- Email: [email protected]
Accuracy of Portable Chest Radiography in the Critical Care Setting
Accuracy of Portable Chest Radiography in the Critical Care Setting* Diagnosis of Pneumonia Based on Quantitative Cultures Obtained From Protected Brush Catheter Michael S. Lefcoe, M .D.; George A Fox, M.D., F.C.C.P.; David]. Leasa, M .D., F.C.C.P.; R. Keith Sparrow , M.D.; and David G. McConnack, M.D. Sixty-six supine portable chest radiographs done on the day of bronchoscopy in 62 critical care unit patients suspected of having pneumonia were examined in a blinded fashion by two radiologists. Quantitative culture results obtained from protected brush catheter (PBC) specimens were compared with chest radiograph scores. For one observer, the sensitivity of the chest radiograph for predicting the presence of positive culture results was 0.60, specificity was 0.29, overall agreement was 0.41, positive predictive value was 0.34, and negative predictive value was 0.55. For the second observer, the values were as follows: sensitivity, 0.64; specificity, 0.27; overall agreement,
N osoeomial bacterial pneumonia in the critical care setting continues to complicate the course of 7 to 41 percent of patients receiving continuous mechanical ventilation . The mortality rate is from 50 to 80 percent. 1 The precise diagnosis of pneumonia in critically ill patients is very difficult and previous studies have suffered from a lack of a consistent gold standard. The usual clinical criteria for pneumonia such as fever, pulmonary infiltrates, purulent sputum, and leukocytosis are not reliable in mechanically ventilated patients. Oropharyngeal and endotracheal tube colonization with pathogenic bacteria can occur within 24 h of intubation and therefore decreases the reliability of expectorated or aspirated tracheal secretions for diagnosis oflower respiratory tract infections. 2 In addition, elevated body temperature and leukocytosis are common with other conditions in critically ill patients. Assessment of the value of chest radiographic findings in diagnosing pneumonia has also suffered from lack of a gold standard. Patients in critical care settings often have pulmonary edema, secretion retention with atelectasis, or pulmonary emboli that make assessment of the significance of radiographic opacities difficult. Wimberley and associates 3 in 1979 introduced a fiberoptic bronchoscopic technique using a plugged *From the Victoria Hospital Corporation, and University Hospital, London, Ontario, Canada Manuscript received May 18. 1993: revision accepted August 19. Repriut requests: Dr. Le{coe, Victoria Hospital. :175 Soutfl Street , Louclou . Outario. Cauacla N6A 4G.S
0.41; positive predictive value, 0.35; and negative predictive value, 0.55. The kappa statistic was calculated at 0.27 indicating marginal interobserver reproducibility. We conclude the portable chest radiograph in the critical care setting is not accurate in predicting the presence of pneumonia when the diagnosis is based on quantitative cultures obtained from protected brush catheter specimens. (Cheat 1994; 105: 885-87) CFU =colony-forming units; ICU PBC z protected bruSh catheter
=intensive
care unit;
double-sheath catheter system to obtain lower respiratory tract secretions uncontaminated by organisms that colonize the upper airways. Further modifications have led to the application of quantitative culture techniques to the protected brush catheter (PBC) specimen. These quantitative bacterial culture techniques have increased the accuracy of diagnosing nosocomial pneumonia by identifying significant bacterial counts(> 1()3 colony forming units [CFU]/ml) in specimens isolated from the lower respiratory tract secretions obtained from bronchoscopy. •-• The validity of this method as a gold standard for the diagnosis of pneumonia is supported by clinical studies in the literature.ll- 11 Chastre and associates 12 obtained PBC specimens from patients shortly after death and verified their results with histologic samples of lung obtained at post mortem examination. There was excellent correlation between cultures of the lung tissue and the protected brush. Further, all patients with histologic evidence of pneumonia had one or more organisms isolated in concentrations of greater than 1
885
Table 1 -Radiographic Categories Suggestive of Pneumonia
Compatible With Pneumonia
Low Probability of Pneumonia
Cavitation Localized fluffy alveolar opaeities without loss of volume Lobar consolidation without loss of volume
Diffuse airspace disease in both lungs Localized fluffy alveolar opaeities with some loss of volume Localized ill-defined pulmonary markings without much volume loss Diffuse interstitial reticulonodular pattern
Pulmonary edema pattern with Kerley-B lines Line shadows
Localized fluf(v alveolar opaeities with pleural effusion
Lobar collapse or segmental collapse Pleural fluid alone Normal lung
of pneumonia" based on radiographic appearance (Table 1).
METHODS
Sixty-six bronchoscopies were performed over an 8-montb period on 62 patients. Patients' ages ranged from l7 to 83 years with the mean age of 65 years. The reference population was all patients admitted to the critical care unit at Victoria Hospital, and the ICU at University Hospital in London, Ontario. Indusion criteria were as follows: (I) patients who were ventilated for greater than 48 h: (2) chest radiographs interpreted by the dinical staff showing pulmonary infiltrates that were new or persisted greater than 24 h: (3) the presence of macroscopic purulent tracheal aspirates: and (4) patients older than 18 years of age. Patients were en·h1ded from the protocol if they ( 1) bad a condition not permitting bronchoscopy. (2) bad a dinical diagnosis of pneumonia on admission to the !CU. (3) were immunocompromised (patients with organ transplants, hematologic malignancy. tumor, or cellular immune deficiency). (4) had gross aspiration as a cause of the pneumonia, or (5) were receiving more than 0.5 mglkg per day of prednisone or its equivalent. Informed consent was obtained from the patient or the next of kin and upon entry into the study, all patients bad lower respiratory tract secretions obtained through flexible bronchoscopy with PBC within 12 h of entry into the study. Specimens were immediately transported to the Microbiology Laboratory and quantitatively cultured on MacConkey agar, 5 percent horse blood agar. Identification of pathogenic bacteria and antibiotic sensitivities were determined according to routine laboratory protocol. Cultures were considered positive if> 103 CFU/ml of one or more potentially pathogenic bacteria were isolated from the PBC. A portable chest radiograph was obtained in all patients on the day of the bronchoscopy prior to the bronchoscopy. The radiographs were done with the patient in a supine position using portable x-ray equipment (average technique 85 kv. 1.5 to 2 mas. Kodak TML film. Lanex regular s~·reens). The radiographs were independently examined in a blinded fashion by two radiologists (K.S., M.L.) and the film was dassified into categories "suggestive for pneumonia," "compatible with pneumonia." or "low probability
The results are not particularly surprising in view of the difficulties involved in radiographic interpretation of portable chest radiographs in the critical care
Table 2 -Portable Chest Radiograph Compared With Protected Brosh Culture Results in Critical Care Patients: Observer (K.S.): n = 66 Observations on 66 Bronchoscopies on 62 Patients*
Table 3 -Portable Chest Radiograph Compared With Protected Brush Culture Results in Critical Care Patients: Observer B (M.L.): n = 66 Observations on 66 Bronchoscopies on 62 Patients*
Chest Radiograph Findings
PBC-VE
PBC+VE
Low probability of pneumonia Suggestive of pneumonia Compatible with pneumonia
12 18 11
10 8
*Sensitivity negative.
886
= 0.60:
Truth
specificity
= 0.29:
+ VE
- VE
A comparison of the chest radiologic category with the PBC results in the 62 patients for the two observers is shown in Tables 2 and 3. Using both the suggestive and compatible radiographs as positive for pneumonia, and the low-probability radiographs as negative for pneumonia, sensitivity for observer A (K.S.) is 0.66, specificity is 0.29, overall agreement is 0.41, positive predictive value is 0.34, and negative predictive value is 0.55. For observer B (M.L.), the values are as follows: sensitivity, 0.64; specificity, 0.27; overall agreement, 0.41; positive predictive value, 0.35; and negative predictive value, 0.55. Interobserver reproducibility is illustrated in Table 4 where the kappa statistic was calculated at 0.27 indicating marginal reproducibility between the observers. Only one patient had an autopsy within 1 month of the bronchoscopy and this patient had the autopsy the day following the bronchoscopy. The radiograph was suggestive of pneumonia and the PBC culture was positive. The pathologic findings confirmed the presence of bilateral bronchopneumonia. DISCUSSI0:-1
Chest Radiograph Findings
Truth PBC-VE
PBC+VE
11
9 3 13
Low probability of pneumonia Suggestive of pneumonia Compatible with pneumonia
7
= positive:
RESULTS
=
*Sensitivity negative.
= 0.64:
specifieity
10
20
= 0.27:
+ VE
= positive:
- VE
=
Accuracy of Portable Chest Radiography in Critical Care (Lefcoe eta/)
Table 4 -lnterobserver Reproducibility•
diagnosis is based on a gold standard of quantitative culture of PBC specimens.
Observer B (M.L.) Obst•rver A (K.S.) Low Probabilitv of Pneumonia Suggestive of Pneumonia Compatible With Pneumonia
Low Probability of Pneumonia
Suggest in· of Pneumonia
Compatible With Pneumonia
10
2
10
22
6
10
10
26
13
18
33
66
4
20
13
REFERENCES
*Kappa= 0.27.
setting. False positives on radiography are probably related to secretion retention that occasionally occurs without much loss of volume and produces fluffy alveolar opacities. Atypical pulmonary edema due to underl}ing emphysema or asymmetrical clearing of pulmonary edema also causes difficulties. It is possible that following the radiographs over a period of days might he of more value than just reading a single radiograph on the date of the bronchoscopy and might increase accuracy although we have not done this in our study. False negatives are likely related technical difficulties in assessing all areas of the lung on a portable supine radiograph. Usng chest computed tomography as the gold standard, Beydon et al 13 recently showed that the sensitivity of the portable chest radiograph in detecting lung consolidations was between 0.33 and 1.00 depending on the lung zone considered. One case in our study had a completely normal portable radiograph and a protected brush culture was quantitatively abnormal. One possible explanation for this situation is that in the very early stages of pneumonia, the presence of organisms in the lung parenchyma may not result in sufficient alveolar exudation to cause a radiographic opacity. Co:-~ct.usioNs
We conclude that the portable chest radiograph in the critical care setting is not accurate in predicting the presence of nosocomial pneumonia when the
2
3
4
5
6
7
8
9
10
11
12
13
Fagon JY. Chastre J. Domart Y, Trouillet JL. Pierre J. Dame C, et al. Nosocomial pneumonia in patients receiving continuous mechanical ventilation. Am Rev Respir Dis 1989; 139:877-84 Barrett-Connor E. The non-value of sputum culture in the diagnosis of pneumococcal pneumonia. Am Rev Respir Dis 1991; 103:845-48 Wimberley N, Faling LJ, Bartlett JG. Afiberoptic bronchoscopy technique to obtain uncontaminated lower airway secretions for bacterial culture. Am Rev Respir Dis 1979; 119:337-43 Fagon JY, Chastre J, Hance AJ, Guiguet M, Trouillet JL, Domart Y, et al. Detection of nosocomial lung infection in ventilated patients. Am Rev Respir Dis 1988; 138:110-16 Bartlett JG, Finegold SM . Bacteriology of expectorated sputum with quantitative culture and wash technique compared to transtracheal aspirated. Am Rev Respir Dis 1978; 117:1019-27 Pollock HM. Hawkins EL, Bonner JR. Sparkman T, Bass JB Jr. Diagnosis of bacterial pulmonary infections with quantitative protected catheter cultures obtained during bronchoscopy. J Clin Microbiol1983; 17:255-59 Meduri GU, Wunderink RG, Leeper KV. Management of bacterial pneumonia in ventilated patients-protected bronchoalveolar lavage as a diagnostic tool. Clin lnv In Crit Care 1992; 101:500-08 Wimberley NW, Bass JB Jr. Boyd BW, Kirkpatrick MB, Serio RA, Pollock HM. Use of a bronchoscopic protected catheter brush for the diagnosis of pulmonary infections. Chest 1982; 81 :556-62 Joshi JH, Want KP, DeJongh CA. Newman KA, Wiemik PH, Schimpff SC. A comparative evaluation of two fiberoptic bronchoscopy catheters: the plugged telescoping catheter versus the single sheathed nonplugged catheter. Am Rev Respir Dis 1982; 126:860-63 Bordelon JY Jr. Legrand P, Gewin WC, Sanders CV. The telescoping plugged catheter in suspected anaerobic infections. Am Rev Respir Dis 1983; 128:465-68 Moser KM, Maurer J. Jassy L, Kremsdorf R, Konopka R, Shure D, et al. Sensitivity. specificity, and risk of diagnostic procedures in a canine model of Streptococcus pneumoniae pneumonia. Am Rev Respir Dis 1982; 125:436-42 Chastre J. Viau F, Brun P, Pierre J. Dauge MC. Bouchama A, et al. Prospective evaluation of the protected specimen brush for the diagnosis of pulmonary infections in ventilated patients. Am Rev Respir Dis 1984; 130:924-29 Beydon L, Saada M, Liu N, Becquemin JP, Harf A, Bonnet F, et al. Can portable chest x-ray examination accurately diagnose lung consolidation after major abdominal surgery? a <:omparison with computed tomography scan. Chest 1992; 102:16971703
CHEST I 105 I 3 I MARCH, 1994
887
N osoeomial bacterial pneumonia in the critical care setting continues to complicate the course of 7 to 41 percent of patients receiving continuous mechanical ventilation . The mortality rate is from 50 to 80 percent. 1 The precise diagnosis of pneumonia in critically ill patients is very difficult and previous studies have suffered from a lack of a consistent gold standard. The usual clinical criteria for pneumonia such as fever, pulmonary infiltrates, purulent sputum, and leukocytosis are not reliable in mechanically ventilated patients. Oropharyngeal and endotracheal tube colonization with pathogenic bacteria can occur within 24 h of intubation and therefore decreases the reliability of expectorated or aspirated tracheal secretions for diagnosis oflower respiratory tract infections. 2 In addition, elevated body temperature and leukocytosis are common with other conditions in critically ill patients. Assessment of the value of chest radiographic findings in diagnosing pneumonia has also suffered from lack of a gold standard. Patients in critical care settings often have pulmonary edema, secretion retention with atelectasis, or pulmonary emboli that make assessment of the significance of radiographic opacities difficult. Wimberley and associates 3 in 1979 introduced a fiberoptic bronchoscopic technique using a plugged *From the Victoria Hospital Corporation, and University Hospital, London, Ontario, Canada Manuscript received May 18. 1993: revision accepted August 19. Repriut requests: Dr. Le{coe, Victoria Hospital. :175 Soutfl Street , Louclou . Outario. Cauacla N6A 4G.S
0.41; positive predictive value, 0.35; and negative predictive value, 0.55. The kappa statistic was calculated at 0.27 indicating marginal interobserver reproducibility. We conclude the portable chest radiograph in the critical care setting is not accurate in predicting the presence of pneumonia when the diagnosis is based on quantitative cultures obtained from protected brush catheter specimens. (Cheat 1994; 105: 885-87) CFU =colony-forming units; ICU PBC z protected bruSh catheter
=intensive
care unit;
double-sheath catheter system to obtain lower respiratory tract secretions uncontaminated by organisms that colonize the upper airways. Further modifications have led to the application of quantitative culture techniques to the protected brush catheter (PBC) specimen. These quantitative bacterial culture techniques have increased the accuracy of diagnosing nosocomial pneumonia by identifying significant bacterial counts(> 1()3 colony forming units [CFU]/ml) in specimens isolated from the lower respiratory tract secretions obtained from bronchoscopy. •-• The validity of this method as a gold standard for the diagnosis of pneumonia is supported by clinical studies in the literature.ll- 11 Chastre and associates 12 obtained PBC specimens from patients shortly after death and verified their results with histologic samples of lung obtained at post mortem examination. There was excellent correlation between cultures of the lung tissue and the protected brush. Further, all patients with histologic evidence of pneumonia had one or more organisms isolated in concentrations of greater than 1
885
Table 1 -Radiographic Categories Suggestive of Pneumonia
Compatible With Pneumonia
Low Probability of Pneumonia
Cavitation Localized fluffy alveolar opaeities without loss of volume Lobar consolidation without loss of volume
Diffuse airspace disease in both lungs Localized fluffy alveolar opaeities with some loss of volume Localized ill-defined pulmonary markings without much volume loss Diffuse interstitial reticulonodular pattern
Pulmonary edema pattern with Kerley-B lines Line shadows
Localized fluf(v alveolar opaeities with pleural effusion
Lobar collapse or segmental collapse Pleural fluid alone Normal lung
of pneumonia" based on radiographic appearance (Table 1).
METHODS
Sixty-six bronchoscopies were performed over an 8-montb period on 62 patients. Patients' ages ranged from l7 to 83 years with the mean age of 65 years. The reference population was all patients admitted to the critical care unit at Victoria Hospital, and the ICU at University Hospital in London, Ontario. Indusion criteria were as follows: (I) patients who were ventilated for greater than 48 h: (2) chest radiographs interpreted by the dinical staff showing pulmonary infiltrates that were new or persisted greater than 24 h: (3) the presence of macroscopic purulent tracheal aspirates: and (4) patients older than 18 years of age. Patients were en·h1ded from the protocol if they ( 1) bad a condition not permitting bronchoscopy. (2) bad a dinical diagnosis of pneumonia on admission to the !CU. (3) were immunocompromised (patients with organ transplants, hematologic malignancy. tumor, or cellular immune deficiency). (4) had gross aspiration as a cause of the pneumonia, or (5) were receiving more than 0.5 mglkg per day of prednisone or its equivalent. Informed consent was obtained from the patient or the next of kin and upon entry into the study, all patients bad lower respiratory tract secretions obtained through flexible bronchoscopy with PBC within 12 h of entry into the study. Specimens were immediately transported to the Microbiology Laboratory and quantitatively cultured on MacConkey agar, 5 percent horse blood agar. Identification of pathogenic bacteria and antibiotic sensitivities were determined according to routine laboratory protocol. Cultures were considered positive if> 103 CFU/ml of one or more potentially pathogenic bacteria were isolated from the PBC. A portable chest radiograph was obtained in all patients on the day of the bronchoscopy prior to the bronchoscopy. The radiographs were done with the patient in a supine position using portable x-ray equipment (average technique 85 kv. 1.5 to 2 mas. Kodak TML film. Lanex regular s~·reens). The radiographs were independently examined in a blinded fashion by two radiologists (K.S., M.L.) and the film was dassified into categories "suggestive for pneumonia," "compatible with pneumonia." or "low probability
The results are not particularly surprising in view of the difficulties involved in radiographic interpretation of portable chest radiographs in the critical care
Table 2 -Portable Chest Radiograph Compared With Protected Brosh Culture Results in Critical Care Patients: Observer (K.S.): n = 66 Observations on 66 Bronchoscopies on 62 Patients*
Table 3 -Portable Chest Radiograph Compared With Protected Brush Culture Results in Critical Care Patients: Observer B (M.L.): n = 66 Observations on 66 Bronchoscopies on 62 Patients*
Chest Radiograph Findings
PBC-VE
PBC+VE
Low probability of pneumonia Suggestive of pneumonia Compatible with pneumonia
12 18 11
10 8
*Sensitivity negative.
886
= 0.60:
Truth
specificity
= 0.29:
+ VE
- VE
A comparison of the chest radiologic category with the PBC results in the 62 patients for the two observers is shown in Tables 2 and 3. Using both the suggestive and compatible radiographs as positive for pneumonia, and the low-probability radiographs as negative for pneumonia, sensitivity for observer A (K.S.) is 0.66, specificity is 0.29, overall agreement is 0.41, positive predictive value is 0.34, and negative predictive value is 0.55. For observer B (M.L.), the values are as follows: sensitivity, 0.64; specificity, 0.27; overall agreement, 0.41; positive predictive value, 0.35; and negative predictive value, 0.55. Interobserver reproducibility is illustrated in Table 4 where the kappa statistic was calculated at 0.27 indicating marginal reproducibility between the observers. Only one patient had an autopsy within 1 month of the bronchoscopy and this patient had the autopsy the day following the bronchoscopy. The radiograph was suggestive of pneumonia and the PBC culture was positive. The pathologic findings confirmed the presence of bilateral bronchopneumonia. DISCUSSI0:-1
Chest Radiograph Findings
Truth PBC-VE
PBC+VE
11
9 3 13
Low probability of pneumonia Suggestive of pneumonia Compatible with pneumonia
7
= positive:
RESULTS
=
*Sensitivity negative.
= 0.64:
specifieity
10
20
= 0.27:
+ VE
= positive:
- VE
=
Accuracy of Portable Chest Radiography in Critical Care (Lefcoe eta/)
Table 4 -lnterobserver Reproducibility•
diagnosis is based on a gold standard of quantitative culture of PBC specimens.
Observer B (M.L.) Obst•rver A (K.S.) Low Probabilitv of Pneumonia Suggestive of Pneumonia Compatible With Pneumonia
Low Probability of Pneumonia
Suggest in· of Pneumonia
Compatible With Pneumonia
10
2
10
22
6
10
10
26
13
18
33
66
4
20
13
REFERENCES
*Kappa= 0.27.
setting. False positives on radiography are probably related to secretion retention that occasionally occurs without much loss of volume and produces fluffy alveolar opacities. Atypical pulmonary edema due to underl}ing emphysema or asymmetrical clearing of pulmonary edema also causes difficulties. It is possible that following the radiographs over a period of days might he of more value than just reading a single radiograph on the date of the bronchoscopy and might increase accuracy although we have not done this in our study. False negatives are likely related technical difficulties in assessing all areas of the lung on a portable supine radiograph. Usng chest computed tomography as the gold standard, Beydon et al 13 recently showed that the sensitivity of the portable chest radiograph in detecting lung consolidations was between 0.33 and 1.00 depending on the lung zone considered. One case in our study had a completely normal portable radiograph and a protected brush culture was quantitatively abnormal. One possible explanation for this situation is that in the very early stages of pneumonia, the presence of organisms in the lung parenchyma may not result in sufficient alveolar exudation to cause a radiographic opacity. Co:-~ct.usioNs
We conclude that the portable chest radiograph in the critical care setting is not accurate in predicting the presence of nosocomial pneumonia when the
2
3
4
5
6
7
8
9
10
11
12
13
Fagon JY. Chastre J. Domart Y, Trouillet JL. Pierre J. Dame C, et al. Nosocomial pneumonia in patients receiving continuous mechanical ventilation. Am Rev Respir Dis 1989; 139:877-84 Barrett-Connor E. The non-value of sputum culture in the diagnosis of pneumococcal pneumonia. Am Rev Respir Dis 1991; 103:845-48 Wimberley N, Faling LJ, Bartlett JG. Afiberoptic bronchoscopy technique to obtain uncontaminated lower airway secretions for bacterial culture. Am Rev Respir Dis 1979; 119:337-43 Fagon JY, Chastre J, Hance AJ, Guiguet M, Trouillet JL, Domart Y, et al. Detection of nosocomial lung infection in ventilated patients. Am Rev Respir Dis 1988; 138:110-16 Bartlett JG, Finegold SM . Bacteriology of expectorated sputum with quantitative culture and wash technique compared to transtracheal aspirated. Am Rev Respir Dis 1978; 117:1019-27 Pollock HM. Hawkins EL, Bonner JR. Sparkman T, Bass JB Jr. Diagnosis of bacterial pulmonary infections with quantitative protected catheter cultures obtained during bronchoscopy. J Clin Microbiol1983; 17:255-59 Meduri GU, Wunderink RG, Leeper KV. Management of bacterial pneumonia in ventilated patients-protected bronchoalveolar lavage as a diagnostic tool. Clin lnv In Crit Care 1992; 101:500-08 Wimberley NW, Bass JB Jr. Boyd BW, Kirkpatrick MB, Serio RA, Pollock HM. Use of a bronchoscopic protected catheter brush for the diagnosis of pulmonary infections. Chest 1982; 81 :556-62 Joshi JH, Want KP, DeJongh CA. Newman KA, Wiemik PH, Schimpff SC. A comparative evaluation of two fiberoptic bronchoscopy catheters: the plugged telescoping catheter versus the single sheathed nonplugged catheter. Am Rev Respir Dis 1982; 126:860-63 Bordelon JY Jr. Legrand P, Gewin WC, Sanders CV. The telescoping plugged catheter in suspected anaerobic infections. Am Rev Respir Dis 1983; 128:465-68 Moser KM, Maurer J. Jassy L, Kremsdorf R, Konopka R, Shure D, et al. Sensitivity. specificity, and risk of diagnostic procedures in a canine model of Streptococcus pneumoniae pneumonia. Am Rev Respir Dis 1982; 125:436-42 Chastre J. Viau F, Brun P, Pierre J. Dauge MC. Bouchama A, et al. Prospective evaluation of the protected specimen brush for the diagnosis of pulmonary infections in ventilated patients. Am Rev Respir Dis 1984; 130:924-29 Beydon L, Saada M, Liu N, Becquemin JP, Harf A, Bonnet F, et al. Can portable chest x-ray examination accurately diagnose lung consolidation after major abdominal surgery? a <:omparison with computed tomography scan. Chest 1992; 102:16971703
CHEST I 105 I 3 I MARCH, 1994
887