This article is mentioned in:

Introduction

The cytological examination of pleural effusion is crucial for treatment planning and prognosis assessment for non-small-cell lung carcinoma (NSCLC). Indeed, positive cytology of the pleural effusion is considered to be a useful prognostic factor and is included as supplemental information in TNM staging (1). However, cancer cells may be present in the thoracic cavity even in the absence of pleural effusion. We previously reported that cancer cells may be detected through pleural lavage cytology (PLC) at the time of thoracotomy in NSCLC patients without pleural effusion (2, 3) and evaluated the prognostic significance of positive cytology. Additionally, several other studies have suggested that PLC findings obtained during surgery are important prognostic indicators (4–13). However, the contribution of PLC to TNM staging has not yet been established. In the present study, we performed re-examinations and additional follow-up using our multi-institutional database, with the aim to assess the prognostic significance of PLC based on more extensive data.

Patients and methods

Patients

Between 1987 and 2004, PLC was performed in 2,043 NSCLC patients without pleural effusion or diffuse adhesions who had undergone primary tumor resection. The study sample included patients who underwent PLC at the Division of Thoracic Surgery, Kobe University (Kobe, Japan) or the Department of Thoracic Surgery, Hyogo Cancer Center (Akashi, Japan). In each of these cases, PLC was performed immediately following thoracotomy and prior to any manipulations in the thoracic cavity. We excluded all cases in which pleural effusion was present and could be aspirated using a syringe. Other exclusion criteria are listed in Fig. 1. The final study sample comprised 1,317 patients.

Figure 1 Flow chart of included cases, excluded cases and dropouts. The final study sample comprised 1,317 patients.

PLC

In all the patients, the pleural cavity was carefully washed with 100 ml of physiological saline solution immediately following thoracotomy and prior to any further manipulation of the pulmonary parenchyma. To ensure a pure sample of desquamated cells, the surgeon avoided coming in contact with the pleural surface. Subsequently, the fluid was placed in a glass bottle with heparin and centrifuged at 1,500 × g for 5 min. The resulting sediment was stained using the Giemsa and Papanicolaou methods (14). The cytological examination results were classified as either negative or positive. Specifically, Papanicolaou classes I-III were considered to be negative, whereas classes IV and V were considered to be positive.

Preoperative evaluation and staging

Preoperative evaluations included a detailed patient history, physical examination, biochemical profile, chest radiography, bronchoscopy and computed tomography of the chest, brain and upper portion of the abdomen. Pathological staging was performed according to the International Staging System for Lung Cancer (1). Intraoperative staging was performed by dissecting intrapulmonary, hilar and mediastinal lymph node stations. The pathologists conducted careful postoperative examinations. The histological type of the tumor was determined according to the World Health Organization classification system (15). The therapeutic strategy, PLC technique and surgical procedure were strictly standardized for both institutes and a single surgeon performed all surgeries.

Statistical analysis

All the statistical analyses were performed using Stata software, version 10.1 (Stata Corp., College Station, USA). Baseline characteristics were reported as percentages for categorical variables and means ± standard deviation (SD) for continuous variables. To investigate the association between PLC status and various clinicopathological parameters, we used the Fisher's exact test or the Student's unpaired t-test. For survival analyses, we used Kaplan-Meier estimates. The differences between survival curves were assessed using the log-rank test. Overall survival (OS) was defined as the time interval between the date of surgery and the date of death. Data were censored at the last visit. The Cox proportional hazards model was used to investigate the associations between OS and potential prognostic factors. All the variables that were found to be significantly associated with OS in the univariate analysis were included in the multivariate Cox analysis. Values of P<0.05 were considered to indicate statistically significant differences in all the analyses.

Results

PLC results

PLC immediately following thoracotomy was positive in 46 of the 1,317 cases (3.5%). Patients with positive cytological findings had a significantly worse prognosis compared to those with negative cytological findings (5-year survival rate: 28 vs. 61%, respectively; P<0.0001; Fig. 2). The median follow-up was 5.7 years (range, 1.0–13.1 years). Patients with positive cytological findings had a better prognosis compared to those with dissemination/effusion (5-year survival rate: 28 vs. 12%, respectively; P=0.0013; Fig. 2).

Figure 2 Kaplan-Meier plot of the overall survival rate in 1,317 patients with NSCLC and its association with the PLC status in all-stage patients. The P-value was determined using the log-rank test.

Association of PLC status and clinicopathological characteristics

Table I presents the associations between the PLC status and various clinicopathological characteristics at the time of surgery. Positive cytology was significantly associated with female gender (P=0.0065) and adenocarcinoma histology (P<0.0001). Pathological stage (stages II and III; P<0.0001), pleural invasion (P2 and P3; P<0.0001), lymphatic invasion (Ly+; P<0.0001) and vascular invasion (V+; P<0.0001) were each significantly associated with positive PLC findings. All the pathological classifications were performed according to the criteria of the International Staging System for Lung Cancer (1).

Table I Associations between pleural lavage cytology status and various clinicopathological characteristics at the time of surgery.

Table I

Associations between pleural lavage cytology status and various clinicopathological characteristics at the time of surgery.

	Pleural lavage cytology
Characteristics	Positive, no.(%) (n=46)	Negative, no.(n=1,271)	Total (n=1,317)	P-value
Age (years)				0.8074
<65	19(3.4)	548	567
≥65	27(3.6)	723	750
Gender				0.0065
Male	23(2.6)	877	900
Female	23(5.5)	394	417
CEA (ng/ml)				0.0561
≤5.0	23(2.8)	811	834
>5.0	23(4.8)	460	483
Procedure				0.9516
Limited resection	11(3.4)	309	320
Lobectomy	35(3.5)	962	997
Pathological stage				<0.0001
I	18	(2.1)	826	844
II	12	(5.2)	219	231
III	16	(6.6)	226	242
Tumor size (mm)				0.0870
≤30	19(2.7)	688	707
>30	27(4.4)	583	610
Histology				<0.0001
Adenocarcinoma	43(5.2)	785	828
Squamous cell carcinoma	3(0.7)	400	403
Othera	0	86	86
Pleural involvement				<0.0001
P0	3(0.4)	711	714
P1	12(3.0)	382	394
P2	18(24.7)	55	73
P3	13(9.6)	123	136
Lymphatic invasion				<0.0001
Negative	17(2.0)	843	860
Positive	29(6.3)	428	457
Vascular invasion				<0.0001
Negative	15(2.0)	725	740
Positive	31(5.4)	546	577

[i] ^aAdenosquamous carcinoma, large–cell carcinoma, pleomorphic carcinoma, large-cell neuroendocrine carcinoma, carcinoid tumor.

Prognostic value of clinicopathological characteristics

We additionally investigated the prognostic value of clinicopathological characteristics. All the variables that were found to be significantly associated with OS in the univariate analysis (log-rank test; data not shown) were entered into a Cox proportional hazards multivariate analysis. Of note, the PLC status was among these significant associations. The multivariate analysis revealed that gender (male), age (≥65 years), preoperative serum carcinoembryonic antigen (>5.0 ng/ml), pathological stage (stages II and III), tumor size (>30 mm in diameter), pleural invasion (+), lymphatic invasion (Ly+), vascular invasion (V+) and PLC status were independent, statistically significant predictors of OS (Table II).

Table II Multivariate analysis of overall survival.

Table II

Multivariate analysis of overall survival.

Factors	Favorable	Unfavorable	Hazard ratio (95% CI)	P-value
Age (years)	<65	≥65	1.788 (1.481–2.170)	<0.0001
Gender	Female	Male	1.612 (1.377–1.891)	<0.0001
CEA (ng/ml)	≤5.0	>5.0	1.452 (1.247–1.691)	<0.0001
Procedure	Lobectomy	Limited resection	1.028 (0.838–1.255)	0.7877
Pathological stage	I	II/III	2.212 (1.878–2.606)	<0.0001
Tumor size (mm)	≤30	>30	1.349 (1.140–1.598)	0.0005
Histology	AD	Non-AD	1.144 (0.968–1.351)	0.1135
Pleural involvement	Negative	Positive	1.206 (1.027–1.416)	0.0222
Lymphatic invasion	Negative	Positive	1.332 (1.130–1.570)	0.0007
Vascular invasion	Negative	Positive	1.259 (1.072–1.479)	0.0054
Pleural lavage cytology	Negative	Positive	1.535 (1.065–2.149)	0.0226

[i] 95% CI, 95% confidence interval; AD, adenocarcinoma.

Of the 844 stage I cases, 18 (2.1%) were classified as PLC-positive (Fig. 3). Among the stage I cases, a log-rank test revealed that the PLC-positive group had a significantly worse OS compared to the PLC-negative group (P<0.0001; Fig. 3).

Figure 3 Kaplan-Meier plot of the overall survival rate in 844 cases with NSCLC and its relationship to the PLC status in stage I patients. P-value was determined using the log-rank test.

Table III presents the recurrence pattern in the 46 PLC-positive cases. Of those patients, 33 (71.7%) exhibited disease recurrence. The local and distant recurrence rates at 5 years were 34.8 and 36.9%, respectively.

Table III Recurrence pattern in PLC positive.

Table III

Recurrence pattern in PLC positive.

	Stages
Recurrence pattern	I(n=18)	II(n=12)	III(n=16)	Total(n=46)
No recurrence	6	3	1	10
Local	6	4	6	16
Distant	6	3	8	17
Unknown	0	2	1	3

Discussion

In 1958, Spjut et al (16) were the first to publish a report on cytological examination for cancer cells in pleural cavity washings at the time of thoracotomy. Eagan et al (17) and Okumura et al (18) later analyzed the clinical significance of this procedure and indicated its prognostic value. We developed our PLC procedure (described in Patients and methods) in the 1980s. Since then, our continuous follow-up survey has demonstrated the prognostic significance of the PLC status (2, 3). In the present study, we conducted an analysis of the most recent data (including 3 years of additional follow-up and 317 new patients), with the aim to provide a more detailed understanding of the clinical significance of the PLC status in NSCLC patients.

PLC immediately following thoracotomy yielded positive results in 46 of the 1,317 patients (3.5%). This percentage is similar to our previous findings (2, 3) and is also similar to the findings reported by other studies (4–13). A 5-year survival rate of 28% was observed among the patients who were found to be PLC-positive. The prognosis of such patients remains poor, similar to our earlier findings. We previously compared the survival of PLC-positive patients to that of patients who had positive pleural effusion findings at the time of lung resection and found no significant survival differences between these two groups (2). The results of the present study also suggest that a positive PLC status may be clinically synonymous with malignant pleural effusion.

In the present study, we also investigated the associations between PLC status and clinicopathological characteristics at the time of surgery. Positive cytology was found to be significantly associated with female gender and adenocarcinoma histology. In 43 of the 46 PLC-positive cases, the histological diagnosis was adenocarcinoma. Conversely, among all adenocarcinoma patients, 5.2% had a positive PLC status. A positive PLC status and adenocarcinoma are more common among patients with NSCLC. Consequently, the PLC procedure may become indispensable for the precise assessment of the clinical outcomes of NSCLC in the near future. Additionally, pleural, lymphatic and vascular invasion were found to be significantly associated with positive PLC findings. Accordingly, we recommend that the PLC status becomes a feature of NSCLC classification as part of TNM staging.

Cytological examination of the fluid has always been recommended for NSCLC cases with pleural effusion. Indeed, the cytological examination results are reflected in the staging of the disease. However, the PLC status has not yet been incorporated in the latest edition of TNM staging for lung cancer (1). Pleural effusion is not an indispensable finding in terms of cancer cell dissemination into the thoracic cavity. Indeed, pleural dissemination is occasionally detected despite a lack of pleural effusion. Therefore, we recommend that PLC is universally performed at the time of thoracotomy to evaluate the spread of NSCLC cells into the thoracic cavity.

A positive PLC status was detected in 18 of the 844 stage I cases (2.1%). The deviation between survival curves (PLC-positive vs. PLC-negative) was significant among stage I patients. A positive PLC status was associated with a high recurrence rate and this finding was significant for patients with stage I disease. These results suggest that the PLC procedure may be crucial in treatment planning for earlier-stage patients. More specifically, PLC findings may enable a more optimal selection of patients who may benefit from postoperative adjuvant therapy, possibly achieving more favorable outcomes following surgical resection in stage I NSCLC patients.

In conclusion, the PLC findings were positive in 3.5% of patients with surgically resected NSCLC and the rate of positive findings was higher among female patients and among patients with adenocarcinoma. Additionally, a positive PLC status was identified as an independent prognostic factor for OS. Together with previous findings, our results indicate that PLC status may be of value during staging and treatment planning in patients with NSCLC. The PLC procedure may contribute to the selection and adaptation of adjuvant therapy for patients with NSCLC, particularly those with early-stage disease.

References