NMR-Metabolomics Reveals a Metabolic Shift after Surgical Resection of Non-Small Cell Lung Cancer

Abstract

Background: Lung cancer can be detected by measuring the patient's plasma metabolomic profile using nuclear magnetic resonance (NMR) spectroscopy. This NMR-based plasma metabolomic profile is patient-specific and represents a snapshot of the patient's metabolite concentrations. The onset of non-small cell lung cancer (NSCLC) causes a change in the metabolite profile. However, the level of metabolic changes after complete NSCLC removal is currently unknown.

Patients and methods: Fasted pre- and postoperative plasma samples of 74 patients diagnosed with resectable stage I-IIIA NSCLC were analyzed using ¹H-NMR spectroscopy. NMR spectra (s = 222) representing two preoperative and one postoperative plasma metabolite profile at three months after surgical resection were obtained for all patients. In total, 228 predictors, i.e., 228 variables representing plasma metabolite concentrations, were extracted from each NMR spectrum. Two types of supervised multivariate discriminant analyses were used to train classifiers presenting a strong differentiation between the pre- and postoperative plasma metabolite profiles. The validation of these trained classification models was obtained by using an independent dataset.

Results: A trained multivariate discriminant classification model shows a strong differentiation between the pre- and postoperative NSCLC profiles with a specificity of 96% (95% CI [86-100]) and a sensitivity of 92% (95% CI [81-98]). Validation of this model results in an excellent predictive accuracy of 90% (95% CI [77-97]) and an AUC value of 0.97 (95% CI [0.93-1]). The validation of a second trained model using an additional preoperative control sample dataset confirms the separation of the pre- and postoperative profiles with a predictive accuracy of 93% (95% CI [82-99]) and an AUC value of 0.97 (95% CI [0.93-1]). Metabolite analysis reveals significantly increased lactate, cysteine, asparagine and decreased acetate levels in the postoperative plasma metabolite profile.

Conclusions: The results of this paper demonstrate that surgical removal of NSCLC generates a detectable metabolic shift in blood plasma. The observed metabolic shift indicates that the NSCLC metabolite profile is determined by the tumor's presence rather than donor-specific features. Furthermore, the ability to detect the metabolic difference before and after surgical tumor resection strongly supports the prospect that NMR-generated metabolite profiles via blood samples advance towards early detection of NSCLC recurrence.

Keywords: NMR spectroscopy; metabolomics; non-small cell lung cancer.

Figure 1

Study flow diagram showing the overview of patient recruitment at Ziekenhuis Oost-Limburg (ZOL) from June 2018 until August 2021 of the NCT03736993 trial with the indication of the three blood sampling time points (control, baseline and effect). The inclusion stop between March and June 2020 due to the COVID-19 pandemic had no further impact on the study participants’ planned blood sampling time points. * Surgery was cancelled or excluded due to patient drop-out. † Postoperative pathological diagnosis of a non-malignant lesion or another type of cancer. *, † Patients were excluded from the NCT03736993 trial and no further blood sampling occurred. ‡ Datasets from at least one of the three time points were missing due to high glycemic values (>200 mg/dL) or exclusion due to patient drop-out. ‡, ¥ Patients were excluded from this explorative sub-study, but not from the NCT03736993 clinical trial.

Figure 2

Overview of the longitudinal study design and three different blood sampling time points. In total, three blood samples were donated by 74 NSCLC patients, who were enrolled in a follow-up period of at least six months after surgery. Clinical evaluation of disease relapse was performed six months after surgery. The preoperative blood sampling at baseline always occurred in the morning on the day of surgery. As indicated in the timeline, the preoperative control and postoperative-3M blood samples were taken within a controlled timespan. B: baseline, preoperative; C: control, preoperative; E: effect, postoperative-3M, three months after surgery; Postoperative-6M: six months after surgical tumor resection, respectively.

Figure 3

Detection of the metabolic shift between baseline/effect (A,B) and control/effect (C,D): the OPLS-DA classification models demonstrate the excellent separation between the pre- and postoperative plasma metabolite fingerprints of NSCLC patients. Each metabolite profile is represented by a specific labeled position in the OPLS-DA score plots. The clear separation on the predictive x-axis, representing the variation between the two groups, illustrates the metabolic shift between the pre- and postoperative metabolite profiles. (A) B/E training classification model (s = 100) using baseline preoperative and postoperative-3M plasma metabolic datasets of 50 NSCLC patients without early disease progression within six months after surgical tumor removal. The trained classifier allows excellent differentiation between the pre- and postoperative metabolite profile with 94% accuracy and shows great predictive accuracy, as demonstrated by the high Q2 value of 0.42. (B) B/E validation of classification (s = 48) based on the trained model using independent baseline preoperative and postoperative-3M plasma metabolic datasets of the remaining 24 patients. The validation confirms the observed metabolic shift between the pre- and postoperative metabolite profiles by discriminating with 90% accuracy between the two groups. (C) C/E training classification model (s = 100) using control preoperative and postoperative-3M plasma metabolic datasets of 50 NSCLC patients. The trained classifier confirms the clear differentiation between the pre- and postoperative metabolite profile with 89% accuracy, as shown in the B/E classifier. (D) C/E validation classification (s = 46) based on the trained model using independent control preoperative and postoperative-3M plasma metabolic datasets of the remaining 23 patients. Validation again confirms the observed metabolic shift between the pre- and postoperative metabolite profiles by showing a differentiation with 93% accuracy between the two groups. B: baseline, preoperative; C: control, preoperative; E: effect, postoperative-3M, three months after surgery; PS: predictive scores.

Figure 4

Receiving operating curves (ROC) demonstrate the excellent area under the curve (AUC) values for the differentiation between the pre- and postoperative B/E and C/E models. Cumulative ROC curves for the B/E and C/E OPLS-DA training and validation classification models show excellent AUC values in contrast to the ROC curves of the preoperative B/C OPLS-DA models. The gray zone represents the 95% confidence interval of the AUC values obtained by internal validation via bootstrapping resampling. AUC: area under the curve; B: baseline, preoperative; C: control, preoperative; E: effect, postoperative-3M, three months after surgery.

Figure 5

Boxplots showing the significant difference between the pre- and postoperative plasma concentration of lactate, cysteine, asparagine and acetate. Postoperative levels of lactate, cysteine and asparagine are elevated, while the postoperative acetate concentration is decreased. (***: significant results with p-value < 0.001, ·:datapoint outside the 1.5 IQR range).