Evaluation of 1H-NMR spectroscopy in predicting platine-related toxicities in lung cancer patients.

Abstract

Introduction: Lung cancer is the leading cause of cancer mortality worldwide. Despite progress in treatment of lung cancer we experience unacceptable toxicities without a significant increase in overall survival (OS) and progression free survival (PFS). Clinical trials do not answer the question of who will experience toxicities of chemotherapy like bone marrow suppression or renal insufficiency. The goal of this study is to find out whether the metabolic profile derived from blood plasma can differentiate between subject groups who do and do not develop toxicities.

Objective : Prediction of severe toxicities of chemotherapy by means of differences in the metabolic profile of plasma. Renal failure, trombopenia, leucopenia, neutropenia and anemia are defined according the common toxicity criteria of the EORTC.

Methods: This project investigates subjects with a new diagnosis of lung cancer (n = 76) independent of stage (IB 1.3 %, IIA, 7.9%, IIB 5.3%, IIIA 31.6%, IIIB 18.4% and stage IV 35.5%) and who are candidate to receive chemotherapy in palliative (n=40) or curative setting (n=36). All patients received a platinum based cytotoxic treatment : carboplatinum (n =16) or cisplatinum (n=60). Participants are asked to provide a fasted sample of blood. Samples are collected before the start of the treatment and analysed by 1H-NMR spectroscopy using a 400 MHz spectrometer. Slightly T2-weighted spectra were acquired using a CPMG pulse sequence and the chemical shifts are referred to TSP at 0.015 ppm. The integration values of 110 spectral integration regions (IRs) are analyzed appropriate statistical methods followed by an orthogonal partial least squares discriminant analysis (OPLS-DA) to investigate whether a discriminating classifier can be constructed. In this way we aim to develop a method that predict which patients will experience toxicities.

Results: There were severe adverse events (grade 3 and 4) in both groups : trombopenia (9%), neutropenia ( 38%), leukopenia (17%) and anemia (8 %). There where no cases of severe renal failure. According to the criteria of toxicity of the ECOG subgroups were created. First of all normal probability plots were constructed and a Kolmogorov-Smirnov test was carried out. Analysis of variance (ANOVA)was performed is the null hypothesis of the Kolmogorov Smirnov test was accepted (sign > 0.05). In the case of rejection of the null hypothesis Kruskall-Wallis was performed. Statistical analyzes demonstrated no significant difference between the groups with and without toxicities. This was true for all the toxicities of interest in this study. Multivariate analyzes by SIMCA (OPLS-DA) was not able to construct a discriminating classifier.

Conclusion: In this limited patient cohort, lung cancer patients experiencing toxicities could not be differentiated from those without toxic effects based on the metabolic profiles obtained by 1H-NMR spectroscopy. However this might be due to the treatment schedule, number of cycles administered, and the rather heterogeneous group of subjects.