. Thus, deciding when to forego or decline therapy may be clearer in adults. The situation for children is further complicated by the curative intent possible in situations that would be otherwise palliative in adult patients. For instance, in children, leptomeningeal spread of disease is not uniformly fatal^3,4 as it is in adults.

Earlier integration of palliative care and hospice should be considered for patients of any age approaching EOL to help facilitate aggressive symptom management, advanced-care planning, and improve quality-of-life. Cost-effective palliative RT at EOL to address active symptoms should be limited to abbreviated courses, as evidenced by recent recommendations by the American Society for Radiation Oncology (ASTRO) regarding palliative treatment of bone metastases. ⁹ While RT is considered to be a “mainstay for the treatment of pain and/or prevention of the morbidity caused by bone metastases,” there are a number of available fractionation schedules that can provide symptom palliation of symptoms. At EOL, in general, briefer is better. We consider RT utilization at EOL to be an overuse metric in adults, but we are analyzing whether hypofractionated regimens (single fractions, perhaps including five-fraction courses in some clinical scenarios) might be acceptable interventions.

Pediatric patients provide a far more complex decision making scenario. The benefit of an additional year of life for a 10-year-old has a different societal value than a similar length of time to a retiree. Further, there are far, far fewer children dealing with EOL issues than adults. Thus, the societal cost of aggressive intervention for these children is a small fraction of the national expense for EOL care.

It may be possible that the role of radiation is less well understood by our pediatric oncology colleagues. In a recent publication, ¹⁰ a survey of Canadian pediatric oncologists revealed that only about 60% considered they had adequate knowledge about clinical indications for palliative RT. Over half of respondents thought that RT was underutilized in palliation of children's symptoms. Their perceived barriers to its more frequent use included (1) patient or family reluctance, (2) travel distance to the cancer center, (3) belief that PRT has little impact on quality of life, and (4) concerns about toxicity. ¹⁰ Thus, given that pediatric patients as a percentage enroll in more trials, and can tolerate more extensive courses of chemotherapy than adults in general, it is possible that they are simply being triaged more commonly toward chemotherapy as the primary palliative option.

With fewer concerns regarding the significant late effects of RT in children treated with curative intent,^11,12 pediatric RT at EOL is an area that requires increased study. The role of PrT is murky at best in the pediatric population in any case. For instance, there is no benefit to using protons in patients requiring whole-brain radiotherapy. While we have proposed that it is ethically sound to refer children requiring craniospinal RT for protons, ¹³ no such case may be made for children with extremity lesions. The national discourses on EOL care for children, and for the role of protons in that timeframe, are just beginning.

Conclusion

Two points may be made about current practice of RT at EOL in children from data presented here: (1) It is far less frequent in children than in adults; and (2) It appears to be no different using PrT compared to conventional RT.