Reviewing more than 1,100 chest CT scans performed on women for various reasons, German researchers have found the imaging incidentally turned up at least one lesion requiring a closer look—i.e., BI-RADS 3 to 5—in nearly 6 percent of the patients. They call for radiologists to be alert for breast cancers when reading all chest CTs.

The German medical journal RöFo published the team’s study report online July 6.

Kathrin Barbara Krug, MD, and colleagues at the University of Cologne looked at all first contrast-enhanced chest CT exams carried out for women at an affiliated care center in 2012. Some 1,170 exams met their study criteria.

While more than half the imaging was done for cancer staging and follow-up (701 patients, 59.9  percent), a broad range of studies were for other indications. These included vascular (190 patients, 16.2  percent), inflammatory (48, 4.1  percent) and pulmonologic (22, 1.9  percent) issues.

Krug and colleagues drew from all available diagnostic data to enhance the existing (reference) radiology reports. They compared these with retrospective readings rendered by a senior radiologist and, in the case of BI-RADS 3 to 5 classifications, by a consensus of two senior rads.

Their key finding was that 68 patients, 5.8  percent, had at least one lesion only retrospectively classified as BI-RADS 3  (probably benign), 4 (suspicious abnormality) or 5 (highly suspicious of malignancy).

The histological potential was known in 57 of these lesions, with 11 proving malignant and 46 benign.

Two of 10 lesions classified as malignant based on the further clinical and radiological course were not mentioned in the written CT reports (0.2  percent), while 13 BI-RADS 4 or 5 consensus assessments (1.1 percent ) were false-positive.

In their discussion, Krug et al. write that the 5.8 percent rate of BI-RADS 3 to 5 findings reflects “the situation encountered in clinical imaging for primarily non-senologic questions,” meaning scans not specifically performed for disorders of the breast, “and therefore differs from what would be expected in a dedicated screening program.”

Meanwhile, they add, the rates of known false-positive BI-RADS 4 or 5 findings in the retrospective evaluations (1.1  percent above) and of false-negative findings in the written CT reports (0.2  percent) “reflect the different diagnostic approaches of image-based senological screening and radiological examinations indicated in order to solve clinical problems not primarily concerning the breast region.”

The authors conclude by urging caution in interpreting their findings while underscoring their import.

“Statements regarding the prevalence of clinically occult breast cancers can only be made with caution in the presented, highly selective group of patients due to the often incomplete visualization of breast tissue and the retrospective approach,” they write.

Nevertheless, the present findings “highlight the importance of paying systematical and targeted attention [to] senological additional findings in CT examinations of the chest in other clinical settings than that of the included patients in a clinic with oncological main focus,” they write.