(Chest. 2000;118:1566-1575.)
© 2000
American College of Chest Physicians
Distribution of Lung Density and Mass in Patients With Emphysema as Assessed by Quantitative Analysis of CT*
Mamadou Hawa Hann Diallo, MD;
Hervé Guénard, MD;
François Laurent, MD;
Pierre Carles, MD, FCCP and
Jacques Giron, MD
*
From the Laboratoire de Physiologie (Drs. Diallo and Guénard), Service D’Exploration Fonctionnelle Respiratoire, Hôpital Pellegrin, Bordeaux, France; the Service de Radiologie (Dr. Laurent), University of Bordeaux II, Bordeaux, France; and the Service de Pneumologie (Dr. Carles) et Radiologie (Dr. Giron), University of Toulouse, France.
Correspondence to: Hervé Guénard, MD, Service d’Exploration Fonctionnelle Respiratoire et Laboratoire de Physiologie, 146 Rue Léo-Saignat, 33760 Bordeaux Cedex, France; e-mail: herve.guenard{at}labphysio.u-bordeaux2.fr
 |
Abstract
|
|---|
Study objective: To assess the effects of emphysema on
the apex-to-base gradient of lung density (D) and lung mass (M)
and to explore the relationship between M and lung function.
Methods: CT scans of whole lungs were performed in 12
healthy subjects and 29 patients who were breathing at functional
residual capacity, after which lung function tests were performed.
Whole D and M and regional D (RLD) and M (RLM) were calculated.
The degree of emphysema was scored.
Results:
The RLM for each height did not differ significantly between patients
with disease and healthy subjects, while RLD was significantly lower in
the patients with disease. A less marked nonlinear, increasing,
craniocaudal gradient of D was observed in the group with disease,
suggesting that the distension increases progressively from the apex to
the base. RLD and RLM in the 40 to 90% lung height differed
significantly among patients in the emphysema group with normal, high,
and low M compared to the healthy subjects. M did not differ
significantly between patients with centrilobular and panlobular
emphysema, which was thought to stem from the marked variations in the
results. Vital capacity was lower in the patients with low M.
Conclusions: The lower RLD in the group with low M was due
to both lung overinflation and to tissue loss, while in the groups with
high or normal M, it was due only to lung
overinflation.
Key Words: emphysema • lung density • lung mass • pulmonary function
|
Introduction
|
|---|
Emphysema
is an anatomic disorder.1
Morphologic and functional
changes develop in relation with the severity of the
disease2
and affect the four major components of the lung,
mainly pulmonary air and tissue volumes, but also interstitial fluid
and blood volumes. At the onset of the disease, the mean linear
intercept increases with lung distension,2
whereas lung
tissue volume (Vtiss) and lung mass (M) change little.3
4
In an animal model of emphysema, M was only observed to increase
sharply after a long interval, exceeding baseline M by
26%.5
CT scans of the lung have revealed an increase in whole M in cases of
human emphysema, although subsequent analyses of subgroups of patients
indicated that normal M or even loss of M was not uncommon in this
condition.6
The decrease in lung density (D) may be
related to a change in the volume of air, or in the mass of tissue, or
a combination of both. Thus, there appears to be a wide range of
alterations in M and D in these patients, reflecting variations
in lung structure in emphysema. In view of the interdependence of M and
mechanical behavior of the lung,7
alterations in regional
D (RLD) or whole D may well have an impact on lung function.
In the lung, the craniocaudal gradient of D has been attributed to
differences in regional blood volume and gas volume, which in turn are
determined by gravity, mechanical stresses, and intrapleural
pressures.7
8
The gradient of D is higher when the subject
is in the vertical rather than in the supine position. In emphysematous
lungs, in addition to the redistribution of lung air and blood content,
the severity of the tissue destruction may be an important additional
factor of this posture-related gradient of D. From a study of regional
lung compliance,9
it has been shown that the vertical
gradient of D observed in emphysematous lungs is related to a change in
local compliance, indicating that this gradient has both tissue and
hydrostatic components. Although there are numerous reports on the
ventrodorsal gradient of D, few studies have been devoted to the
craniocaudal density gradient of human lungs in subjects with
emphysema.9
10
To our knowledge, the interrelationship
between both the severity and extent of emphysema (Xt) and the
distributions of tissue mass and air volume in this gradient have not
been investigated previously.
The aims of the present study were the following: (1) to assess the
craniocaudal distribution of D as well as its determinants as measured
in patients with emphysema while in the supine posture and in a group
of healthy subjects; (2) to relate alterations in D to either
distension or to changes in M; and (3) to correlate these alterations
with lung function as well as with indexes of the degree of emphysema
(ED).
|
Materials and Methods
|
|---|
Patient Selection
Twenty-nine patients in stable condition (27 men, 2 women) with
mild-to-severe emphysema (mean age, 55 years; age range, 40 to 70
years) and 12 healthy subjects (mean age, 38 years; age range, 26 to 50
years) devoid of pulmonary disease who gave their informed consent were
included in the study. Only patients showing macroscopic emphysematous
lesions on high-resolution CT (HRCT) images were retained for the
study. Screening clinical and functional examinations preceded CT
studies of the lungs. Those examinations showing reversible airflow
(eg, an increase in FEV1 of > 15%
during bronchodilator tests) were excluded.
Pulmonary Function Testing
Thoracic gas volume at functional residual capacity (FRC) was
measured by means of a body plethysmograph (constructed in the
laboratory) using the occlusion technique, followed by full inspiration
to total lung capacity (TLC) and subsequent expiration to residual
volume (RV). Vital capacity (VC), TLC, RV, and the RV/TLC ratio were
computed. The study was completed by three FVC maneuvers with display
of flow-volume loops. The best loop based on the best sum
(FEV1 + FVC) was retained. The
FEV1/FVC ratio was calculated. Airway resistance
(Raw) during panting, and quasistatic, transpulmonary pressure-volume
curves were obtained using the integrated signal from a
pneumotachograph at the mouth, while esophageal pressure was measured
by an esophageal balloon, during slow exhalation from TLC. Quasistatic
compliance (Cqs) was taken as the slope of the quasi-static
transpulmonary pressure-volume curve between FRC and FRC + 0.5
L. Measurements of the single-breath diffusing capacity of the lung for
carbon monoxide (DLCOsb) yielded values of the lung
transfer factor for carbon monoxide (MasterLab; Jaeger; Wurzberg,
Germany). All results were expressed as percent predicted, using
the reference values (Table 1)
.11
PaO2 and
PaCO2 were measured on a sample of
arterial blood (Corning 168; Chiron; East Walpole, MA).
CT Study
Physical Parameters and Procedures:
Lungs were all scanned
lengthwise in both conventional (ie, 120 kV, 100 mA) and
high-resolution mode (ie, 137 kV, 255 mA) using a
fourth-generation rotating device (Somatom DRH; Siemens; Erlangen,
Germany) with a 4-s scanning time. The device was calibrated
every 7 h using an air phantom. The ability of this system to
accurately measure D, volume, and mass was ascertained by scanning
standard phantoms of various densities ranging between air and water
(Appendix 1).
The subjects underwent scanning at their FRC using a homemade gating
device. Contiguous 8-mm-thick scans were obtained from the apex to the
base of the lungs by incremental movement of the scanning table, the
position of which could be specified and read with an accuracy of
± 0.5 mm (SD). Scans of 1-mm thickness were obtained at the
same imaging session at 10-mm intervals. Images were analyzed using a
350-mm field of view and a 256 x 256 reconstruction matrix in
conventional mode. In high-resolution mode, a 240-mm field of view,
512 x 512 reconstruction matrix, and a high spatial-frequency
algorithm were used. All images were obtained at window widths and
levels for the appropriate enhancement of parenchymal
structures. The type of emphysema was determined by the results
of previous studies from the HRCT images12
and was
referred to as centrilobular emphysema (CLE), panlobular emphysema
(PLE), or mixed emphysema (ME), when both types were present in the
lung.
Data Analysis and Regional Parameters:
For the
calculation of individual scan density and mass, thick scan images were
displayed on the screen and were processed via a computer linked
on-line with the display screen. A fast contour detection program using
a high-frequency algorithm automatically rejected high-density hilar
structures and isolated pulmonary tissue on each individual scan. The
regional radiologic density and volume of the enclosed area on
individual images were determined using the dedicated software.
Twenty-five to 35 images were processed in this fashion for the left
and right lungs. Regional Hounsfield units (HU) were converted later
into regional physical density from the linear relationship between
density and HU13
14
over the range between air (-1,000
HU) and water (0 HU). Regional M (RLM) was calculated from each
individual scan image as RLD times regional radiologic
volume.6
These data then were used to calculate the M and
D of the whole lung and those of the left and right lungs. The regional
volume of air in the lungs was calculated assuming that the density of
the parenchyma alone was equal to unity; ie, that the
regional Vtiss was equal to the RLM irrespective of units:
 |
where Vairr is regional lung air volume and Vtissr is regional
lung tissue volume. To allow comparison of the distributions of RLD and
RLM over total lung height between healthy subjects and patients,
8-mm-thick scans were sampled from apex to base at approximately 10%
height spacings. Therefore, only about 30% of the scans were taken
into account. The first and the last scans were added to the previous
nine scans so as to fit the total height of the lungs. The origin of
the left pulmonary artery was used as an anatomic marker to divide the
lungs into apex and base scans. This corresponded closely to the 50%
of the apex-to-base distance of the lungs in patients with emphysema.
Therefore, the scans between 0% and 50% apex-to-base distance were
assigned to the apex, and those between 60% and 100% were assigned to
the lung base.
Analysis of the Patient Groups
Patients were classified into three groups according to the
ratio of total M to predicted M at predicted FRC (FRCo) (Mref)
(ie, the M that a subject should have with FRCo and with the
predicted D at FRCo [Do]). The Do was determined from the mean D at
FRCo in healthy subjects, and the FRCo was determined from reference
values.11
Using the Do and the FRCo, we calculated
the Mref (Mref = [FRCo (Do/(1 - Do)]) (see Appendix 2). Patients
were assigned to one of three groups according to their M/Mref ratio.
Patients with M/Mref ratios > 1.12 were considered to have high M,
those with M/Mref ratios of < 0.92 were considered to have low M, and
those with M/Mref ratios within the 1.12 to 0.92 range were considered
to have a normal M. These M/Mref ratios were chosen so as to obtain
three groups of nearly equal size. The RLD, RLM, and Vairr in patients
with high, low, and normal M and in the healthy subjects were plotted
against lung height at 10% spacing from apex to base.
Quantification of the ED
To calculate the ED, the severity of lung
destruction (SLD) and the Xt were quantified on the basis of the
previously selected 8- mm-thick scans, using a small grid and the
direct observational method of Sakai et al.15
Both SLD and
Xt were quantified by two observers. Twenty-two scans of the right and
left lung were read for SLD and Xt in each patient. Briefly, the Xt on
each scan was measured by overlaying a small plastic grid on the CT
images as the count of the squares falling on emphysematous surface
area relative to the surface area of the whole scan under
consideration. The data were expressed as the percentage of lung
involvement and were converted later to numeric values of 1
(< 25% involvement), 2 (25 to 50% involvement), 3 (50 to 75%
involvement), and 4 (> 75% involvement). The SLD was read
directly from the CT images and also was scored on four levels (0 to
3). Final scores were recorded by consensus between the two readers.
The SLD and the Xt were mixed according to the method of Sakai et
al15
to provide an index of the ED.
Data Analysis:
Data were analyzed using a standard
statistical package (NCSS6.0; Statistical Solutions Lim; Cork,
Ireland). Whole M, D, and lung function were tested and compared
between healthy subjects and patients with emphysema using the
nonpaired t test. Intragroup and intergroup differences
in lung function and also differences in the ED, RLD, RLM,
and Vairr in the three groups of patients with emphysema and in healthy
subjects were assessed by multivariate analysis of variance (MANOVA)
for two factors; first, for the craniocaudal lung height and second, at
each given 10% level of this craniocaudal distance. Once the MANOVA
test found a term significant, a univariate analysis of variance was
used to determine which of the variables and factors were responsible
for the significance. The differences among the groups then were
assessed using Fisher’s Exact Test least significant difference
multiple-comparison procedure. The correlations between whole D and
whole M, and between whole M and RLM for a given lung height, were
analyzed to assess the specific effect of M found on apex or base scans
on the increase or decrease in M. The correlations between the
radiologic and the functional data and between RLM and ED
were tested in the total patient population. In all cases, the level of
significance was set at 0.05.
|
Results
|
|---|
M and D
Patients had CLE (n = 15), PLE (or base-dominant lesions in one
ZZ patient n = 10), or ME (n = 4). There were seven patients with
CLE and four patients with PLE in the group with high M; six patients
with CLE, one patient with PLE, and three patients with ME in the group
with normal M; and two patients with CLE, five patients with PLE, and
one patient with ME in the group with low M. The mean (± SD) M was
greater in patients with CLE than in those with PLE both in the group
with high M (CLE, 1,302 ± 148 g; PLE, 1,273 ± 196 g) and in the
group with low M (CLE, 906 ± 277 g; PLE, 813 ± 175 g) without
reaching a significant level.
The mean D was significantly lower in the group with emphysema
(0.209 ± 0.048 g/mL; p < 0.001) than in healthy subjects
(0.322 ± 0.062 g/mL). The mean M, although higher in the group with
emphysema (1,084 ± 238 g), did not differ significantly from that in
healthy subjects (987 ± 202 g). Of the 29 emphysema patients, 11 had
normal M, 10 had high M, and 8 had low M. The right lung was heavier
than the left lung, although the difference was not significant (NS)
either in the group of healthy subjects or in the group of
patients with emphysema. The correlation between D and M was
significant in the group of patients with emphysema (D = 0.074 x M
[kg] + 0.128) (r = 0.379; p < 0.05) but was not in
healthy subjects.
RLM did not differ significantly between the healthy subjects and the
patients with emphysema (Fig 1
, top, a). However, RLDs at each 10% of lung
height of the apex-to-base distance were significantly lower in the
group of patients with emphysema (Fig 1
, bottom,
b) than in the healthy subjects.
Excluding the first scan, the gradient of D vs height from the apex to
the base could be split into the following two parts: one part nearly
flat, and the other part steeper in the bases (Fig 1
,
bottom, b). The slope of the regression
line of RLD as a function of lung height (percentage) was significantly
(p < 0.05) different from zero in the group of healthy subjects but
not in the group of emphysema patients in the upper two thirds of the
lung (0 to 70% height). In the lower third (70 to 100% height), the
gradient of RLD was 20% higher in the healthy subjects compared to the
group of emphysema patients, and the slopes of the regression lines,
RLD as a function of lung height percentage, differed
significantly from zero both in the group of emphysema patients
(p < 0.05) and in the group of healthy subjects (p < 0.02). The
absolute difference in D between the healthy subjects and the group of
emphysema patients (Fig 1 , bottom, b [right-side
axis]) at each 10% spacing level increased significantly
(r = 0.866; p < 0.001) from the apex to the base.
To discern the position in the lung in which differences in RLM might
account for the overall differences in whole M between patients, linear
regression between RLM for a given height and whole M were calculated.
The slopes of the regression lines increased from top to bottom, as
shown at 20% and 60% lung height (Fig 2
, bottom, b, and top, a,
respectively). Slopes were different from zero from 30 to 90% lung
height and increased significantly and progressively from 1.6 to 4.9%
at 30% and 90% lung height, respectively.
RLM, RLD, and Vairr
The mean differences of RLM (Fig 3
, top, a), RLD (Fig 3
, middle,
b), and Vairr (Fig 3
, bottom, c)
between healthy subjects and patients with emphysema, and among the
emphysema patient groups, at each 10% spacing level in the overall
lung height can be seen in Figure 3
. The shape of the distributions of
RLM, RLD, and Vairr over the total lung height was similar in the
healthy subjects and in the three groups of emphysema patients. The
maximal values of RLM were observed at 80% height in the three groups
of emphysema patients; the maximum volume of Vairr was observed at
about 50% lung height (Fig 3
, top, a, and
bottom, c, respectively).
The ED in the Apex-to-Base Distance
The average ED was significantly (p < 0.05) lower
in the group with high M (1.065 ± 0.66) than in the groups with low
M (1.41 ± 0.63) and normal M (1.49 ± 0.66). There were
significant (p < 0.001) negative correlations between ED
and the percent of lung height in the groups with high and normal M
(Fig 4
), while a positive correlation was noted in the group with low M. At
each height, regression equations of RLM and RLD vs ED were
calculated in the total patient population. All regression equations,
except the one for the first apex scan, had negative and significant
slopes (p < 0.001). Slopes increased from the apex to the base,
although the data were more scattered at the base. There was no
significant correlation between RLD and ED.
Functional Data
In healthy subjects, there were significant correlations between M
and height (r = 0.692; p < 0.05) and between D and Raw
(r = -0.609; p < 0.05). In the whole group of patients
with emphysema, M was correlated significantly with height
(r = 0.468; p < 0.02) and FEV1
(r = 0.393; p < 0.05), while D was correlated with
FEV1 (r = 0.476; p < 0.01),
FEV1/FVC ratio (r = 0.608;
p < 0.01), and RV/TLC ratio (r = -0.495; p < 0.02).
The striking difference
among the groups of patients with emphysema
(Table
2) was in VC. There were significant differences in VCs among the three
groups (Table 2)
. The group with low M had lower than normal VC, while
VC was close to normal in the group with high M and was in the normal
range in the group with normal M. Distension (high RV) and obstruction
(low FEV1/FVC ratio) were more marked in patients
in the group with low M, who had, however, normal pulmonary compliance.
The decrease in VC was correlated significantly with the resultant
following four factors: low FEV1/VC ratio and
body weight; and high RV and Raw. DLCOsb and
PaO2 did not differ among
the groups. The group with low M thus appeared to exhibit a stepwise
trend in the alteration of lung function.
|
Discussion
|
|---|
The main findings of this study are the following: (1) the
significant positive correlation between D and M in patients with
emphysema, the average M not being different from that in the healthy
subjects; (2) an increased difference in the apex-to-base gradient of D
in patients with emphysema compared to the healthy subjects; (3) the
lower RLD observed in patients with normal and high M compared to the
healthy subjects was a result of distension (ie, an increase
in regional air content), whereas in the patients with low M, lower RLD
stemmed from both distension and loss of lung tissue; (4) the
alterations in whole M in the patients with low and high M were a
result of alterations in RLM in the bases but not in the apices; and
(5) ED was more severe in the patients with low
M, which is in agreement with the fact that RLM, regardless of height,
was negatively correlated with ED.
Whole M and D
In a previous study6
including 24 patients, we found
a slightly but significantly greater M in patients with emphysema than
in healthy subjects. In the present study, which includes more
patients, the difference in M between the group of patients with
emphysema and the healthy subjects was NS. The interesting point is
that M was not reduced in most patients with emphysema, which might be
expected from the definition of emphysema1
as a
destructive process, which suggests a loss of tissue (ie, a
loss of mass). Overall M in patients with pulmonary emphysema results
from the balance between lung destruction and secondary healing or
remodeling.16
17
This would indicate that lung remodeling
outweighs lung destruction in the emphysema patients with relatively
high M, and vice versa for the patients with low M. Therefore, M might
represent a useful index of the following two types of emphysema: one
that is mostly destructive with little remodeling and light lungs (low
M); and the other dominated by the remodeling process with heavy
lungs (high M). In healthy subjects, at a given lung volume
(ie, at FRC), D is constant regardless of M (ie,
the volume of air is directly proportional to M). In patients with
emphysema, the decrease in D with M cannot be wholly attributed to the
decrease in M, as the slope of the D = f(mass) relationship was too
low, but, rather, it suggests that patients with low M have lungs that
are relatively more distended than those in patients with
high M. Since D depends directly on lung volume, our emphysema patients
and healthy subjects all underwent CT scanning at end-expiratory level
(ie, at FRC).
With the subject in the supine position, FRC tends to be lower than
when the subject is in the sitting position by several hundred
milliliters in healthy individuals, depending on the displacement of
the abdominal contents toward the thorax. In emphysema patients, FRC in
the supine position is less likely to change, owing to their distension
and/or obstruction. This would have enhanced slightly the observed
difference in D between the healthy subjects and the emphysema
patients.
RLD and RLM as a Function of Lung Height Percentage (Healthy
Subjects vs Emphysema Patients)
The means of the distributions of RLM with lung height were
comparable in the healthy subjects and in the group of patients with
emphysema, suggesting that the balance between destruction and
remodeling was the same throughout the emphysematous lungs. The shape
of the distribution of RLD with lung height (Fig 1
, top,
a) in emphysema patients resembled that of healthy subjects,
albeit shifted downward. Excluding the uppermost part of the apex, the
distribution of RLD in healthy subjects in the craniocaudal axis
appeared nonlinear, increasing slightly from the top to the hilum, and
sharply from the hilum to the base. This increase in RLD with height in
the base could be due to alterations in parenchymal or blood masses. In
the supine position, as when standing, the lung is subjected to
gravity,7
9
and the weight of the posterior parts of the
bases will stress both the anterior part of the bases and the apices.
Furthermore, in the supine position, the stress in the bases will be
reduced by the abdominal mass, displacing the lung toward the apex,
thereby reducing the volume of air in the bases and increasing its
density. Blood volume also would be expected to be greater in the
underhanging part of the lung (ie, the posterior parts of
the bases).18
19
Independently of the greater
vascularization and higher capillary density of basal lung regions
compared to the apex,19
20
distension in the supine
position would tend to attenuate the gradient in patients with
emphysema (Fig 1
, bottom, b), as it would in
healthy subjects during inspiration. This interpretation is supported
by the fact that Vairrs were increased in all compartments in the three
diseased groups (Fig 3
, bottom, c). Whereas
static compliances were in the normal range in most emphysema
patients, the lower gradient of D in these patients suggests a
severe loss of elasticity in their lung bases, especially in the group
with low M.
The difference in D between healthy subjects and the group of patients
with emphysema increased with lung height, suggesting either a greater
loss of tissue or a greater distension in the lung bases of the
patients. As there was no difference in the distribution of RLM between
the two groups, distension would appear to be the major factor in the
increasing difference in RLD with height.
Differences Between the Patient Groups
The overall analysis of a rather heterogeneous population tended
to obscure differences between the patients with high, low, or normal
M. Thus, the patients were allocated into three groups as a function of
their M/Mref ratios. Threshold values of the M/Mref ratio of 0.92 and
1.12 were chosen to obtain groups with equal numbers of patients.
Patients also could have been split up by taking the mean value of M of
the healthy subject group ± 1 SD as the normal range or by using a
clustering statistical analysis. However, our healthy group was too
small to define the SD of a healthy population accurately, and
clustering would need a larger series of patients. Furthermore, M is
dependent on the size of the lung, which in turn depends on height.
Therefore, we thought it more appropriate to calculate the M that the
patients should have had if they had a normal FRC and a normal D, as D
has been shown to be independent of age,6
21
22
and to
classify patients according to this Mref. The groups of emphysema
patients with high M and normal M had a slightly positive but
nonsignificant craniocaudal gradient of D between the apex and 80%
lung height, whereas the group with low M had an absence of gradient
(Fig 3 , middle, b). Although the gradient of D
observed at low lung volume9
10
23
in healthy subjects as
well as in emphysema patients has been shown to be reduced with lung
inflation at TLC, our observations indicated that this gradient might
be abolished even at FRC in emphysema patients who have severe,
destructive losses of lung tissue.
There were significant differences in absolute regional lung air
volume between all diseased groups and healthy subjects, except
at 0% and 10% lung height. The absolute values of regional lung air
content were lower in the patients with low M compared with the
patients with normal and high M, although the values were consistently
higher than in the healthy subjects, indicating that the low D in the
group with low M was due not only to the loss of lung tissue but also
to lung distension.
ED Score and RLM
The slope of the relationship between the RLM of a given
compartment and the whole M of the group of patients with emphysema
showed an increasing slope between 20% and 90% lung height, except
for the end scan at the base. Therefore, the increase or decrease in M
observed in the groups of patients with emphysema was mostly due to
changes in M in the bases and was not evenly distributed, as was
suggested by the analysis of the emphysema patients as a whole. Such
differences could be accounted for in terms of the anatomic differences
between the apex and the lung base. Blood and relative tissue volumes
are greater in the bases19
21
and would be expected to be
more sensitive to structural alterations like the loss of tissue or to
the diffuse inflammation commonly found in emphysema.24
25
This was supported by the findings from our three groups of emphysema
patients that, whereas RLM was related to ED in the apex
and in the base, only a high score of ED in the base, as
reported in the group patients with low M, was followed by severe loss
of tissue.
The average regression equation of RLM vs ED in the overall
group of patients with emphysema indicated that with increasing
severity of emphysema there was significant loss of lung tissue. Since
M did not differ between the two types of emphysema in either of the
patient groups having high M or low M, it could be argued that the
increase or loss of M stemmed more from the localization than the type
of the emphysema. Coxson et al26
studied the relationship
between whole M and the ED assessed from the lower -910-HU
level of the D curve. The lack of a significant correlation between
910-HU level and M in their results was partly related to the fact
that the low D areas that are assumed to quantify the extent of
emphysema reflected lung overdistension rather than true loss of
tissue. The present results are in agreement with those of Coxson et
al26
as no significant correlation was observed between Xt
and RLM, although we did observe a correlation between the SLD indexes
with both RLM and ED (which is a pool of both indexes). The
lack of correlation between RLD and ED indicates that RLD
depends on both RLM and Vairr and that the overdistension observed in
many patients in the disease group was not directly related to the
ED.
The dissimilar pattern of distribution of emphysema between the patient
groups with high M and normal M compared to the patient group with low
M was related to the type of emphysema in each group. Among the 21
patients of the groups with high M and normal M, 13 had CLE, 5 had PLE,
and emphysema was more severe in the apex; in the group with low M, 2
patients had CLE, 5 patients had PLE, and emphysema was more severe in
the base. As the remaining patients had a mixed form of emphysema, the
variation in emphysematous involvement noted in our three groups was
consistent with the recognized distribution pattern of the disease
according to the type of emphysema.27
Lung Function in the Diseased Groups
The present data indicate that the group with low M with the most
severe forms of obstruction and distension had lower D and M. Previous
studies based on a lung model have established the important role
played by M on lung mechanics and function.7
10
The
present findings extend these observations and indicate that severe
loss of M up to 20%, as observed in our group of emphysema patients
with low M, may occur in patients with emphysema and would be
accompanied by a marked alteration in lung function. It would be of
interest to make a longitudinal study of lung function and alterations
in M in individual patients. This might indicate whether patients with
low M were previously in the group of patients with high M and normal
M. Indirect evidence that the preservation of M has a beneficial effect
on lung function is supported by a study undertaking the pathologic
examination of bronchioles and bronchiolar wall thickness in patients
suffering from emphysema.28
These authors found
that the patients with the greatest increase in bronchiolar tissue
volume were the least obstructed. In contrast, the patients with losses
of bronchiolar tissue volume, which were akin to those in our group of
patients with low M, had more severe obstruction, lending support to
the idea that the mechanical behavior of the lung in emphysema is
related to whole M.7
The lower VC in the group of emphysema patients with low M than in
groups with high M and normal M may derive from several factors acting
alone or in combination. Our patients with low M (1) were leaner than
those with normal M (52 ± 10 kg vs 67 ± 11 kg, respectively); (2)
had lower FEV1/FVC; (3) had higher Raw; and (4) had higher
RV. Although these four factors did not differ individually
among the patient groups, their combined effect on VC limitation was
statistically significant. As it has been shown that lung function was
better related to the lung base than to the lung apex, it is likely
that the location of the severe loss of tissue in the lung base in
patients with low M contributed to the marked functional alterations
observed in this group.29
It would be of interest to know
whether the low M observed in the patients with low M corresponds to a
particular form of emphysema with low remodeling or whether it is the
result of a general process of becoming more lean, as suggested by the
lower body weight of these patients.
|
Conclusion
|
|---|
In conclusion, the M of emphysematous lungs is, on average, not
different from normal. In the patients with emphysema, the mean of the
distribution of M appeared to be normal, although D was lower and the
apex-to-base gradient of D was lower than that in the healthy subjects.
However, there were considerable differences among the groups of
patients. Tissue mass in the lung base appeared to be the main
determinant of the alteration in whole M in patients with emphysema. In
the patient group with high M, the decrease in D can be attributed to a
distension overshadowing the overweight, while in the group with low M,
the loss in D was due both to distension and to loss in lung weight.
The patients with low M appeared to be the most functionally impaired.
|
Appendix 1
|
|---|
To check the ability of CT scanning to determine density
displacement (Dw), volume (Vlw), and mass (Mw), as measured by water
displacement, of material and to determine the calibration
curve of our scanner, we performed a companion study of various
phantoms. Seven parallelepiped phantoms constructed from
polystyrene (0.0340 g/mL), chestnut (0.664 g/mL), poplar (0.396 g/mL),
oak (0.602 g/mL), framiré (0.602 g/mL),
iroko (0.638 g/mL), and niangon (0.727 g/mL) were
scanned lengthwise using contiguous 8-mm scans. The results of the
end-most slices were discarded to exclude voxels averaging at the
wood-air interface. Air and water phantoms were scanned later to cover
the overall density range of pulmonary tissue. The true densities of
the phantoms were determined from the ratio of their Mw to their
volumes Vlw, as determined by weighing and water displacement,
respectively. Taking into account the exclusion of the end slices, CT
underestimated Dw, Mw, and Vlw of the wood phantoms by 4 ± 5%,
6 ± 7%, and 8 ± 6%, respectively. A plot of HU vs Dw gave a
regression line very close to the identity line over the range of
densities between air (1 mg/mL) and water (1 g/mL). Some scatter was
observed but only toward the higher densities (ie, > 0.80
g/mL) and still with a < 3% error.
|
Appendix 2
|
|---|
The relationship between the Vtiss and M of a subject breathing at
FRC could serve to determine the Mref from the FRCo and the Do at FRCo:
where dtiss is lung tissue density (air excluded).
If dtiss = 1
|
Acknowledgements
|
|---|
The authors are indebted to P. Drouillard for use
of the facilities in the Diagnostic Radiology Scanning Department at
Haut Lévêque Hospital. The authors also thank
F. Denise for technical assistance with the CT measurements, D. Audet
for skillful secretarial assistance, and Dr. M. Pujazon for her
contribution to the functional investigation.
|
Footnotes
|
|---|
Abbreviations:
CLE = centrilobular emphysema; Cqs = quasistatic compliance;
D = lung density; DLCOsb = single-breath diffusing
capacity of the lung for carbon monoxide; Do = predicted lung density
at predicted functional residual capacity; Dw = density of biological
material as measured by water displacement; ED = degree
of emphysema; FRC = functional residual capacity; FRCo = predicted
functional residual capacity; HRCT = high-resolution CT;
HU = Hounsfield units; M = lung mass; MANOVA = multivariate
analysis of variance; ME = mixed emphysema; Mref = predicted lung
mass at predicted functional residual capacity; Mw = mass of
biological material as measured by water displacement; NS = not
significant; PLE = panlobular emphysema; Raw = airway resistance;
RLD = regional lung density; RLM = regional lung mass;
RV = residual volume; SLD = severity of lung destruction;
TLC = total lung capacity; Vairr = regional lung air volume;
VC = vital capacity; Vlw = volume of biological material as
measured by water displacement; Vtiss = lung tissue volume;
Xt = extent of emphysema
Received for publication August 17, 1999.
Accepted for publication July 17, 2000.
|
References
|
|---|
-
Snider, GL, Kleinerman, J, Thurlbeck, WM, et al (1985) The definition of emphysema: report of a National Heart, Lung and Blood Institute, Division of Lung Diseases, Workshop. Am Rev Respir Dis 132,182-185[ISI][Medline]
-
Eidelman, DH, Bellofiore, S, Chiche, D, et al (1985) Behavior of morphometric indices in pancreatic elastase-induced emphysema in rats. Lung 168,159-168
-
Petty, TL, Silvers, GW, Stanford, RE (1987) Mild emphysema is associated with reduced elastic recoil and increased lung size but not with airflow limitation. Am Rev Respir Dis 136,867-871[ISI][Medline]
-
Marvin, L, Maria, LP, Christopher, C (1992) Induction of emphysema in hamster by intratracheal instillation of Cathepsin B. Am Rev Respir Dis 145,661-668[ISI][Medline]
-
Snider, GL, Sherter, CB (1977) A one year study of the elastase induced emphysema in hamsters. J Appl Physiol 43,721-729[Free Full Text]
-
Guenard, H, Diallo, MHH, Laurent, F, et al (1992) Lung density and lung mass in emphysema. Chest 102,198-203[Abstract/Free Full Text]
-
West, JB (1971) Distribution of mechanical stress in the lung, a possible factor in localization of pulmonary disease. Lancet 24,839-841
-
Genereux, GP (1985) Computed tomography and the lung: review of anatomic and densitometric features with their clinical application. J Can Assoc Radiol 36,88-102[ISI][Medline]
-
Millar, AB, Denison, DM (1990) Vertical gradient of lung density in supine subjects with fibrosing alveolitis or pulmonary emphysema. Thorax 45,602-605[Abstract]
-
Millar, AB, Fromson, B, Strickland, B, et al (1986) Computed tomography based estimates of regional gas and tissue volume of the lung in supine subjects with chronic airflow limitation or fibrosing alveolitis. Thorax 41,932-939[Abstract]
-
Standardized lung function testing. Quanjer, H eds. Bull Eur Physiopathol Respir 1983;19(suppl),45-51
-
Foster, WL, Pratt, PC, Roggli, VL, et al (1986) Centrilobular emphysema: CT-pathologic correlation. Radiology 159,27-32[Abstract/Free Full Text]
-
Phelps, ME, Gado, MH, Hoffman, EJ (1975) Correlation of effective atomic number and electron density with attenuation coefficient measured with polychromatic X rays. Radiology 117,585-588[Abstract]
-
Mull, RT (1984) Mass estimates by computed tomography: physical density from CT numbers. AJR Am J Roentgenol 143,1101-1104[Abstract/Free Full Text]
-
Sakai, F, Gamsu, G, Im, J, et al (1987) Pulmonary function abnormalities in patients with CT-determined emphysema. J Comput Assist Tomogr 11,963-968[ISI][Medline]
-
Fukuda, Y, Masuda, Y, Ishizaki, M, et al (1989) Morphogenesis of abnormal elastic fibers in lungs of patients with panacinar and centriacinar emphysema. Hum Pathol 20,652-659[CrossRef][ISI][Medline]
-
Roman, JR, Mcdonald, JA (1991) Cellular processes in lung repair. Chest 100,245-248[Free Full Text]
-
Brudin, LH, Rhodes, CG, Valind, SO, et al (1992) Regional structure-function correlation in chronic obstructive lung disease measured with positron emission tomography. Thorax 47,914-921[Abstract]
-
Butler, C, II, Kleinerman, J (1970) Capillary density: alveolar diameter, a morphometric approach to ventilation and perfusion. Am Rev Respir Dis 102,886-894[ISI][Medline]
-
Brudin, LH, Rhodes, CG, Valind, SO, et al (1987) Regional lung density and blood volume in non smoking and smoking subjects measured by PET. J Appl Physiol 63,1324-1334[Abstract/Free Full Text]
-
Rosenblum, LJ, Maugerl, RA, Wellenstein, DE, et al (1980) Density patterns in the normal lung as determined by computed tomography. Radiology 137,409-416[Abstract/Free Full Text]
-
Wollmer, P, Albrechtsson, U, Brauer, K, et al (1986) Measurements of pulmonary density by means of x-rays computerized tomography: relation to pulmonary mechanics in normal subjects. Chest 90,387-391[Abstract/Free Full Text]
-
Goddard, PR, Nicholson, EM, Laszlo, EM, et al (1982) Computed tomography in pulmonary emphysema. Clin Radiol 33,379-387[CrossRef][ISI][Medline]
-
Berend, N (1981) Lobar distribution of bronchiolar inflammation in emphysema. Am Rev Respir Dis 124,218-220[ISI][Medline]
-
Wright, JL, Weggs, BJ, Hogg, JC (1984) Airways disease in upper and lower lobe in lungs of patients with and without emphysema. Thorax 39,282-285[Abstract]
-
Coxson, H, Whittal, KP, Moore, BJ, et al (1996) Measurement of regional lung density in emphysema using computed tomography [abstract]. Am Rev Respir Dis 153,A325
-
Thurlbeck, WM (1962) The incidence of pulmonary emphysema with observations on the relative incidence and spatial distribution of various types of emphysema. Am Rev Respir Dis 87,206-215
-
Nagai, A, West, WW, Thurlbeck, WM (1985) The National Institutes of Health intermittent positive pressure breathing trial: pathology studies: II. Correlation between morphologic findings and evidence of expiratory flow obstruction. Am Rev Respir Dis 132,946-953[ISI][Medline]
-
Haraguchi, M, Shimura, S, Hida, W, et al (1998) Pulmonary function and regional distribution of emphysema as determined by high-resolution computed tomography. Respiration 65,125-129[CrossRef][ISI][Medline]
TOP
Abstract
Introduction
) and in
emphysema patients (
), and RLD (bottom,
b) in healthy subjects (•) and in emphysema patients
(
) are sampled at a 10% spacing level from the apex (0%) to the
base (100%) of the lung. Note the relative loss of lung tissue at 70%
and 80% lung height, attaining 20% in the latter. The absolute
difference of D between the healthy subjects and the emphysema patients
(right-side axis) was shown as a linear regression line of increasing
gradient of D toward the bottom of the lung.
), and low M
(