Chemoresistance,
i.e., tumor insensitivity to chemotherapy, shortens
life expectancy of cancer patients. Despite the availability of new
treatment options, initial systemic regimens for solid tumors are
dominated by a set of standard chemotherapy drugs, and alternative
therapies are used only when a patient has demonstrated chemoresistance
clinically. Chemoresistance predictors use laboratory parameters measured
on tissue samples to predict the patient’s response to chemotherapy
and help to avoid application of chemotherapy to chemoresistant patients.
Despite thousands of publications on putative chemoresistance predictors,
there are only about a dozen predictors that are sufficiently accurate
for precision oncology. One of the major reasons for inaccuracy of
predictors is inaccuracy of analytical methods utilized to measure
their laboratory parameters: an inaccurate method leads to an inaccurate
predictor. The goal of this study was to identify analytical challenges
in chemoresistance-predictor development and suggest ways to overcome
them. Here we describe principles of chemoresistance predictor development
via correlating a clinical parameter, which manifests disease state,
with a laboratory parameter. We further classify predictors based
on the nature of laboratory parameters and analyze advantages and
limitations of different predictors using the reliability of analytical
methods utilized for measuring laboratory parameters as a criterion.
Our eventual focus is on predictors with known mechanisms of reactions
involved in drug resistance (drug extrusion, drug degradation, and
DNA damage repair) and using rate constants of these reactions to
establish accurate and robust laboratory parameters. Many aspects
and conclusions of our analysis are applicable to all types of disease
biomarkers built upon the correlation of clinical and laboratory parameters.