Emphysema

Study objectives: To determine the ability of quantitative CT, with special reference to area of emphysema, to predict early postoperative oxygenation capacity and outcome after lung lobectomy for cancer.

Methods: Sixty-two consecutive patients scheduled to undergo lung lobectomy for cancer were enrolled in this study. The area of emphysema (< - 910 Hounsfield units) was measured on a three-dimensional CT lung model. Arterial oxygen saturation (Sa[O.sub.2]) was calculated from Pa[O.sub.2] measured 1 day before and 1 day after surgery with patients at rest breathing room air. A patient was considered to have recovered at the completion of a standardized management regimen.

Results: Postoperative Sa[O.sub.2] (postSa[O.sub.2]) was predicted by the baseline value and the area of emphysema with the use of a regression equation. Ten of the 62 patients (16%) had postoperative cardiopulmonary complications (CPCs). The median time to postoperative recovery was 3 days (range, 1 to 17 days). Predicted postSa[O.sub.2] and predicted postoperative FE[V.sub.1] were shown to be significant independent predictors of postoperative CPCs as well as postoperative recovery time.

Conclusion: Determining the area of emphysema by quantitative CT is useful in predicting early postoperative oxygenation capacity. Predicted oxygenation capacity and predicted ventilatory capacity independently affect perioperativc outcomes. Therefore, using quantitative CT in combination with spirometry may improve risk prediction in patients undergoing lung lobectomy for cancer. However, the role of quantitative CT in grading nonemphysematous lung diseases, such as interstitial lung diseases, must be investigated.

Key words: complication; hypoxemia; low attenuation area; pulmonary, resection; quantitative CT; spirometry

Abbreviations: CPC = cardiopulmonary complication; 3D = three dimensional; HU = Hounsfield units; %LAA = extent of low-attenuation area; postSa[O.sub.2] = postoperative arterial oxygen saturation; ppoFE[V.sub.1] = predicted postoperative FE[V.sub.1]; ppoSa[O.sub.2] = predicted postoperative arterial oxygen saturation; preSa[O.sub.2] = preoperative arterial oxygen saturation. Sa[O.sub.2] = arterial oxygen saturation; VC = vital capacity

**********

Recent advances in computer-assisted quantification of CT images have allowed the accurate identification of morphologie lung abnormalities. This identification is objective and reproducible because the quantified variables reflect the entire lung and can be obtained semiautomatically. (1) Emphysema is the most common underlying lung disease in patients with lung cancer, and it adversely affects perioperative outcome after lung resection. An extent of the abnormally low attenuation area is used widely to determine the severity of emphysema because this variable accurately reflects pulmonary function and exercise capacity. (2) In addition, this variable is superior to pulmonary function tests in detecting early stage emphysema. (3) However, the ability of quantitative CT to aid in identifying selection criteria for surgery remains controversial. The aim of this prospective study was to determine the ability of quantitative CT to predict perioperative outcome after lung lobectomy for cancer. We made predictions of early postoperative oxygenation capacity and then identified predictors of postoperative recovery and postoperative complications.

MATERIALS AND METHODS

Patients

During the period April 2003 through September 2004, 62 consecutive patients scheduled to undergo lung lobectomy for diagnosed or suspected lung malignancy at our institution were enrolled in the study. This study was approved by the institutional review board of the Yamaguchi University School of Medicine. Operability was determined according in the existing guidelines for pulmonary resection. (4) Criteria for resection included PaC[O.sub.2] [less than or equal to] 50 mm Hg, mean pulmonary arterial pressure < 30 mm Hg, and calculated predicted postoperative FE[V.sub.1] (ppoFE[V.sub.1]) > 500 mL. Patient data obtained prior to surgery included age, sex, smoking habits, body mass index, spirometric variables, arterial blood gas values, and extent of the low-attenuation area. Smoking data included pack-years smoked (smoking index, the average number of packages of cigarettes smoked per day multiplied by the number of years smoked). The extent of the low-attenuation area was determined by quantitative CT (described below).

Patient characteristics are summarized in Table 1. Mean smoking index was 39 [+ or -] 38 pack-years (range, 0 to 180 packycars), and mean [+ or -] SD resected segments were 4.1 [+ or -] 1.2. Twenty patients complained preoperatively of slight breathlessness (inability to keep up with healthy persons of equivalent age on hills or stairs (5)), and 17 patients were restricted with respect to physically strenuous activity' (Eastern Cooperative Oncology Group performance status grade 1). (6) Thin-slice high-resolution CT images showed varying degrees of local low attenuation, representative of emphysematous changes, with or without bullae, in 33 patients. Diffuse interstitial lung infiltrates were seen in three patients, manifesting as pulmonary fibrosis in one patient and as idiopathic interstitial pneumonitis in the other two patients. None of the other underlying lung diseases, including pneumoconiosis, old tuberculosis, or chronic bronchitis, were detected clinically before surgery, except in one patient with bronchial asthma.

Preoperative Pulmonary Assessment

Spirometric variables were obtained within the preoperative month (Morgan Transfer Test Model C; PK Morgan Ltd; Kent, England) and included vital capacity (VC) and FE[V.sub.1]. Arterial blood gases (Pa[O.sub.2], PaC[O.sub.2]) were measured within 5 min after sampling in patients resting and breathing room air (BGE; Instrumentation Laboratory: Paris, France). PpoFE[V.sub.1] was calculated using the Nakahara formula (7): ppoFE[V.sub.1] = FE[V.sub.1] (1 - [b - n]/[42 - n]), where n is the number of obstructed subsegments, and b is the total number of removed subsegments. The total number of pulmonary subsegments is 42, with 10 each of the left upper lobe and left lower lobe, 6 of the right upper lobe, 4 of the right middle lobe, and 12 of the right lower lobe. The n value was calculated from the findings of preoperative bronchofibroscopy or CT. VC, FE[V.sub.1], and ppoFE[V.sub.1] were expressed as percentages of predicted values for age, gender, and height.

Assessment of Hypoxemia

Arterial blood gases were studied within the first 24 h after surgery, 20 min after supplementary oxygen was discontinued. Additional oxygen was provided via facemask during the first postoperative night. A decrease in Pa[O.sub.2] is observed in a large majority of patients in the first postoperative days, so we defined hypoxemia on the basis of the Pa[O.sub.2] status within the first 24 postoperative h. In linear regression analysis of postoperative hypoxemia, Pa[O.sub.2] was converted to arterial oxygen saturation (Sa[O.sub.12]) lay means of the Kelman formulas because Sa[O.sub.2] is linearly dependent on arterial oxygen consumption.

CT Technique and Measurement of Low-Attenuation Areas

CT was performed with a Siemens Volume Zoom Scanner (Siemens-Asahi Medical Technologies; Tokyo, Japan) in the helical mode in the absence of intravascular contrast material. With the patient in the supine position, scans were obtained during full inspiration with the following parameters: 120 to 140 kilovolt peak, 280 to 320 mA, 10-mm collimation, and a pitch of 1.5. Scan volumes extended from the thoracic inlet to the lung base and were acquired in one-breathhold periods. Lung images were reconstructed in a 512 x 512 matrix with a lung algorithm. The digital imaging and communications in medicine data of these images were transferred electronically to the teleradiologic workstation, and volume-rendering three-dimensional (3D) models of the lungs were reconstructed with imaging software (M900 QUADRA; ZioSoft K.K.; Osaka, Japan). Threshold limits of -600 to -1,024 Hounsfield units (HU) were applied to segment the lung parenchyma and to exclude soft tissue surrounding the lung and large vessels within the lung. The 3D model was viewed as a shaded surface display at multiple angles to ensure that the model was valid. The trachea, mainstem bronchi, and GI structures were manually and selectively removed from the model using volume-adding and multifusion imaging techniques (Fig 1). The total number of voxels with any selected specific attenuation numbers (in HU) in the lung model could be counted automatically by the computer. The ratio of the number of voxels with attenuation values < -910 HU to the toted numbers of voxels of the entire lung was taken as the extent of the low-attenuation area (%LAA) because low-attenuation thresholds that have been used most often to identify emphysema on conventional 10-mm-thick CT sections are -900 HU or -910 HU. (9-12) On the 3D lung model, the density-masked lung section images, which highlighted lung areas with attenuation values of [less than or equal to] 910 HU in red, could be displayed at any selected orthogonal (transaxial, coronal, or sagittal) planes.

[FIGURE 1 OMITTED]

Postoperative Management

All patients were managed postoperatively as follows. Chest tubes were removed the day after surgery if no air leak was detected, regardless of pleural drainage. Oxygen support was discontinued on the morning after surgery and was reintroduced only if the patient complained of dyspnea or the saturation level was < 95%. IV antibiotic (cefazolin) administration was discontinued on the morning after surgery. Meal intake and ambulation were initiated on the day after surgery, and IV infusion was discontinued when the patient was able to take meals. Urinary catheterization and epidural analgesia were discontinued when the patient was able to walk without assistance. We previously confirmed that this regimen is useful for patients who have undergone pulmonary resection; postoperative recovery is achieved when all these steps are completed. (13)

Cardiopulmonary Complications

Cardiopulmonary complications (CPCs) were defined as follows: pneumonia (body temperature > 38[degrees] for 48 h, purulent sputum production, and infiltrates on chest radiograph), lobar atelectasis, pulmonary embolus, pulmonary edema, need for mechanical ventilation due to ventilatory inefficiency, tracheotomy, air leak for > 7 days, symptomatic arrhythmia requiring drug administration, myocardial ischemia or infarct, and clinical cardiac insufficiency.

Statistical Analysis

Univariate and multivariate regression analyses were used to determine the relation between patient characteristics and postoperative Sa[O.sub.2] (postSa[O.sub.2]). Univariate and multivariate proportional hazards analyses were used to determine the influence of preoperative variables on postoperative recovery. Stepwise variable selection was used in the multivariate analyses. Univariate and multivariate logistic regression analyses were used to determine the risk of postoperative CPCs; p < 0.05 was considered significant.

RESULTS

All patients were extubated postoperatively in the operating room. A total of 13 CPCs occurred in 10 of the 62 patients (16%): ventilatory inefficiency (n = 1), pneumonia (n = 2), lobar atelectasis (n = 2), atrial fibrillation (n = 2), and air leak > 7 days (n = 6). Two of these patients required reoperation: one patient for prolonged air leak on postoperative day 14, and the other patient for completion pneumonectomy on postoperative day 4. The median time to postoperative recovery was 3 days (range, 1 to 17 days).

Correlation between patient characteristic variables and postSa[O.sub.2] are shown in Table 2. By simple regression analysis, postSa[O.sub.2] was shown to be linearly dependent on the smoking index (r = -0.291, p = 0.038), preoperative Sa[O.sub.2] (preSa[O.sub.2]) [r = 0.48, p < 0.001; Fig 2], and %LAA (r = -0.39, p = 0.002; Fig 3). By stepwise multiple regression analysis, preSa[O.sub.2] (F = 30.0) and %LAA (F = 12.8) were identified as the best independent predictors of postSa[O.sub.2]. Predicted postoperative Sa[O.sub.2] (ppoSa[O.sub.2]) was calculated by a regression equation obtained by stepwise multiple regression analysis (regression equation: ppoSa[O.sub.2] = -51 + 1.5 preSa[O.sub.2] - 0.21%LAA). ppoSa[O.sub.2] showed a strong relation to postSa[O.sub.2] (r = 0.64, p < 0.0001; Fig 4), and ppoSa[O.sub.2] was included with the preoperative variables in subsequent analyses.

[FIGURES 2-4 OMITTED]

By univariate proportional hazards analysis, smoking index, VC, FE[V.sub.1] ppoFE[V.sub.1], %LAA, preSa[O.sub.2], and ppoSa[O.sub.2] were identified as significant predictors of postoperative recovery time (p < 0.05 for all; Table 3). By multivariate analysis, ppoFE[V.sub.1] (relative probability, 1.028; p = 0.0439) and ppoSa[O.sub.2] (relative probability, 1.171; p = 0.0097) were shown to be the best independent determinant of postoperative recovery time.

By univariate logistic regression analysis, FE[V.sub.1], ppoFE[V.sub.1], %LAA, and ppoSa[O.sub.2] were shown to be significant risk factors for postoperative CPCs (Table 4). As with prediction of postoperative recovery, ppoFE[V.sub.1] and ppoSa[O.sub.2] were strongly associated with postoperative CPCs. Dependence between ppoSa[O.sub.2] and ppoFE[V.sub.1] is shown in Figure 5. All cases with complications (closed circles) showed ppoFE[V.sub.1] < 60% and ppoSa[O.sub.2] < 93%. Sensitivity and specificity of identifying risk of CPCs are 29% and 100%, respectively, when ppoFE[V.sub.1] < 60% alone is taken into consideration, whereas sensitivity and specificity are 44% and 100%, respectively, when both ppoFE[V.sub.1] < 60% and ppoSa[O.sub.2] < 93% are taken into consideration. By multiple logistic regression analysis, ppoFE[V.sub.1] (odds ratio, 0.866; p = 0.0124) and ppoSa[O.sub.2] (odds ratio, 0.631; p = 0.0157) were independently associated with postoperative CPC.

[FIGURE 5 OMITTED]

Comment

The ability of quantitative CT to predict perioperative outcome of patients undergoing lung lobectomy has not yet been evaluated. Spirometry relies on the detection of increased airway resistance and decreased surface area of the alveolar-capillary membrane. (14,15) This test is relatively insensitive and nonspecific with respect to identifying the cause of airway obstruction. Quantitative CT is an objective technique that uses computer software to distinguish voxels with abnormally low attenuation, representing emphysema, from those representing normal lung. Because emphysema is defined by pathologic criteria as abnormal permanent enlargement of air spaces distal to the terminal bronchioles, accompanied by destruction of the alveolar walls and without obvious fibrosis, (16) the severity of emphysema may be predicted as well by the extent of the low-attenuation area as by spirometry.

Loss of respiratory function (ventilation, gas exchange, and perfusion) induced by pulmonary resection frequently leads to CPCs if the functional reserve is limited. Therefore, prediction of postoperative respiratory function is clinically valuable for the assessment of risk prior to lung resection. Postoperative residual capacity for ventilation, as well as that for perfusion, can be predicted by the baseline capacity and the extent of lung resection. In contrast, postoperative capacity for gas exchange is usually not affected by the extent of lung resection, unless the resection is excessive, but is affected by the underlying pulmonary disease. (17,18) Is Postoperative gas exchange abnormalities are often attributed to general anesthesia, which temporarily decreases functional residual volume by reducing the muscle tone of the chest wall. (19) In patients with underlying lung disease, this impairment results in intrapulmonary shunting or reduction of the ventilation/perfusion ratio by the closure of small airways. (19) In the present study, we found that early postoperative loss of gas exchange capacity was affected by %LAA, as quantified by 3D CT. The low-attenuation area, corresponding to overinflated lung, may also contribute to reduce functional lung volume induced by general anesthesia because such diseased lung remains inflated even postoperatively due to air trapping. With the use of %LAA, a regression equation for postSa[O.sub.2] was generated by stepwise regression analysis. The ppoSa[O.sub.2] and ppoFE[V.sub.1] were found to independently affect postoperative CPCs. For instance, 10 of 34 patients (29%) with ppoFE[V.sub.1] < 60% had CPCs, whereas 10 of 23 patients (44%) with ppoFE[V.sub.1] < 60% and ppoSa[O.sub.2] < 93% had CPCs (Fig 5). These results suggest that quantitative CT in combination with spirometry may improve risk prediction for patients undergoing lung lobectomy for cancer.

Regardless of existing definitions for postoperative CPCs, it can be difficult to accurately detect CPCs induced by lung resection surgery because some respiratory abnormalities are likely to be overlooked in patients with preserved pulmonary function, whereas minimal impairment can be exacerbated in patients with limited pulmonary function. In our series, 84% of the patients did not experience defined postoperative CPCs, although some may have experienced subclinical CPCs that resulted in prolonged hospitalization. To assess perioperative outcomes in these patients, we measured postoperative recovery time. As we have reported, (13) postoperative recovery time was a more accurate reflection of the perioperative physical status than postoperative hospital stay because patients in Japan elect to remain in the hospital long after recovery from surgery. As we expected, ppoSa[O.sub.2[ and ppoFE[V.sub.1] were the best determinants of postoperative recovery by multivariate analysis. This further supports the use of quantitative CT in risk assessment prior to lung lobectomy.

The consecutive patients undergoing lung resection for cancer included those with varying degrees of emphysema and those without any evidence of emphysema. Because the %LAA value in the non-emphysematous lung is close to zero, the %LAA may not play a part in predicting postoperative hypoxemia in patients without any evidence of lung emphysema. For example, in our series, the %LAA was predictive of postoperative hypoxemia in patients with a smoking history but not in those without a smoking history. However, the regression equation for postSa[O.sub.2] yielded accurate prediction of the values both in patients with and those without a smoking history. The correlation coefficient (r) for patients with a smoking history was similar to that for patients without a smoking history (0.68 vs 0.64, respectively). These results suggest that the regression equation defined in the present study may be applicable for patients scheduled to undergo lung lobectomy for cancer, with or without emphysema.

It should be noted that the %LAA does not reflect the severity of interstitial lung disease that occasionally causes postoperative hypoxemia because the proliferating parenchyma in this disease usually enhances lung density. In our series, the %LAA in all three patients with diffuse interstitial lung disease was < 4%, and their preSa[O.sub.2] and postSa[O.sub.2] levels (94.7 [+ or -] 0.5% and 88.1 [+ or -] 6.6%, respectively) were similar to those of patients with a %LAA < 4% but without any underlying lung disease (95.8 [+ or -] 1.2% and 89.4 [+ or -] 4.6%; both p > 0.1). However, it remains controversial whether the present regression equation for postSa[O.sub.2[ is applicable to patients with more progressing interstitial lung diseases than ours. Further effort is needed to clarify the role of quantitative CT in grading interstitial lung diseases and predicting postoperative complications, such as acute exacerbation.

CT provides transverse anatomic images in which the value of each picture element (pixel) corresponds to the radiographic attenuation of a defined volume of tissue (voxel). With software that is available with most modern scanners, voxel attenuation values can now be measured without the need to transfer data to other computer systems, and 3D lung models can be generated with volumetric data acquired in the helical mode. The lungs are easily differentiated from soft tissue structures and from structures with lower attenuation, such as the trachea, mainstem bronchi, and esophagus, and minimal postprocessing of the shaded surface display is required. With this model, the proportion of voxels with attenuation values within the range of that representing emphysema is readily determined by moving a boundary line to a defined threshold. %LAA can be obtained within a few minutes without need for extensive technical training. The results of quantitative CT are reproducible across viewers of varying expertise and experience and across institutions, allowing for accurate comparisons of results among different centers. (2) Computer-assisted quantification of low-attenuation area is thus a time-saving and cost-saving method for objective and reproducible risk assessment prior to lung lobectomy.

CONCLUSION

Determining the %LAA by quantitative CT is valuable for predicting early postoperative oxygenation capacity in patients undergoing lung lobectomy for cancer. Predicted postoperative oxygenation capacity and predicted postoperative ventilatory capacity independently affect postoperative cardiopulmonary complications as well as postoperative recovery time. Therefore, using quantitative CT in combination with spirometry may improve risk prediction in patients undergoing lung lobectomy for cancer. However, the role of quantitative CT in grading nonemphysematous lung diseases, such as interstitial lung diseases, must be investigated.

REFERENCES

(1) Park KJ, Bergin CJ, Clansen JL. Quantitation of emphysema with three-dimensional CT densitometry: comparison with two-dimensional analysis, visual emphysema scores, and pulmonary function test results. Radiology 1999; 211:541-547

(2) Madani A, Keyzer C, Gevenois PA. Quantitative computed tomography assessment of lung structure and function in pulmonary emphysema. Eur Respir J 2001; 18:720-730

(3) Sanders C, Nath PH, Bailey WC. Detection of emphysema with computed tomography: correlation with pulmonary function tests and chest radiography. Invest Radiol 1988; 23:262-266

(4) Celli BR. What is the value of preoperative pulmonary function testing? Med Clin North Am 1993; 77:309-325

(5) Stoller JK, Ferranti R, Feinstein AB. Further specification and evaluation of a new clinical index for dyspnea. Am Rev Respir Dis 1986; 134:1129-1134

(6) 0ken MM, Creech RH, Tormey DC, et al. Toxicity and response criteria of the Eastern Cooperative Ontology Group. Am J Clin Oncol 1982; 5:649-655

(7) Nakahara K, Monden Y, Ohno K, et al. A method for predicting postoperative lung function and its relation to postoperative complications in patients with lung cancer. Ann Thorac Surg 1985; 39:260-265

(8) Kelman CR. Digital computer subroutine for the conversion of oxygen tension into saturation. J Appl Physiol 1966; 21:1375-1376

(9) Park KJ, Bergin CJ, Clausen JL. Quantitation of emphysema with three-dimensional CT densitometry: comparison with two-dimensional analysis, visual emphysema scores, and pulmonary function test results. Radiology 1999; 211:541-547

(10) Kinsella M, Muller NL, Abboud RT, et al. Quantitation of emphysema by computed tomography using a "density mask" program and correlation with pulmonary function tests. Chest 1990; 97:315-321

(11) Lamers RJ, Thelissen CR, Kessels AG, et al. Chronic obstructive pulmonary disease: evaluation with spirometrically controlled CT lung densitometry. Radiology 1994; 193:109-113

(12) Knudson RJ, Standen JR, Kaltenborn WT, et al. Expiratory computed tomography for assessment of suspected pulmonary emphysema. Chest 1991; 99:1357-1366

(13) Ueda K, Kaneda Y, Sakano H, et al. Obstacles for shortening hospitalization after video-assisted pulmonary resection for lung cancer. Ann Thorac Surg 2003; 76:1816-1820

(14) Thurlbeck MW, Henderson JA, Fraser RG, et al. Chronic obstructive lung disease: a comparison between clinical, roentgenologic, functional and morphological criteria in chronic bronchitis, emphysema, asthma and bronchiectasis. Medicine (Baltimore) 1970:49:81-145

(15) Berend N, Woolcock AJ, Martin GE. Correlation between the function and structure of the lung in smokers. Am Rev Respir Dis 1979; 119:695-705

(16) American Thoracic Society. Chronic bronchitis, asthma, and pulmonary emphysema by the Committee on Diagnostic Standards for Nontuberculosis Respiratory Disease. Am Rev Respir Dis 1962; 85:762-768

(17) Mlezoch J, Zutter W, Keller R, et al. Influence of lung resection on pulmonary circulation and lung function at rest and on exercise. Respiration 1975; 32:424-435

(18) Bolliger CT, Jordan P, Soler M, et al. Pulmonary function and exercise capacity after lung resection. Eur Respir J 1996; 9:415-421

(19) Jones JG, Sapsford DJ, Wheatley RG. Postoperative hypoxaemia: mechanisms and time course. Anaesthesia 1990; 45:566-573

* From the First Department of Surgery, (Drs. Ueda, Kaneda, Sudoh, Mitsutaka, and Hamano) and the Department of Radiology (Drs. Tanaka and Suga), Yamaguchi University School of Medicine, Ube Yamaguchi, Japan.

Manuscript received March 29, 2005; revision accepted June 6, 2005.

Reproduction of this article is prohibited without written permission from the American College of Chest Physicians (www.chestjournal. Org/misc/reprints.shtml).

Correspondence to: Kazuhiro Ueda, MD, First Department of Surgery, Yamaguchi University School of Medicine, 1-1-1 Minami-Kogushi, Ube Yamaguchi 755-8505, Japan; e-mail: kaueda@c-able.ne.jp

COPYRIGHT 2005 American College of Chest Physicians
COPYRIGHT 2005 Gale Group