Intraoperative Pleural Lavage Cytology in Lung Cancer Patients Meinoshin Okumura, MD, Senya Ohshima, MD, Yahiro Kotake, MD, Hideo Morino, MD, Masanori Kikui, MD, and Tsutomu Yasumitsu, MD Departments of Surgery and Pathology, Osaka Prefectural Habikino Hospital, Osaka, Japan
Cytology of intraoperative pleural lavage was examined in 164 lung cancer patients who underwent pulmonary resections. None of the patients had any pleural effusion or dissemination. Cytology was performed three times: (1)at thoracotomy, (2)immediately after resection, and (3) after washing the pleural cavity with 5,000 mL of physiological saline solution just before closure of the chest wall. Twenty-three patients (14%) had more than one positive cytological finding. The frequency of positive cytological findings was significantly correlated with pathological T classification, pleural status, and pathological stage. The positive lavage group had a significantly higher recurrence rate than the negative lavage
group in patients with stage I or stage I1 cancer. Four patients in the positive lavage group (17.4%) had recurrence in pleura or pericardium whereas only 1 patient in the negative lavage group (0.7%) had a recurrence in pericardium. The positive cytological finding of pleural lavage has more important meaning as a prognostic factor in stage I and stage I1 and indicates a greater possibility of recurrence in pleura or pericardium, but further examinations to evaluate the viability of detected malignant cells are required so that the positive cytological findings of pleural lavage can be regarded as subclinical pleural dissemination. (Ann Thorac Surg 2991;51:599-604)
S
pleural lavage cytology to confirm subclinical pleural dissemination and its utility as a prognostic factor.
pjut and associates [I] reported the cytological findings of pleural cavity washings in lung cancer patients and referred to the possibility of implantation of malignant cells by incisional biopsy. Eagan and associates [2] studied the cytological findings from pleural lavage after curative pulmonary resection in 135 lung cancer patients and reported the correlation of positive cytological findings with cell type, lymph node status, visceral pleural status, and pathologcal stage. Kondo and associates (31 reported that the positive cytological findings were significantly correlated with pleural status and lymphatic permeation but not with histology, stage, or lymph node involvement in pleural lavage just after thoracotomy. Buhr and associates [4] reported the results of pleural lavage cytology before and after pulmonary resections and the poor prognosis of patients with positive cytological findings, and insisted that positive cytological findings of pleural lavage should be added to the criteria for the staging system. Results from previous reports differ, and the importance of pleural lavage cytology as a prognostic factor or an indicator of subclinical pleural dissemination in lung cancer patients has not been understood completely. In this report, we also present the results of cytological findings obtained from pleural lavage fluid and try to clarify the relationship of positive cytological findings with several pathological factors and results of preoperative transthoracic needle biopsy. The other aim of this study is to refer to the possibility of Accepted for publication Nov 29, 1990. Address reprint requests to Dr Okumura, First Department of Surgery, Osaka University Medical School, 1-50 1-chome Fukushima Fukushima-ku Osaka-City, 553, Japan. 0 1991by The Society of Thoracic Surgeons
Material and Methods The study group consists of 164 lung cancer patients who had intraoperative pleural lavage cytology examined in our institute between September 1985 and December 1988. None of the study group had any pleural effusion or pleural dissemination. All patients of the study group had three intraoperative pleural lavages: at thoracotomy, immediately after lobectomy or pneumonectomy with extensive sampling of mediastinal lymph nodes, and after routine washing of the pleural cavity with 5,000 mL of physiological saline solution just before closure of the chest wall. At each lavage, the pleural cavity was filled with 500 mL of physiological saline solution and all the saline solution was aspirated into a glass bottle with heparin. About 400 mL of the supernatant of the lavage fluid was discarded after several hours’ settlement. The rest of the saline solution, including the precipitates, was shaken and centrifuged at 1,500 rpm for 5 minutes. The sedimented material was stained by the Papanicolaou method. One hundred ninety-eight patients with lung cancer underwent pulmonary resection during the period of this study in our institute, but 34 patients were excluded from the study group: 1 had apparent pleural dissemination at thoracotomy, 14 had pleural effusion at thoracotomy, and 19 lacked more than one lavage cytology specimen. Class 1 and class 2 were regarded as the negative cytological finding, class 4 and class 5 as the positive cytological finding. The patients whose three lavage specOOO3-4975/91/$3.50
600
OKUMURA ET AL PLEURAL LAVAGE CYTOLOGY
Ann Thorac Surg 1991;51:599-604
Table 1. Patients With Positive Cytological Findings Patient No. 1
2 3 4 5 6 7 8 9 10 11
Cytology 1 -
+ -
+ f
+ + + + +
2
3
Histology
TIVM
P
Stage
+ + +
-
Sq sq Sq sq sq sq Ad Ad Ad Ad Ad
T1 N O MO T2 N O MO T2 Ell MO T3 N O MO T3 hll MO T4 N2 MO T2 h13 MO T2 N O MO T1 NO MO T2 h0 MO T2 hO MO T1 hO MO
PO PO PO P3 P3 P3 PO P2 PO P1 P2 P1
I I I1 IIIa IIIa IIIb IIIb I I I
Ad Ad Ad Ad Ad Ad Ad Ad Ad Ad La
T2 NO MO T2 NO MO T2 NO MO T2 N2 MO T2 N2 MO T2 N2 MO T2 N2 MO T3 N2 MO T3 N2 MO T3 N3 MO T3 NO MO
PO P2 P2 PO PO P2 P2 P3 P3 P3 P3
I I I IIla IIIa IIIa IIIa IIIa IIIa
-
+ + -
+ -
+ -
-
12
-
+
13 14 15 16 17 18 19 20 21
+ + + + +
-
+
+
+
-
-
+
-
22
-
23
-
-
+ + + + +
f
-
+ -
+ -
+
-
sq
1
I
11%
IIIa
Follow-up (mo)
Site of Recurrence
35, NED 27, DAD 8, DT 23, DT 39, DT 24, DT 17, DT 22, DT 32, NED 25, NED 30, alive 30, CEA elevation 35, NED 7, DT 14, DT 25, NED 31, alive 13, DT 10, DT 14, DT 18, DT 13, DT 17, NED
None None Pericardium Bone Lung Medias tinum Lung Brain, lung None None Lung
... None Pleura Pleura None Lung, bone SCLN Liver Brain, lung Lung Pericardium None
Ad = adenocarcinoma; CEA = cardnoembryonic antigen; DAD = died of another disease; La = large cell DT = died of tumor; carcinoma; NED = no evidence of disease; P = pleural status; SCLN = supraclavicular lymph node; Sq = squamous cell carcinoma.
imens were all negative were defined as the negative lavage group, and the patients who had more than one specimen with a positive cytological finding were defined as the positive lavage group. Differences in histology, pathological TNM classification, pathological stage, pIeural status, and presence or absence of diagnostic preoperative transthoracic needle aspiration biopsy were examined between these two groups. TNM classification and stages were determined according to the Staging System
of the Union Internationale Contre le Cancer [5]. The general rules for clinical and pathological staging of lung cancer by the Japanese Lung Cancer Society [6] were adopted to divide pN1 into two subgroups: pNla, which has involvement in the intralobar lymph nodes, and pNlb, which has involvement in the hilar lymph nodes, and also to classify pleural status into four groups as follows: PO, visceral pleura is not invaded by tumor; P1, tumor has reached but not invaded through the visceral
Table 2. Patients With Cytological Findings of Class 3 Patient No.
Cytology
1
2
3
1 2 3 4
f
-
-
-
f
t
f f
-
5
-
6
f
-
f
-
Histology
TNM
P
Stage
sq Sq sq
T2 NO MO T2 NC MO T2 N2 MO T2 NO MO T2 N2 MO T1 N1 MO
P1 P1 P1
I I IIIa
PO
I IIIa
Ad Ad La
DT = died of tumor; Ad = adenocarcinoma; DAD = died of another disease; disease; I' = pleural status; Sq = squamous cell carcinoma.
PO P1
I1
Follow-up (mo)
Site of Recurrence
27, DT 28, DT 12, DT 15, NED 25, NED 4, DAD
Lung Bone Lung None None None
La = large cell carcinoma;
NED = no evidence of
601
OKUMURA ET AL PLEURAL LAVAGE CYTOLOGY
Ann Thorac Surg 1991;51:599-604
Table 3. Histology and Pleural Lavage Cytology
Cytology
Ad
sq
La
Sm
Ca
Total
Negative Positive
63 15 (19.2%)
61 7 (10.3%)
7 1(12.5%)
2 0
2 0
135 23
Total
78
68
8
2
2
158
Ad = adenocarcinoma;
Ca = carcinoid;
La
=
large cell carcinoma;
pleura; P2, tumor has invaded the visceral pleura but not involved the parietal pleura; and P3, tumor has invaded the parietal pleura and chest wall. All patients of the study group had an exact diagnosis preoperatively by transbronchial or transthoracic biopsy, so none of them underwent diagnostic intraoperative tumor biopsy. The last follow-up was done in April 1990, and sites of recurrence were confirmed in all patients with recurrence except 1 whose site has not been found yet despite rapid elevation in the level of carcinoembryonicantigen. The x 2 test was used for check for statistical significance and the method of Kaplan-Meier was used to calculate survival rates. The statistical significance of differences between the survival curves was examined by the generalized Wilcoxon test.
Results The positive lavage group consisted of 23 patients, 14%of the study group. Patients of the positive lavage group are listed in Table 1. The negative lavage group consisted of 135 patients. There were 6 patients whose worst cytological findings were class 3. These 6 patients are listed in Table 2, but they were not included in either the positive or the negative lavage group.
Histology Frequencies of positive cytological findings on pleural lavage in patients with adenocarcinoma, squamous cell carcinoma, and large cell carcinoma are presented in Table 3. Positive cytological findings were most common in adenocarcinoma, but there was no significant difference between cell types.
Pathological T Classification Patients with more advanced T classification had a higher frequency of positive cytological findings. There was a significant relationship between cytological findings and T classification (Table 4).
Sm = small cell carcinoma;
Sq = squamous cell carcinoma.
Pathological N Classification Frequencies of positive cytological findings in each N classification are presented in Table 5. No significant differences were seen except between pN1 and pN3. Thirteen of 103 patients (12.6%)with pNO or pNla whose nodal involvement did not extend to the hilar lymph nodes had positive cytological findings, and 10 of 55 patients (18.2%) with pNlb, pN2, or pN3 whose nodal involvement extended to hilar or mediastinal lymph nodes had positive cytological findings, but there was no significant difference.
Pleural Status Frequencies of positive cytological findings in patients with PO and P1 were almost the same value, that with P2 was the second highest, and that with P3 was the highest. There was a significant relationship between frequency of positive cytological findings and pleural status (Table 6).
Pathological Stage Frequencies of positive cytological findings in stage I, stage 11, stage IIIa, and stage IIIb were 12.7%, 3.7%, 20.5%, and 42.9%, respectively. Positive cytological findings were significantly more frequent in patients with stage IIIa, stage IIIb, or stage IV (23.1%)than in those with stage I or stage II (10.4%)(Table 7).
Preoperative Transthoracic Needle Aspira tion Biopsy Preoperative transthoracic needle aspiration biopsy was performed in 39 patients. Positive cytological findings were obtained in 8 patients with it (20.5%) and in 15 patients without it (12.6%). Presence or absence of preoperative transthoracic needle aspiration biopsy had no significant relationship with the frequency of positive cytological findings.
Prognosis The survival rates for 1 year, 2 years, and 3 years were 88.9%, 78.3%, and 69.2% in the negative lavage group
Table 4. T Classification and Pleural Lavage Cytology" Cytolow
T1
T2
T3
T4
Total
Negative Positive
51 3 (5.6%)
75 13 (14.8%)
8 6 (42.9%)
1
1(50.0%)
135 23
Total
54
88
14
2
158
TI versus T3 ,$ = 13.471,p < 0.005; 12.198,p < 0.005.
a
TI versus T4: ,$ = 5.744,p < 0.01;
T2 versus T3:,$ = 6.285,p < 0.01;
TI
+ T2 versus T3 + T 4 ,$ =
602
OKUMURA ET AL PLEURAL LAVAGE CYTOLOGY
Table 5 .
Ann Thorac Surg 1991;51:599-W4
N Classification and Pleural Lavage Cytology"
Cytology
NO
N1
N2
N3
Total
Negative Positive
74 12 (14.0%)
29 (Nla, 16; Nlb, 13) 2 (6.5%)(Nla, 1; Nlb, 1)
29 7 (19.4%)
3 2 (40.0%)
135 23
Total
86
31
36
5
158
a
N1 versus N 3 ,$ = 4.906, p < 0.05;
NO
+ Nla versus N l b + N2 + N3: ,$ = 0.891, not significant.
Table 6 . Pleural Status and Pleural Lavage Cytology"
Cytology
Po
P1
P2
P3
Total
Negative Positive
86 8 (8.5%)
25 2 (7.4%)
15 6 (28.6%)
9 7 (43.8%)
135 23
Total
94
27
21
16
158
PO versus P 2 ,$ = 6.461, p < 0.025; 16.449, p < 0.005.
a
PO versus P3: 2 = 14.417, p < 0.005;
and 87.0%, 50.2%, and 38.8% in the positive lavage group, respectively. The survival rate for the negative lavage group was significantly better than that of the positive lavage group (Fig 1).Forty-seven patients in the negative lavage group (34.8%)have had recurrence. On the other hand, 16 patients in the positive lavage group (69.6%)have had recurrence. The recurrence rate for the positive lavage group is significantly higher than that for the negative lavage group, but the fact that the proportion of higher stages is greater in the positive lavage group than in the negative lavage group should not be neglected; therefore, differences in recurrence rate between the two groups must be discussed in each stage. In stage I, stage 11, and stage IIIa, the recurrence rate was higher in the positive lavage group; in stage IIIb they were the same (Table 8). In patients with stage I or stage I1 disease, the recurrence rate for the positive lavage group (54.5%)was significantly higher than that for the negative lavage group (20%). In patients with stage IIIa, stage 11% or stage IV disease, there was no significant difference between the results of lavage cytology and the recurrence rate (Table 9). Sites of recurrence were classified as carcinomatous pleuritis, carcinomatous pericarditis, mediastinal or cervical lymph nodes, and distant metastases including lungs. There were 2 patients with carcinomatous pleuritis and 2 patients with carcinomatous pericarditis in the positive lavage group, though there was no patient with carcino-
PI versus P3: ,$ = 8.018, p < 0.005;
PO + PI versus P2 + P3: ,$ =
matous pleuritis and only 1 patient with carcinomatous pericarditis in the negative lavage group (Table 10). Of 6 patients whose worst cytologic findings were class 3, there were 3 patients (50%)who had recurrence. Their sites of recurrence, however, were not pleura or pericardium.
Comment We obtained the cytological findings from pleural lavage fluid at three times intraoperatively. Thirteen patients had positive cytological findings in the first lavage, 12 patients in the second lavage, and 6 patients in the third lavage. As the number of patients with positive cytological findings in each lavage is small, we considered it less meaningful to discuss the results of each pleural lavage cytology; therefore, patients who had at least one positive cytological finding were grouped together into the positive lavage group. Twenty-three of 164 patients in the study group (14.0%) had positive cytological findings, the incidence of which had a significant relationship with pathological T classification, pleural status, and pathological stage. There was no significant relationship of positive cytological findings with lymph node involvement or histology in this study group. Previous investigators did not comment on the effect of preoperative transthoracic needle aspiration biopsy. Its presence or absence has no significant relation-
Table 7 . Pathological Stage and Pleural Lavage Cytology"
Cytology
Stage I
Stage I1
Stage IIIa
Stage IIIb
Stage IV
Total
Negative Positive
69 10 (12.7%)
26 1(3.7%)
35 9 (20.5%)
4 3 (42.9%)
1 0
135 23
Total
79
1
158
a
27
44
7
stage I1 versus stage IIIa: ,$ = 3.880, p < 0.05; Stage I versus stage 11%: ,$ = 4.570, p < 0.05; stage I + stage I1 versus stage IIIa + stage IIIb + stage N: ,$ = 4.764, p < 0.05.
0.005;
stage I1 versus stage IIIb: ,$
=
8.209, p
Cytology of intraoperative pleural lavage was examined in 164 lung cancer patients who underwent pulmonary resections. None of the patients had any pleural effusion or dissemination. Cytology was performed three times: (1)at thoracotomy, (2)immediately after resection, and (3) after washing the pleural cavity with 5,000 mL of physiological saline solution just before closure of the chest wall. Twenty-three patients (14%) had more than one positive cytological finding. The frequency of positive cytological findings was significantly correlated with pathological T classification, pleural status, and pathological stage. The positive lavage group had a significantly higher recurrence rate than the negative lavage
group in patients with stage I or stage I1 cancer. Four patients in the positive lavage group (17.4%) had recurrence in pleura or pericardium whereas only 1 patient in the negative lavage group (0.7%) had a recurrence in pericardium. The positive cytological finding of pleural lavage has more important meaning as a prognostic factor in stage I and stage I1 and indicates a greater possibility of recurrence in pleura or pericardium, but further examinations to evaluate the viability of detected malignant cells are required so that the positive cytological findings of pleural lavage can be regarded as subclinical pleural dissemination. (Ann Thorac Surg 2991;51:599-604)
S
pleural lavage cytology to confirm subclinical pleural dissemination and its utility as a prognostic factor.
pjut and associates [I] reported the cytological findings of pleural cavity washings in lung cancer patients and referred to the possibility of implantation of malignant cells by incisional biopsy. Eagan and associates [2] studied the cytological findings from pleural lavage after curative pulmonary resection in 135 lung cancer patients and reported the correlation of positive cytological findings with cell type, lymph node status, visceral pleural status, and pathologcal stage. Kondo and associates (31 reported that the positive cytological findings were significantly correlated with pleural status and lymphatic permeation but not with histology, stage, or lymph node involvement in pleural lavage just after thoracotomy. Buhr and associates [4] reported the results of pleural lavage cytology before and after pulmonary resections and the poor prognosis of patients with positive cytological findings, and insisted that positive cytological findings of pleural lavage should be added to the criteria for the staging system. Results from previous reports differ, and the importance of pleural lavage cytology as a prognostic factor or an indicator of subclinical pleural dissemination in lung cancer patients has not been understood completely. In this report, we also present the results of cytological findings obtained from pleural lavage fluid and try to clarify the relationship of positive cytological findings with several pathological factors and results of preoperative transthoracic needle biopsy. The other aim of this study is to refer to the possibility of Accepted for publication Nov 29, 1990. Address reprint requests to Dr Okumura, First Department of Surgery, Osaka University Medical School, 1-50 1-chome Fukushima Fukushima-ku Osaka-City, 553, Japan. 0 1991by The Society of Thoracic Surgeons
Material and Methods The study group consists of 164 lung cancer patients who had intraoperative pleural lavage cytology examined in our institute between September 1985 and December 1988. None of the study group had any pleural effusion or pleural dissemination. All patients of the study group had three intraoperative pleural lavages: at thoracotomy, immediately after lobectomy or pneumonectomy with extensive sampling of mediastinal lymph nodes, and after routine washing of the pleural cavity with 5,000 mL of physiological saline solution just before closure of the chest wall. At each lavage, the pleural cavity was filled with 500 mL of physiological saline solution and all the saline solution was aspirated into a glass bottle with heparin. About 400 mL of the supernatant of the lavage fluid was discarded after several hours’ settlement. The rest of the saline solution, including the precipitates, was shaken and centrifuged at 1,500 rpm for 5 minutes. The sedimented material was stained by the Papanicolaou method. One hundred ninety-eight patients with lung cancer underwent pulmonary resection during the period of this study in our institute, but 34 patients were excluded from the study group: 1 had apparent pleural dissemination at thoracotomy, 14 had pleural effusion at thoracotomy, and 19 lacked more than one lavage cytology specimen. Class 1 and class 2 were regarded as the negative cytological finding, class 4 and class 5 as the positive cytological finding. The patients whose three lavage specOOO3-4975/91/$3.50
600
OKUMURA ET AL PLEURAL LAVAGE CYTOLOGY
Ann Thorac Surg 1991;51:599-604
Table 1. Patients With Positive Cytological Findings Patient No. 1
2 3 4 5 6 7 8 9 10 11
Cytology 1 -
+ -
+ f
+ + + + +
2
3
Histology
TIVM
P
Stage
+ + +
-
Sq sq Sq sq sq sq Ad Ad Ad Ad Ad
T1 N O MO T2 N O MO T2 Ell MO T3 N O MO T3 hll MO T4 N2 MO T2 h13 MO T2 N O MO T1 NO MO T2 h0 MO T2 hO MO T1 hO MO
PO PO PO P3 P3 P3 PO P2 PO P1 P2 P1
I I I1 IIIa IIIa IIIb IIIb I I I
Ad Ad Ad Ad Ad Ad Ad Ad Ad Ad La
T2 NO MO T2 NO MO T2 NO MO T2 N2 MO T2 N2 MO T2 N2 MO T2 N2 MO T3 N2 MO T3 N2 MO T3 N3 MO T3 NO MO
PO P2 P2 PO PO P2 P2 P3 P3 P3 P3
I I I IIla IIIa IIIa IIIa IIIa IIIa
-
+ + -
+ -
+ -
-
12
-
+
13 14 15 16 17 18 19 20 21
+ + + + +
-
+
+
+
-
-
+
-
22
-
23
-
-
+ + + + +
f
-
+ -
+ -
+
-
sq
1
I
11%
IIIa
Follow-up (mo)
Site of Recurrence
35, NED 27, DAD 8, DT 23, DT 39, DT 24, DT 17, DT 22, DT 32, NED 25, NED 30, alive 30, CEA elevation 35, NED 7, DT 14, DT 25, NED 31, alive 13, DT 10, DT 14, DT 18, DT 13, DT 17, NED
None None Pericardium Bone Lung Medias tinum Lung Brain, lung None None Lung
... None Pleura Pleura None Lung, bone SCLN Liver Brain, lung Lung Pericardium None
Ad = adenocarcinoma; CEA = cardnoembryonic antigen; DAD = died of another disease; La = large cell DT = died of tumor; carcinoma; NED = no evidence of disease; P = pleural status; SCLN = supraclavicular lymph node; Sq = squamous cell carcinoma.
imens were all negative were defined as the negative lavage group, and the patients who had more than one specimen with a positive cytological finding were defined as the positive lavage group. Differences in histology, pathological TNM classification, pathological stage, pIeural status, and presence or absence of diagnostic preoperative transthoracic needle aspiration biopsy were examined between these two groups. TNM classification and stages were determined according to the Staging System
of the Union Internationale Contre le Cancer [5]. The general rules for clinical and pathological staging of lung cancer by the Japanese Lung Cancer Society [6] were adopted to divide pN1 into two subgroups: pNla, which has involvement in the intralobar lymph nodes, and pNlb, which has involvement in the hilar lymph nodes, and also to classify pleural status into four groups as follows: PO, visceral pleura is not invaded by tumor; P1, tumor has reached but not invaded through the visceral
Table 2. Patients With Cytological Findings of Class 3 Patient No.
Cytology
1
2
3
1 2 3 4
f
-
-
-
f
t
f f
-
5
-
6
f
-
f
-
Histology
TNM
P
Stage
sq Sq sq
T2 NO MO T2 NC MO T2 N2 MO T2 NO MO T2 N2 MO T1 N1 MO
P1 P1 P1
I I IIIa
PO
I IIIa
Ad Ad La
DT = died of tumor; Ad = adenocarcinoma; DAD = died of another disease; disease; I' = pleural status; Sq = squamous cell carcinoma.
PO P1
I1
Follow-up (mo)
Site of Recurrence
27, DT 28, DT 12, DT 15, NED 25, NED 4, DAD
Lung Bone Lung None None None
La = large cell carcinoma;
NED = no evidence of
601
OKUMURA ET AL PLEURAL LAVAGE CYTOLOGY
Ann Thorac Surg 1991;51:599-604
Table 3. Histology and Pleural Lavage Cytology
Cytology
Ad
sq
La
Sm
Ca
Total
Negative Positive
63 15 (19.2%)
61 7 (10.3%)
7 1(12.5%)
2 0
2 0
135 23
Total
78
68
8
2
2
158
Ad = adenocarcinoma;
Ca = carcinoid;
La
=
large cell carcinoma;
pleura; P2, tumor has invaded the visceral pleura but not involved the parietal pleura; and P3, tumor has invaded the parietal pleura and chest wall. All patients of the study group had an exact diagnosis preoperatively by transbronchial or transthoracic biopsy, so none of them underwent diagnostic intraoperative tumor biopsy. The last follow-up was done in April 1990, and sites of recurrence were confirmed in all patients with recurrence except 1 whose site has not been found yet despite rapid elevation in the level of carcinoembryonicantigen. The x 2 test was used for check for statistical significance and the method of Kaplan-Meier was used to calculate survival rates. The statistical significance of differences between the survival curves was examined by the generalized Wilcoxon test.
Results The positive lavage group consisted of 23 patients, 14%of the study group. Patients of the positive lavage group are listed in Table 1. The negative lavage group consisted of 135 patients. There were 6 patients whose worst cytological findings were class 3. These 6 patients are listed in Table 2, but they were not included in either the positive or the negative lavage group.
Histology Frequencies of positive cytological findings on pleural lavage in patients with adenocarcinoma, squamous cell carcinoma, and large cell carcinoma are presented in Table 3. Positive cytological findings were most common in adenocarcinoma, but there was no significant difference between cell types.
Pathological T Classification Patients with more advanced T classification had a higher frequency of positive cytological findings. There was a significant relationship between cytological findings and T classification (Table 4).
Sm = small cell carcinoma;
Sq = squamous cell carcinoma.
Pathological N Classification Frequencies of positive cytological findings in each N classification are presented in Table 5. No significant differences were seen except between pN1 and pN3. Thirteen of 103 patients (12.6%)with pNO or pNla whose nodal involvement did not extend to the hilar lymph nodes had positive cytological findings, and 10 of 55 patients (18.2%) with pNlb, pN2, or pN3 whose nodal involvement extended to hilar or mediastinal lymph nodes had positive cytological findings, but there was no significant difference.
Pleural Status Frequencies of positive cytological findings in patients with PO and P1 were almost the same value, that with P2 was the second highest, and that with P3 was the highest. There was a significant relationship between frequency of positive cytological findings and pleural status (Table 6).
Pathological Stage Frequencies of positive cytological findings in stage I, stage 11, stage IIIa, and stage IIIb were 12.7%, 3.7%, 20.5%, and 42.9%, respectively. Positive cytological findings were significantly more frequent in patients with stage IIIa, stage IIIb, or stage IV (23.1%)than in those with stage I or stage II (10.4%)(Table 7).
Preoperative Transthoracic Needle Aspira tion Biopsy Preoperative transthoracic needle aspiration biopsy was performed in 39 patients. Positive cytological findings were obtained in 8 patients with it (20.5%) and in 15 patients without it (12.6%). Presence or absence of preoperative transthoracic needle aspiration biopsy had no significant relationship with the frequency of positive cytological findings.
Prognosis The survival rates for 1 year, 2 years, and 3 years were 88.9%, 78.3%, and 69.2% in the negative lavage group
Table 4. T Classification and Pleural Lavage Cytology" Cytolow
T1
T2
T3
T4
Total
Negative Positive
51 3 (5.6%)
75 13 (14.8%)
8 6 (42.9%)
1
1(50.0%)
135 23
Total
54
88
14
2
158
TI versus T3 ,$ = 13.471,p < 0.005; 12.198,p < 0.005.
a
TI versus T4: ,$ = 5.744,p < 0.01;
T2 versus T3:,$ = 6.285,p < 0.01;
TI
+ T2 versus T3 + T 4 ,$ =
602
OKUMURA ET AL PLEURAL LAVAGE CYTOLOGY
Table 5 .
Ann Thorac Surg 1991;51:599-W4
N Classification and Pleural Lavage Cytology"
Cytology
NO
N1
N2
N3
Total
Negative Positive
74 12 (14.0%)
29 (Nla, 16; Nlb, 13) 2 (6.5%)(Nla, 1; Nlb, 1)
29 7 (19.4%)
3 2 (40.0%)
135 23
Total
86
31
36
5
158
a
N1 versus N 3 ,$ = 4.906, p < 0.05;
NO
+ Nla versus N l b + N2 + N3: ,$ = 0.891, not significant.
Table 6 . Pleural Status and Pleural Lavage Cytology"
Cytology
Po
P1
P2
P3
Total
Negative Positive
86 8 (8.5%)
25 2 (7.4%)
15 6 (28.6%)
9 7 (43.8%)
135 23
Total
94
27
21
16
158
PO versus P 2 ,$ = 6.461, p < 0.025; 16.449, p < 0.005.
a
PO versus P3: 2 = 14.417, p < 0.005;
and 87.0%, 50.2%, and 38.8% in the positive lavage group, respectively. The survival rate for the negative lavage group was significantly better than that of the positive lavage group (Fig 1).Forty-seven patients in the negative lavage group (34.8%)have had recurrence. On the other hand, 16 patients in the positive lavage group (69.6%)have had recurrence. The recurrence rate for the positive lavage group is significantly higher than that for the negative lavage group, but the fact that the proportion of higher stages is greater in the positive lavage group than in the negative lavage group should not be neglected; therefore, differences in recurrence rate between the two groups must be discussed in each stage. In stage I, stage 11, and stage IIIa, the recurrence rate was higher in the positive lavage group; in stage IIIb they were the same (Table 8). In patients with stage I or stage I1 disease, the recurrence rate for the positive lavage group (54.5%)was significantly higher than that for the negative lavage group (20%). In patients with stage IIIa, stage 11% or stage IV disease, there was no significant difference between the results of lavage cytology and the recurrence rate (Table 9). Sites of recurrence were classified as carcinomatous pleuritis, carcinomatous pericarditis, mediastinal or cervical lymph nodes, and distant metastases including lungs. There were 2 patients with carcinomatous pleuritis and 2 patients with carcinomatous pericarditis in the positive lavage group, though there was no patient with carcino-
PI versus P3: ,$ = 8.018, p < 0.005;
PO + PI versus P2 + P3: ,$ =
matous pleuritis and only 1 patient with carcinomatous pericarditis in the negative lavage group (Table 10). Of 6 patients whose worst cytologic findings were class 3, there were 3 patients (50%)who had recurrence. Their sites of recurrence, however, were not pleura or pericardium.
Comment We obtained the cytological findings from pleural lavage fluid at three times intraoperatively. Thirteen patients had positive cytological findings in the first lavage, 12 patients in the second lavage, and 6 patients in the third lavage. As the number of patients with positive cytological findings in each lavage is small, we considered it less meaningful to discuss the results of each pleural lavage cytology; therefore, patients who had at least one positive cytological finding were grouped together into the positive lavage group. Twenty-three of 164 patients in the study group (14.0%) had positive cytological findings, the incidence of which had a significant relationship with pathological T classification, pleural status, and pathological stage. There was no significant relationship of positive cytological findings with lymph node involvement or histology in this study group. Previous investigators did not comment on the effect of preoperative transthoracic needle aspiration biopsy. Its presence or absence has no significant relation-
Table 7 . Pathological Stage and Pleural Lavage Cytology"
Cytology
Stage I
Stage I1
Stage IIIa
Stage IIIb
Stage IV
Total
Negative Positive
69 10 (12.7%)
26 1(3.7%)
35 9 (20.5%)
4 3 (42.9%)
1 0
135 23
Total
79
1
158
a
27
44
7
stage I1 versus stage IIIa: ,$ = 3.880, p < 0.05; Stage I versus stage 11%: ,$ = 4.570, p < 0.05; stage I + stage I1 versus stage IIIa + stage IIIb + stage N: ,$ = 4.764, p < 0.05.
0.005;
stage I1 versus stage IIIb: ,$
=
8.209, p
