Login / Signup

Comparative evaluation of the STANDARD M10 and Xpert C . difficile assays for detection of toxigenic Clostridioides difficile in stool specimens.

Hyun-Woo LeeHui-Jin YuHeejung KimSun Ae YunEunsang SuhMinhee KangTae Yeul KimHee Jae HuhNam Yong Lee
Published in: Journal of clinical microbiology (2024)
This study compared the performance of two commercial molecular assays, the STANDARD M10 Clostridioides difficile assay (M10) and the Xpert C. difficile assay (Xpert), for detecting toxigenic C. difficile in stool specimens. A total of 487 consecutive stool specimens submitted for routine C. difficile testing between June and November 2023 were included. Following routine testing using C. DIFF QUIK CHEK COMPLETE (QCC), M10 and Xpert were tested in parallel, alongside toxigenic culture (reference standard). Additionally, two-step algorithms, using QCC on the first step and either M10 or Xpert on the second step, were assessed. Both M10 and Xpert demonstrated a sensitivity and negative predictive value (NPV) of 100%. M10 exhibited significantly higher specificity and positive predictive value (PPV; 91.9% and 64.2%, respectively) than Xpert (90.3% and 59.8%, respectively). Both two-step algorithms showed a sensitivity and NPV of 98.4% and 99.8%, respectively. The specificity and PPV of the two-step algorithm using M10 (95.2% and 75.0%, respectively) were slightly higher than those of the one using Xpert (94.8% and 73.2%, respectively), without statistical significance. Receiver operating characteristic curve analysis, assessing the predictive ability of cycle threshold (Ct) values for the detection of free toxin, exhibited an area under the curve of 0.825 for M10 and 0.843 for Xpert. This indicates the utility of Ct values as predictors for the detection of free toxin in both assays. In conclusion, M10 proves to be an effective diagnostic tool with performance comparable to Xpert, whether utilized independently or as part of a two-step algorithm.
Keyphrases
  • clostridium difficile
  • machine learning
  • high throughput
  • deep learning
  • escherichia coli
  • computed tomography
  • magnetic resonance imaging
  • magnetic resonance
  • contrast enhanced
  • dual energy
  • single molecule