Lung Adenocarcinomas Diagnosed by

Open Lung or Thoracoscopic vs

Bronchoscopic Biopsy*
Hiroaki Nomori, MD; Hirotoshi Horio , MD; Gentarou Fuyuno, MD;
Ryuichirou Kobayashi, MD; Shojiroh Morinaga, MD; and
Keiichi Suemasu, MD, FCCP

Study objectives: To examine the characteristics of peripheral lung adenocarcinomas diagnosed

by open lung or video-assisted thoracoscopic surgery (VATS) biopsy.
Design: We used retrospective analysis to compare tumor stage, pleural involvement, central
tumor fibrosis, and the number of bronchi or vessels involved with tumors of small peripheral
lung adenocarcinomas diagnosed by bronchoscopic biopsy.
Patients: Subjects had lung adenocarcinomas diagnosed by open lung or VATS (n=22) and those
diagnosed by bronchoscopic biopsy (n=22), which were matched by tumor size.
Results: The TINOMO tumor was notably more frequent in the open lung or VATS group (77.3%)
than in the bronchoscopic biopsy group (36.4%) (p<O.Ol). Tumors invading beyond the pleural
surface were less frequent in the open lung or VATS group (4.5%) than in the bronchoscopic
biopsy group (40.9%) (p<O.Ol). The grade of in-tumor central fibrosis-a malignancy factor and
the cause of bronchi or vessel involvement with tumors-was significantly lower in the open lung
or VATS group than in the bronchoscopic biopsy group (p<O.Ol). The number of bronchi or
vessels involved with tumors was significantly fewer in the open lung or VATS group than in the
bronchoscopic biopsy group (p<O.OOI).
Conclusions: (I) Lung adenocarcinomas diagnosed by open lung or VATS biopsy were more
frequently TINOMO than those diagnosed by bronchoscopic biopsy, which was caused by low
grade of central tumor fibrosis rather than small tumor size. (2) They were infrequently
diagnosed by hronchoscopic biopsy because few bronchi were involved by tumors due to the low
grade of central tumor fibrosis. (3) Small pulmonary nodules not diagnosable by hronchoscopic
biopsy should he diagnose d aggressively using VATS biopsy to detect early-stage lung cancer.
(CHEST 1998; 114:40-44)

Key words: adenocarcinoma; biopsy; bronchoscopy; non -small cell lung cancer; video-assisted thoracoscopic surgery

Abbreviations: PNAC= percutaneous needle aspiration cytology; VATS = video-assisted thoracoscopic surgery

Q pen lung biopsy was conventionally conducted
to diagnose small peripheral pulmonary nodules
but has now been largely r eplaced by video-assisted
thoracoscopic surgery (VATS ) biopsies,1 -3 which
have the advantages of 100% diagnostic accuracy and
less invasiveness than open lung biopsy. VATS biopsy
diagnoses smaller pulmonary nodules and thus help
detect early-stage lung cancer more frequently th an
bronchoscopic biopsy, but we wondered whether
*From th e Departments of Thoracic Surge1y (Drs. Nomori,
Horio, and Suemasu), Medicine (Drs . Fuyuno and Kobayashi),
and Pathology (Dr. Morinaga), Saiseikai Central Hospital,
Tokyo.
Manuscript received July 3 1, 1997; revision accepted D ecembe r
9, 1997.
Reprint ·requests: Hiroaki Nomori, MD, Dept of Surge11j, Saisei-
kai Central Hospital, 1-4-17 Mita, Mitato-ku, Tokyo 108-0073,
Japan
small lung cancers diagnosed by VATS was a factor
in early diagnosis. To clarify this, we compared
tumor stage, pleural involvement, grade of central
fibrosis , and number of bronchi or vessels involved
with tumors between peripheral lung adenocarcino-
mas diagnosed by open lung or VATS biopsy and
those diagnosed by bronchoscopic biopsy matched
for tumor size.
MATERIALS AND METHODS
From May 1988 to April 1997, 92 patients underwent open
lung ( n =22) or VATS biopsy (n = 70) for peripheral pulmonary
nodules at Saiseikai Central Hospital because bronchoscopic
biopsy could obtain non e of th e diagnosti c materials. Conditions
included primary lung cancer in 35 pati ents, metastatic lu ng
cance r in 13, inflammatory nodules in 30, benign lung tumor in
7, and miscellaneous in 7. Of th e 3 5primary lung cancer subjects,

40 Clinical Investigations
31 had adenocarcinoma. To determine the adenocarcinoma
stage, we selected 22 patients undergoing lobectomy and medi-
astinal lymph node dissection, and excluded 9 patients undergo-
ing limited surgery. The subjects selected underwent presurgical
bronchoscopic biopsy one to three times (mean, 2.2), but no
tumors were indicated. Three of these patients also underwent
presurgical CT-guided percutaneo us needl e aspiration biopsy,
but no tumors were indicated. Cases diagnosed using percutane-
ous needle aspiration biopsy were excluded from th e present
study.
Control subjects were 22 patients with small peripheral lung
adenocarcinoma diagnosed by bronchoscopic biopsy during the
same period (from 1988 to 1997), which were selected from the
smallest lesion. The bronchoscopic biopsy procedure included
forceps biopsy, brushing, and bronchial washing. Tumor size was
classified as the maximum dimension on thin-section CT (5 mm
per section).
Disease stage was based on the UICC TNM classification.4
Tumors exposed on the pleural surface were categorized as T2.
Carcinomas with minute, satellite nodules found incidentally
within the same lobe of the resected specimen were categorized
as Ml.
Pleural involvement was cbssified into four grades based on
the Japan Lung Cancer Society classification:5 (1) pO: tumor with
no pleural involvement or reaching th e visceral pleural but not
extending beyond the elastic layer; (2) p1: tumor extending
beyond the e lastic and the visceral pleural layer but not to the
pleural surface; (3) p2: tumor exposed on the pleural surface but
not involving parietal pleura; and (4) p3: tumor involving parietal
pleura or organs adjacent to the lung.
Central fibrosis within tumors was classified into four grades:"-9
(1) grade 1: no or minimal desmoplasia; (2) grade 2: fibroblastic
tissue with a small amount of collagen; (3) grade 3: fibroblastic
tissue with either amode rate or large amount of collagen; and (4)
grade 4: fibroblastic tissue with hyalinization.
The number of bronchi or vessels involved with tumors was
counted based on thin-section CT (5 mm per section; window
level: -500; \vindow width: 1,500).
Tumor size, tumor stage, pleural involvement, central tumor
fibrosis, and number of bronchi or vessels involved with tumors
were compared between those diagnosed b y open lung or VATS
biopsy and those diagnosed by bronchoscopic biopsy.
Statistical Analysis
All data were analyzed for significance using the two-tailed
Student's t test. Values of p< 0.05 were accepted as significant.
All values in text and tables are given as means::'::SD.
RESULTS
Both open lung and VATS biopsy localized all
pulmonary nodules and enabled them to be resected
Table 1-TNM of Lung Adenocarcinomas Diagnosed by Open Lung or VATS vs Bronchoscopic Biopsy
Tli\'OMO Tl-2NlMO
No. of cases
Open lung or VATS biopsy 17 3 l
Bronchoscopic biopsy 8 1
Total 25 4 7
without postoperative complication. The mean age
of patients whose conditions were diagnosed by open
lung or VATS biopsy (14 male, 8 female) was 61:±:9
years, and that of those whose conditions were
diagnosed by bronchoscopic biopsy (13 male, 9
female) was 59:±:12 years. The tumor diameter diag-
nosed by open lung or VATS biopsy ranged from 7 to
30 mm (mean, 17.7:±: 6.9 mm ), while that diagnosed
by bronchoscopic biopsy ranged from 14 to 23 mm
(mean, 19.2:±:2.7 mm ). There was no significant
difference in mean tumor diameter between the two
groups.
TNM classification (Table 1) in the open lung or
VATS group was NO (T1NOMO) in 17 cases, N1 in 3
(T1N1MO in 2 and T2N1MO in 1), and N2
(T2N2MO) and M1 in 1 each; in the bronchoscopic
group , it was NO (T1NOMO) in 8 cases, N1
(T2N1MO) in 1, N2 in 6 (T1N2MO in 5 and T2N2MO
in 1), T4 in 3, and M1 in 4. All T2 cases were tumors
exposed on the pleural surface and <30 mm in
diameter. The percentage of T1NOMO cases diag-
nosed by the open lung or VATS group (77.3%) was
thus significantly higher than that in the broncho-
scopic group (36.4%) (p=0.006). The percentage of
T4 or M1 in the open lung or VATS biopsy group
(4.5%) was significantly lower than that in the bron-
choscopic group (31.8%) (p=0.02).
The grade of pleural involvement (Table 2) in the
open lung or VATS group was pO in 15 cases, p1 in
6, and p2 in 1; in the bronchoscopic group, it was pO
in 9 cases, p1 in 4, and p2 in 9. The percentage of p2
in the open lung or VATS group (4.5%) was signifi-
cantly lower than that in the bronchoscopic group
(40.9%) (p=0.004).
The grade of central fibrosis (Table 3) in the open
lung or VATS group was grade 1 in 7 cases, grade 2
in 10, grade 3 in 4, and grade 4 in 1; in the
bronchoscopic group, grade 1 was seen in 1 case,
grade 2 in 10, and grade 3 in 11. The percentage of
grade 1 fibrosis in the open lung or VATS group
(31.8%) was thus significantly higher than that in the
bronchoscopic group (4.5%) (p= 0.02).
Tumors diagnosed by open lung or VATS biopsy
often showed fewer bronchi or vessels involved with
tumors than those diagnosed by bronchoscopic bi-
Ti\'M
Tl-2N2MO T4 :tv!l Total
0 22
6 3 4 22
3 5 44
CHEST I 114 I 1 I JULY, 1998 41
 Table 2-Grade of Pleural Involvement of Lung
Adenocarcinoma Diagnosed by Open Lung or VATS vs
Bronchoscopic Biopsy
Grade of Pleural
Involvement

pO p1 p2 Total
No. of cases
Open lung or VATS biopsy 15 6 22
Bronchoscopic biopsy 9 4 9 22
Total 24 10 10 44

opsy specimens (Figs 1 and 2). Table 4 showed the

number of bronchi or vessels involved with tumors in
relation to the central fibrosis grade in the two
groups. The number of bronchi or vessels involved in
the open lung or VATS group (range, 2 to 5; mean,
3.1 ± 1.1) was significantly lower than in the broncho-
scopic group (range, 2 to 6; mean, 4.6±l.2)(p=4.3X
10- .s). The mean number of bronchi or vessels in
both groups was 2.9 in grade 1, 3.5 in grade 2, 4.8 in
grade 3, and 5.0 in grade 4. The higher the central
fibrosis grade, the more bronchi or vessels were
involved with tumors. The number for grade 1
(mean, 2.9±1.0) was significantly lower than in
grades 2 to 4 (mean, 4.1±1.3) (p=0.012). The
number in grade 1 or 2 (mean, 3.3±1.2) was signif- FIGURE l. Lung adenocarcinoma diagnosed by VATS biopsy.
icantly lower than in grade 3 or 4 (mean, 4.8±1.0) Top: findings show few bronchi involved with tumor (arrow).
(p= 0.00013 ). Bottom: histologic findings show minimal desmoplasia in the
cen tral portion of the tumor (grade 1 fibrosis) (hematoxylin-
Table 5 shows the correlation between the grade eosin, original magnification X75). The grade of pleural involve-
of central fibrosis and TNM in both groups. While all ment is pO, and the tumor was TlNOMO.
8 (100%) tumors with grade 1 were TlNOMO, 17 of
36 (47.2%) tumors with grades 2 to 4 were T1NOMO.
Grade 1 tumors were ·significantly more frequently grade of central fibrosis , the more frequently the
T1NOMO than grades 2 to 4 (p=0.006). Table 6 tumor involved the pleura.
shows the correlation between the grades of central
fibrosis and pleural involvement in both groups.
While all8 (100%) tumors with grade 1 were pO, 16 DISCUSSION
of 36 (44.4%) tumors with grades 2 to 4 were pO (p
=0.004) . While 19 of 28 (67.9%) tumors with grade Open lung or VATS biopsy generally diagnoses
1 or 2 were pO, 5 of 16 (31.3%) tumors with grades smaller peripheral lung cancers than bronchoscopic
3 or 4 were pO (p = 0.02 ). Namely, the higher the biopsy. We therefore selected small peripheral lung
adenocarcinomas diagnosed by bronchoscopic bi-
opsy as the control. Results showed no significant
Table 3- Grade of Central Fibrosis of Lung difference in tumor size between the two groups.
Adenocarcinoma Diagnosed by Open Lung or VATS vs Our results showed that lung adenocarcinoma diag-
Bronchoscopic Biopsy nosed by open lung or VATS biopsy had significantly
higher T1NOMO frequency and less pleural involve-
Grade of Central
Fibrosis
ment than that diagnosed by bronchoscopic biopsy,
regardless of tumor size.
2 3 4 Total Shimosato et al6 and others 7 -LO reported that cen-
No. of cases tral fibrosis within lung adenocarcinoma can occur
Open lung or VATS biopsy 7 10 4 1 22 during tumor growth, and that more advanced cen-
Bronchoscopic biopsy 1 10 11 0 22 tral fibrosis is associated with increased involvement
Total 8 20 15 44
of surrounding bronchi or vessels, higher grade of

42 Clinical Investigations
 Table 5- Correlation Between Grades of Central
Fibrosis and TNM in All Cases

G rade of TNM
C e ntral
Fi b rosis Tl NOMO Tl -2N 1MO Tl-2N2MO T4 M1 Total

No. of cases
1 8 0 0 0 0 8
2 lO 3 3 1 3 20
3 7 0 4 2 2 15
4 0 1 0 0 1 1
Total 25 4 7 3 0 44

central fibrosis and involved f ewerbronchi or vessels

FIGURE 2. Lung a d e nocarc ino ma_ diagnosed by bronchoscopic
bi opsy. Top: C T findings show a few bron chi involve d with th e
tumo r with conve rge nce (arrow). Bott om: Histo logic findin gs
show a moderate collagen in th e central tu mor (grad e 3 fi bros is)
(he m a toxylin-eosin, original magnification X 75). The g rad e o f
ple ural involvem e n t is p2, and the tu mo r was T 2NOM l.
pleural involvement, higher incidence of lymph node
metastasis, and poorer p rognosis. In addition, they
reported that central fibrosis also plays a ma j or role
in involvement of many bronchi or vessels with
tumors .6 Our results also showed similar findings, ie,
the higher the grade of central fibrosis , the greater
number of bronchi or vessels were involved with
tumors, the higher the g rade of pleural involvement,
and the more frequent the tumor advancement. We
also found that tumors diagnosed b y open lung o r
VATS biopsy had a significantly lower grade o f
than those diagnosed by ·bronchoscopic biopsy. We
therefore conclude that lung a denocarcinoma diag-
nosed b yopen lung o r VATS biopsy would probably
be difficult by bronchoscopic procedure alone be-
cause too few bronchi are involved with tumors to be
diagnosed b y bronchoscopic procedures, caused b y
the lower grade of central fibrosis and resulting in
greater d etection of early-stage disease byopen lung
or VATS biopsy.
Lung adenocarcinomas we diagnosed b ybroncho-
scopic biopsy were somewhat frequently advanced
despite small tumor size. We thus assume that small
peripheral lung adenocarcinoma diagnosable by
bronchoscopic biopsy often shows a high grade of
central fibrosis, resulting in many bronchi being
involved with tumors facilitating bronchoscopic di-
agnosis, a high degree of malignancy, and thus
advanced tumor stage.
Percutaneous needle aspiration cytology (PNAC)
is also fefective in the diagnosis of pulmonary nod-
ules not diagnosable by bronchoscopic biopsy.I 1- 13
PNAC takes a transthoracic approach, which is sim-
ilar t o VATS biopsy, so weexcluded lung cancers
diagnosed b y PNAC from the present study. PNAC
is repmted to have a false-negative rate of 3 to
11%, 11 however, Layfield et aP 2 reported that its
diagnostic accuracy decreased to 60% for lesions :::; 1
em. Kashiwabara et all 3 reported that th e positive
diagnostic rate for nonmalignant lesions by PNAC
was only 56%. During the same period as the present
study, 119 patients with pulmonary nodules at our

Table 4-Number of Bronchi or Vessels Involved With Tumors and Grade of Central Fibrosis *
Grade o f Cent ral Fi b rosis

2 3 4 Total

No. of bronchi or vessels

Ope n lung or VATS biopsy 2.7 ::+:: 1.0 2.9 ::':: 0.9 4. 0 ::+:: 0.8 5.0 ::+:: 0 3. 1::':: l.l
Bronchoscopic biopsy 4.0 ::':: 0 3.6 ::':: 1.0 5. 1 ::':: 1.0 4.6::':: 1.2
Total 2.9 ::':: 1.0 3.5 ::':: 1.3 4.8 ::':: 1.1 5.0 ::':: 0 3.9::':: 1.4
*Values are mean::+:S D .

CHEST / 114 / 1 / JULY, 1998 43

 Table 6-Correlation Between the Grades of Central
Fibrosis and Pleural Involvement in all Cases
Grade of Pleural
Involvement
Grade of
Central Fibrosis pO p1 p2 Total
No. of cases
1 8 0 0 8
2 11 4 5 20
3 5 5 .5 15
4 0 0 1
Total 24 10 10 44
hospital underwent PNAC. Of these, 61 (51%) were
diagnosed as having lung carcinoma, and 25 among
them underwent surgery. Eight patients (6.7%)
could not have their conditions diagnosed by PNAC,
but were diagnosed as having lung carcinoma
through open lung or VATS biopsy. The mean tumor
size was 18.0±10.1 mm in these 8 cases, which was
significantly smaller than the 30.0± 14.8 mm in the
25 cases diagnosable by PNAC (p< 0.01 ). Tumors
< 20 mm were found in 3 cases (14%) of the 25 cases
and 5 cases (63%) of the 8 cases, which was a
significant difference (p<O.Ol). We therefore con-
clude that pe1ipheral pulmonary nodules that are
< 20 mm will be difficult to diagnose using PNAC.
In Japan , bronchoscopic biopsy conventionally
costs 29,500¥ ($227), CT-guided PNAC costs
30,400¥ ($234), and VATS biopsy costs 323,000¥
($2,485). These prices include coverage by national
insurance plans. Bronchoscopic biopsy has low mor-
bidity, PNAC involves the risk of pneumothorax, and
VATS biopsy involves morbidity related to anesthesia
and pulmonary wedge resection. For the reasons of
cost, the morbidity associated with each procedure,
and false-negative diagnosis in bronchoscopic and
PNAC procedures, we recommend the following
strategy in diagnosing small pulmonmy nodules: (1)
bronchoscopy should be conducted first for both
biopsy and examination of the central tracheobron-
chial lumen; (2) CT-guided PNAC may be valuable
for tumors > 20 mm in diameter and located where
needle aspiration is easy; and (3) VATS biopsy should
be promptly performed if the first bronchoscopic
biopsy or PNAC attempt fails to diagnose nodules.
In conclusion, VATS biopsy diagnoses peripheral
44
lung cancer at an earlier tumor stage than broncho-
scopic biopsy, which is caused by low grade of
central tumor fibrosis rather than small tumor size.
Small pulmonary nodules virtually not diagnosable
by bronchoscopic biopsy or PNAC should be diag-
nosed aggressively using VATS biopsy to detect
early-stage lung cancer.
REFERENCES
1 Mack MJ, Gordon MJ, Postma TW, et al. Percutaneous
localization of pulmonary nodules for thoracoscopic lung
resection. Ann Thorac Surg 1992; 53:1123-24
2 Nomori H, Horio H. Colored collagen is a long-lasting point
marker for small pulmonary nodules in thoracoscopic opera-
tions. Ann Thorac Surg 1996; 61:1070-73
3 Nomori H, Horio H. Endofinger for tactile localization of
pulmonary nodules during thoracoscopic resection. Thorac
Cardiovasc Surg 1996; 44:50-.53
4 Hermanek P, Sobin LH, eels. UICC TNM classiflcation on
malignant tumors. 4th eel, 2nd rev. Be rlin: Springer, 1992
5 The Japan Lung Cancer Society. General rule for clinical and
pathological record of lung cancer. 3rd eel. Tokyo: Kanehara,
1987
6 Shimosato Y, Hashimoto T, Kodama T, e t al. Prognostic
implications of fibrotic focus (scar) in small peripheral lung
cancers. Am J Surg Pathol 1980; 4:365-73
7 Takise A, Kodama T, Shirnosato Y, et a!. Histopathologic
prognostic factors in adenocarcinomas of the peripheral lung
less than 2 em in diameter. Cancer 1988; 61:2083-88
8 Noguchi M, Morikawa A, Kawasaki M, et la. Small adenocar-
cinoma of the lung. Cancer 1995; 7.5:2844-52
9 Nomori H, Hirohashi S, Noguchi M, et al. Tumor cell
heterogeneity and subpopulations with metastatic ability in
differentiated adenocarcinoma of the lung. Chest 1991; 99:
934-40
10 Nomori H, Matsuno Y, Noguchi M, et al. Adenocarcinoma of
th e lung with selective metastasis to the lung: clinical, histo-
logic and DNA-cytofluorometric analyses. Jpn J Cancer Res
1992; 83:93-100
11 Sisle r GE. Malignant tumors of th e lung: role of video-
ass isted thoracic surgery. Chest Surg Clin North Am 1993;
3:307-17
12 Layfield LJ, Coogan A, Johnston W'vV, et a!. Transthoracic
fine needle aspiration biopsy: sensitivity in relation to guid-
ance technique and lesion size and location . Acta Cytol 1996;
40:687-90
13 Kashiwabara K, Nakamura H, Fukai Y, et al. Availability of
diagnosis by percutaneous needle aspiration cytology of the
lung in cases who showed a peripheral solitary tumorous
shadow on chest Xp and diagnostic rate of transbronchial
approach. Jpn J Thorac Dis 1993; 31:1426-31
Clinical Investigations