High-resolution CT in the evaluation of clinically suspected Pneumocystis carinii pneumonia in AIDS patients with normal, equivocal, or nonspecific radiographic findings.

Abstract

We prospectively studied AIDS patients with a high clinical pretest probability of Pneumocystis carinii pneumonia (PCP) in whom chest radiographic findings were normal, equivocal, or nonspecific with high-resolution CT (HRCT) to determine the incidence of PCP in these patients, to assess the diagnostic accuracy of HRCT for the presence or absence of PCP, to evaluate the role of HRCT in patient management, and to determine the clinical outcome of all patients 1 month after evaluation. All patients were referred to the Division of Pulmonary and Critical Care Medicine for diagnosis of clinically suspected PCP. Thirty-three patients were prospectively evaluated with HRCT within 24 hr of diagnostic bronchoalveolar lavage; 18 other patients who underwent HRCT were managed according to the HRCT interpretation and followed up clinically. All HRCT scans were independently reviewed by three chest radiologists; patchy or nodular ground-glass attenuation was considered to indicate "possible PCP." The incidence of PCP was 12% (6/51). The sensitivity of HRCT was 100%; specificity, 89%; and accuracy, 90% (p < .005). We had five false-positive and no false-negative interpretations. Some form of "airways disease" (n = 23) was the single most common HRCT interpretation. HRCT may allow exclusion of PCP in patients with findings that are normal, equivocal, or nonspecific on chest radiographs. Empiric therapy or immediate bronchoscopy can be avoided in many patients on the basis of the HRCT findings.