Radiology: Tumor detection on chest x-ray is observer dependent
For lung cancer tumors with a diameter of 6.8 to 50.7 mm found during CT screenings, a detection rate of 94 percent could be achieved with chest radiography only at the expense of a high false-positive rate and an excessive number of work-up CT examinations, according to the findings of a study published in the May edition of Radiology.
In the evaluation of digital chest radiography for the detection of lung cancer, Bartjan de Hoop, MD, From the Department of Radiology and Image Sciences Institute at University Medical Center in Utrecht, the Netherlands, and colleagues said that lung cancer screening studies by traditional chest radiography have yielded disappointing results. However, “the use of modern digital equipment with highly quantum-efficient detectors and elaborate processing tools improves visualization of pulmonary structures with chest radiography and may, therefore, be a more suitable screening tool than is conventional chest radiography,” they wrote.
The researchers included 55 patients with at least one malignant nodule for inclusion in their study after receiving ethics committee approval. They also included a control cohort of 72 individuals in whom CT did not show nodules larger than 5 mm in diameter and in whom chest radiography had been performed within two months from CT screening for reasons other than a suspicious lung nodule.
According to the authors, four radiologists with varying degrees of experience ranging from 20 years of experience to a third-year resident identified and localized potential cancers on chest radiographs utilizing a confidence scale of level one or “no lesion,” up to level five or “definite lesion.” Assuming that suspicious lesions detected during chest radiography would lead to further work-up with CT, the number of work-up CT examinations per detected cancer was calculated at various confidence levels for the screening population, they explained.
The diameter of the tumor sizes ranged from 6.8 to 50.7 mm, with an average size of 11.8 mm. The researchers noted that detection rates “substantially varied” with observer experience and confidence level. “At a confidence level of five, detection rates ranged from 18 percent at one CT examination per cancer to 53 percent at 13 CT examinations per cancer,” the authors wrote. “At a confidence level of two or higher, detection rates ranged from 94 percent at 62 CT examinations per cancer to 78 percent at 44 CT examinations per cancer.”
The authors said that the number of detected malignant lesions ranged from 37 percent to 78 percent depending on the reader, and only in the lesions that were rated as probably or definitely present were considered to require further diagnostic work-up. They explained that this would result in 22 to 63 percent of lung cancers being missed at a stage of disease in which they could be detected with CT, if only chest radiography was utilized.
De Hoop and colleagues concluded: “Even in the era of digital imaging, detection of lung cancer with chest radiography is a challenging task that shows high interreader variability.”
In the evaluation of digital chest radiography for the detection of lung cancer, Bartjan de Hoop, MD, From the Department of Radiology and Image Sciences Institute at University Medical Center in Utrecht, the Netherlands, and colleagues said that lung cancer screening studies by traditional chest radiography have yielded disappointing results. However, “the use of modern digital equipment with highly quantum-efficient detectors and elaborate processing tools improves visualization of pulmonary structures with chest radiography and may, therefore, be a more suitable screening tool than is conventional chest radiography,” they wrote.
The researchers included 55 patients with at least one malignant nodule for inclusion in their study after receiving ethics committee approval. They also included a control cohort of 72 individuals in whom CT did not show nodules larger than 5 mm in diameter and in whom chest radiography had been performed within two months from CT screening for reasons other than a suspicious lung nodule.
According to the authors, four radiologists with varying degrees of experience ranging from 20 years of experience to a third-year resident identified and localized potential cancers on chest radiographs utilizing a confidence scale of level one or “no lesion,” up to level five or “definite lesion.” Assuming that suspicious lesions detected during chest radiography would lead to further work-up with CT, the number of work-up CT examinations per detected cancer was calculated at various confidence levels for the screening population, they explained.
The diameter of the tumor sizes ranged from 6.8 to 50.7 mm, with an average size of 11.8 mm. The researchers noted that detection rates “substantially varied” with observer experience and confidence level. “At a confidence level of five, detection rates ranged from 18 percent at one CT examination per cancer to 53 percent at 13 CT examinations per cancer,” the authors wrote. “At a confidence level of two or higher, detection rates ranged from 94 percent at 62 CT examinations per cancer to 78 percent at 44 CT examinations per cancer.”
The authors said that the number of detected malignant lesions ranged from 37 percent to 78 percent depending on the reader, and only in the lesions that were rated as probably or definitely present were considered to require further diagnostic work-up. They explained that this would result in 22 to 63 percent of lung cancers being missed at a stage of disease in which they could be detected with CT, if only chest radiography was utilized.
De Hoop and colleagues concluded: “Even in the era of digital imaging, detection of lung cancer with chest radiography is a challenging task that shows high interreader variability.”