Diagnosis of peripheral lung cancer with three echoic features via endobronchial ultrasound

Chest. 2007 Sep;132(3):922-9. doi: 10.1378/chest.06-3106. Epub 2007 Jul 23.

Abstract

Background: Endobronchial ultrasonography (EBUS) is useful in localizing peripheral lung lesions. Previous reports have revealed that several characteristic echoic patterns correlate well with the histopathologic findings of benign and malignant lesions. Therefore, EBUS may be also useful in the differential diagnosis of malignant lesions of the lung.

Objective: To assess the feasibility of EBUS in the differential diagnosis between malignant and benign lesions by the following three characteristic echoic features indicating malignancy: continuous margin; absence of a linear-discrete air bronchogram; and heterogeneous echogenicity.

Method: EBUS images from 224 patients who undergone bronchoscopy for a peripheral lung lesion were analyzed. The sensitivity and specificity for each echoic feature or in combination in diagnosing malignancy or benignity were determined.

Result: Continuous margin, absence of linear-discrete air bronchogram, and heterogeneous echogenicity are three echoic features indicating malignancy. The absence of linear-discrete air bronchogram has the highest sensitivity in the diagnosis of malignancy (91.9%), but the lowest specificity (62.4%). By contrast, a well-defined margin has the highest specificity (93.1%), but the lowest sensitivity (27.6%). The sensitivity and specificity for heterogeneous echogenicity are intermediate (65.0% and 90.1%, respectively). The negative predictive value for the malignancy of a lesion with none of these three echoic features is 93.7%. The positive predictive value for malignancy of a lesion with any two of the three echoic features is 89.2%.

Conclusion: These results indicate that EBUS is useful as an adjunct in lung cancer diagnosis, especially when peripheral lung lesions are not visible in traditional bronchoscopy.

MeSH terms

  • Bronchoscopy*
  • Carcinoma / diagnostic imaging*
  • Carcinoma / pathology
  • Cohort Studies
  • Diagnosis, Differential
  • Endosonography*
  • Humans
  • Lung Diseases / diagnostic imaging*
  • Lung Diseases / pathology
  • Lung Neoplasms / diagnostic imaging*
  • Lung Neoplasms / pathology
  • Predictive Value of Tests
  • Reproducibility of Results