Improved Lung Cancer Detection in Cardiovascular Outpatients by the Pulmonologist-based Interpretation of Chest Radiographs

Abstract

Objective Pulmonologists and cardiologists view chest radiographs differently. Lung cancer may therefore go undetected in patients referred to cardiovascular departments. We aimed to determine the clinical benefit of the additional interpretation of chest radiographs by pulmonologists in study involving cardiovascular outpatients. Methods A retrospective review of chest radiographs of outpatients attending a Japanese cardiovascular hospital between April 2000 and March 2010 was conducted. Lung cancer patients were categorized into 3 groups: group C, patients in whom tumors were detected by a cardiologist at the first visit; group P, patients in whom tumors were detected by the additional interpretation of a chest radiographs by a pulmonologist after a lesion was missed by a cardiologist; and group H, patients from an earlier period in which chest radiographs were only examined by a cardiologist. Results Cardiologists detected 9 cases of lung cancer in groups C and H from 2,430 and 2,288 radiographs, respectively. Pulmonologists detected 10 cases of lung cancer (group P) and 3 other malignancies that were previously undetected, giving a miss rate of 52.6% for the cardiologists. Tumor diameters were significantly smaller in group P than in group C or H. Furthermore, a significantly higher number of the tumors in group P were of an early stage and resectable, with more superposing structures than in groups C or H. Conclusion The additional pulmonologist-based interpretations significantly increased the detection rate of operable tumors that mostly corresponded to the early T1 stage; this serves offers a potential clinical benefit in reducing the period of time from patient presentation to the diagnosis of lung cancer.

Computer-assisted diagnosis for an early identification of lung cancer in chest X rays

“…Assessing also the impact of delayed diagnosis, Sakai and cols. showed in a study on five thousand patients that the identification rate of pulmonary nodules in chest x-rays when examined by clinicians not trained in the assessment of lung images was fifty percent below trained readers, who diagnosed smaller tumours which were more often surgically-treated 12 . The use of a CAD algorithm designed for the identification of pulmonary nodules would facilitate the recognition of pulmonary nodules in x-rays read by untrained physicians, pointing to the images with the highest probability which need to be checked by a radiologist.…”

“…Chest x-ray is one of the most complex imaging modalities, with up to twenty percent discrepancy in their interpretation between radiologists 14 . This difficulty determines that pulmonary nodules which are early-stage LC may be missed when the chest x-ray is examined by an untrained physician, a situation that has been repeatedly reported to be related to the underdiagnosis of LC 11 , 12 , and has a negative impact on survival 12 , 13 , 15 – 20 .…”

Application of Artificial Intelligence in Lung Cancer

“…Furthermore, mistakes in reading CXR or LDCT images occur, and it constitutes a large number of malpractice law suits [83]. Though experts were shown to detect more pulmonary nodules on CXRs [84], approximately 20% of lung nodules <3 cm are missed by radiologists [85]. In the 21st century, the prediction accuracy of pulmonary nodules on CXRs has improved with the computer-aided diagnosis systems or AI-based programs.…”