Validation of the prognostic model for palliative radiotherapy in older patients with cancer

Abstract

BACKGROUND

Older patients are more likely to have a poor performance status and comorbidities. There is a reluctance to extensively investigate and treat older patients. As elderly individuals and patients with neoplasms each increase in number, palliative treatment of older patients is expected to grow as an issue.

AIM

To investigated the role of palliative radiotherapy in older patients and patients who were expected to demonstrate a therapeutic effect.

METHODS

From February 2019 to February 2022, 33 patients aged ≥ 80 years underwent palliative radiotherapy. The prognosis in palliative care study predictor (PiPS), palliative prognostic index (PPI), and delirium-palliative prognostic score (D-PaP) models were used for prognosis prediction. D-PaP scores calculated according to the doctor's prediction of clinical prediction of survival (CPS) were excluded and then analyzed for comparison. Radiation was prescribed at a dose of 2.5-7 Gy per fraction, up to a median of 39 Gy₁₀ (range, 28-75 Gy₁₀).

RESULTS

The median follow-up was 2.4 months (range, 0.2-27.5 months), and 28 patients (84.8%) showed subjective symptom improvements following treatment. The 2- and 6-month survival rates of all patients were 91.5% and 91.5%, respectively. According to regression analysis, the performance status index, symptom type, and radiation dose all showed no significant correlation with the treatment response. When survival was expected for > 55 days in the PiPS model, the 2-month survival rate was 94.4%. For patients with PPI and D-PaP-CPS values of 0-3.9 points, the 2-month survival rates were 90.0% and 100%, respectively. For patients with a score of ≥ 4 points, the 2-month survival rates were 37.5% and 0%, respectively.

CONCLUSION

This study shows that the prognosis prediction model used in palliative care can be used to identify patients suitable for treatment.

Key Words: Elderly; Neoplasm; Palliative radiotherapy; Prognostic factors; Survival

Core Tip: This is a retrospective study to investigate the role of palliative radiotherapy in older patients and patients who were expected to demonstrate a great therapeutic effect. The prognosis in palliative care study predictor, palliative prognostic index, and delirium-palliative prognostic score models were used for prognosis prediction. Most of patients showed subjective symptom improvements following treatment. The prognosis prediction model showed good correlation with survival. In order to increase the therapeutic effectiveness in palliative radiotherapy, it is necessary to assess a patient's exact prognosis and select appropriate patients accordingly.

INTRODUCTION

The incidence of cancer is high among individuals 60-69 years old and is 11 times greater among those ≥ 65-years-old compared to those < 65-years-old. For this reason, about half of all cancer cases are diagnosed in individuals aged ≥ 70 years, and older patients account for a large portion of the total population regarding the prevalence of cancer[1]. Cancer is one of the most significant diseases in older patients. About 60% of all cancer-related deaths occur in older patients aged 70 years[1,2]. Moreover, cancer accounts for about one-third of the causes of death in the elderly population[1,2]. When choosing a cancer treatment, both the characteristics of the cancer and the overall health status of the patient, such as their general condition and any underlying diseases, should be considered[2]. Older patients have a shorter life expectancy than younger patients; moreover, they typically have many accompanying underlying diseases and have a poorer general condition. For this reason, older patients are often rejected from receiving active testing and treatment services. Therefore, even if other factors, such as the underlying disease, are the same in young and old patients, older patients typically receive less treatment due to the simple fact that they are older[3].

Palliative treatment is a treatment approach that improves the pain and symptoms of a patient and their quality of life. Although palliative treatment is applicable regardless of patient age and the type and severity of their disease, most patients requiring palliative treatment are cancer patients. Palliative radiotherapy is relatively effective for cancer patients and tends to be a well-tolerated treatment. Although some studies have reported the usefulness of palliative radiotherapy in elderly patients, a large number of patients and caregivers are not receiving treatment due to fears of treatment, the risks of side effects, and doubts about treatment effectiveness[1]. Since actual age is not always associated with physical ability, the determination of treatment based solely on age can be an obstacle preventing appropriate treatment opportunities. The importance of palliative care is increasing due to the recent growth of the elderly population, as well as, the increase in cancer incidence, and the changes in traditional views or perceptions, such as a growing acceptance of the pursuit of a dignified death[4]. Therefore, in this study, we investigated the role of palliative radiotherapy in older patients and in patients who are expected to show a great therapeutic effect.

MATERIALS AND METHODS

Patients and initial evaluations

This study was approved by the Institutional Review Board (DKUH 2024-09-006). From February 2019 to February 2022, a total of 353 patients received palliative radiotherapy. Palliative radiotherapy was recommended: (1) When symptoms developed due to advanced cancer; and (2) When other treatment options, such as surgery or chemotherapy, were unavailable due to general conditions and underlying diseases. Among the 353 total patients, 39 were ≥ 80 years old, and 33 received planned palliative radiotherapy. Four of the six patients who did not receive the planned treatment did not complete treatment for reasons such as deterioration of their general condition. Patients who underwent same-site re-irradiation were excluded from the analysis. Prognosis in palliative care study predictor (PiPS), palliative prognostic index (PPI), and delirium-palliative prognostic score (D-PaP) models were used to predict patient prognosis[5]. For a more objective comparison, scores based on the physician's prediction of survival (CPS) were excluded from the D-PaP model and then compared and analyzed. The following data were collected for the PiPS model: Cognitive function, pulse, anorexia, dyspnea, dysphagia, fatigue, performance status index Eastern Cooperative Oncology Group (ECOG) score, global health status, leukocyte count, platelet count, urea, alanine aminotransferase, albumin, and C-reactive protein. Total scores were calculated using a computer-based interface, and the estimated survival time was expressed as days (0-13 days), weeks (14-55 days), or months (> 55 days). For the D-PaP model, the following data were collected: Dyspnea, anorexia, Karnofsky performance status (KPS), leukocyte count, lymphocyte percentage, and delirium. The total D-PaP score ranged from 0-19.5 points, of which the CPS contributed 0-8.5 points; however, these points were excluded at the time of score calculation. For the PPI model, the degree of dietary intake, swelling, the presence or absence of breathing difficulties at rest, delirium, and the palliative performance scale scores were collected. The total PPI score ranged from 0-15 points.

Treatment

Simulation computed tomography imaging was performed for treatment planning in all patients. Radiotherapy was prescribed at a dose of 2.5-7 Gy per fraction and up to the median biologically effective dose of 39 Gy₁₀ (range, 28-75 Gy₁₀). In general, for symptom relief, 30 Gy and 10 fractions, 20 Gy and 5 fractions, or 7 Gy and 1 fraction were prescribed. There is no known difference in treatment effect between treatment schedules[6]. Based on these results, the treatment dose and number of fractions were prescribed in consideration of the treatment site and the patient's general condition, as well as the patient's residence, treatment cost, and treatment preference. Primarily, if the patient’s general condition was good and there was no discomfort during movement, a long-term treatment regimen was prescribed, such as 30 Gy and 10 fractions. For patients with discomfort during movement or who lived at a far distance from the hospital, a short-term treatment regimen, such as 20 Gy and 5 fractions or 7 Gy and 1 fraction, was prescribed.

Surveillance and statistical analyses

Patient symptom evaluation and physical examinations were performed during and at the end of radiotherapy. Given the patient's general condition and life expectancy, imaging tests at the treatment site were not performed at the time of follow-up. The treatment-response evaluation results were recorded according to the patient's subjective evaluation. Each patient was evaluated according to their main symptoms. The pain level was evaluated by the change in the numerical rating scale (NRS), whereas bleeding was evaluated via the presence or absence of gross bleeding and changes in the amount of bleeding. Neurological symptoms, dyspnea, and obstructive symptoms were evaluated based on improvements in symptoms. The observation period was defined as the time to death or censoring from the start of radiotherapy. Survival durations were calculated using the Kaplan-Meier method and log-rank test. Logistic regression analysis was used for independent prognostic factor analysis. Statistical analyses were performed using SPSS version 22.0 (IBM Corporation, Armonk, NY, United States).

RESULTS

Patient characteristics

Patient characteristics are described in Table 1. The median age of all patients was 82 years (range, 80-92 years), and more than half of the patients were male (69.7%). The ECOG score was 1 point for 112 patients (36.4%), 2 points for 13 patients (36.4%), and 3 points for seven patients (29.4%). Lung cancer was the most common primary tumor. Most patients (72.7%) had metastatic tumors at the time of radiotherapy. Seventeen patients (51.5%) received palliative radiotherapy as the first treatment after tumor diagnosis, and the remaining patients had received other treatments, such as surgery or chemotherapy, after the first diagnosis.

Table 1 Patients characteristics, n (%).

Characteristics	Total (n = 33)
Age (years)
Median	82
Range	80-92
Gender
Male	23 (69.7)
Female	10 (30.3)
ECOG
0	1 (3.0)
1	12 (36.4)
2	13 (39.4)
3	7 (21.2)
4	0
Primary site of neoplasm
Lung	13 (39.4)
Gastrointestinal	7 (21.2)
Genitourinary	5 (15.2)
Others	11 (24.2)
Distant metastasis
No	9 (27.3)
Yes	24 (72.7)
Previous treatment
No	17 (51.5)
Surgery	6 (18.2)
Chemotherapy	6 (18.2)
Radiotherapy	1 (3.0)
Others	3 (9.1)

ECOG: Eastern Cooperative Oncology Group.

The distribution of patients according to the prognostic prediction model is described in Table 2. According to the PiPS model, most patients (81.8%) were expected to survive for > 55 days. When considering the D-PaP model, which excludes the PPI and CPS values, about one-third of the patients scored ≥ 4 points.

Table 2 Distribution of prognosis in palliative care study, palliative prognostic index and delirium-palliative prognostic score without clinical prediction, n (%).

Variable	Total (n = 33)
PiPS
Months+ (> 55 days)	27 (81.8)
Weeks (14-55 days)	6 (18.2)
Days (0-13 days)	0
PPI
< 4	28 (84.8)
≥ 4	5 (15.2)
D-PaP without CPS
< 4	28 (84.8)
≥ 4	5 (15.2)

PiPS: Prognosis in palliative care study; PPI: Palliative prognostic index; D-PaP: Delirium-palliative prognostic score; CPS: Clinical prediction of survival.

Treatment outcomes and prognostic factors

There were no patients who demonstrated changes in ECOG scores before and after treatment. The median follow-up duration was 2.4 months (range, 0.2-27.5 months), and five patients (15.2%) died during follow-up. A total of 31 patients had symptoms before treatment. Two patients who had no symptoms were treated with palliative radiotherapy for metastatic brain tumors. Subjective symptom improvement was confirmed in 26 patients (78.8%), whereas the remaining five patients (15.2%) experienced no symptom improvement. Among the patients who were treated for pain, 89.5% (17/19) experienced improvements in their NRS scores. Moreover, all three patients who were treated for bleeding experienced decreased bleeding. Among the patients who were treated for obstructive symptoms, 75% (3/4) experienced improvements in dysphagia, whereas 100% (2/2) experienced improvements in dyspnea. One-third of patients who were treated for neurologic symptoms reported improvements in headaches, cramps, and altered mental status. A total of 6 patients (18.2%) were transferred to another hospital within one week after the end of treatment, with long-term follow-up data consequently being unavailable; however, among these patients, four patients showed improvements in symptoms until the last follow-up date after treatment, whereas two patients demonstrated no improvement in symptoms. The survival rates at 2 and 6 months of all patients were 91.5% and 91.5%, respectively.

Logistic regression analysis showed that the age, ECOG score, type of symptoms, treatment location, and palliative radiotherapy dose were not significantly associated with treatment response (Table 3). The prognostic prediction model revealed no significant association with treatment response but did reveal a significant association with survival (Figure 1). The 2-month survival rate was 94.4% among the 27 patients who were expected to survive for > 55 days in the PiPS model. Among the six patients who were expected to survive ≤ 55 days in the PiPS model, one patient died at about one month after the end of treatment, whereas the remaining five patients were lost to follow-up. The 2-month survival rate of 28 patients with a PPI score of < 4 points was 100%. Of the remaining five patients who scored ≥ 4 points in this model, three were lost to follow-up, and two died at 1 month and 1.5 months after the end of treatment (P < 0.001). After excluding the CPS, the 2-month survival rates in the D-PaP model were 100% in the < 4 points group and 0% in the ≥ 4 points group (P < 0.001).

Figure 1 Overall survival according to prognostic groups. A: Overall survival according to prognosis in palliative care study model; B: Overall survival according to palliative prognostic index score; C: Overall survival rate according to delirium-palliative prognostic score model excluding CPS. PiPS: Prognosis in palliative care study; PPI: Palliative prognostic index; D-PaP: Delirium-palliative prognostic score; CPS: Clinical prediction of survival.

Table 3 Prognostic factors affecting treatment response upon logistic regression analysis.

Characteristics	OR (95%CI)	P value
Age	0.807 (0.579-1.126)	0.207
ECOG	0.956 (0.303-3.022)	0.940
Symptoms (pain vs others)	2.833 (0.389-20.179)	0.298
Treatment site (bone vs brain vs others)	0.834 (0.299-2.325)	0.728
Radiation dose, BED	1.021 (0.942-1.106)	0.610
PiPS (Months vs Weeks)	1.375 (0.120-15.721)	0.798
PPI (< 4 vs ≥ 4)	5.111 (0.592-44.146)	0.138
D-PaP without CPS (< 4 vs ≥ 4)	1.375 (0.120-15.721)	0.798

ECOG: Eastern Cooperative Oncology Group; BED: Biologically effective dose; PiPS: Prognosis in palliative care study; PPI: Palliative prognostic index; D-PaP: Delirium-palliative prognostic score; CPS: Clinical prediction of survival.

DISCUSSION

Palliative treatment aims to prevent and alleviate pain. It is a treatment approach that improves the quality of life of a patient by diagnosing and evaluating their pain, as well as other physical and psychosocial problems[7]. Cancer patients complain of symptoms due to various causes[8]. Palliative radiotherapy has been used for symptom relief in all types of tumors that cause symptoms since the early 1900s, and its effectiveness has been proven through a number of studies[9]. Approximately half of advanced cancer patients receive radiotherapy, and 30%-40% of radiotherapy are used to relieve symptoms in cancer patients[10]. However, in older patients, the rate of cancer treatment, including palliative radiotherapy, is lower than that of younger patients[3]. In general, older patients have poor general conditions and multiple underlying diseases, making treatment decisions difficult. These patients also tend to avoid thorough laboratory or imaging evaluations and aggressive treatment[1,3]. Unlike the fear of treatment, many investigators have reported good results of chemotherapy or radiotherapy with less toxicity[1,11-13]. Zachariah et al[11] reported the results of radical or palliative radiotherapy in cancer patients ≥ 80 years of age; specifically, most of their patients finished the planned treatment, with just 3.6% of patients who received radical radiotherapy and 9.6% of patients who received palliative radiotherapy being unable to receive the planned treatment. About 9.8% of patients in the radical radiotherapy group experienced treatment-related toxicity of ≥ 3 degrees, whereas none of those in the palliative radiotherapy group experienced such toxicity. Similarly, in our study, the treatment response rates were 77% in the radical radiotherapy group and 81% in the palliative radiotherapy group. A separate study by Campos et al[1] reported improvements in symptoms after palliative radiotherapy in about 50% of patients with metastatic bone tumors. There was no difference in the improvement of these symptoms according to age, with older patients ≥ 75 years of age exhibiting improvement in symptoms similar to those of patients < 65 years of age. The KPS index showed significant association with symptom improvement rather than the age.

Similarly, age was not a prognostic factor for survival in older patients treated with palliative radiotherapy[14]. Nieder et al[14] reported that one-year survival rates after palliative radiotherapy were 35% and 33% in patients who were 80 years or older and who were younger than 80 years, respectively. Brain metastasis and poor performance status were significant prognostic factors for survival. These results have also been confirmed in other studies[12,15]. Katano et al[15] analyzed the clinical outcomes of palliative radiotherapy in patients with head and neck squamous cell carcinoma. The results showed that the one-year survival rate was 35.6%, and the KPS was the only significant prognostic factor for survival. A study published by Hickish et al[12] analyzed 290 patients who received palliative chemotherapy for lung cancer and reported that performance status was a significant prognostic factor for survival, regardless of patient age.

Although there is a beneficial effect of palliative radiotherapy, if immediate death is expected, the expected treatment side effects outweigh symptom relief, or treatment compliance is low, the treatment effect will not be significant[16]. Therefore, it is important to select appropriate patients who can be expected to experience treatment effects while minimizing side effects via prognosis prediction[17]. However, it is not easy to predict a patient's prognosis and determine the best treatment[18]. Several studies have reported that medical staff tend to overestimate patient prognosis, and the estimated survival duration in one study was 12.3 weeks longer than the actual duration[17]. Similarly, in another study, the expected survival period was > 2 months in 40%-50% of patients with an actual survival period of ≤ 1 week[19]. Therefore, an objective evaluation tool is necessary for ensuring adequate treatment. Many studies have reported the prognostic significance of performance status in older patients who are treated with palliative radiotherapy. However, it is difficult to use performance status alone when deciding on palliative treatment. Generally, in cancer patients with terminal disease and a short life expectancy, the stage and type of cancer are less predictive of prognosis[17]. Among these patients, the general condition and the presence or absence of life-threatening diseases or conditions, such as shortness of breath, embolism, or cachexia, are more important factors, and a prognostic prediction model using performance status, anorexia, respiratory failure, edema, white blood cell count, and lymphocyte percentage is employed[8,20]. Various models have been proposed to predict patient prognosis, including PiPS, PPI, and D-PaP, which were used in this study[17]. These prognostic models use patient symptoms and, signs and laboratory data including performance status[21,22].

These prognostic prediction models have several limitations. There is controversy with respect to which model most accurately predicts prognosis; however, it can be appropriate to apply different models depending on the clinical situation and available variables[5]. A prognostic prediction model can be applied to patients who receive palliative radiotherapy, as performed in this study. With proper prognosis prediction, unnecessary treatment can be avoided in patients with a short life expectancy, and appropriate patient selection for palliative radiotherapy is possible. Another limitation is that the models use patient symptoms such as dyspnea, dysphagia, and edema. These symptoms can be relieved after treatment. Therefore, physician can still be important, and caution is required when these prediction models are used.

This study has several limitations. First, the number of patients was low, and the number of patients who were lost to follow-up loss was high. Therefore, the statistical robustness of the study is weak. In addition, it was difficult to evaluate treatment toxicity and treatment responses through long-term follow-up. Second, due to the inherent nature of retrospective data, the types of cancer varied, and the patient group were inhomogeneous. In addition, a subjective assessment or improvement of symptoms by the patient was used to evaluate treatment response instead of employing a standardized evaluation tool or objective evaluation index. Therefore, further studies using a more objective evaluation tool for therapeutic toxicity and adverse effects are necessary in the future. Additionally, a multicenter, large-scale study is necessary to increase the sample size and extend the follow-up period.

CONCLUSION

This study shows that palliative radiotherapy can be used as an effective palliative treatment in older patients. In order to increase the therapeutic effectiveness of palliative radiotherapy, it is necessary to assess a patient's exact prognosis and select appropriate patients accordingly. Although no factors were significantly correlated with symptom improvement, this study showed that good performance status and dietary intake were correlated with improved prognosis. Therefore, it is important to consider these factors in determining the use of palliative radiotherapy in such patients. This study confirmed the possibility that the existing prognostic prediction model can be applied to patients who receive palliative radiotherapy and that more appropriate treatment can consequently be provided to patients through prognostic prediction.