Fig. 6: Reduction of TTR to 2 h retains high performance for CZA. | Nature Communications

Fig. 6: Reduction of TTR to 2 h retains high performance for CZA.

From: Accurate and rapid antibiotic susceptibility testing using a machine learning-assisted nanomotion technology platform

Fig. 6

a To develop a model for CZA on 2 h nanomotion recordings, the MIC distribution (left) of 46 E. coli and K. pneumoniae strains was analysed at 32 µg/ml CAZ (conc. Dashed line) and 4 µg/ml AVI. Right: Classification according to nanomotion and a 6-SP model, with MIC strip serving as a reference. b Independent testing of the 6-SP model in (a) on 21 different E. coli and K. pneumoniae isolates, left, MIC distribution, right, scores for each recording. c Training and testing performance of the 2-h model for CZA, accuracy [(TP + TN)/n], sensitivity [TP/(TP + FN)], and specificity [TN/((TN + FP)]. Data are available in the source data file. Single SP values and scores are available in Supplementary Data 2.

Back to article page