This article is mentioned in:

Introduction

Distant metastases at the time of presentation of non-small-cell lung cancer (NSCLC) are a frequent clinical problem. Approximately 30–40% of NSCLC patients present with metastatic disease at the time of diagnosis (1, 2). The most common metastatic site is bone, followed by the lungs, brain, liver and adrenal glands. Clinically, treatment for metastatic NSCLC consists only of systemic therapy using cytotoxic and/or molecularly-targeted agents and palliative radiotherapy for symptomatic metastases. It is generally considered that the life expectancy of NSCLC patients depends on the extent of the disease and the response to chemotherapy. Performance status (PS), gender and weight loss have been evaluated as prognostic factors (3). Specific organ metastasis also appears to exert an effect on survival, although this has not been clearly determined (4–6). In this study, using our database of NSCLC patients, we investigated whether specific organ metastasis at the time of presentation was of prognostic significance in NSCLC patients.

Materials and methods

Patient staging

All the consecutive patients who were diagnosed with metastatic NSCLC in our Divisions of Respiratory Medicine of the University of Tsukuba Hospital and Tsukuba Medical Center Hospital between 1999 and 2012 were retrospectively analyzed. In all the patients, the diagnosis of NSCLC was confirmed with pathological and/or cytological specimens. The pathological diagnosis of NSCLC was determined according to the World Health Organization classification (7). A staging procedure was performed for all the patients prior to any treatment, using head computed tomography (CT) or magnetic resonance imaging (MRI), bone scintigraphy, as well as ultrasonography and/or abdominal CT. The patients were staged according to the Union for International Cancer Control (7th edition) TNM classification (8).

Metastatic locations

We evaluated the clinical significance of each organ metastasis for survival. In addition, in order to evaluate the clinical significance of metastases to different body regions, the metastases were divided according to region, namely brain/head, thorax, abdomen and extremities. Brain/head metastases included those in the brain, eyes, nasal sinuses and gingiva. Thoracic metastases included the lungs, pleura, pericardium and accumulation of pleural/pericardial fluid. Abdominal metastases included the liver, adrenal glands, pancreas, spleen, kidneys, gastrointestinal tract, peritoneum and abdominal lymph nodes. Finally, metastases in the extremities included muscle, bone, bone marrow and skin metastases.

Statistical analysis

The Kaplan-Meier method was used to assess survival curves and the log-rank test was used in the univariate analysis to evaluate the statistical significance of survival between two groups. The survival time was defined as the interval (in months) between the date of initial therapy or supportive care until the date of death or the date of the last follow-up. Significant variables identified in the univariate analysis were included in the multivariate survival analysis using the Cox proportional hazards model to investigate the effects of clinicopathological factors on survival (9). Logistic regression analysis was applied in order to determine whether the metastatic location was associated with PS and the response to chemotherapy. All the statistical analyses were performed using StatView software for Windows, version 5.0 (SAS Institute Inc., Cary, NC, USA) and P<0.05 was considered to indicate a statistically significant difference.

Results

Patient characteristics

Among the consecutive 1,542 NSCLC patients, 729 patients had pathologically and/or cytologically confirmed NSCLC with distant metastasis. The characteristics of these patients are summarized in Table I. There were 509 (69.8%) men and 220 women, with a median age of 69 years (range, 21–96 years). Among the patients, 250 (34.3%), 234 (32.1%), 207 (28.4%), 122 (16.7%), 98 (13.4%) and 69 (9.5%) had bone, lung, brain, adrenal gland, liver and extrathoracic lymph node metastases, respectively, whereas 283 (38.8%) patients presented with pleural and/or pericardial fluid effusion.

Table I Characteristics of 729 patients with metastatic non–small–cell lung cancer.

Table I

Characteristics of 729 patients with metastatic non–small–cell lung cancer.

Characteristics	No. (%)
Age (years)
Median	69
Range	21–96
Gender
Male	509 (69.8%)
Female	220 (30.2%)
Smoking history
Never	192 (26.3%)
Current or former	537 (73.7%)
ECOG PS
0–1	409 (56.1%)
2–4	320 (43.9%)
Histology
Adenocarcinoma	543 (74.5%)
Squamous	162 (22.2%)
Large–cell	19 (2.6%)
Other	5 (0.7%)
Treatment
Operation	14 (1.9%)
Radiotherapy	12 (1.6%)
Chemotherapy, CRT	456 (62.6%)
Best supportive care	247 (33.9%)
Metastatic site
Pleural/pericardial fluid	283 (38.8%)
Bone	250 (34.3%)
Lungs	234 (32.1%)
Brain	207 (28.4%)
Adrenal glands	122 (16.7%)
Liver	98 (13.4%)
Extrathoracic lymph nodes	69 (9.5%)
Pleura	41 (5.6%)
Othera	40 (5.5%)

[i] ^aPeritoneum, 10 (0.6%); skin, 9 (0.6%); muscle, 6 (0.4%); spleen, 6 (0.4%); pancreas, 3 (0.2%); kidney, 3 (0.2%); bone marrow, 2 (0.1%); stomach, 2 (0.1%); duodenum, 1 (0.1%); jejunum/ileum, 1 (0.1%); colon, 1 (0.1%); nasal cavity, 1 (0.1%); gingiva, 1 (0.1%); retina, 1 (0.1%); heart, 1 (0.1%). ECOG, Eastern Cooperative Oncology Group; PS, performance status; CRT, CHEMORADIOTHERAPY.

Significance of each organ metastasis for survival

In the univariate analysis, patients with metastasis to the bone (P=0.024), adrenal glands (P<0.001), liver (P<0.001) and extrathoracic lymph nodes (P=0.014) exhibited a shorter survival compared to those without metastases to these sites. In a multivariate analysis using the Cox proportional hazards model, the presence of metastasis to the adrenal glands and the liver was associated with a poorer survival (P<0.001) (Table II).

Table II Survival analysis by organ/site.

Table II

Survival analysis by organ/site.

A, Univariate analysis
	Median OS (months)
Organ/site	With metastasis	Without metastasis	P-value
Pleural/pericardial fluid	8.4	7.4	0.389
Bone	6.2	8.5	0.024
Lungs	6.9	7.9	0.469
Brain	6.7	8.3	0.071
Adrenal glands	3.9	8.9	<0.001
Liver	3.8	8.7	<0.001
Extrathoracic lymph nodes	5.8	8.0	0.014
Pleura	6.1	7.9	0.887
Other	4.9	8.0	0.020
B, Multivariate analysis
Organ/site	Hazard ratio	95% CI	P-value
Bone	1.16	0.97–1.38	0.102
Adrenal glands	1.83	1.47–2.28	<0.001
Liver	1.55	1.22–1.96	<0.001
Extrathoracic lymph nodes	1.31	0.99–1.73	0.056

[i] OS, overall survival; CI, confidence interval.

Significance of metastasis to different body regions for survival

In the univariate analysis, patients with metastases to the brain/head (P=0.046), abdomen (P<0.001) and extremities (P=0.024) exhibited a poorer survival compared to those without metastases to these regions. However, metastasis to the thorax was not found to be an unfavorable prognostic factor (P=0.226). In a multivariate analysis using the Cox proportional hazards model, only the presence of metastasis to the abdomen was associated with a poorer survival (P<0.001), whereas the presence of metastases to the brain/head (P=0.248) and/or the extremities (P=0.140) were not identified as unfavorable prognostic factors (Table III).

Table III Survival analysis by metastatic region.

Table III

Survival analysis by metastatic region.

A, Univariate analysis
	Median OS (months)
Body region	With metastasis	Without metastasis	P-value
Brain/head	6.6	8.4	0.046
Thorax	8.4	7.1	0.226
Abdomen	4.5	9.8	<0.001
Extremities	6.1	8.5	0.024
B, Multivariate analysis
	Body region	Hazard ratio	95% CI	P-value
Brain/head	1.11	0.93–1.34	0.248
Abdomen	1.74	1.44–2.09	<0.001
Extremities	1.14	0.96–1.36	0.140

[i] OS, overall survival; CI, confidence interval.

Association between metastatic region and PS

Using logistic regression analysis, metastases to the adrenal glands and pleural and/or pericardial fluid effusion were correlated with a poor PS of 2–4 (Table IV). As regards metastatic location, the presence of abdominal metastasis was found to be associated with a poor PS. However, no association was identified between specific organ metastases and response to chemotherapy in the logistic regression analysis (Table V).

Table IV Logistic regression analysis for factors associated with a poor performance status (PS) (2–4).

Table IV

Logistic regression analysis for factors associated with a poor performance status (PS) (2–4).

Variables	Hazard ratio	95% CI	P-value
Metastatic organ
Pleural/pericardial fluid	1.51	1.09–2.09	0.013
Bone	1.13	0.82–1.56	0.456
Lungs	1.14	0.83–1.58	0.420
Brain	1.03	0.73–1.46	0.878
Adrenal glands	2.23	1.47–3.36	<0.001
Liver	1.41	0.90–2.20	0.132
Extrathoracic lymph nodes	1.09	0.65–1.83	0.745
Pleura	0.97	0.51–1.87	0.935
Metastatic region
Brain/head	1.03	0.73–1.45	0.875
Thorax	1.38	0.99–1.92	0.059
Abdomen	1.80	1.29–2.52	<0.001
Extremities	1.15	0.84–1.58	0.396

[i] CI, confidence interval.

Table V Logistic regression analysis for factors associated with an unfavorable response to chemotherapy (SD + PD).

Table V

Logistic regression analysis for factors associated with an unfavorable response to chemotherapy (SD + PD).

Variables	Hazard ratio	95% CI	P-value
Metastatic organ
Pleural and/or pericardial fluid	0.84	0.53–1.32	0.441
Bone	1.00	0.64–1.56	0.994
Lungs	1.43	0.90–2.29	0.130
Brain	1.31	0.78–2.19	0.301
Adrenal glands	1.15	0.61–2.15	0.671
Liver	1.32	0.68–2.59	0.412
Extrathoracic lymph nodes	0.81	0.81–0.42	0.528
Pleura	0.60	0.24–1.50	0.271
Metastatic region
Brain/head	1.48	0.89–2.47	0.132
Thorax	1.09	0.69–1.72	0.716
Abdomen	0.99	0.61–1.61	0.978
Extremities	1.09	0.70–1.69	0.706

[i] CI, confidence interval; SD, STABLE DISEASE; PD, PROGRESSIVE DISEASE.

Discussion

In previous studies, the site of involvement did not appear to affect survival (10–12). However, other researchers reported that metastasis to specific organs may be associated with a poor prognosis (3–6). Finkelstein et al (3) reported that bone and liver metastases were identified as independent prognostic factors in 893 metastatic NSCLC patients. Sorensen et al (4) also demonstrated that NSCLC patients with brain metastasis exhibited a shorter survival compared to those without brain metastasis. In addition, Hoang et al (5) reported that skin and liver metastases were unfavorable prognostic factors in 1,436 patients with locally advanced or metastatic NSCLC. Recently, Bauml et al (6) reported that, among 376 NSCLC patients treated with systemic therapy, those with bone metastasis had a poor prognosis.

In the present study, the incidence of distant metastasis at the time of initial diagnosis of NSCLC was 47.3% and the most common metastatic sites were bone, lungs, brain, adrenal glands and liver. We investigated whether specific organ metastasis at presentation was of prognostic significance in NSCLC patients. In the multivariate analysis, the adrenal gland and liver metastases were identified as unfavorable prognostic factors. In the logistic regression analysis, adrenal gland metastasis and pleural and/or pericardial fluid effusion exhibited a statistically significant association with poor PS. When we evaluated metastasis to the head and neck, thorax, abdomen and extremities, only metastasis to abdominal organs was found to be an unfavorable prognostic factor in the uni- and multivariate analyses. In our logistic regression analysis, metastases to abdominal organs exhibited a statistically significant association with poor PS. No metastasis other than abdominal was found to be associated with an unfavorable PS.

The results of this study may have important implications in clinical practice. First, brain and bone metastases were not unfavorable prognostic factors in our study; however, several previous studies reported brain metastasis to be one of the unfavorable prognostic factors, as neurological symptoms due to metastasis may be irreversible (4, 13, 14). Bone metastasis has also been considered to be associated with poor survival due to skeletal-related events, such as pathological fractures, spinal cord compression and hypercalcemia of malignancy (15). The improved survival of patients with brain metastasis may be attributed to the discovery of early metastatic lesions by brain MRI, the introduction of stereotactic radiation therapy and the introduction of of epidermal growth factor receptor tyrosine-kinase inhibitors for the treatment of NSCLC patients. Zoledronic acid may also explain the improved survival in patients with bone metastases (16).

Second, liver metastasis was an unfavorable prognostic factor in our study, which was consistent with previously reported results (3, 5). Liver metastatic lesions are rarely associated with severe symptoms, although the majority of NSCLC patients had multiple nodules morphologically (17). It is well known that the majority of NSCLC patients with liver metastasis do not respond well to chemotherapy (18, 19). NSCLC may cause biliary tract obstruction by metastasizing to the lymph nodes in the porta hepatis or the hepatic parenchyma. The administration of chemotherapy may be complicated by liver metastasis regarding the activation or metabolism of several cytotoxic agents commonly used in the treatment of NSCLC. There have been several patients with liver metastases who were unable to continue chemotherapy due to liver dysfunction (20).

Third, adrenal gland metastasis was one of the unfavorable prognostic factors and was associated with a poor PS in this study, although adrenal metastasis per se rarely presents with severe symptoms (21) and only few patients with adrenal gland metastasis eventually develop adrenal insufficiency (22). It has not been determined whether patients with adrenal gland metastasis also present with unfavorable several organ metastases or with unfavorable metastatic patterns. However, adrenal gland metastasis per se was found to be an unfavorable factor, even in the multivariate analysis. The precise etiology of adrenal metastasis being an unfavorable prognostic factor remain unclear. In the future, we aim to evaluate metastatic factors using a cluster analysis to elucidate the reasons underlying this finding. Further studies are required to confirm the effects of these organ metastases as observed in the present study.

Despite these significant findings, there were certain limitations to this study. First, the design of the study was retrospective and, therefore, complicated by lead time and length time biases. Second, there was lack of information regarding the development of distant metastases during the clinical course of the disease. However, despite these limitations, our findings may be of clinical value for the future management of NSCLC patients from unselected groups. Our results confirmed that the therapeutic approach to the treatment of NSCLC patients with distant metastases is complicated.

In summary, our results suggest that liver and adrenal gland metastases adversely affect the outcome of NSCLC, whereas brain and bone metastases do not. Therefore, when deciding on the administration of agressive therapy, which may increase treatment-related mortality, the individual patient's medical condition, including coexistence of such metastases, should be taken into consideration.

References