Abstract
Purpose
Cancer of the upper digestive tract (uGI) is a major contributor to cancer-related death worldwide. Due to a rise in occurrence, together with poor survival rates and a lack of diagnostic or prognostic clinical assays, there is a clear need to establish molecular biomarkers.
Experimental design
Initial assessment was performed on urine samples from 60 control and 60 uGI cancer patients using MS to establish a peak pattern or fingerprint model, which was validated by a further set of 59 samples.
Results
We detected 86 cluster peaks by MS above frequency and detection thresholds. Statistical testing and model building resulted in a peak profiling model of five relevant peaks with 88% overall sensitivity and 91% specificity, and overall correctness of 90%. High-resolution MS of 40 samples in the 2–10 kDa range resulted in 646 identified proteins, and pattern matching identified four of the five model peaks within significant parameters, namely programmed cell death 6 interacting protein (PDCD6IP/Alix/AIP1), Rabenosyn-5 (ZFYVE20), protein S100A8, and protein S100A9, of which the first two were validated by Western blotting.
Conclusions and clinical relevance
We demonstrate that MS analysis of human urine can identify lead biomarker candidates in uGI cancers, which makes this technique potentially useful in defining and consolidating biomarker patterns for uGI cancer screening.
Cancer of the upper digestive tract (uGI) is a major contributor to cancer-related death worldwide. Due to a rise in occurrence, together with poor survival rates and a lack of diagnostic or prognostic clinical assays, there is a clear need to establish molecular biomarkers.
Experimental design
Initial assessment was performed on urine samples from 60 control and 60 uGI cancer patients using MS to establish a peak pattern or fingerprint model, which was validated by a further set of 59 samples.
Results
We detected 86 cluster peaks by MS above frequency and detection thresholds. Statistical testing and model building resulted in a peak profiling model of five relevant peaks with 88% overall sensitivity and 91% specificity, and overall correctness of 90%. High-resolution MS of 40 samples in the 2–10 kDa range resulted in 646 identified proteins, and pattern matching identified four of the five model peaks within significant parameters, namely programmed cell death 6 interacting protein (PDCD6IP/Alix/AIP1), Rabenosyn-5 (ZFYVE20), protein S100A8, and protein S100A9, of which the first two were validated by Western blotting.
Conclusions and clinical relevance
We demonstrate that MS analysis of human urine can identify lead biomarker candidates in uGI cancers, which makes this technique potentially useful in defining and consolidating biomarker patterns for uGI cancer screening.
| Original language | English |
|---|---|
| Pages (from-to) | 586-596 |
| Number of pages | 11 |
| Journal | PROTEOMICS-Clinical Applications |
| Volume | 9 |
| Issue number | 5-6 |
| DOIs | |
| Publication status | Published - 8 May 2015 |
