Accuracy of low-dose chest CT in intensive care patients.

In this prospective study, we set out to determine the accuracy of low-dose computerized tomography (LDCT) of the chest in intensive care patients. Fifteen adult intensive care patients were examined with a standard-dose CT protocol (average radiation dose = 6.7 mSv), chosen as the reference standard, followed by a non-contrast-enhanced LDCT protocol (average radiation dose = 0.59 mSv). Each examination was then read by two separate groups of radiologists blinded to both the purpose and the protocol of the study. In the small group examined, the results showed 100% accuracy in the diagnosis of pneumomediastinum, pericardial effusion, and pleural effusion, and 90% accuracy in the diagnosis of pneumothorax and consolidation. There were no false-positive findings, and the few false-negative findings were unlikely to lead to any clinical interventions. Our examination protocol, while providing a tenfold reduction of the radiation dose, nevertheless remained accurate enough for resolving certain clinical questions common in the intensive care patient. Thus, we suggest that protocols aimed at reducing the radiation dose in chest CT could be applied to the intensive care patient for resolving some specific questions, without compromising the diagnostic yield of the examinations.