The Mycobacterium Tuberculosis Direct (MTD) Test is widely used for detection of Mycobacterium tuberculosis complex (MTBC) in respiratory specimens. Based on high specificity, it is accepted that a positive MTD result, in proper clinical context, "rules in" a diagnosis of tuberculosis. Test utility for "ruling out" tuberculosis has been limited by lower sensitivity and negative predictive value (NPV), based in part on poorly defined reaction inhibitors in clinical specimens. After discovering that specimen dilution could decrease concentrations of reaction inhibitors without reaching the MTBC detection limit, the Massachusetts State Laboratory Institute in 2003 began routinely testing paired undiluted and diluted samples in parallel. We retrospectively analyzed results from 105 respiratory specimens (87 smear-positive, 18 smear-negative) tested consecutively over 23 months using this dilution protocol. MTD results for each sample pair were merged into a sum result (positive or negative) and compared with culture outcome as the "gold standard." We found that the MTD sum result always matched the culture outcome, even for smear-negative specimens. Accordingly, the MTD using the dilution protocol had sensitivity, specificity, positive predictive value, and NPV of 100% vs culture.