The Sensitivity and Specificity of a new Scoring System Using High Resolution Computed Tomography to Diagnose Lung Cancer

Abstract

Aim To find a non-invasive diagnostic method for lung cancer with the results almost as accurate as histopathological examinations using HRCT scoring system.

Method This study was conducted from December 2006 until March 2008. A total of 55 persons, comprised of 40 male and 15 female patients suspected of having lung cancer, underwent high resolution computed tomography with and without contrast as well as CT-guided transthoracic needle aspiration to obtain cytology specimens.

Results Histopathological examinations revealed the existence of lung cancer in 43 patients (78%) and benign lesions in 12 patients (22%). A scoring system was then made based on the similarities of findings from CT and histopathologic examinations. In the the age group of 49 years and above, tumor volume of more than 68 cm3, HU of more than 21, spicula positive, angiogram positive and lymph node positive had a value of respectively 20, 19, 10, 24, 18, and 17 (total score 108).

Conclusion This new method to diagnose lung cancer is constantly as accurate as histological findings with a sensitivity of 97.7% and a specificity of 83,3% and a cut-off score of 35. According to this system, a score of less than 35 indicates that the lesions were benign while a score higher than 35 is suggestive malignant. (Med J Indones 2009;18:179-86)