Validation of a multiomic model of plasma extracellular vesicle PD-L1 and radiomics for prediction of response to immunotherapy in NSCLC.

BACKGROUND: Immune-checkpoint inhibitors (ICIs) have showed unprecedent efficacy in the treatment of patients with advanced non-small cell lung cancer (NSCLC). However, not all patients manifest clinical benefit due to the lack of reliable predictive biomarkers. We showed preliminary data on the predictive role of the combination of radiomics and plasma extracellular vesicle (EV) PD-L1 to predict durable response to ICIs. MAIN BODY: Here, we validated this model in a prospective cohort of patients receiving ICIs plus chemotherapy and compared it with patients undergoing chemotherapy alone. This multiparametric model showed high sensitivity and specificity at identifying non-responders to ICIs and outperformed tissue PD-L1, being directly correlated with tumor change. SHORT CONCLUSION: These findings indicate that the combination of radiomics and EV PD-L1 dynamics is a minimally invasive and promising biomarker for the stratification of patients to receive ICIs.

Associating T1-Weighted and T2-Weighted Magnetic Resonance Imaging Radiomic Signatures With Preoperative Symptom Severity in Patients With Cervical Spondylotic Myelopathy.

BACKGROUND: Preoperative symptom severity in cervical spondylotic myelopathy (CSM) can be variable. Radiomic signatures could provide an imaging biomarker for symptom severity in CSM. This study utilizes radiomic signatures of T1-weighted and T2-weighted magnetic resonance imaging images to correlate with preoperative symptom severity based on modified Japanese Orthopaedic Association (mJOA) scores for patients with CSM. METHODS: Sixty-two patients with CSM were identified. Preoperative T1-weighted and T2-weighted magnetic resonance imaging images for each patient were segmented from C2-C7. A total of 205 texture features were extracted from each volume of interest. After feature normalization, each second-order feature was further subdivided to yield a total of 400 features from each volume of interest for analysis. Supervised machine learning was used to build radiomic models. RESULTS: The patient cohort had a median mJOA preoperative score of 13; of which, 30 patients had a score of >13 (low severity) and 32 patients had a score of ≤13 (high severity). Radiomic analysis of T2-weighted imaging resulted in 4 radiomic signatures that correlated with preoperative mJOA with a sensitivity, specificity, and accuracy of 78%, 89%, and 83%, respectively (P < 0.004). The area under the curve value for the ROC curves were 0.69, 0.70, and 0.77 for models generated by independent T1 texture features, T1 and T2 texture features in combination, and independent T2 texture features, respectively. CONCLUSIONS: Radiomic models correlate with preoperative mJOA scores using T2 texture features in patients with CSM. This may serve as a surrogate, objective imaging biomarker to measure the preoperative functional status of patients.

Pre-operative MRI radiomics model non-invasively predicts key genomic markers and survival in glioblastoma patients.

PURPOSE: Although glioblastoma (GBM) is the most common primary brain malignancy, few tools exist to pre-operatively risk-stratify patients by overall survival (OS) or common genetic alterations. We developed an MRI-based radiomics model to identify patients with EGFR amplification, MGMT methylation, GBM subtype, and OS greater than 12 months. METHODS: We retrospectively identified 235 patients with pathologically confirmed GBMs from the Cancer Genome Atlas (88; TCGA) and MD Anderson Cancer Center (147; MDACC). After two neuroradiologists segmented MRI tumor volumes, we extracted first-order and second-order radiomic features (gray-level co-occurrence matrices). We used the Maximum Relevance Minimum Redundancy technique to identify the 100 most relevant features and validated models using leave-one-out-cross-validation and validation on external datasets (i.e., TCGA). Our results were reported as the area under the curve (AUC). RESULTS: The MDACC patient cohort had significantly higher OS (22 months) than the TCGA dataset (14 months). On both LOOCV and external validation, our radiomics models were able to identify EGFR amplification (all AUCs > 0.83), MGMT methylation (all AUCs > 0.85), GBM subtype (all AUCs > 0.92), and OS (AUC > 0.91 on LOOCV and 0.71 for TCGA validation). CONCLUSIONS: Our robust radiomics pipeline has the potential to pre-operatively discriminate common genetic alterations and identify patients with favorable survival.

A general skull stripping of multiparametric brain MRIs using 3D convolutional neural network.

Accurate skull stripping facilitates following neuro-image analysis. For computer-aided methods, the presence of brain skull in structural magnetic resonance imaging (MRI) impacts brain tissue identification, which could result in serious misjudgments, specifically for patients with brain tumors. Though there are several existing works on skull stripping in literature, most of them either focus on healthy brain MRIs or only apply for a single image modality. These methods may be not optimal for multiparametric MRI scans. In the paper, we propose an ensemble neural network (EnNet), a 3D convolutional neural network (3DCNN) based method, for brain extraction on multiparametric MRI scans (mpMRIs). We comprehensively investigate the skull stripping performance by using the proposed method on a total of 15 image modality combinations. The comparison shows that utilizing all modalities provides the best performance on skull stripping. We have collected a retrospective dataset of 815 cases with/without glioblastoma multiforme (GBM) at the University of Pittsburgh Medical Center (UPMC) and The Cancer Imaging Archive (TCIA). The ground truths of the skull stripping are verified by at least one qualified radiologist. The quantitative evaluation gives an average dice score coefficient and Hausdorff distance at the 95th percentile, respectively. We also compare the performance to the state-of-the-art methods/tools. The proposed method offers the best performance.The contributions of the work have five folds: first, the proposed method is a fully automatic end-to-end for skull stripping using a 3D deep learning method. Second, it is applicable for mpMRIs and is also easy to customize for any MRI modality combination. Third, the proposed method not only works for healthy brain mpMRIs but also pre-/post-operative brain mpMRIs with GBM. Fourth, the proposed method handles multicenter data. Finally, to the best of our knowledge, we are the first group to quantitatively compare the skull stripping performance using different modalities. All code and pre-trained model are available at: https://github.com/plmoer/skull_stripping_code_SR .