A microRNA Prognostic Signature in Patients with Diffuse Intrinsic Pontine Gliomas through Non-Invasive Liquid Biopsy

Abstract

Diffuse midline gliomas (DMGs) originate in the thalamus, brainstem, cerebellum and spine. This entity includes tumors that infiltrate the pons, called diffuse intrinsic pontine gliomas (DIPGs), with a rapid onset and devastating neurological symptoms. Since surgical removal in DIPGs is not feasible, the purpose of this study was to profile circulating miRNA expression in DIPG patients in an effort to identify a non-invasive prognostic signature with clinical impact. Using a high-throughput platform, miRNA expression was profiled in serum samples collected at the time of MRI diagnosis and prior to radiation and/or systemic therapy from 47 patients enrolled in clinical studies, combining nimotuzumab and vinorelbine with concomitant radiation. With progression-free survival as the primary endpoint, a semi-supervised learning approach was used to identify a signature that was also tested taking overall survival as the clinical endpoint. A signature comprising 13 circulating miRNAs was identified in the training set (n = 23) as being able to stratify patients by risk of disease progression (log-rank p = 0.00014; HR = 7.99, 95% CI 2.38-26.87). When challenged in a separate validation set (n = 24), it confirmed its ability to predict progression (log-rank p = 0.00026; HR = 5.51, 95% CI 2.03-14.9). The value of our signature was also confirmed when overall survival was considered (log-rank p = 0.0021, HR = 4.12, 95% CI 1.57-10.8). We have identified and validated a prognostic marker based on the expression of 13 circulating miRNAs that can shed light on a patient's risk of progression. This is the first demonstration of the usefulness of nucleic acids circulating in the blood as powerful, easy-to-assay molecular markers of disease status in DIPG. This study provides Class II evidence that a signature based on 13 circulating miRNAs is associated with the risk of disease progression.

Keywords: circulating miRNAs; diffuse intrinsic pontine gliomas; neuro-oncology; prognosis.

Figure 1

Prognostic signature based on ct-miRNAs from the training data set. (A) Joint density estimation for the ct-miRNA signature and time-to-event variable in the training set (x-axis = miRNA model; y-axis = time-to-event; z-axis = density kernel estimate). (B) Heatmap of the expression levels related to the 13 miRNAs entering the signature. On the horizontal axis are the respective miRNAs; on the vertical axis are the patient samples (n = 23). The samples are ranked based on the signature score, having the dividing threshold at 0.007481 defining those of a low or high risk, and the line plot above the heatmap summarizes the score value per sample. (C) Kaplan–Meier survival curves for patients predicted to be at high (blue, n = 11) or low (red, n = 12) risk of cancer progression. High-risk patients had a shorter PFS (progression-free survival) than those at low risk (log-rank test, p = 0.00014; hazard ratio (HR) = 7.99, 95% confidence interval (CI) 2.38–26.87). The permutation test (based on 100 permutations) had a p-value of 0.03, indicating a low probability of overfitting for the above-mentioned log-rank analysis. The Schoenfeld individual test was assessed to test Cox regression assumption and to discard any violation considering the fast dip to zero trend for the high-risk cases in contrast to the low-risk cases. Since the Schoenfeld individual test reaches p = 0.911, the test is not statistically significant and, therefore, we can assume the proportional hazards.

Figure 2

Validation of the ct-miRNA signature. (A) Joint density estimation for the signature and time-to-event variable in the validation data set (x-axis = miRNA model; y-axis = time-to-event; z-axis = density kernel estimate). (B) Heatmap of the expression levels of the 13 miRNAs comprising the signature. On the horizontal axis are the respective miRNAs; on the vertical axis are the patient samples (n = 24). Samples are ranked based on the signature score; even if the rank order differs somewhat between the two heatmaps (training vs. validation), the division between the miRNAs remains clear; the line plot above the heatmap summarizes the score value per sample. (C) Kaplan–Meier survival curves for patients predicted to be at a high (blue, n = 12) or low (red, n = 12) risk of progression, taking PFS (progression-free survival) as the endpoint. High-risk patients had a significantly shorter PFS than those at a low risk (log-rank test, p = 0.00026; hazard ratio (HR) = 5.51, 95% confidence interval (CI) 2.03–14.9). (D) Kaplan–Meier survival curves taking OS (overall survival) as the clinical endpoint. When risk stratification by the ct-miRNA model was tested for OS, it was significantly shorter for high-risk than low-risk patients (log-rank, p = 0.0021, hazard ratio (HR) = 4.12, 95% confidence interval (CI) 1.57–10.81). High-risk in red, low-risk in blue.

Figure 2

Validation of the ct-miRNA signature. (A) Joint density estimation for the signature and time-to-event variable in the validation data set (x-axis = miRNA model; y-axis = time-to-event; z-axis = density kernel estimate). (B) Heatmap of the expression levels of the 13 miRNAs comprising the signature. On the horizontal axis are the respective miRNAs; on the vertical axis are the patient samples (n = 24). Samples are ranked based on the signature score; even if the rank order differs somewhat between the two heatmaps (training vs. validation), the division between the miRNAs remains clear; the line plot above the heatmap summarizes the score value per sample. (C) Kaplan–Meier survival curves for patients predicted to be at a high (blue, n = 12) or low (red, n = 12) risk of progression, taking PFS (progression-free survival) as the endpoint. High-risk patients had a significantly shorter PFS than those at a low risk (log-rank test, p = 0.00026; hazard ratio (HR) = 5.51, 95% confidence interval (CI) 2.03–14.9). (D) Kaplan–Meier survival curves taking OS (overall survival) as the clinical endpoint. When risk stratification by the ct-miRNA model was tested for OS, it was significantly shorter for high-risk than low-risk patients (log-rank, p = 0.0021, hazard ratio (HR) = 4.12, 95% confidence interval (CI) 1.57–10.81). High-risk in red, low-risk in blue.

Figure 3

Assessment of the model’s performance. (A) Prediction error curves for Brier scores based on the miRNA signature stratification, and also on estimates for all patients without any stratification applied (reference scenario curve). For a single patient, the Brier score at the time t is defined as the squared difference between the observed survival status and the predicted outcome probability. Red dotted line = IBS for the ct-miRNA model; black dotted line = IBS for the reference scenario. (B) Calibration plot for PFS (progression-free survival) at the landmark follow-up time point of 8.5 months. The plot shows the predictions obtained by the model on the x-axis and the observed outcomes on the y-axis. (C) Area under the ROC curves (AUC) based on our ct-miRNA model’s fitting of the PFS, obtained from a time-dependent ROC analysis. (D) DCA (decision curve analysis) for the model’s efficacy at predicting progression in DIPG, and to assess the clinical utility of the proposed miRNA model. DCA shows, graphically, the clinical usefulness of the ct-miRNA model based on a continuum of potential thresholds for patients’ risk of progression (x-axis) and the net benefit of using the ct-miRNA model to stratify patients (y-axis). The horizontal black line indicates the all-true negative rate (corresponding to the risk assumption that no patients had disease progression at 8.5 months) and the diagonal gray line indicates the all-true positive rate (corresponding to the risk assumption that all patients had disease progression at 8.5 months). The dotted line indicates an improvement in the prediction achieved by the ct-miRNA model.