Direct Probe Ionisation Mass Spectrometry applied to biomarker discovery in pancreatic cancer

Significance of the Topic

Pancreatic ductal adenocarcinoma (PDAC) remains a leading cause of cancer mortality due to late diagnosis and limited treatment options. Identification of serum metabolite biomarkers can enable earlier detection and personalized interventions. Direct probe ionisation mass spectrometry (DPiMS) offers an ultrafast, minimal-prep approach to capture metabolic alterations associated with PDAC.

Objectives and Study Overview

This study evaluated the performance of high-resolution DPiMS on Q-TOF instrumentation to distinguish metabolic profiles of PDAC patient serum versus healthy controls. Key aims included rapid MS1 biomarker discovery, narrow-band ion mobility data-independent acquisition (iDIA) for metabolite annotation, and cross-validation against conventional LC-MS/MS data.

Methodology and Instrumentation

• Sample cohort: 30 PDAC patients and 30 matched healthy controls; ethical approval obtained from a pancreatic cancer tissue bank committee.
• Sample preparation: 50 µL serum mixed with 950 µL 1:1 ethanol:water, centrifugation at 16 000×g, 10 µL supernatant analyzed directly.
• DPiMS analysis: Shimadzu LCMS-9030 Q-TOF system equipped with a direct probe; solid needle tip (~1 µm radius) oscillating at ~3 Hz; total analysis time 2 minutes with polarity switching.
• Data acquisition: MS1 TOF scan (m/z 100–1500) followed by iDIA MS/MS (1 Da precursor windows over sequential 200 Da segments); positive and negative mode scans acquired per cycle.
• Data processing: Spectra binned at 5 mDa, features present in ≥80% of samples, intensity threshold >1000 counts; statistical workflows included volcano plots (FDR-corrected p<0.05, fold change>2), lasso regression (λ=0.1, 10-fold cross-validation), and random forest (1000 trees).

Main Results and Discussion

• Feature selection: 21 significant features in positive mode and 30 in negative mode distinguished PDAC from controls.
• Candidate biomarkers: Several lipid classes (LPC 18:1, LPC 18:2, LPC 20:5; LPE 18:2), amino acids (histidine, glutamine), and hydroxybutyric acid showed differential abundance.
• Statistical validation: Combined volcano, lasso, and random forest analyses confirmed robust separation between PDAC and healthy groups.
• Metabolite annotation: Narrow-band iDIA MS/MS spectra matched closely with LC-DDA-MS/MS references, enabling high-confidence identification despite isomeric ambiguity in lyso-phospholipids.

Benefits and Practical Applications

• Rapid throughput: 2 minutes per sample versus 20–30 minutes for typical LC-MS/MS workflows.
• Minimal sample handling: Direct analysis with negligible suppression and high tolerance to salt matrices.
• High sensitivity on small volumes, facilitating large-scale clinical screening.
• Compatibility with existing spectral libraries for streamlined metabolite verification.

Future Trends and Potential Applications

• Integration with advanced machine learning algorithms to refine biomarker panels and predictive models.
• Extension of DPiMS methodologies to other cancers and biofluids (urine, tissue extracts).
• Technological advancements for isomer resolution and expanded metabolite coverage.
• Development of portable DPiMS platforms for point-of-care or intraoperative diagnostics.

Conclusion

DPiMS combined with narrow-band iDIA on a high-resolution Q-TOF system delivers a fast, reliable workflow for serum metabolomic profiling in PDAC biomarker discovery. This approach eliminates the need for chromatographic separation while maintaining high identification confidence.

References

• Mandel et al. Application of Probe Electrospray Ionization Mass Spectrometry (PESI-MS) to Clinical Diagnosis: Solvent Effect on Lipid Analysis. J. Am. Soc. Mass Spectrom. 2012, 23(11):2043–2047.
• Chung et al. Using probe electrospray ionization mass spectrometry and machine learning for detecting pancreatic cancer with high performance. Am J Transl Res. 2020;12(1):171–179.