ORIGINAL ARTICLES: GENERAL THORACIC
B m
C m
A
H ki M
H , K
an
Sec , Un
B
the
inc
of
Health Organization (WHO) classification in 1999. The
clinical features of patients with BAC diagnosed accord-
ing to the recently revised WHO classification have not
yet been clarified. In this retrospective study, we inves-
tig
for
com
no
M
we
cm
Jap
we
no
tum
les
no
acc
R
BA
no
pa
Lymph node involvement was seen for 30 lesions (36%)
of adenocarcinoma other than BAC, but not for any BAC
lesions. The median duration of follow-up after surgery
was 5.1 years. There was no recurrence in the postoper-
ati
fre
su
C
de
WH
ev
pa
cu
pa
B
its
ma
been distinct from those of other adenocarcinomas [1–3].
Bronchioloalveolar carcinoma has shown aerogenous,
rat
sit
tha
aft
ap
16]
acc
(W
O
his
W
an adenocarcinoma with no evidence of stromal, vascu-
lar, or pleural invasion. An excellent prognosis would
justifiably be expected for BACs, which are noninvasive
noma other than BAC, as evaluated by a repeat pathology
review.
Acc
Ad
Fac
cho, Nakakoma-gun, Yamanashi 409–3898, Japan; e-mail: sakuraihm@
ybb.ne.jp.
© 2
Pu
G
EN
ER
A
L
T
H
O
R
A
C
IC
004 by The Society of Thoracic Surgeons 0003-4975/04/$30.00
blished by Elsevier Inc doi:10.1016/j.athoracsur.2004.05.017
her than lymphatic, spread [4–9], and radiologic diver-
y from focal to diffuse disease [10–12]. It was reported
t there was an increase in the incidence of BAC year
er year [13, 14]. In a recent series, BAC accounted for
proximately 20% of adenocarcinomas of the lung [15,
. However, the definition of BAC has been revised
ording to the most recent World Health Organization
HO) classification in 1999 [17].
adenocarcinomas. However, it is difficult to compare the
results across many previous studies because of the
changes in the WHO definition of BAC. Consequently,
there is little information available on the clinicopatho-
logic significance of BAC defined as a pathologic nonin-
vasive lesion.
Therefore, we retrospectively examined the pattern of
recurrence and survival outcome in patients with re-
sected BAC, which was defined as a noninvasive adeno-
carcinoma based on the revised WHO classification,
compared with the pattern in patients with adenocarci-
epted for publication May 4, 2004.
dress reprint requests to Dr Sakurai, Second Department of Surgery,
ulty of Medicine, University of Yamanashi, 1110, Shimokato, Tamaho-
ated the pattern of recurrence and survival outcome
patients with resected BAC by pathology review,
pared with those in patients who had adenocarci-
ma other than BAC.
ethods. From 1985 through 2002, 108 patients under-
nt surgical resection for pulmonary adenocarcinoma 3
or less in diameter at the University of Yamanashi,
an. All of the resected specimens of these 108 patients
re pathologically reviewed again to confirm the diag-
sis as BAC or adenocarcinoma other than BAC. The
or was defined as BAC when the adenocarcinoma
ion had a pure bronchioloalveolar growth pattern and
evidence of stromal, vascular, or pleural invasion
ording to the WHO classification (third edition).
ronchioloalveolar carcinoma (BAC) of the lung is a
subtype of adenocarcinoma that is characterized by
unique growth pattern along the alveolar wall. So far
ny of the clinical and radiologic features of BAC have
ve course in patients with BAC for a 5-year disease-
e survival rate of 100%, whereas the 5-year disease-free
rvival rate for other adenocarcinoma was 63.5%.
onclusions. The patients with resected BAC, which is
fined as a noninvasive adenocarcinoma by the revised
O classification, had an excellent prognosis. How-
er, these results may depend on a strictly accurate
thology diagnosis as BAC. Limited resection might be
rative in patients with focal BAC based on evidence of
thologic noninvasive features.
(Ann Thorac Surg 2004;78:1728–33)
© 2004 by The Society of Thoracic Surgeons
n the other hand, the most recent revision of the
tologic classification of lung and pleural tumors by
HO [17] has clearly limited the designation of BAC to
ronchioloalveolar Carcino
entimeters or Less in Dia
ssessment
iroyuki Sakurai, MD, Yoh Dobashi, MD, Ei
irochika Matsubara, MD, Shoji Suzuki, MD
d Masahiko Matsumoto, MD, PhD
ond Department of Surgery and First Department of Pathology
ackground. Bronchioloalveolar carcinoma (BAC) of
lung is a subtype of adenocarcinoma for which the
idence is actually rising, but the histologic definition
BAC has been recently changed by the revised World
a of the Lung 3
eter: A Prognostic
izutani, MD,
unio Takano, MD, Shunya Shindo, MD,
iversity of Yamanashi, Yamanashi, Japan
esults. Twenty-five patients (23%) had a diagnosis of
C, and 83 (77%) had a diagnosis of other adenocarci-
ma. There was a female predominance among both
tients with BAC and those with other adenocarcinoma.
M
Pa
Fo
200
lun
Su
cou
no
Am
ha
dia
pa
tio
Th
or
stu
ter
ad
his
rec
sta
the
ser
me
for
Op
the
Pa
Fo
sta
aci
cu
by
mi
(th
cin
wa
ha
ev
inv
wa
ve
Th
exp
dis
sli
og
car
his
tio
vid
ad
lar
su
an
na
ev
tum
lym
Ple
tum
tum
lar
pre
lum
A
sh
79
(33
op
7 p
tec
su
Su
Sta
A
Ta
cm
Ch
No
Ag
M
R
Sex
M
F
Op
L
S
Lym
M
H
Su
C
I
His
B
A
P
S
M
Pat
S
S
S
BA
niz
1729Ann Thorac Surg SAKURAI ET AL
2004;78:1728–33 BRONCHIOLOALVEOLAR CARCINOMA PROGNOSIS
G
EN
ER
A
L
T
H
O
R
A
C
IC
aterial and Methods
tients
r the 18-year period from January 1985 to December
2, there were 210 pulmonary resections for primary
g adenocarcinoma at the Department of Thoracic
rgery, University of Yamanashi, Japan. These ac-
nted for 61% of all resections for primary lung carci-
mas resected during the same period (344 resections).
ong the 210 patients with adenocarcinoma, 108 (51%)
d adenocarcinoma that measured 3.0 cm or less in
meter and was located in the periphery of the lung
renchyma, excluding patients with a history of resec-
n of other malignancy or preoperative treatment.
ese 108 patients with peripheral adenocarcinoma 3 cm
less in diameter were the focus of this retrospective
dy. Clinicopathologic records were reviewed to charac-
ize patients with BAC compared with patients who had
enocarcinoma other than BAC, and included age, sex,
tory, serum tumor marker level, mode of resection,
urrence, and survival. The surgical and postsurgical
ges were determined according to the TNM system of
UICC (Union Internationale Conte le Cancer) [18]. The
um tumor marker carcinoembryonic antigen (CEA) was
asured preoperatively in all patients. The normal range
the CEA level at our institute was 5.0 ng/mL or less.
erative death was defined as any death within 30 days of
operation or during hospitalization.
thology Examinations
rmaldehyde-fixed, paraffin-embedded blocks were
ined using hematoxylin-eosin, elastica, and periodic
d-Schiff, and the entire nodule, including the largest
t surface of the tumor, was retrospectively examined
light microscopy. The histologic subtype was deter-
ned based on the WHO classification revised in 1999
ird edition) as BAC, acinar, papillary, solid with mu-
, or adenocarcinoma with mixed subtypes. The tumor
s defined as BAC when the adenocarcinoma lesion
d a pure bronchioloalveolar growth pattern and no
idence of stromal, vascular, or pleural invasion. If
asive foci were present within the tumor, the tumor
s classified as adenocarcinoma mixed bronchioloal-
olar subtype (adenocarcinoma with mixed subtypes).
ese findings were independently investigated by two
erienced pulmonary pathologists (Y.D., H.S.), and
crepancies were resolved by joint examination of the
des under a two-headed microscope. The two pathol-
ists were masked with respect to outcomes when they
ried out the classification of adenocarcinoma into the
tologic subtype based on the revised WHO classifica-
n. For the purpose of comparison, patients were di-
ed into two groups: those with BAC and those with
enocarcinoma other than BAC, such as acinar, papil-
y, solid, or mixed subtypes. In addition, BAC was
bdivided into three groups: nonmucinous, mucinous,
d mixed mucinous and nonmucinous or indetermi-
te. The following histopathologic findings were also
aluated in the same slides: tumor size (maximum
or dimension), pleural involvement, vascular/ fre
phatic permeation, and lymph node involvement.
ural involvement was classified as positive when the
or was exposed on the pleural surface or when the
or invaded the parietal pleura or chest wall. Vascu-
/lymphatic permeation was evaluated according to the
sence of identifiable tumor cells in the blood vessel
en or lymphatic lumen, respectively.
clinicopathologic summary of all 108 patients is
own in Table 1. The patients ranged in age from 34 to
years with a mean age of 65.3 years. Thirty-six patients
%) were male and 72 (67%) were female. The mode of
eration was lobectomy in 101 patients (94%). The other
atients underwent limited resection such as segmen-
tomy or wedge resection because of being at poor risk
ch as impaired pulmonary or cardiovascular disease.
rgical curability was complete in all patients.
tistics
�2 test or Fisher’s exact test was used to compare the
ble 1. Characteristics of Patients With Adenocarcinoma 3
or Less in Diameter
aracteristic Data
. of patients 108
e (years)
ean 65.3
ange 34–79
ale 36 (33%)
emale 72 (67%)
erative mode
obectomy 101 (94%)
egmentectomy/partial 7 (6%)
ph node dissection
ediastinohilar 98 (91%)
ilar only/none 10 (9%)
rgical curability
omplete 108 (100%)
ncomplete 0 (0%)
tologic subtype by the WHO classification
AC 25 (23%)
cinar 9 (8%)
apillary 10 (9%)
olid with mucin 7 (7%)
ixed subtypes 57 (53%)
hological stage by TNM category
tage I Total � 77
T1N0M0 73 (67%)
T2N0M0 4 (4%)
tage II Total � 14
T1N1M0 13 (12%)
T2N1M0 1 (1%)
tage III Total � 17
T1N2M0 14 (13%)
T2N2M0 2 (2%)
T4N1M0 1 (1%)
C � bronchioloalveolar carcinoma; WHO � World Health Orga-
ation.
quencies between subgroups. A survival curve was
est
ini
rec
be
tes
con
Re
Cl
Tw
83
na
mi
cla
BA
sh
be
ag
na
CE
BA
oth
en
cin
ple
wh
me
Ly
in
gro
the
ad
the
No
Pr
Th
ye
ea
cin
5-y
an
100
BA
cou
ad
un
no
pa
en
cau
les
Ta
Cli
Ch
Ag
M
R
Sex
M
F
CE
�
�
Tu
M
R
Ple
N
P
Va
p
N
P
Lym
i
N
N
Pat
I
I
BA
oni
Ta
WH
Cy
No
Mu
Mi
BA
niz
Fig
1730 SAKURAI ET AL Ann Thorac Surg
BRONCHIOLOALVEOLAR CARCINOMA PROGNOSIS 2004;78:1728–33
G
EN
ER
A
L
T
H
O
R
A
C
IC
imated by the Kaplan-Meier method using the date of
tial resection as the starting point and the date of
urrence or last follow-up as the endpoint. Differences
tween survival curves were assessed by the log-rank
t. Deaths by causes other than lung cancer were
sidered censored.
sults
inicopathologic Features
enty-five patients (23%) had a diagnosis of BAC, and
(77%) had a diagnosis of other adenocarcinoma (aci-
r, n � 9; papillary, n � 10; solid with mucin, n � 7;
xed subtypes, n � 57), according to the 1999 WHO
ssification. The clinicopathologic characteristics of
C compared with those of other adenocarcinomas are
own in Table 2. There were no significant differences
tween BAC and other adenocarcinoma with regard to
e (p � 0.423) or sex (p � 0.519). There was a predomi-
nce of female patients in both groups. An elevated
A level was not observed in any of the patients with
C, although this was observed in 18% of patients with
ble 2. Relationship Between Histologic Subtype and
nicopathologic Factors in Pulmonary Adenocarcinoma
aracteristic
Adenocarcinoma
p Value
BAC
(n � 25)
Other
Adenocarcinoma
(n � 83)
e (years) 0.423
ean 64.0 65.7
ange 47–78 34–79
0.519
ale 7 (28%) 29 (35%)
emale 18 (72%) 54 (65%)
A (ng/mL) 0.022
5.0 25 (100%) 68 (82%)
5.0 0 (0%) 15 (18%)
mor size (cm) 0.086
ean 1.9 2.1
ange 0.5–3.0 0.9–3.0
ural involvement 0.001
egative 25 (100%) 56 (67%)
ositive 0 (0%) 27 (33%)
scular/lymphatic
ermeation
0.000
egative 25 (100%) 48 (58%)
ositive 0 (0%) 35 (42%)
ph node
nvolvement
0.000
0 25 (100%) 53 (64%)
1–3 0 (0%) 30 (36%)
hological stage 0.000
25 (100%) 53 (64%)
I–IV 0 (0%) 30 (36%)
C � bronchioloalveolar carcinoma; CEA � serum carcinoembry-
c antigen level.
er adenocarcinoma. There were no significant differ-
nom
rat
ces in tumor size between BACs and other adenocar-
omas (p � 0.086). None of the BAC lesions showed
ural involvement or vascular/lymphatic permeation,
ereas other adenocarcinomas showed pleural involve-
nt in 33%, and vascular/lymphatic permeation in 42%.
mph node involvement was seen for 30 lesions (36%)
the other adenocarcinoma group, but not in the BAC
up. The pathologic stage was IA in all of the lesions in
BAC group, whereas 50 lesions (60%) in the other
enocarcinoma group were stage IA. The distribution of
cytologic subtype for BAC lesions is shown in Table 3.
nmucinous subtype accounted for 82% of all BACs.
ognosis
e postoperative median duration of follow-up was 5.1
ar. There were no operative deaths. The 5-year dis-
se-free survival rate in all 108 patients with adenocar-
oma 3 cm or less in diameter was 70.0% (Fig 1). The
ear disease-free survival rates for patients with BAC
d those with adenocarcinoma other than BAC were
% and 63.5%, respectively (Fig 2). No patients with
C had recurrence in their postoperative clinical
rse. One of the patients with BAC had a contralateral
enocarcinoma 16 years after the initial resection and
derwent a second resection. This lesion was an ade-
carcinoma with a mainly bronchioloalveolar growth
ttern. Although we could not morphologically differ-
tiate second primary from pulmonary metastasis be-
se of the histologic growth pattern similar to BAC, this
ion was diagnosed as second primary adenocarcinoma
ble 3. Cytologic Subtype of 25 BACs According to the
O Classification
tologic subtype No. of Patients
nmucinous 21 (84%)
cinous 2 (8%)
xed mucinous and nonmucinous 2 (8%)
C � bronchioloalveolar carcinoma; WHO � World Health Orga-
ation.
1. Survival among all 108 patients with pulmonary adenocarci-
a 3.0 cm or less in diameter. The 5-year disease-free survival
e was 70.0% (95% confidence interval, 57.6% to 82.3%).
ba
[19
Co
Be
no
(ac
ina
as
gro
ter
ev
W
A
of
be
gro
mu
typ
un
an
If
cla
(an
BA
exc
mo
vis
ne
rep
83%
pa
rec
cla
BA
ho
for
as
cri
ne
pre
be
I
sio
log
aci
bla
Tu
me
fin
og
joi
mi
the
les
gro
inv
tum
cin
or
W
wi
bu
the
int
ing
de
Se
str
tec
[25
mo
erv
3,
wo
cin
of
we
les
pa
gio
C
be
Fig
nom
veo
BA
Ta
Wi
Au
Bre
Vo
Eb
Re
Fu
Pre
BA
Ov
1731Ann Thorac Surg SAKURAI ET AL
2004;78:1728–33 BRONCHIOLOALVEOLAR CARCINOMA PROGNOSIS
G
EN
ER
A
L
T
H
O
R
A
C
IC
sed on the previous report by Martini and Melamed
].
mment
fore the 1999 revised WHO classification, adenocarci-
mas of the lung were classified into four subtypes
inar, papillary, BAC, or solid) based on their predom-
nt histologic growth pattern. Then, BAC was regarded
an adenocarcinoma with a striking alveolar replacing
wth pattern, and so far its clinicopathologic charac-
istics had been discussed in many researches. How-
er, the definition of BAC was changed by the 1999
HO classification [1–8, 10–13, 15].
ccording to the revised WHO histologic classification
lung and pleural tumors in 1999, adenocarcinoma can
grouped into five subtypes according to the histologic
wth pattern, ie, acinar, papillary, BAC, solid with
cin formation, and adenocarcinoma with mixed sub-
es. Among these, BAC is an adenocarcinoma with a
ique replacing growth pattern along the alveolar wall
d no evidence of stromal, vascular, or pleural invasion.
an invasive component is identified, the tumor is
ssified as adenocarcinoma mixed bronchioloalveolar
d acinar or papillary, if present) subtype rather than
C. Patients with BAC should be expected to have an
ellent prognosis because BACs are noninvasive tu-
rs. Nevertheless, in recent reports based on the re-
ed WHO classification, resected stage I BACs were not
cessarily associated with excellent 5-year survival. The
orted 5-year disease-free survival rates were 73% to
(Table 4) [14, 16, 20–22]. These reports included
tients with distant metastasis as well as locoregional
urrence. Perhaps these tumors should have been
ssified as invasive adenocarcinoma rather than as
C. These reports did not provide pathologic details on
w closely they followed the WHO criteria. The issue
a rigorous pathologic evaluation of tumors regarded
2. Survival according to the histologic subtype of adenocarci-
a (Ad). The 5-year survival rates were 100% for bronchioloal-
lar carcinoma (BAC) and 63.5% for adenocarcinoma other than
C.
BAC must be addressed. A standardized morphologic ae
teria defining what constitutes invasive features are
eded now. Unless the pathologic diagnosis as BAC is
cise, the true clinicopathologic characteristics would
equivocal.
n the present study, the diagnosis of histologic inva-
n was based on the 1999 WHO classification. Histo-
ic invasion was suggested by tumor cells arranged in
nic, papillotubular structures or solid nests in a fibro-
stic stroma, often accompanied by collagenization.
mors were classified as BAC only if the entire lesion
t the WHO description. In the present study, these
dings were independently investigated by two pathol-
ists (Y.D., H.S.), and discrepancies were resolved by
nt examination of the slides under a two-headed
croscope. Thus, with respect to the tumor histology, all
tumors without any invasive features within the
ion, ie, the tumors showing a pure bronchioloalveolar
wth pattern, had no pleural, vascular, or lymphatic
olvement. In other words, an adenocarcinoma with
or histology showing a pure bronchioloalveolar car-
oma could be regarded as having no pleural, vascular,
lymphatic invasion, although BAC according to the
HO classification was defined as an adenocarcinoma
th not only a pure bronchioloalveolar growth pattern
t also no evidence of vascular or pleural invasion. On
other hand, several reports have supposed that the
erobserver concordance was poor for pathologic find-
s in lung adenocarcinoma [23, 24]. The prevalent
finition of histologic “invasion” would be essential.
veral investigators have reported that disruption of the
omal elastic framework, which is the supporting archi-
tural structure of the lung, indicated tumor invasion
, 26]. We pathologically examined not only the tumor
rphology by hematoxylin-eosin staining but also pres-
ation of the elastic framework by elastic staining (Fig
A and B). Elastic staining highlights the elastic frame-
rk quite well. The tumors were regarded as adenocar-
omas other than BAC based on evidence of disruption
the elastic framework. Consequently, it seemed that
could accurately diagnose BAC as a noninvasive
ion from a prognostic perspective, since none of the
tients with BAC in our study had postoperative locore-
nal recurrence or distant metastasis.
onversely, mucinous BACs have been suggested to
more likely than nonmucinous BACs to develop an
ble 4. Postoperative 5-Year Survival Rates for Patients
th Stage I BAC
thor (Year) Prognosis
athnach et al (2001) 74%, DFS
lpino et al (2001) 74.9% (stage IA), DFS
right et al (2002) 83.3%, Overall
na et al (2003) 81%, DFS
rak et al (2003) 71%/77% (stage IA/IB), Overall
sent study 100%, DFS
C � bronchioloalveolar carcinoma; DFS � disease-free survival;
erall � overall survival.
rogenous spread, which leads to an unfavorable prog-
no
de
de
oth
sis
wa
les
tha
dif
BA
mu
occ
F
tre
can
a p
req
wi
res
an
con
T1
be
res
Lu
tha
wa
rec
su
siv
res
fea
mi
in
tha
sec
inv
fib
ev
sec
fea
sec
fro
as
fea
res
iss
T
no
pro
a l
cla
ad
pr
“m
alt
im
exp
go
R
su
gro
att
ma
(C
pa
cou
GG
an
pa
his
no
wi
I
BA
pa
wh
cor
Lim
Fig
lin
sta
bro
1732 SAKURAI ET AL Ann Thorac Surg
BRONCHIOLOALVEOLAR CARCINOMA PROGNOSIS 2004;78:1728–33
G
EN
ER
A
L
T
H
O
R
A
C
IC
sis [3, 5, 6, 27–29]. Immunohistochemical studies have
monstrated that mucinous BAC was more likely to
tach from the underlying basement membrane than
er subtypes [30, 31]. In the present study, the progno-
of resected BAC, whether mucinous or nonmucinous,
s excellent although the number of mucinous BAC
ions was quite small. Many reports have s