13
Respiratory Tract
Table 13.19 Diagnostic Accuracy of Sputum and Bronchial Material
Investigator
Year
Specimen type
Sensitivity (%)
Specificity (%)
Koss et al.
1964
Sputum x 3
89
99.8
Sputum x 5
96.1
Erozan and Frost
1970
Bronchial washings x 1
61
NS
Sputum x 1
41
Sputum x 3
82
Sputum x 5
91
Bibbo et al.
1973
Bronchial brushings
70
98
Bedrossian et al.
1976
Sputum x 3
57
NS
Bronchial washings
76
Bronchial brushings
76
Johnston and Bossen
1981
Sputum x 1
27
99.9
Bronchial washings/
brushings x 1
22
Sputum/bronchial
brushings x 5
87
Pilotti et al.
1982
Sputum x 3
57
NS
Bronchial brushings
67
Ng and Horak
1983
Sputum x 3
83
NS
Bronchial washings
74
Truong et al.
1985
Sputum
60
99.9
Bronchial washings
66
Bronchial brushings
77
Tanaka
1985
Sputum x 1
27
NS
NS = not supplied.
1976 reported a sensitivity of 56% in cancer detection when
three sputum samples were examined.535 This rate increased to
76% when either bronchial brushings or bronchial washings
were used. In 1982, Pilotti and colleagues reported for sputum
an overall sensitivity of 58% and for bronchial brushings a 67%
rate.525,526 Ng and Horak reported in 1983 an overall sensitivity
of 74% for bronchial washings and 83% for three sputum sam-
ples.523,524 Truong and associates, in their 1985 study, determined
that the overall sensitivities of sputum, bronchial washings, and
bronchial brushings were 60, 66, and 77%, respectively.531 Their
false-positive rate was 2.8%. Tanaka and associates examined
the accuracy of cytologic diagnosis and typing in 154 patients.
Central lesions were detected in 57-64% of the cases by either
3-day pooled sputum or aerosol-induced specimens.530
Sputum has shown the highest levels of sensitivity in detect-
ing the more centrally located tumors, but this sensitivity has
also declined drastically for the peripheral cancers. Bronchial
brushing techniques for these peripheral lesions have improved
diagnostic accuracy in cancer detection up to the levels of 70-
88% of cases.511,514,532,536-545
In 1973, Bibbo and associates reported 693 specimens
obtained by a fluoroscopically controlled bronchial brushing
technique.537 The series included 224 confirmed primary tumors
and 30 metastatic tumors. For primary tumors, the average
diagnostic yield (sensitivity) was 70% and 53% for metastatic
lesions. In 160 cases, sputum samples taken before brushings
showed tumor cells in only 7% of cases; however, sputum sam-
ples after brushing showed an increase to 66% tumor detection
rate. Nine false-positive diagnoses were recorded and reported
as a 2% rate. Bibbo has emphasized the excellence of cellular
preservation and the increased amounts of tumor cells arranged
in irregular sheets as compared with sputum and bronchial
washings.536,546
In 1975, Johnston and Bossen reported on the results of
examination of 9892 consecutive specimens of sputum sam-
ples, bronchial washings, and bronchial brushings. The patient
groups on which these studies were based are described in Tables
13.20 and 13.21. Specimens from 363 patients were interpreted
as diagnostic of cancer. Of these, 232 also had a histopathologic
diagnosis of cancer. Some 125 patients were treated for lung
cancer on the basis of the clinical, radiographic, and cytologic
findings. In six patients, follow-up studies failed to confirm a
cytologic diagnosis of cancer. Thus, these six patients were con-
sidered to have been given false-positive diagnoses.
Table 13.21 lists the patient group on which histopatho-
logic diagnoses of primary or metastatic malignant neoplasms
were made. A total of 633 patients had recorded histopatho-
logic diagnoses of malignant neoplasms involving the lungs or
adjacent respiratory areas. Of those, 590 patients had histologi-
cally confirmed primary malignant neoplasms and 43 patients
353
previous page 350 ComprehensiveCytopathology 1104p 2008 read online next page 352 ComprehensiveCytopathology 1104p 2008 read online Home Toggle text on/off