Noninvasive quantification of heterogeneous lung growth following extensive lung resection by high-resolution computed tomography

Abstract

To quantify the in vivo magnitude and distribution of regional compensatory lung growth following extensive lung resection, we performed high-resolution computed tomography at 15- and 30-cmH₂O transpulmonary pressures and measured air and tissue (including microvascular blood) volumes within and among lobes in six adult male foxhounds, before and after balanced 65% lung resection (∼32% removed from each side). Each lobe was identified from lobar fissures. Intralobar gradients in air and tissue volumes were expressed along standardized x,y,z-coordinate axes. Fractional tissue volume (FTV) was calculated as the volume ratio of tissue/(tissue + air). Following resection compared with before, lobar air and tissue volumes increased 1.8- to 3.5-fold, and whole lung air and tissue volumes were 67 and 90% of normal, respectively. Lobar-specific compliance doubled post-resection, and whole lung-specific compliance normalized. These results are consistent with vigorous compensatory growth in all remaining lobes. Compared with pre-resection, post-resection interlobar heterogeneity of FTV, assessed from the coefficient of variation, decreased at submaximal inflation, but was unchanged at maximal inflation. The coefficient of variation of intralobar FTV gradients changed variably due to the patchy development of thickened pleura and alveolar septa, with elevated alveolar septal density and connective tissue content in posterior-caudal and peripheral regions of the remaining lobes; these areas likely experienced disproportional mechanical stress. We conclude that HRCT can noninvasively and quantitatively assess the magnitude and spatial distribution of compensatory lung growth. Following extensive resection, heterogeneous regional mechanical lung strain may exceed the level that could be sustained solely by existing connective tissue elements.

surgical lung resection by pneumonectomy (PNX) is a robust model of compensatory lung growth. Following PNX, blood flow to the remaining lung increases by a factor of (1/fraction of lung removed), and the negative intrathoracic pressure causes the remaining lobes to expand by a similar factor. These mechanical stresses, and possibly other nonmechanical stimuli, are thought to induce adaptation in the remaining lobes, including recruitment of alveolar-capillary reserves, remodeling of existing structure, and reinitiated growth of new alveolar tissue (8). It is likely that the distribution of post-PNX tissue stresses is nonuniform, evidenced by asymmetric differential lobar expansion (19). Following 42% lung resection by left PNX, four lobes remain in the right lung: the cranial (upper) and the most caudal (infracardiac) remaining lobes expand nearly twofold, while others (middle and caudal lobes) show little to no expansion. Following 58% lung resection by right PNX, three lobes remain in the left lung; all expand twofold or more with persistent interlobar differences (19). The greater increase in air and tissue volumes of the remaining lobes corresponds to a higher fractional increase in lung diffusing capacity following 58% rather than 42% resection (12). Comparisons suggest that regional anatomical response becomes more uniform as more lung units are resected; uniformity may contribute to better functional outcome. Our laboratory recently observed that, following more extensive (∼70%) lung resection, fractional increase in physiological lung diffusing capacity was even greater than that following 58% resection (9), suggesting even more vigorous compensatory lung growth and perhaps more uniform regional anatomical responses.

Anatomical compensation is traditionally quantified by postmortem morphometry; this approach requires sufficient and unbiased sampling of tissue blocks from each lobe to account for spatial heterogeneity (23). In addition, postmortem assessment of fixed lung tissue does not reflect in vivo changes in dynamic anatomical variables, such as alveolar air and capillary blood volumes that interact with alveolar tissue to determine the magnitude of physiological compensation. An alternative approach, using high-resolution computed tomography (HRCT), reliably detects in vivo changes in regional lung attenuation. Our group (18, 19) and others (2, 3) have used HRCT to quantify lung air and tissue volumes during somatic maturation, as well as following lung resection. HRCT-derived estimates of whole lung air and tissue volumes correlate significantly with independent physiological estimates obtained by an acetylene rebreathing method (15, 16) and postmortem by morphometry (1, 19), indicating their potential use as noninvasive indexes of lung growth. In addition, HRCT-derived estimates include air, tissue, and microvascular blood components of the gas exchange parenchyma and allow detailed quantification of three-dimensional spatial distributions. By identifying and separating individual lobes via their natural boundaries, we could follow the same regions over successive scans, even in the presence of gross distortion. Previous studies have shown heterogeneous changes in average lobar air and tissue volumes following PNX (18, 19), consistent with nonuniform regional responses in lung expansion, perfusion, and/or generation of new alveolar tissue; however, intralobar gradients have not been quantified.

The objectives of this study were twofold: first, we mapped intralobar HRCT-derived indexes [air volume, tissue+blood volume, fractional tissue volume (FTV), and specific lung compliance] along standardized x,y,z-coordinate axes to allow quantitative comparisons of corresponding regions in the presence of gross anatomical distortion. Second, we applied this technique to examine air and tissue volume distributions before and after extensive lung resection to test the hypothesis that lobar compensatory lung growth is uniform following extensive resection. Our results are consistent with vigorous compensatory lung growth in all remaining lobes. However, regional responses remain heterogeneous, at least partly reflecting patchy accumulation of alveolar septal connective tissue within the remaining lobes.

MATERIALS AND METHODS

Animals.

The Institutional Animal Care and Use Committee approved all procedures. Six litter-matched male mixed breed hounds (∼9 mo old) were studied before and after undergoing bilateral surgical lung resection by a two-stage procedure performed 1 mo apart. The surgical procedure and physiological measurements have been described in detail elsewhere (9). Briefly, under general anesthesia, a lateral thoracotomy was performed. The appropriate lobar arteries and veins were ligated and cut. The appropriate lobar bronchi were stapled and the lobe removed. The stumps were immersed in warm saline to check for leaks. The chest wall was closed in layers, and a chest tube with a one-way valve was placed for 24 h. In the first procedure, a left lateral thoracotomy was performed, and the left caudal lobe removed. Following recovery, 1 mo later, a right lateral thoracotomy was performed, where the right middle, caudal, and infracardiac lobes were removed. The remaining right cranial, left cranial, and left middle lobes comprised, on average, 17.8 ± 0.4, 11.3 ± 1.8, and 6.2 ± 1.3% (means ± SD), respectively, of the original lung units (9). Because roughly similar amounts of lung remained in each side, the mediastinum maintained its midline position, and these animals tolerated the resection well and without complications. HRCT was performed 1 mo before and 3 mo after surgery.

HRCT.

Animals were fasted overnight and premedicated with acepromazine and glycopyrrolate. A peripheral intravenous catheter was inserted. Anesthesia was induced with an intravenous bolus of ketamine and diazepam and maintained by an infusion of the same medication titrated to effect. The animal was intubated with a large cuffed endotracheal tube, placed supine on the computed tomography (CT) table and mechanically ventilated to eliminate spontaneous breathing efforts. Initially, a GE High Speed CTi was used (3 × 3 mm collimation, 120 kV, 250 mA, pitch 1.0, and rotation time 0.8 s). Images were reconstructed at consecutive 1-mm intervals, resulting in 300–325 images per animal. Later, a GE Lightspeed 16 scanner became available and was used to obtain consecutive images at 1.25-mm thickness. Scout images were obtained to ensure the field of scan included the entire lung from the apex to the costophrenic angle. Before each imaging sequence, the lungs were hyperinflated with three tidal breaths, followed by passive expiration to functional residual capacity. Then the lungs were inflated to a transpulmonary pressure (Ptp) of 15 or 30 cmH₂O. The breath was held for ∼40 s, while CT images were obtained, after which the animal was reconnected to the respirator.

Data analysis.

Images in DICOM format were transferred to a personal computer, reconstructed, and analyzed using a customized, semiautomatic image analysis program, developed using Microsoft Visual C++ 6.0. Details of the image analysis have been described previously (18, 19). The area occupied by lung tissue on each image was outlined by attenuation thresholding, which excluded conducting structures larger than 1–2 mm in diameter. The trachea and next three generations of large-conducting airways were manually marked and excluded. Lung volume of each image is equal to the product of its area and slice thickness. Total lung volume is the sum of lung volume on all images. To identify individual lobes, interlobar fissures were identified and fitted with cubic splines; the splines were adjusted through serial images using a semiautomated algorithm.

The CT attenuation of extrathoracic air was set at −1,000 Hounsfield units (HU) and water at 0 HU. The CT attenuation for intrathoracic air (CT_air) was sampled and averaged from three regions (5 mm above the carina, 5 mm below the end of the endotracheal tube, and halfway between the two points) to obtain a mean CT_air in each animal (−975 HU, range −994 to −949 HU). Assuming the average CT value for air-free lung tissue and uncontrasted blood (CT_tissue) equals that of muscle, we averaged three muscles in each animal (infraspinatus, supraspinatus, and pectoralis at the level just above the carina); the mean CT_tissue values were relatively stable (+69 HU, range +64 to +76 HU) (19). A small variability in CT attenuation gradient causes little error and was neglected in further computation. HRCT-derived tissue volume includes the volume of tissue in the distal lung, as well as extraseptal structures (airways and vessels) that are <1–2 mm in diameter, as well as the blood within these small vessels. These values, CT_air and CT_tissue, were used for partition of the total lobar volume (V_lobe) into air (V_air) and tissue+blood (V_tissue) volumes using the following equations:

$${\text{V}}_{\text{tissue}}=\frac{{\text{CT}}_{\text{lobe}}-{\text{CT}}_{\text{air}}}{{\text{CT}}_{\text{tissue}}-{\text{CT}}_{\text{air}}}\times {\text{V}}_{\text{lobe}}$$ (1)

$${\text{V}}_{\text{tissue}}={\text{V}}_{\text{lobe}}-{\text{V}}_{\text{tissue}}$$ (2)

$$\text{FTV}=\frac{{\text{V}}_{\text{tissue}}}{{\text{V}}_{\text{air}}-{\text{V}}_{\text{tissue}}}$$ (3)

where CT_lobe is the average CT value (in HU) of a lobe.

Within each lobe, the voxel FTV gradients were calculated along the three coordinate axes: x (medial to lateral), y (posterior to anterior), and z (cephalad to caudal) and classified into bins, according to their relative position along the span of each axis, 0–10, 11–30, 31–50, 51–70, 71–90, and 91–100% of total span, and analyzed with respect to the average position of the bin along a given axis, i.e., 5, 20, 40, 60, 80, and 95% of total span. Regional lung compliance was calculated from the regional change in volume at two levels of Ptp. In this way, each lobe served as its own reference space for pre- to post-resection comparisons.

Statistical analysis.

Measurements were expressed as means ± SD. Comparisons of regional gradients before and after lung resection and between two inflation volumes were performed using repeated-measures analysis of variance. Comparisons among lobes were performed using one-way ANOVA with post hoc analysis by Fishers protected least significant difference. The coefficient of variation (CV = SD/mean) in regional FTV was calculated to assess heterogeneity within and among lobes. We used a commercial statistical package (STATVIEW version 5.0, SAS Institute, Cary, NC). A P value ≤ 0.05 was considered significant.

RESULTS

Body weight did not change significantly before and after lung resection (24.6 ± 4.1 and 25.6 ± 3.9 kg, respectively, means ± SD). Systemic hemoglobin concentration was also similar (14.7 ± 2.1 and 14.3 ± 0.8% pre- and post-resection, respectively). Representative three-dimensional reconstructed lobar images pre- and post-resection (Fig. 1) illustrate marked enlargement of all of the remaining lobes. Absolute lobar volumes and specific lobar compliance are shown in Table 1. In normal (pre-resection) lobes, distribution of lung inflation from Ptp of 15 to 30 cmH₂O was markedly nonuniform, with the caudal lobes receiving 32–35% of total increase in air volume, while other lobes received 2–11% (Fig. 2). Specific compliance of the caudal lobes was 2.7- to 5.6-fold of that of middle and cranial lobes. As expected, lobar tissue+blood volume did not change significantly with inflation state. FTV ranged from 0.10 to 0.12 at Ptp = 15 cmH₂O and decreased to 0.85–0.95 at Ptp = 30 cmH₂O (Table 1).

Fig. 1.

Three-dimensional high-resolution computed tomography (HRCT) reconstruction of 7 pre-resection (PRE) normal lobes and 3 post-resection (POST) remaining lobes in one animal, shown in 4 orientations.

Table 1.

Lobar volumes and specific compliance

Resection	Ptp, cmH2O	Lobe	Air Volume, ml	Tissue+Blood Volume, ml	FTV	Specific Compliance, ml·(cmH2O·ml)−1
PRE	15	R cranial	440±69	49±7	0.101±0.005†	0.007±0.002
		R middle	240±42	24±4	0.092±0.006‡	0.003±0.002
		R caudal	482±76†	66±11	0.121±0.009†	0.018±0.006
		R infracardiac	178±31	23±4	0.115±0.010†	0.013±0.005
		L cranial	280±47	32±5	0.104±0.005†	0.006±0.003
		L middle	154±46	17±5	0.098±0.005†	0.008±0.003
		L caudal	510±95†	70±12	0.122±0.013†	0.021±0.012
		R lung	1,340±209†	162±25	0.108±0.006†	0.011±0.003
		L lung	944±155†	119±20	0.112±0.008†	0.014±0.007
		Whole lungs	2,284±359†	282±44	0.110±0.006†	0.012±0.004
PRE	30	R cranial	490±88	48±7	0.090±0.005
		R middle	253±51	23±4	0.085±0.005
		R caudal	628±115	65±9	0.094±0.005
		R infra-cardiac	216±41	22±3	0.095±0.008
		L cranial	307±57	31±5	0.092±0.006
		L middle	174±56	16±5	0.087±0.004
		L caudal	679±142	68±11	0.092±0.005
		R lung	1,587±285	159±21	0.092±0.005
		L lung	1,160±234	116±18	0.092±0.005
		Whole lungs	2,747±517	275±39	0.092±0.005
POST	15	R cranial	742±257*	120±20*	0.154±0.058*	0.013±0.011
		L cranial	431±99*	69±13*	0.141±0.031*	0.011±0.002*
		L middle	330±113*	58±22*	0.152±0.028*‡	0.015±0.002*
		Whole lung	1,503±322*	247±24	0.145±0.032*	0.012±0.005
POST	30	R cranial	888±257*	122±15*	0.129±0.043§
		L cranial	521±109*	69±13*	0.120±0.025*
		L middle	429±150*	58±22*	0.121±0.024*
		Whole lung	1,838±345*	248±30	0.122±0.027*

Values are means ± SD. PRE, pre-resection; POST, post-resection; Ptp, transpulmonary pressure; R, right; L, left; FTV, fractional tissue volume.

^*P < 0.05 and

^§P = 0.07 vs. PRE;

^†P < 0.05 and

^‡P = 0.06 vs. 30 cmH₂O.

Fig. 2.

Distribution of inflation volume from 15- to 30-cmH₂O transpulmonary pressure among lobes PRE and POST resection. Values are means ± SD; nos. above bars are mean values of each lobe. P < 0.05 vs. *same lobe PRE resection; †left (L) cranial, ‡L middle, and §right (R) middle lobes at the same point, and ¶L and R caudal lobes vs. all other lobes PRE resection, by factorial ANOVA.

Relative changes were determined from post- to pre-resection ratios (Table 2). The amount of lung removed constituted ∼65% of total pre-resection lung volume at Ptp = 30 cmH₂O. Following resection compared with pre-resection, whole lung air volume decreased by ∼32%, while whole lung tissue volume decreased by ∼10%. Air volume of the remaining lobes increased 1.6- to 2.5-fold. Lobar FTV increased by 30–57% and lobar tissue+blood volume increased 2.2- to 3.6-fold, such that total tissue+blood volume was restored to ∼90% of total pre-resection value. Lobar-specific compliance increased nearly twofold in all remaining lobes, such that whole lung specific compliance was restored to the pre-resection value (0.012 ml·[cmH₂O·ml]⁻¹).

Table 2.

Volume and compliance ratios (POST/PRE resection)

Ptp, cmH2O	Lobe/Control Lobe, or Whole Lung/2 Normal Lungs	Air Volume POST/PRE	Tissue+Blood Volume POST/PRE	FTV POST/PRE	Specific Compliance POST/PRE
15	R cranial	1.75±0.66*	2.49±0.60*	1.55±0.69*	1.95±1.11*
	L cranial	1.59±0.50*	2.16±0.48*	1.37±0.34*	2.08±1.11*
	L middle	2.13±0.32*	3.57±0.87*	1.57±0.31*	2.22±0.71*
	Whole lung	0.67±0.17*	0.89±0.13*	1.33±0.34*	1.21±0.43
30	R cranial	1.88±0.63*	2.58±0.45*	1.45±0.55*
	L cranial	1.75±0.50*	2.26±0.41*	1.30±0.29*
	L middle	2.46±0.36*	3.54±0.79*	1.39±0.27*
	Whole lung	0.68±0.15*	0.91±0.11*	1.34±0.32*

Values are means ± SD.

^*P < 0.05 vs. 1.0 (one-tailed).

Intralobar distribution of FTV is shown along each coordinate axis at two inflation levels pre- and post-resection in Fig. 3. Before resection, significant intralobar FTV gradients were observed in all lobes, with larger magnitudes and gradients in the caudal and infracardiac lobes than the middle or cranial lobes. With increasing Ptp, FTV in most regions declined; the greatest decline was in the caudal and posterior regions of the caudal and infracardiac lobes. Following resection at a given Ptp compared with before, absolute FTV was higher in most regions of the remaining lobes, except at the apex of the cranial lobes and the most lateral region of the left middle lobe; FTV was highest in the posterior-caudal end of the remaining lobes. The normal intralobar gradient of FTV in the cephalad-to-caudal direction was also exaggerated following resection. With increasing Ptp from 15 to 30 cmH₂O, the expected decline in FTV was greater following resection than before in most regions, except at the apical cranial lobes and the most lateral region of the left middle lobe (Fig. 3). Intralobar distribution of specific compliance is shown in Fig. 4. Following resection, specific compliance increased in some regions of the left cranial and middle lobes, but was unchanged in the right cranial lobe.

Fig. 3.

Regional distribution of fractional tissue volume (FTV) along each coordinate axis at 15- and 30-cmH₂O transpulmonary pressures PRE (PRE-15 and PRE-30) and POST (POST-15 and POST-30). Values are means ± SD. P < 0.05: *POST-15 vs. PRE-15 and †POST-30 vs. PRE-30 by repeated-measures ANOVA.

Fig. 4.

Regional distribution of specific lung compliance along each coordinate axis PRE and POST. Values are means ± SD. *P < 0.05 and §P = 0.07 vs. PRE by repeated-measures ANOVA.

All animals developed patchy areas of high FTV, most pronounced in the posterior caudal aspects of the remaining lobes and illustrated in surface (pleural) (Fig. 5) and cross-sectional (Fig. 6) Tissue blocks from subpleural regions corresponding to average or higher than average FTVs (areas A and B, respectively, in Fig. 7) were selected, embedded in glycol methacrylate, sectioned at 4 μm, stained with Masson's trichrome, and examined under a light microscope. The regions of high-surface FTV exhibited pleural thickening, while regions of high-parenchyma FTV showed grossly higher density of alveolar septa, as well as higher septal content of connective tissue compared with regions of average FTV.

Fig. 5.

Three-dimensional surface color maps of FTV distribution are shown in one animal PRE and POST resection. Peripheral regions showed generally increased FTV POST compared with PRE, particularly at the caudal ends of the remaining lobes.

Fig. 6.

HRCT-derived color maps of FTV from the right cranial lobe of one animal PRE and POST resection, showing patchy areas of increased FTV. Areas shown in white in color map represent excluded blood within large blood vessels.

Fig. 7.

Top: representative color FTV map from the right cranial lobe POST resection is compared with the fixed lung cross section obtained at approximately the same location in the same animal. Middle and bottom: histological sections from areas A (average FTV) and B (high FTV) were stained with Masson's trichrome and shown at two magnifications (middle: bar = 200 μm; bottom: bar = 50 μm).

As indexes of spatial heterogeneity, we calculated the CV in FTV within intralobar regions and among lobes. Following resection, the CV of average FTV among lobes at 15-cmH₂O Ptp was significantly lower compared with pre-resection (0.077 vs. 0.114, P < 0.05), indicating greater uniformity among lobes at submaximal inflation; the CV of lobar FTV at full inflation (30-cmH₂O Ptp) was unchanged post-resection compared with pre-resection (0.050 and 0.048, respectively, P > 0.05). Consistent with the development of patchy areas of alveolar septal thickening, CV of intralobar regional FTV gradients changed variably among lobes following resection (Fig. 8).

Fig. 8.

Heterogeneous intralobar regional response to lung resection is assessed from the coefficient of variation (CV) in FTV gradients, shown at transpulmonary pressure of 30 cmH₂O. Each line connects the CV from PRE to POST. *P ≤ 0.05 and †P = 0.07, PRE- vs. POST by repeated-measures ANOVA.

DISCUSSION

Summary of results.

This is the first study to examine in vivo regional anatomical compensation following extensive lung resection. Our laboratory previously used HRCT to follow the increases in average lobar air and tissue (including microvascular blood) volumes in vivo during developmental and compensatory lung growth (18, 19). We now describe a detailed quantitative approach to map the magnitude and distribution of regional lung tissue volume during compensatory lung growth after 65% of lung units are resected. We used natural lobar boundaries to separate functional lung units, allowing comparisons among and within the distorted lobes in the same animals before and after resection, and a simple quantitative approach to analyze the distribution of regional FTV and specific compliance within and among lobes referenced to external three-dimensional coordinates. The major findings are as follows. 1) Significant heterogeneities normally exist in the distributions of air and tissue+blood volumes and specific compliance within and among lobes. 2) Following lung resection, inflation volume was more uniformly distributed among the remaining lobes compared with pre-resection. 3) Lobar tissue+blood volume doubled post-resection, and whole lung tissue plus capillary blood volume was restored to normal, consistent with a) increased perfusion to, and b) robust alveolar-capillary growth in the remaining lobes. 4) FTV increased in all remaining lobes and was highest in the periphery of each remaining lobe, particularly in the posterior-caudal regions. Regions of high FTV exhibited patchy areas of elevated alveolar septal volume density and septal connective tissue content on histology, suggesting that they were sites of exaggerated mechanical strain. These regions of high FTV contributed to intralobar heterogeneity post-resection. Interlobar heterogeneity of average FTV decreased or remained unchanged post-resection at submaximal and maximal inflation, respectively. 5) Specific lung compliance of each remaining lobe doubled following resection, and whole lung-specific compliance was restored to the pre-resection baseline. Thus compensatory lung growth normalized the overall static mechanical properties of alveolar tissue.

Critique of the methods.

Bilateral lung resection was performed in two stages; this approach is associated with no mortality and little morbidity. In contrast, morbidity and mortality of simultaneous bilateral resection removing 65–70% of total lung mass would be prohibitive. During the first-stage resection, only one lobe is removed (up to 24% of lung mass). Expansion of the remaining lobes is modest, and compensatory alveolar growth is not stimulated until significantly more than 42% of lung is removed (10, 19).

Lobar boundary detection programs have been developed by others (20, 25) and by us (18, 19). Since the remaining lobes became grossly enlarged and distorted following resection, we did not attempt to match voxels on serial scans before and after resection. Instead, we divided each lobe into regions according to an external coordinate system, allowing comparison of FTV and compliance among regions, according to their relative positions within the lobe. Without contrast enhancement, HRCT could not distinguish parenchyma tissue from microvascular blood. This fact does not detract from in vivo assessment using HRCT, because functional compensation requires balanced increases in both gas-exchange tissue and capillary blood. Mean specific lung compliance was calculated between Ptp of 15 and 30 cmH₂O, near the top of the pressure-volume curve; hence the absolute value (0.012 ml·[cmH₂O·ml]⁻¹) is lower than that estimated from functional residual capacity, but the present estimates are close to the expected value in the same pressure range in adult canine lungs based on our laboratory's previous studies (0.013 ml·[cmH₂O·ml]⁻¹) (22). Because of the difficulty in reliably separating anatomical boundaries at low lung volumes, we did not scan the lung at functional residual capacity.

The HRCT images and histological sections could not be exactly matched. As much as possible, we matched the lobar orientation for postmortem sectioning to that during HRCT and selected representative HRCT images and fixed tissue surfaces at the same fractional position along the span of a given lobe for qualitative comparison. Areas of higher septal density on the fixed cut surfaces were often grossly discernable.

Intralobar parenchymal heterogeneity.

Significant heterogeneities exist in the distributions of FTV within and among normal canine lobes. In the normal cranial and caudal lobes, FTV was higher in the posterior-dependent regions than in anterior nondependent regions, consistent with morphometric data in canine lung, where alveolar size and surface density were larger in the nondependent than the dependent region, while the width of alveolar capillaries was larger in the dependent region (5–7). Post-resection inflation volume was more uniformly distributed among the remaining lobes than pre-resection, and specific lung compliance was relatively stable within the remaining lobes. However, contrary to our hypothesis, inter- and intralobar heterogeneity remained following extensive lung resection. FTV was highest in the lung periphery, consistent with independent data from our laboratory and others (4, 17) that show greater cell proliferative activities in the peripheral than central lung regions during postnatal, as well as post-PNX, lung growth. Bands of increased FTV also developed particularly in the posterior-caudal regions of the remaining lobes, corresponding to regions of elevated alveolar septal density and grossly increased septal connective tissue content. Since both caudal lobes were surgically removed and the remaining lobes expanded primarily in the caudal direction, the caudal regions likely experienced excessive mechanical strain compared with other regions, which could have induced an adaptive increase in connective tissue content and explains the cephalad-to-caudal gradient of FTV (Fig. 3). Despite the increased septal connective tissue, specific lobar compliance increased in most regions post-resection, an observation consistent with vigorous growth of septal tissue and capillaries.

Interlobar heterogeneity.

In multiple previous cohorts, we observed reproducible volume partition between the left and right normal canine lungs (42–45 and 55–58%, respectively) with modest variability among lobes. In the present cohort, the lobes removed constituted 65% of total lung volume at 30-cmH₂O Ptp, slightly less than that estimated in a previous cohort studied where the same lobes constituted 70% of total lung volume (9). If compensatory lung growth restored the normal alveolar architecture in the expanded remaining lobes, we expect proportional increases in regional air and tissue volume and a normal FTV. If the remaining lobes expanded without alveolar tissue growth, we expect FTV to be reduced. In previous studies following 42% resection, CT-derived tissue volume increased 75–100% in the remaining cranial and infracardiac lobes, but not in other lobes, resulting in a 27% net increase for the remaining right lung. Air volume was up to 120% higher in cranial and infracardiac lobes, but was not increased in other lobes, resulting in a post-resection decline of FTV in all lobes, consistent with the lack of significant overall compensatory alveolar tissue growth (10, 19). Following 58% resection, CT-derived tissue volume increased 95–166% in the remaining lobes, resulting in 118% net increase for the remaining left lung. Lobar air volume increased 87–193%, a net increase of 114% for the remaining lung. As a result of roughly proportional increases in air and tissue volume, lobar FTV changed only mildly (6–8%), consistent with balanced compensatory alveolar growth and restoration of normal alveolar architecture (11, 19). In the present study following 65% resection, the compensatory increase in CT-derived tissue volume (116–250%) exceeded the increase in air volume (59–113%); consequently, the increase in FTV (37–57%) was higher compared with that observed following 58% resection (18, 19). Whole lung tissue volume was restored to 90% of normal, but whole lung air volume was partially restored to ∼68% of normal at a given Ptp. These data suggest that 65% resection elicited more vigorous tissue response than 58% resection, involving the generation of more gas exchange tissue and capillary blood, as well as nongas exchange interstitial components.

We interpreted the consistent development of patchy alveolar septal thickening in all post-resection animals as a response to heterogeneous mechanical strain imposed on the remaining lung tissue. Our laboratory has previously shown a direct correlation between post-PNX lung expansion and alveolar cellular response, growth, and function (13, 14, 24, 26); however, tissue mechanical strain was not directly measured. It is likely that, in some regions, post-resection mechanical strain exceeded that which could be sustained by the remaining connective tissue elements, leading to the generation of more septal connective tissue, which strengthens alveolar walls but also causes septal thickening; the latter change is not expected to benefit gas exchange and may detract from the benefit of alveolar-capillary growth. Regions of alveolar connective tissue accumulation were not observed following 58 or 42% resection, suggesting the likelihood that a limit of anatomical compensation was reached following 65–70% resection, where the need to augment gas exchange tissue and surfaces is balanced by the need to maintain alveolar structural integrity.

In summary, following bilateral resection of 65% of total lung tissue, distinct regional gradients in the intensity of parenchymal response within and among the remaining lobes could be followed by HRCT. While the magnitude of CT-derived indexes of lung growth was greater following 65% than 58% resection, compensatory increases in air and tissue volumes were discordant following 65% resection, suggesting limitation of lobar expansion. Specific lung compliance was normalized, despite the patchy areas of pleural and septal thickening and connective tissue accumulation that developed in the remaining parenchyma in all animals, an observation consistent with adaptation to heterogeneous excessive chronic mechanical strain. We conclude that quantitative analysis of regional HRCT attenuation is a powerful tool for serial evaluation of parenchymal response to resection and during compensation. The approach facilitates directed tissue sampling for histological-radiological correlations in the presence of lobar distortion and heterogeneity and can be readily applied to injury repair models, as well as to human studies.

GRANTS

The research was supported by National Heart, Lung, and Blood Institute Grants R01 HL40070, HL45716, and HL62873.

DISCLAIMER

The contents of this article are solely the responsibility of the authors and do not necessarily represent the official views of the National Heart, Lung, and Blood Institute or of the National Institutes of Health.