Importance: Gold standard specimens, gastric aspirate (GA) and induced sputum (IS), for tuberculosis diagnosis in young children are invasive and rarely collected in resource-limited settings. We identified a far less invasive approach to TB diagnostic testing in children <5 years as sensitive as current reference standards. Objective: To characterize the sensitivity of specimen and assay combinations, preferably minimally-invasive, relative to maximum observed yield from all specimens and assays combined. Design: Prospective cross-sectional diagnostic study. The reference standard was a panel of up to two samples of each of six specimen types tested for Mycobacterium tuberculosis complex by Xpert MTB/RIF assay and mycobacteria growth indicator tube culture (MGIT). Multiple different combinations of specimens and tests were evaluated as index tests. Setting: Inpatient and outpatient care settings in Kisumu County, Kenya (October 2013–August 2015). Participants: Consecutive sample of children <5 years with symptoms of TB (unexplained cough, fever, malnutrition, or cervical lymphadenopathy) and parenchymal abnormality on chest radiograph. Children with >1 evaluable specimen for >4 primary study specimen types were included in analysis. Main Outcome Measures: Cumulative and incremental diagnostic yield of combinations of specimen types and tests relative to the maximum observed yield. Results: Of 300 children enrolled, 294 met criteria for analysis (median age 2.0 years, interquartile range [IQR] 1.0-3.6; 50.3% female). Of 31 participants with confirmed tuberculosis (maximum observed yield), 24 (77%) had positive results for up to two GA and 20 (64%) for up to two IS samples. The yield of two NPA (23/31 [74%]), of one NPA and one stool (22/31 [71%]), or of one NPA and one urine sample (21.5/31 [69%]) was similar to reference standard specimens. Combining up to two each of GA and NPA had an average yield of 28/31 (90%). Conclusions: NPA, in duplicate or in combination with stool or urine specimens, was readily obtainable and had diagnostic yield comparable to reference standard specimens. This combination could improve tuberculosis diagnosis among children in resource-limited settings. Combining GA and NPA had greater yield than that of the current standards and may be useful in certain clinical and research settings.