Solitary pulmonary nodules (SPN) can be visualized by endobronchial ultrasound (EBUS) in at least 75%. It has been described that by a complex analysis of different structural items the histology is suggested. In a preliminary study we demonstrated that computerized analysis might be a simple and reliable alternative method.
AIM OF STUDY:
In a prospective study on 50 SPN the distribution of grey scale signals on the EBUS-image were calculated by a computer program (“histogram”) and compared to the histological findings. By computer analysis of the greyscale distribution in EBUS-images of solitary pulmonary nodules the nature could be predicted in the majority of cases. The consistency of these data had to be confirmed by an independent study.
EBUS images of SPN in 50 consecutive cases in Kumamoto (25 benign, 25 malignant) were sent to Heidelberg for computerised analysis of their internal structure without knowing the final histology. After decision for a benign or malignant nature according to the grey scale distribution the results were compared to the histology. In 46 of 50 cases (92%) the diagnosis by EBUS was correct. Two benign lesions (1 Cryptococcosis and 1 Aspergilloma, 2%) had a malignant appearance, whereas two small cell lung cancers (2%) appeared as inflammatory infiltrates.
The study confirmed our primary assumption that computerized analysis of the EBUS image is a simple and reliable method to predict the histology of SPN and might be useful in decision for treatment of these
H.D. Becker, None.