Palliative radiotherapy during the last month of life: Predictability for referring physicians and radiation oncologists
- Authors:
- Published online on: September 2, 2015 https://doi.org/10.3892/ol.2015.3656
- Pages: 3043-3049
Abstract
Introduction
In numerous high-income countries with well-developed health care systems, regulatory bodies and other stakeholders participate in efforts towards improving the system, by optimizing access for patients requiring treatment, while also avoiding over-treatment and the use of non-cost-effective interventions (1). Adequate access for minorities and/or underserved regions and populations requires considerable health resources; such resources must therefore not be misspent by doctors prescribing therapeutic measures with unproven or doubtful benefit (2). Cancer treatment provides an example of the difficulties that arise when attempting to select the most appropriate treatment for a patient (3). Ideally, every patient should receive the optimum treatment and number of chemotherapy cycles or radiation fractions during the phase of disease where such treatment is meaningful to prolong life or maintain good functional status. By contrast, in the phase of unavoidable progression of the disease, the focus should be placed on easily tolerable palliative measures and avoidance of hospitalization or active anticancer therapy (4).
While palliative radiotherapy undoubtedly benefits patients with cancer in various stages of the disease, in terms of improving symptoms and potentially prolonging survival, its use during the last month of life has recently been questioned (5–7). Clinicians are often excessively optimistic when estimating the life expectancy of patients with advanced cancer, which may lead to treatment decisions that eventually become a burden for patients and caregivers, without improving the quality of life of the patients during the terminal phase of the disease (8,9). A recent audit of clinical practice in Nordland Hospital (Bodø, Nordland, Norway) between 2007 and 2009 revealed that 9% of all palliative radiotherapy was administered to patients during their last month of life (10). In other words, the physician's decision-making was appropriate in ~9 out of 10 patients, based on the assumption that patients with extremely short survival time are unlikely to experience any significant improvement in symptoms or quality of life (11). Due to the lack of patient-reported outcomes, the clinical benefit of palliative radiotherapy during the last month of life was not analyzed in the present study.
The appropriateness of using an arbitrary cut-off to define short survival/end of life (such as the 1 month cut-off used in previous publications) may also be debated, as a number of patients may survive only few days longer than this period (12). In order to avoid administering palliative radiotherapy during the last month of life, or to make informed decisions in cases where patients request treatment, despite limited expectation of survival, a prediction tool was previously developed and validated by Angelo et al (10). This prediction model was based on the following baseline parameters: i) Patients with lung or bladder cancer of any histological type; ii) those with an Eastern Cooperative Oncology Group performance status score of 3 or 4 (ECOG PS 3/4); iii) those presenting progressive disease outside the target volume(s) of radiotherapy; iv) those exhibiting levels of hemoglobin below the institutional limit of normal; v) those who received opioid analgesics at the start of radiotherapy; and vi) those who received steroids at the start of the radiotherapy treatment.
Based on this model, 75% of patients with all six of the aforementioned characteristics received radiotherapy during their last month of life. This percentage reduced to 74% in the validation dataset. However, the disadvantage of this model is that it is solely applicable to patients with lung or bladder cancer, and not to those suffering from other malignancies. Therefore, for the present study, a larger database was created, containing information from 2007 to 2011, which provides a higher likelihood of identifying predictive factors that may lead to a more powerful prediction model. The primary aim of the present study was to analyze the prognostic factors of survival time for patients with advanced cancer, and to evaluate in detail the clinical records of patients who had received palliative radiotherapy during the last month of life.
Patients and methods
Patients
The records of 873 consecutive patients with metastatic or otherwise incurable cancer receiving palliative radiotherapy at a single hospital were retrospectively reviewed. Patients with hematological or primary brain malignancies were not included in the analysis, due to the different biological behavior of these types of cancer.
The patients included in the study had commenced treatment during the period from June 20th, 2007 (the opening date of the radiotherapy facility at Nordland Hospital) to December 31st, 2011. Curative radiotherapy was not administered at Nordland Hospital during this period. All of the medical records of the patients, details of their treatment, and date of mortality were available on the electronic patient record (EPR) system of Nordland Hospital.
The survival status and date of mortality or last follow-up of the these patients were obtained from their corresponding EPRs during September 2014, resulting in ≥2.5 years of follow-up for the surviving patients. The survival time was measured from day 1 of palliative radiotherapy. If a patient was subjected to a second course of radiotherapy, the follow-up was censored at that time, and another record was created for the next radiation treatment, as each course of radiotherapy carries its own probability of being administered during the last month of life. This methodology was used in our previous prediction model (10).
Statistical analysis
IBM SPSS Statistical software, version 21 (IBM SPSS, Armonk, NY, USA), was used to evaluate the association between survival and potential prognostic factors, including blood biochemical and hematological parameters, such as levels of C-reactive protein (CRP), leukocytes, thrombocytes, hemoglobin, creatinine, lactate dehydrogenase (LDH), albumin, alkaline phosphatase and calcium. The institutional upper and lower limits of the normal values were applied: CRP, <10 mg/l; leukocytes 3.5–11.0×109/l; thrombocytes, 130–400×109/l; hemoglobin, 11.7–15.3 g/dl (females) and 13.4–17.0 g/dl (males); creatinine, 45–90 µmol/l; LDH, <205 U/l; albumin, 36–48 g/l; alkaline phosphatase, <105 U/l; and calcium 2.15–2.55 mmmol/l. Only those blood test results that were obtained within 1 week prior to the start of radiotherapy were included in the analysis. The ECOG PS of the patients at the time of consultation regarding radiotherapy, characteristics of the radiotherapy treatment received, and other baseline factors are presented in Table I.
Actuarial survival curves were generated using the Kaplan-Meier method and compared by log-rank test (univariate analysis performed for all baseline factors). For the multivariate analysis of survival from the start of the radiotherapy treatment, a Cox regression analysis was used (forward stepwise method). All factors with a significant P-value identified by the univariate log-rank test were considered in the subsequent multivariate regression analysis. The associations between the different variables of interest were assessed with the χ2 and Fisher's exact probability tests. P<0.05 was considered to indicate a statistically significant difference in two-sided tests.
Results
Baseline characteristics and risk in metastatic vs. non-metastatic disease
The median age of the patients included in the analysis was 68 years (range, 23–97 years). The median time from diagnosis of cancer to receipt of palliative radiotherapy was 24 months (range, 0–386 months). In patients with distant metastases, the median time from the detection of the first metastasis to the administration of palliative radiotherapy was 6 months (range, 0–149 months). Further details are presented in Table I. The majority of patients received radiotherapy for painful bone (54%) or brain metastases (15%), or for thoracic symptoms resulting from lung cancer (12%). The most common fractionation regimen was 3 Gy × 10 (43%), followed by 5–7 fractions of 4 Gy (22%). Stereotactic radiotherapy was not available. Depending on the anatomical site of the tumor and the total dose received, 2- or 3-dimensional treatment planning was used. As indicated in Table II, fewer patients with non-metastatic cancer received radiotherapy during their last month of life compared with patients with metastatic cancer (7 vs. 13%, respectively; P=0.12).
![]() | Table II.Rate of palliative radiotherapy during the last month of life according to the type of cancer. |
Prognostic factors for survival
The median survival time for the patients included in the analysis was calculated to be 6 months, and the 1- and 2-year survival rates were 35 and 27%, respectively. Multivariate analysis identified a number of significant prognostic factors, including ECOG PS; presence of brain, liver or bone metastases; disease progression outside the irradiated field, despite receiving systemic therapy; unavailability of systemic treatment due to age or comorbidity; history of >1 type of cancer diagnosis; use of opioid analgesics; and levels of CRP >30 mg/l (P=0.0001). Other factors also contributed to the regression model, including leukocytosis (based on the institutional upper limit of normal, 11.0×109 cells/l; P=0.006), use of steroids (P=0.007), and pleural metastases and/or effusion (P=0.01). Notably, the number of metastatic sites and the age of the patient were not observed to be significant. The most important risk factor for predicting short survival time was ECOG PS 4 (median survival time, 14 days; 1-year survival rate, 0%), followed by ECOG PS 3 (median survival time, 64 days; 1-year survival rate, 8%), as represented in Fig. 1 (P=0.0001). All patients with ECOG PS 4 were hospitalized when receiving radiotherapy. Of the 30 patients with ECOG PS 4, 22 received radiotherapy during the last month of life (73%). This rate was lower (30%) for patients with ECOG PS 3.
Similarly to the previously reported method (10), a scoring system was created that included all identified significant independent prognostic factors, weighted according to their influence on survival (Tables III and IV). Due to their significantly different hazard ratios for survival, ECOG PS 3 was assigned 3 points, while ECOG PS 4 was assigned 5 points. The other parameters were assessed in a present/absent format. Patients with metastatic cancer and points sum ≥21 were at high risk of receiving radiotherapy during their last month of life (77%). Comparable results were observed in patients without metastases and points sum ≥12 (75%). However, this score did not clearly outperform our previous prediction model (10) (77/75 vs. 75%, respectively). Based on this previous model, which requires a diagnosis of lung or bladder cancer, among other selection criteria, avoiding radiotherapy would have been recommended in 15% of patients with lung or bladder cancer, as this proportion of patients fulfilled all of the criteria. In the present study, 290 patients presented lung or bladder cancer, and therefore 43.5 patients (15%) should not be considered for treatment, if adhering to the old model. By contrast, the score calculated in the present study would lead to a recommendation against treatment in only 34 patients (4.6%) with lung or bladder cancer, which includes patients with metastatic disease and point sum of ≥21 and patients with non-metastatic disease and point sum of ≥12. Considering that 105 patients received radiotherapy during the last month of life, the majority of patients would still proceed to treatment, according to this model. These findings emphasize the limitations of the scoring method in the present study.
![]() | Table III.Score predicting radiotherapy during the last month of life in 641 patients with metastatic cancer. |
![]() | Table IV.Score predicting radiotherapy during the last month of life in 98 patients with non-metastatic cancer. |
Therefore, a pragmatic approach would be to extend the score of the previous model, which includes patients with lung or bladder cancer, and develop further disease-specific scores for the remaining patients. This possibility was explored, and a disease-specific score for patients with prostate or breast cancer was considered, since these groups contained >100 cases each, and presented comparable survival outcomes. The median survival times were observed to be 12.3 and 13.8 months for patients with prostate and breast cancer, respectively (P=0.57). None of the other groups exhibited comparable survival figures (the maximum survival time was 9.7 months in the case of patients with kidney cancer). In the combined prostate/breast cancer dataset, there were 21 instances of palliative radiotherapy during the last month of life, all among patients with metastatic disease (Table II). Multivariate analysis of prognostic factors revealed as significant the following parameters: ECOG PS (P=0.0001), elevated levels of serum LDH (P=0.0001), and presence of liver (P=0.0001), adrenal gland (P=0.01) and pleural metastases and/or effusion (P=0.036). The number of metastatic sites and the age of the patient were not observed to be significant. A score was assigned based on all of the above significant factors (Table V). However, due to the insufficient number of events, it was not possible to obtain firm conclusions from the analysis. Furthermore, despite the possibility of larger studies confirming that patients with points sum ≥10 present extremely short survival times and are poor candidates for radiotherapy, according to this model, treatment would be withheld in 1% of patients with prostate or breast cancer; therefore, the majority of patients with poor prognosis would still proceed to radiotherapy.
![]() | Table V.Score predicting radiotherapy during the last month of life in patients with breast or prostate cancer. |
Frequency of unexpected mortalities
Details of the clinical course were reviewed for the 105 patients included in the present study who had received palliative radiotherapy during the last month of life. The analysis identified a number of patients who had died unexpectedly, as follows: i) A 57-year old male patient with ECOG PS 2, presenting hepatocellular cancer with painful bone metastases and abdominal progression following systemic therapy with sorafenib, who succumbed to hepatic and renal failure due to erroneous drainage of excessively large volumes of ascites; ii) a 76-year old male patient with ECOG PS 2, affected by prostate cancer and painful bone metastases, who succumbed to small bowel ischemia; iii) a 67-year old female patient with ECOG PS 2, exhibiting non-small cell lung cancer with adrenal gland metastases and dyspnea from thoracic progression, who was receiving second-line systemic therapy with erlotinib, and died from a pneumothorax as a result of pleurodesis; iv) a 38-year old female patient with ECOG PS 2, presenting human epidermal growth factor receptor 2-positive breast cancer and ulcerated skin metastases, who was receiving third-line systemic therapy, and succumbed to right ventricular failure, possibly due to extensive lymphangitis carcinomatosa; v) a 68-year old male patient with ECOG PS 2, non-small cell lung cancer and pulmonary metastasis, who was undergoing sequential chemoradiotherapy, succumbed to sudden cardiac death; vi) a 51-year old female patient with ECOG PS 2 and supposedly limited pelvic relapse of small cell bladder cancer post-surgery, who died from rapid progression of novel distant metastases that resulted in hypercalcemia; and vii) a 72-year old male patient with ECOG PS 2, prostate cancer and painful bone metastases, who succumbed to myocardial infarction. All remaining cases presented a combination of adverse prognostic features, including ECOG PS 3/4, extensive metastatic disease, documented disease progression outside of the irradiated field(s), or lack of further systemic treatment options.
Discussion
The most important finding from the present comprehensive retrospective chart review is that numerous instances of palliative radiotherapy during the last month of life appear to be preventable if physicians are aware of the prognosis of the patient. Unexpected events, such as unforeseeable cardiac events or iatrogenic complications were unusual, whereas adverse prognostic factors indicating limited survival time were frequent. However, no simple prognostic model is capable of predicting mortality within 30 days with high accuracy (13). As discussed in the present study, numerous prognostic factors identified as significant in univariate analysis remained significant in the multivariate regression model. In the prediction model, complex prediction scores displayed some potential; however, the majority of patients who received radiotherapy during the last month of life were not be identified a priori. It is important to highlight the clinical dilemma of decision-making in favor of or against palliative radiotherapy in patients with limited survival expectations (14). Considering that short-course regimens with no or low-grade side effects exist, which often improve symptoms such as pain, dyspnea and hemoptysis, clinicians are wary of withholding a meaningful therapeutic measure for patients with terminal cancer (15–18). Therefore, prediction tools must not predict short survival times in patients who survive long enough to experience a reduced burden of symptoms, and must also identify the majority of patients will succumb to the disease too early to benefit from the treatment.
According to the results of the present study, disease-specific models may possess promising potential. In addition to patients with lung, bladder, prostate and breast cancer, relatively high rates of palliative radiotherapy during the last month of life were also observed in patients with metastatic kidney, colorectal and pancreatic cancer (Table II). However, further studies of disease-specific models will require larger databases.
Previous studies have reported data in agreement with the findings presented in the current study. In this regard, Anshushaug et al (19) also observed that ECOG PS 3/4 was strongly associated with palliative radiotherapy during the last month of life. The impact of performance status as a prognostic factor in patients with brain metastases is well known (20). In a Canadian study, palliative patients with cancer presenting ECOG PS 4 or 3 exhibited a median survival time of 25.5 or 55.0 days, respectively (21). These figures are in agreement with the results obtained in the present study, in which ECOG PS 4 was observed to perform almost as well as the complex scores in the prediction model. In the current study, 73% of patients with ECOG PS 4 had received treatment during the last month of life. However, the majority of patients who received radiotherapy during the last month of life presented ECOG PS 3, and a number of patients presented ECOG PS 2. As described in Fig. 1, certain patients with ECOG PS 3 may experience beneficial effects from radiotherapy, including prolonged survival. Consequently, they should be regarded possible candidates for therapy. However, clinicians should assess these patients comprehensively and thoroughly, and make individual decisions accordingly. A clear definition and communication to the patient of the goals of the treatment is mandatory, in order to avoid inaccurate beliefs about the effects of the treatment (22). Other factors associated with the treatment, including the toxic side-effects of the therapy, time for the patient to travel to the hospital, cost of the therapy and absence of family members to provide support for the patient, must also be considered (23). Furthermore, if radiotherapy is to be prescribed, it is important to select easily tolerable regimens and avoid lengthy treatment courses (24). The policy of Nordland Hospital is to assess the above mentioned scores for patients with lung/bladder and breast/prostate cancer. In this sense, patients at high-risk, and all those patients with other primary tumors and ECOG PS 3/4, receive information about the assessment and participate in the decision-making process for or against radiotherapy.
The limitations of the present study include incomplete baseline information in certain cases, particularly regarding the results of blood tests, and the limited number of events in the disease-specific analyses. In addition, not all patients included in the study underwent complete restaging during the month prior to radiotherapy and, therefore, the metastatic burden may have been greater than suspected in certain cases. Furthermore, data on the palliative efficacy of radiotherapy was not collected. Regardless, the present study provides important stimuli for further research towards the development of decision-making tools that may reduce subjectivity in the daily clinical assessment of the suitability of palliative radiotherapy for patients with terminal cancer. Therefore, future studies which assess the clinical benefit of palliative radiotherapy in patients with terminal cancer and stratify patients by primary tumor type are required. The additional inclusion of quality of life parameters and symptom severity into prognostic models may also be useful.
References
Scalo JF and Rascati KL: Trends and issues in oncology costs. Expert Rev Pharmacoecon Outcomes Res. 14:35–44. 2014. View Article : Google Scholar : PubMed/NCBI | |
Enewold L, Horner MJ, Shriver CD and Zhu K: Socioeconomic disparities in colorectal cancer mortality in the United States, 1990–2007. J Community Health. 39:760–766. 2014. View Article : Google Scholar : PubMed/NCBI | |
Dienstmann R, Salazar R and Tabernero J: Personalizing colon cancer adjuvant therapy: Selecting optimal treatments for individual patients. J Clin Oncol. 33:1787–1796. 2015. View Article : Google Scholar : PubMed/NCBI | |
Randén M, Helde-Frankling M, Runesdotter S and Strang P: Treatment decisions and discontinuation of palliative chemotherapy near the end-of-life, in relation to socioeconomic variables. Acta Oncol. 52:1062–1066. 2013. View Article : Google Scholar : PubMed/NCBI | |
Guadagnolo BA, Liao KP, Elting L, Giordano S, Buchholz TA and Shih YC: Use of radiation therapy in the last 30 days of life among a large population-based cohort of elderly patients in the United States. J Clin Oncol. 31:80–87. 2013. View Article : Google Scholar : PubMed/NCBI | |
Murphy JD, Nelson LM, Chang DT, Mell LK and Le QT: Patterns of care in palliative radiotherapy: A population-based study. J Oncol Pract. 9:e220–e227. 2013. View Article : Google Scholar : PubMed/NCBI | |
Kapadia NS, Mamet R, Zornosa C, Niland JC, D'Amico TA and Hayman JA: Radiation therapy at the end of life in patients with incurable nonsmall cell lung cancer. Cancer. 118:4339–4345. 2012. View Article : Google Scholar : PubMed/NCBI | |
Gripp S, Mjartan S, Boelke E and Willers R: Palliative radiotherapy tailored to life expectancy in end-stage cancer patients: Reality or myth? Cancer. 116:3251–3256. 2010. View Article : Google Scholar : PubMed/NCBI | |
Hartsell WF, Desilvio M, Bruner DW, Scarantino C, Ivker R, Roach M III, Suh J, Demas WF, Movsas B, Petersen IA and Konski AA: Can physicians accurately predict survival time in patients with metastatic cancer? Analysis of RTOG 97–14. J Palliat Med. 11:723–728. 2008. View Article : Google Scholar : PubMed/NCBI | |
Angelo K, Norum J, Dalhaug A, Pawinski A, Aandahl G, Haukland E, Engljähringer K and Nieder C: Development and validation of a model predicting short survival (death within 30 days) after palliative radiotherapy. Anticancer Res. 34:877–885. 2014.PubMed/NCBI | |
Langley RE, Stephens RJ, Nankivell M, Pugh C, Moore B, Navani N, Wilson P, Faivre-Finn C, Barton R, Parmar MK and Mulvenna PM: QUARTZ Investigators: Interim data from the Medical Research Council QUARTZ Trial: Does whole brain radiotherapy affect the survival and quality of life of patients with brain metastases from non-small cell lung cancer? Clin Oncol (R Coll Radiol). 25:e23–e30. 2013. View Article : Google Scholar : PubMed/NCBI | |
Nieder C, Andratschke N, Angelo K, Haukland E and Grosu AL: Development of a score predicting survival after palliative reirradiation. J Oncol. 2014:1282402014. View Article : Google Scholar : PubMed/NCBI | |
Ramchandran KJ, Shega JW, Von Roenn J, Schumacher M, Szmuilowicz E, Rademaker A, Weitner BB, Loftus PD, Chu IM and Weitzman S: A predictive model to identify hospitalized cancer patients at risk for 30-day mortality based on admission criteria via the electronic medical record. Cancer. 119:2074–2080. 2013. View Article : Google Scholar : PubMed/NCBI | |
Nieder C, Norum J, Dalhaug A, Aandahl G and Pawinski A: Radiotherapy versus best supportive care in patients with brain metastases and adverse prognostic factors. Clin Exp Metastasis. 30:723–729. 2013. View Article : Google Scholar : PubMed/NCBI | |
Cameron MG, Kersten C, Vistad I, Fosså S and Guren MG: Palliative pelvic radiotherapy of symptomatic incurable rectal cancer - a systematic review. Acta Oncol. 53:164–173. 2014. View Article : Google Scholar : PubMed/NCBI | |
Laugsand TS, Kaasa S, Romundstad P, Johannesen TB and Lund JÅ: Radiotherapy for bone metastases: Practice in Norway 1997–2007. A national registry-based study. Acta Oncol. 52:1129–1136. 2013. View Article : Google Scholar : PubMed/NCBI | |
van Oorschot B, Schuler M, Simon A, Schleicher U and Geinitz H: Patterns of care and course of symptoms in palliative radiotherapy: A multicenter pilot study analysis. Strahlenther Onkol. 187:461–466. 2011. View Article : Google Scholar : PubMed/NCBI | |
van Oorschot B, Rades D, Schulze W, Beckmann G and Feyer P: Palliative radiotherapy - new approaches. Semin Oncol. 38:443–449. 2011. View Article : Google Scholar : PubMed/NCBI | |
Anshushaug M, Gynnild MA, Kaasa S, Kvikstad A and Grønberg BH: Characterization of patients receiving palliative chemo- and radiotherapy during end of life at a regional cancer center in Norway. Acta Oncol. 54:395–402. 2015. View Article : Google Scholar : PubMed/NCBI | |
Leth T, von Oettingen G, Lassen-Ramshad YA, Lukacova S and Høyer M: Survival and prognostic factors in patients treated with stereotactic radiotherapy for brain metastases. Acta Oncol. 54:107–114. 2015. View Article : Google Scholar : PubMed/NCBI | |
Jang RW, Caraiscos VB, Swami N, Banerjee S, Mak E, Kaya E, Rodin G, Bryson J, Ridley JZ, Le LW and Zimmermann C: Simple prognostic model for patients with advanced cancer based on performance status. J Oncol Pract. 10:e335–e341. 2014. View Article : Google Scholar : PubMed/NCBI | |
Chen AB, Cronin A, Weeks JC, Chrischilles EA, Malin J, Hayman JA and Schrag D: Expectations about the effectiveness of radiation therapy among patients with incurable lung cancer. J Clin Oncol. 31:2730–2735. 2013. View Article : Google Scholar : PubMed/NCBI | |
Sharma S, Hertan L and Jones J: Palliative radiotherapy: Current status and future directions. Semin Oncol. 41:751–763. 2014. View Article : Google Scholar : PubMed/NCBI | |
Lutz ST, Jones J and Chow E: Role of radiation therapy in palliative care of the patient with cancer. J Clin Oncol. 32:2913–2919. 2014. View Article : Google Scholar : PubMed/NCBI |