Evaluation of mediastinal lymph node segmentation of heterogeneous CT data with full and weak supervision.

Accurate lymph node size estimation is critical for staging cancer patients, initial therapeutic management, and assessing response to therapy. Current standard practice for quantifying lymph node size is based on a variety of criteria that use uni-directional or bi-directional measurements. Segmentation in 3D can provide more accurate evaluations of the lymph node size. Fully convolutional neural networks (FCNs) have achieved state-of-the-art results in segmentation for numerous medical imaging applications, including lymph node segmentation. Adoption of deep learning segmentation models in clinical trials often faces numerous challenges. These include lack of pixel-level ground truth annotations for training, generalizability of the models on unseen test domains due to the heterogeneity of test cases and variation of imaging parameters. In this paper, we studied and evaluated the performance of lymph node segmentation models on a dataset that was completely independent of the one used to create the models. We analyzed the generalizability of the models in the face of a heterogeneous dataset and assessed the potential effects of different disease conditions and imaging parameters. Furthermore, we systematically compared fully-supervised and weakly-supervised methods in this context. We evaluated the proposed methods using an independent dataset comprising 806 mediastinal lymph nodes from 540 unique patients. The results show that performance achieved on the independent test set is comparable to that on the training set. Furthermore, neither the underlying disease nor the heterogeneous imaging parameters impacted the performance of the models. Finally, the results indicate that our weakly-supervised method attains 90%- 91% of the performance achieved by the fully supervised training.

Phase I study of intraventricular infusions of autologous ex vivo expanded NK cells in children with recurrent medulloblastoma and ependymoma.

BACKGROUND: Recurrent pediatric medulloblastoma and ependymoma have a grim prognosis. We report a first-in-human, phase I study of intraventricular infusions of ex vivo expanded autologous natural killer (NK) cells in these tumors, with correlative studies. METHODS: Twelve patients were enrolled, 9 received protocol therapy up to 3 infusions weekly, in escalating doses from 3 × 106 to 3 × 108 NK cells/m2/infusion, for up to 3 cycles. Cerebrospinal fluid (CSF) was obtained for cellular profile, persistence, and phenotypic analysis of NK cells. Radiomic characterization on pretreatment MRI scans was performed in 7 patients, to develop a non-invasive imaging-based signature. RESULTS: Primary objectives of NK cell harvest, expansion, release, and safety of 112 intraventricular infusions of NK cells were achieved in all 9 patients. There were no dose-limiting toxicities. All patients showed progressive disease (PD), except 1 patient showed stable disease for one month at end of study follow-up. Another patient had transient radiographic response of the intraventricular tumor after 5 infusions of NK cell before progressing to PD. At higher dose levels, NK cells increased in the CSF during treatment with repetitive infusions (mean 11.6-fold). Frequent infusions of NK cells resulted in CSF pleocytosis. Radiomic signatures were profiled in 7 patients, evaluating ability to predict upfront radiographic changes, although they did not attain statistical significance. CONCLUSIONS: This study demonstrated feasibility of production and safety of intraventricular infusions of autologous NK cells. These findings support further investigation of locoregional NK cell infusions in children with brain malignancies.

A Coclinical Radiogenomic Validation Study: Conserved Magnetic Resonance Radiomic Appearance of Periostin-Expressing Glioblastoma in Patients and Xenograft Models.

PURPOSE: Radiomics is the extraction of multidimensional imaging features, which when correlated with genomics, is termed radiogenomics. However, radiogenomic biological validation is not sufficiently described in the literature. We seek to establish causality between differential gene expression status and MRI-extracted radiomic-features in glioblastoma. EXPERIMENTAL DESIGN: Radiogenomic predictions and validation were done using the Cancer Genome Atlas and Repository of Molecular Brain Neoplasia Data glioblastoma patients (n = 93) and orthotopic xenografts (OX; n = 40). Tumor phenotypes were segmented, and radiomic-features extracted using the developed radiome-sequencing pipeline. Patients and animals were dichotomized on the basis of Periostin (POSTN) expression levels. RNA and protein levels confirmed RNAi-mediated POSTN knockdown in OX. Total RNA of tumor cells isolated from mouse brains (knockdown and control) was used for microarray-based expression profiling. Radiomic-features were utilized to predict POSTN expression status in patient, mouse, and interspecies. RESULTS: Our robust pipeline consists of segmentation, radiomic-feature extraction, feature normalization/selection, and predictive modeling. The combination of skull stripping, brain-tissue focused normalization, and patient-specific normalization are unique to this study, providing comparable cross-platform, cross-institution radiomic features. POSTN expression status was not associated with qualitative or volumetric MRI parameters. Radiomic features significantly predicted POSTN expression status in patients (AUC: 76.56%; sensitivity/specificity: 73.91/78.26%) and OX (AUC: 92.26%; sensitivity/specificity: 92.86%/91.67%). Furthermore, radiomic features in OX were significantly associated with patients with similar POSTN expression levels (AUC: 93.36%; sensitivity/specificity: 82.61%/95.74%; P = 02.021E-15). CONCLUSIONS: We determined causality between radiomic texture features and POSTN expression levels in a preclinical model with clinical validation. Our biologically validated radiomic pipeline also showed the potential application for human-mouse matched coclinical trials.