Interrelationships Among Sensitivity, Precision, Accuracy, Specificity and Predictive Values in Bioassays,...

Area of Science:

• Biomarker validation
• Clinical diagnostics
• Medical device technology

Background:

• Existing methods for biomarker cutoff validation may overlook crucial diagnostic parameters.
• The UBC Rapid assay for bladder cancer detection was quantitatively re-evaluated.

Purpose of the Study:

• To introduce accuracy- and precision-ROC curves alongside sensitivity-specificity (SS) and positive-predictive value (PV) ROC curves.
• To present a novel method for profiling biomarker characteristics and validating optimal cutoffs for clinical decision-making.

Main Methods:

• Quantitative analysis of used test cassettes from the UBC Rapid assay using a photometric reader.
• Construction of ROC curves (SS, PV, accuracy, precision) at various concentration thresholds (5–300 µg/L).
• Development of an ROC index cutoff diagram for integrated analysis of diagnostic parameters.

Main Results:

• A common optimal cutoff value was identified using the ROC index cutoff diagram, enhancing specificity.
• Accuracy, precision, and predictive values provide disease-related information, unlike sensitivity-specificity ROC curves alone.
• The novel multi-parameter cutoff-index diagram allows quantitative comparison of ROC curve results.

Conclusions:

• Accuracy- and precision-ROC curves offer a more transparent method for identifying appropriate bioassay cutoffs compared to single-parameter methods.
• A multi-parameter diagnostic profile and cutoff-index diagram provide superior diagnostic information than single-point determinations.
• The proposed method, including the novel index cutoff AOX, enables quantitative comparison of multi-parameter ROC curves.