|
Abstract INTRODUCTION:
Ion ratios are routinely used in clinical LC–MS/MS assays to assess analytical specificity. Traditionally, a measured ion ratio is compared against an expected value, with deviations beyond a prescribed tolerance interpreted as evidence of interference. However, clinical guidelines no longer define universal tolerance limits, acknowledging that acceptable ion ratio variability is assay and analyte dependent. Establishing objective tolerances is challenged by intrinsic sources of variability, including heteroskedastic signal response, non linear detector behavior, and instrumental drift. Here, we present case studies for both small molecules and peptides illustrating why a single tolerance is inadequate and propose data driven considerations for defining meaningful ion ratio limits.
METHODS:
Quantitative LC–MS/MS analyses were performed on an Aria Transcend TLX 4 multiplexing HPLC system coupled to Sciex 5500 or 7500 triple quadrupole mass spectrometers operated in SRM mode. For each analyte and stable isotope labeled internal standard, two transitions (quantifier and qualifier) were monitored. Ion ratios were calculated by dividing the peak area of the qualifier by that of the quantifier. For small molecule assays, between sample ion ratio bias was defined as the percent difference between a test sample’s ion ratio and the expected value (average ion ratio measured in calibrators in the same analytical batch/run). For peptide assays, within sample ion ratio bias was assessed by comparing analyte and internal standard ion ratios within the same sample. Bias across analytical runs was summarized to evaluate systematic and concentration dependent effects.
RESULTS:
Ion ratio performance evaluated across multiple small molecule assays showed several sources of variability that complicate the use of universal tolerance limits. Ion ratios for Vitamin K1 and its internal standard were monitored across 44 analytical runs (N > 3000). Mean ion ratio biases of −0.3% and −1.4%, respectively, indicated no persistent interference; however, vitamin K1 ion ratios exhibited significantly greater variability at lower concentrations (F test, p < 0.05), consistent with heteroskedastic LC MS/MS response. As such, the acceptable ion ratio biases were concentration dependent with 95% confidence intervals of ±13.0% below 0.5 ng/mL and ±8.0% above 0.5 ng/mL.
Ion ratio monitoring of MPA, MPAG, and matching internal standards assessed over more than 30 runs (N > 3000) showed linearity differences between two mass spectrometers (Sciex 5500 and 7500). At higher MPA concentrations (>12.5 µg/mL), ion ratio bias was greater on the 5500 system (+4.1%) compared with the 7500 (+1.7%), reflecting a more pronounced non linear response of the quantifying transition on the 5500. [D3] MPAG ion ratios also demonstrated concentration dependent non linearity, producing systematically higher bias at elevated MPAG concentrations due to isotopic contamination of the qualifying transition. Notably, this effect was more pronounced on the 7500 system (+14.2%) than on the 5500 (+5.7%), underscoring instrument specific influences on ion ratio behavior.
For 10 MHD, lacosamide, and their internal standards monitored over 18 analytical runs (N > 2800), non linear ion ratio bias was again observed at higher analyte concentrations. For [D3] 10 MHD, isotopic contribution from analyte into internal standard quantifying and qualifying transitions (2.0% and 5.3%, respectively) was consistent with the observed ion ratio bias (+2.9%). In contrast, [13C,D3] lacosamide demonstrated negligible isotopic contribution (≤0.2%), yet consistent non linear ion ratio bias across calibrators and unknowns ruled out sample specific interference. Ongoing analyses indicate thermal drift as a likely contributor, and supporting data will be presented.
Peptide ion ratio behavior in which analysis was performed within each sample by comparing analyte and internal standard ion ratios further demonstrated that ion ratio tolerances must account for assay format and underlying sources of variability rather than relying on fixed acceptance limits.
CONCLUSION:
These case studies showed ion ratios in clinical LC–MS/MS assays to be assay specific and subject to concentration‑dependent heteroskedasticity, detector non‑linearity, isotopic contributions, and thermal drift. As such, these studies support setting ion ratio tolerances empirically for each assay, concentration range, instrument platform, and assay format rather than attempting to establish a universal tolerance. Data-driven context‑specific confidence intervals provide a more reliable way to evaluate analytical specificity and reduce the risk of inappropriate result rejection.
|
= Discovery stage. (57.21%, 2026)
= Translation stage. (23.38%, 2026)
= Clinically available. (19.40%, 2026)
