Efficacy and Survival after Palliative Radiotherapy for Malignant Pulmonary Obstruction

Abstract

Introduction:

The purpose of this study was to determine the efficacy of palliative radiotherapy (PRT) for patients with pulmonary obstruction from advanced malignancy and identify factors associated with lung re-expansion and survival.

Materials and Methods:

We reviewed all patients treated with PRT for malignant pulmonary obstruction (n = 108) at our institution between 2010 and 2018. Radiographic evidence of lung re-expansion was determined through review of follow-up CT or chest X-ray. Cumulative incidence of re-expansion and overall survival (OS) were estimated using competing risk methodology. Clinical characteristics were evaluated for association with re-expansion, OS, and early mortality. Treatment time to remaining life ratio (TT:RL) was evaluated as a novel metric for palliative treatment.

Results:

Eighty-one percent of patients had collapse of an entire lung lobe, 46% had Eastern Cooperative Oncology Group (ECOG) performance status 3–4, and 64% were inpatient at consultation. Eighty-four patients had follow-up imaging available, and 25 (23%) of all patients had lung re-expansion at median time of 35 days. Rates of death without re-expansion were 38% and 65% at 30 and 90 days, respectively. Median OS was 56 days. Death within 30 days of PRT occurred in 38%. Inpatients and larger tumors trended toward lower rates of re-expansion. Notable factors associated with OS were re-expansion, nonlung histology, tumor size, and performance status. Median TT:RL was 0.11 and significantly higher for subgroups: ECOG 3–4 (0.19), inpatients (0.16), patients with larger tumors (0.14), those unfit for systemic therapy (0.17), and with 10-fraction PRT (0.14).

Conclusion:

One-fourth of patients experienced re-expansion after PRT for malignant pulmonary obstruction. Survival is poor and a significant proportion of remaining life may be spent on treatment. Careful consideration of these clinical factors is recommended when considering PRT fractionation.

Introduction

Lung cancer is the leading cause of cancer-related death worldwide and most patients present with advanced stage disease.¹ In addition, the lung is a common site of metastasis and progression for patients with nonlung primary malignancies. Malignant airway obstruction secondary to tumor growth occurs in nearly a third of such patients and portends a very poor prognosis, with median survival rates measured in months.² Malignant airway obstruction often results in symptoms such as dyspnea, hemoptysis, chest pain, hoarseness, cough, and pneumonia. Postobstructive pneumonia can be life-threatening and is often recurrent or refractory to treatment with broad spectrum antibiotics.³ Persistent infection can not only progress to sepsis and respiratory failure, but also severely limits patient performance status and can preclude patients from receiving systemic therapy. Thus, reversing malignant obstructions in an effective and timely manner is paramount.

A consensus on the optimal palliative treatment of malignant airway obstruction is lacking and management of this condition is often determined by provider preference and clinical factors (patient prognosis, performance status, procedural risks, etc.).⁴ Options include bronchoscopic intervention with tumor debulking, APC (argon plasma coagulation), stent placement, endobronchial brachytherapy, or palliative radiotherapy (PRT).^5–7 Some evidence suggests improved outcomes with a combined modality approach⁶ although this appears to be at cost of increased rate of adverse events.⁸ PRT is commonly used due to wide availability and minimal treatment-related risks. PRT fractionation varies significantly and typically ranges from 1 to 10 fractions, and the optimal regimen is not clear. Relatively few trials have reported the effectiveness of this modality and rates of lung re-expansion, with lung re-expansion rates ranging from 54% to 79% in patients with good performance status.^9,10 There are little data examining outcomes in hospitalized or critically ill patients, or those with poor performance status. Occasionally, these patients expire during or shortly after completing radiation treatment.¹¹ To facilitate evidence-based decision making in this palliative clinical scenario, we aimed to determine rates of re-expansion after PRT, identify clinical factors influencing patient survival and re-expansion, and to assess the impact of PRT fractionation on the temporal burden on patients' remaining life.

Materials and Methods

After obtaining approval from our Institutional Review Board (IRB), we queried our radiation oncology electronic medical records (Mosaiq, Elekta AB) for all patients treated with palliative radiation for malignant airway obstruction at our academic institution between January 2010 and December 2018. We collected clinical data on presenting symptoms, radiation dose and fractionation, survival, and reviewed radiographic records to determine the approximate timing of obstruction/atelectasis and re-expansion.

Time to re-expansion was defined, only in patients with post-PRT follow-up diagnostic chest imaging, as the duration between start of PRT and clear radiographic evidence of increased aeration of the affected lung on chest CT (or chest X-ray if no other imaging available). The determination of lung re-expansion was based on direct review of the imaging and required that the radiographic report make specific mention of lung re-expansion, improved aeration, reduced consolidation, or reduced obstruction. Review of the imaging was conducted by panel of three coauthors (A.G.J., R.T.H., and M.K.F), including faculty radiation oncologists with specialty in thoracic radiation therapy (M.K.F.), and palliative radiation therapy (R.T.H.), and a consensus of agreement on re-expansion was required. The extent of obstruction was categorized as “lobar” if the obstructing malignant mass resulted in consolidation and/or collapse of a segmental lung volume or entire lobe of a lung (Fig. 1). Extent of obstruction was categorized as “entire lung” if the obstructing malignant mass resulted in complete consolidation and/or collapse of an entire lung (Fig. 2). Overall survival (OS) was defined as the time from start of PRT to the date of death from any cause (event) or last follow-up (right censor). A treatment time to remaining life ratio (TT:RL) was calculated using the total number of treatment days divided by days of remaining life from the start of PRT. This novel metric was created to quantify and compare the burden of time under treatment and has no precedent in the literature. Early mortality was defined as death within 30 days of start of PRT.

FIG. 1.

Representative case of obstruction resulting in lobar collapse. Patient with metastatic NSCLC admitted for respiratory failure and postobstructive pneumonia from a right hilar mass with consolidation of the lower lobe of the right lung (A). Palliative radiation with 20 Gy in 5 fractions was delivered through AP/PA fields. Treatment plan isodose curves on coronal and axial imaging were as shown (B). Imaging at one month follow-up revealed significantly improved aeration of the right lung (C). AP/PA, anterior-posterior and posterior-anterior (radiation field arrangement).

FIG. 2.

Representative case of obstruction resulting in entire-lung collapse. Patient with metastatic squamous cell carcinoma of the lung presented with hemoptysis, respiratory distress, and malignant obstruction resulting in collapse of the entire left lung (A). Palliative radiation with 8 Gy in 1 fraction was delivered to the left hilum through AP/PA fields. Treatment plan isodose curves on coronal and axial imaging were as shown (B). Follow-up CT imaging two months after treatment revealed significantly improved lung aeration, decreased size of the hilar mass, and re-expansion of the left lung (C).

Descriptive analyses were performed using count (frequency) and median (interquartile range) for categorical and continuous variables, respectively. These were compared between groups using chi-square and Kruskal–Wallis test, respectively. Cumulative incidence of re-expansion with death without re-expansion as a competing risk was estimated using competing risk methodology and compared using Gray's test.¹² To adequately account for the significant competing risk of death without further radiographic assessment, death without follow-up imaging was also considered a competing risk. Multivariable competing risk regression modeling was used to estimate the single-variable subdistribution hazard of various factors on re-expansion. OS was estimated using the Kaplan–Meier method; multivariable Cox proportional hazards modeling was performed to identify predictors of OS. Bivariate logistic regression models were generated to identify factors associated with early mortality, defined as death within 30 days of PRT start. All statistical analyses were performed using R version 3.6 (R Foundation for Statistical Computing, Vienna, Austria).

Results

Patient population

In total, 108 patients treated with palliative radiotherapy for malignant pulmonary obstruction were identified. Mean age was 62, Eastern Cooperative Oncology Group (ECOG) performance status was 3–4 in 46%, and 81% of patients had lobar collapse, and median duration of obstruction before treatment was 17.5 days (Table 1). Most patients (65%) initiated PRT within 30 days of presenting with airway obstruction. A large (≥6 cm) mass was present in 68% of patients. The level of care at the time of consultation was most frequently inpatient medical-surgical service (58%), followed by outpatient (36%), and inpatient intensive care unit (6%, all of whom were intubated). Average length of stay was 12.3 days. Of the 69 patients hospitalized at the time of consultation, 71% had PRT initiated before discharge. Of these patients receiving inpatient radiation (n = 49), 33 had sufficient stabilization for discharge and 16 died during hospitalization. Fourteen patients underwent prior lung-directed intervention to temporarily relieve pulmonary obstruction, typically bronchoscopic tumor debulking and bronchial stent placement. Only 47% of patients received systemic therapy at any point—22% received systemic therapy only before PRT, 10% received systemic therapy after PRT, and 15% received systemic therapy both before and after PRT. The most common fractionation regimen was 30 Gy in 10 fractions (55%) followed by 8 Gy in 1 fraction (26%), 20 Gy in 5 fractions (10%), and others (Table 1). Median follow-up from PRT start was 25 days (confidence interval [95% CI] 1–380).

Table 1.

Baseline Patient and Treatment Characteristics

	Total (n = 108)
Age, mean (range)	64 (32–90)
ECOG performance status, n (%)
0	0 (0)
1	23 (21)
2	36 (33)
3	33 (31)
4	16 (15)
Histology
Adenocarcinoma	24 (22)
Squamous cell carcinoma	19 (18)
NSCLC NOS	31 (29)
Small cell carcinoma	13 (12)
Metastasis/other	21 (19)
Disease status
New diagnosis	61 (56)
Recurrence of prior cancer	17 (16)
Progression of known disease	30 (28)
Extent of obstruction
Lobar	88 (81)
Entire lung	20 (19)
Maximal tumor size, median (range), cm	6.5 (2.5–16)
<6 cm	35 (32%)
≥6 cm	73 (68%)
Clinical setting at consultation, n (%)
Outpatient	39 (36)
Inpatient, floor	62 (57)
Inpatient, intensive care	7 (6)
Intubated due to obstruction, n (%)	7 (6)
Lung-directed intervention	14 (13)
Systemic therapy
Before RT	40 (37)
After RT	27 (25)
Any	51 (47)
Duration of pre-RT obstruction, median (IQR)	17.5 (8–41)
≤30 days	70 (65)
>30 days	38 (35)
Prescription dose (Gy), median (IQR)	30 (8–30)
30 Gy/10 Fx	59 (55)
8 Gy/1 Fx	28 (26)
20 Gy/5 Fx	11 (10)
16 Gy/2 Fx	7 (7)
17 Gy/2 Fx	2 (2)
8.5 Gy/1 Fx	1 (1)

ECOG, Eastern Cooperative Oncology Group performance status; NOS, not otherwise specified; NSCLC, nonsmall cell lung cancer; RT, radiotherapy; SCLC, small cell lung cancer.

Re-expansion and symptomatic response after palliative radiotherapy

In the treated population, 84 patients had post-treatment imaging available for analysis of the primary endpoint and 24 patients died without follow-up imaging. Among the patients with radiographic follow-up, 25 (23%) experienced lung re-expansion in the follow-up period. Median time to re-expansion was 35 days (95% CI 2–180).

Cumulative incidence of re-expansion for the entire cohort was 6% at 30 days, 17% at 60 days, 20% at 90 days, and 23% at 180 days (Fig. 3). Death without re-expansion rates at corresponding time points were 41%, 54%, 63%, and 71%, respectively. Multivariable analysis of re-expansion found no significant factors associated with re-expansion (Table 2). Inpatient setting and tumor size were nonsignificantly associated with re-expansion (p = 0.08). Cumulative incidence of re-expansion at 30, 60, 90, and 180 days was 5%, 21%, 31%, and 36% for outpatients and 7%, 14%, 14%, and 16% for inpatients, respectively (Fig. 4).

FIG. 3.

Cumulative incidence plot of pulmonary re-expansion and death without re-expansion.

Table 2.

Multivariable Models of Lung Re-Expansion and Mortality after Palliative Radiotherapy for Pulmonary Obstruction

	Re-expansiona			Overall mortalityb
	Hazard ratio	95% confidence interval	p	Hazard ratio	95% confidence interval	p
Age	0.99	0.96–1.02	0.60	1.00	0.98–1.03	0.77
Histology (vs. NSCLC)
SCLC	0.83	0.21–3.23	0.78	1.06	0.50–2.25	0.88
Metastasis	0.89	0.31–2.57	0.84	0.34	0.14–0.82	0.02
ECOG 3–4 (vs. 1–2)	0.71	0.30–1.70	0.44	2.70	1.57–4.66	<0.01
Extent of obstruction (entire lung vs. lobar)	1.76	0.50–6.12	0.38	0.89	0.43–1.84	0.75
Tumor size <6 cm (vs. ≥6 cm)	2.16	0.92–5.09	0.08	0.53	0.29–0.95	0.03
Inpatient (vs. outpatient)	0.44	0.17–1.12	0.08	1.74	0.94–3.20	0.08
Obstruction 30+ days (vs. <30 days)	0.58	0.24–1.41	0.23	0.51	0.29–0.93	0.03
No. of fractions (vs. 1)
2–5	2.31	0.54–9.96	0.26	0.34	0.15–0.77	0.01
10	2.41	0.69–8.36	0.17	0.24	0.11–0.50	<0.01
Re-expansion	—	—	—	0.28	0.15–0.52	<0.01

^a Competing risk regression.

^b Cox proportional hazards.

FIG. 4.

Cumulative incidence of re-expansion by outpatient versus inpatient status at the time of consultation.

In total, baseline dyspnea was present in 87 patients, not present in 10, and unknown in 11. Of the 87 patients with dyspnea, 67 (77%) had documented symptom follow-up, and improvement in dyspnea was noted in 26 (39%). Baseline hemoptysis was present in 24 patients, absent in 49, and unknown in 35. Of 18 patients with baseline hemoptysis and adequate follow-up, 13 (72%) reported improvement. Cough was present in 49, absent in 28, and unknown in 31. Of those with cough and follow-up (n = 40), 13 (33%) noted improvement. Of those with fatigue at baseline and adequate follow-up (n = 25), none improved after PRT. Similarly, of six patients with hoarseness, four had documented follow-up and none improved.

Seven patients were treated in the setting of severe respiratory failure requiring intubation and mechanical ventilation. Six patients were treated with single-fraction or 2-fraction regimens and one patient received 3 fractions of a 30 Gy in 10 fraction regimen before decompensating. None of the intubated patients achieved significant improvement in respiratory status after treatment to permit planned extubation, and all eventually underwent compassionate extubation before expiring in the hospital or inpatient hospice. One intubated patient did have lung re-expansion on follow-up imaging three weeks after treatment; however, this did not affect their clinical course and they succumbed to systemic illness and died during hospitalization.

Survival

Median OS for the entire cohort was 56 days (95% CI 39–92) as depicted in Figure 5. Re-expansion was associated with improved median OS (157 days vs. 25 days, p < 0.01). Multivariable analysis identified re-expansion, nonlung histology, ECOG performance status 3–4, smaller tumor size, obstruction >30 days before PRT, and number of fractions of PRT as significant predictors of OS (Table 2).

FIG. 5.

Kaplan–Meier plot of overall survival.

Forty-one patients (38%) died within 30 days of start of PRT. Predictors of early mortality included ECOG 3–4 (odds ratio [OR] 1.25), tumor size <6 cm (OR 0.72), inpatient status (OR 1.25), and number of fractions (10 vs. 1, OR 0.71) (Table 3). Re-expansion was not included in the model due to the significant number of re-expansion events occurring after the 30-day early mortality threshold.

Table 3.

Multivariable Logistic Regression Models of Predictors of Death within 30 Days of Palliative Radiotherapy Start

	Death within 30 days
	Odds ratio	95% confidence interval	p
Age (continuous)	1.00	1.00–1.01	0.33
Histology (vs. NSCLC)
SCLC	1.08	0.85–1.37	0.54
Metastasis	1.02	0.83–1.26	0.82
ECOG 3–4 vs. 1–2	1.25	1.05–1.47	0.01
Extent of obstruction (entire lung vs. lobar)	0.97	0.78–1.22	0.81
Tumor size <6 cm (vs. ≥6 cm)	0.72	0.61–0.85	<0.01
Inpatient (vs. outpatient)	1.25	1.04–1.50	0.02
Obstruction 30+ days (vs. <30 days)	0.93	0.79–1.09	0.37
No. of fractions (vs. 1)
2–5	0.84	0.66–1.06	0.15
10	0.71	0.58–0.86	<0.01

Median TT:RL was 0.11 (IQR 0.05–0.26). Patients with an ECOG performance status 3–4 had a significantly higher TT:RL than those with performance status of 1–2 (median: 0.19 vs. 0.07, p < 0.01), as did those seen while inpatient (median: 0.16 vs. 0.07, p < 0.01), patients with large tumors (0.14 vs. 0.07, p < 0.01), those not treated with systemic therapy (0.17 vs. 0.07, p < 0.01), and those treated with 10 fractions (median: 0.14 for 10 fractions, 0.09 for 5 fractions, and 0.06 for 1 fraction, p = 0.04). Histology and extent of obstruction were not associated with TT:RL.

Discussion

Patients undergoing palliative radiation for locally advanced and/or metastatic disease with progressive malignant airway obstruction have an exceedingly poor prognosis. We found a significant competing risk of death (without re-expansion) in our patient cohort, with ultimately only 23% of patients surviving and experiencing lung re-expansion after treatment, much lower than that seen in prior literature (54%–79%).^9,10 Re-expansion rates as high as 79% were seen in a report by Lee et al., although all patients included had good performance status (ECOG of 0–2), one third had localized (cM0) disease, and 83% received treatment to a biological equivalent dose (BED) ≥39 (10–15 fractions).^9,10 Our study examined a significantly sicker patient population with 64% of patients hospitalized and ECOG performance status of 3–4 in nearly 50%. Our patient cohort also included seven intubated ICU patients, none of whom had significant clinical improvement after PRT. We believe this cohort is generalizable to most institutions and contributes to our understanding of PRT outcomes in patients with little physiological reserve and limited treatment options. Our multivariable models for re-expansion and mortality found a nonsignificant association between lung re-expansion and outpatient status and smaller tumor size. Small tumor size has been associated with improved response to radiation treatments and increased likelihood of re-expansion in other studies as well.^9,10 Notably, more protracted PRT fractionation (5 or 10 fraction vs. 1) was not associated with re-expansion in our model, suggesting that a prolonged treatment course does not improve chances of lung re-expansion.

Predictors of improved survival identified in our study are in line with other studies that indicate that patients with larger tumors and nonresponders to PRT have significantly worse survival rates.^9,10 Many studies have shown small but significant survival benefits for good performance status patients with longer treatment courses^2,13–15 and/or higher BED of treatment.^9,16 For our patients, poor performance status (ECOG 3–4) was the strongest predictor of mortality, whereas prolonged (10 fraction) treatment course was most protective. The “protective” effect of protracted treatment may be confounded by selection bias, as our providers defer to shorter treatment courses for sicker patients and favor the institutional standard (10 fractions) for healthier patients.

Median survival after PRT was limited and 30-day mortality was higher in our cohort than seen in prior palliative literature—which ranges from 10% to 25% depending on the series.^11,17–21 For a population with such limited life expectancy, efficiency of PRT delivery is critical. To these ends, we devised a novel metric to evaluate the temporal burden of the treatment course on patients' remaining life and found a median TT:RL of 0.11 (11% of a patient's life from the start of PRT was spent on treatment). This information can have practical significance for patients and providers who may favor a shorter PRT course for inpatients, patients with large tumors (>6 cm), those unfit for chemo or with poor performance status. It should be noted that TT:RL is a novel metric has no precedent in prior literature and requires future validation.

Regarding symptomatic response, the data presented are limited by the follow-up documentation in the medical record. Thus, our patients had lower symptom palliation rates than typically seen in palliative radiation literature. Fairchild et al. carried out a meta-analysis of data across 13 RCTs comparing fractionation schemes and evaluating symptom palliation and survival. Response rates for hemoptysis were highest, ranging from 45% to 92% but most often in the 80% to 90% range.^{14,16,22–29} Palliation of cough ranged from 7% to 76% but most typically ranged 40% to 60%.^14,22–29 The difference in our symptom control can be partially explained by retrospective limitations—as inconsistent documentation of symptoms likely masked the true response rate. Although symptomatic improvement as a function of PRT fractionation was not a primary focus of this study, prior randomized studies have shown no difference in symptomatic improvement between short- and long-course PRT schedules.^13,15,22,26

This study has limitations. First, its retrospective nature limits us to the available documentation and imaging in the medical record, thus restricting our ability to systematically evaluate re-expansion rates and identify changes in patient-reported symptoms between evaluation and follow-up. As a result, this study is limited to hypothesis generation and requires prospective validation. Unfortunately, given the poor survival and high symptom and comorbidity burden evident in this patient population, randomized trials powered for lung re-expansion (or the resultant improvement in respiratory status) would be challenging to perform. We find these data enlightening and thought-provoking despite these limitations.

Conclusion

Palliative external beam radiotherapy for malignant pulmonary obstruction effectively achieved lung re-expansion in only 23% of patients with a median time to effect of 35 days. Median survival is limited and 30-day mortality was high. We identify clinical factors associated with re-expansion and survival that may guide evidence-based practice in this clinical situation. Careful evaluation of baseline and disease-related factors should be considered when evaluating a patient for PRT. In the interest of preserving quality of remaining life for such patients, shorter treatment courses should be considered. The treatment time to remaining life ratio may be a useful metric to quantify the value of PRT near the end of life and requires external validation.

Data Sharing

Research data are available on request.

Funding Information

No funding was received for this article.

Author Disclosure Statement

No competing financial interests exist.