Wu, Teresa

Abstract Recurrent glioblastoma is challenging to distinguish from so called “treatment effect” on routine clinical imaging. Further, within tumor heterogeneity reveals that some regions can be histologically dominated by tumor progression whilst others can be dominated by secondary effects of treatment response. Apparent tumor progression on MRI can be very difficult to manage clinically as it is unclear the degree to which the imaging changes are actually tumor progression vs response to treatment (including inflammatory response and necrosis). In this analysis, we study a unique cohort of patients for whom image localized-biopsies reveal heterogeneity in response vs progression. Our dataset included 70 biopsy samples from 32 patients with GBM each histolopatholgically characterized for tumor abundance vs immune infiltrate. Six multiparametric MRI contrasts were available, including T1, T1gd, T2, FLAIR, SWI, and ADC. Images were co-registered. Radiomic (statistical + texture) features were extracted from the region of six image contrasts locally matched with each biopsy sample. Machine learning models were built to predict each biomarker using radiomic features. Leave-one-out cross validation was used to evaluate the prediction accuracy. Radiomic features were found to be informative to the prediction of biomarkers. ANOVA tests show significant improvement of using radiomic features compared with the null model. The prediction accuracy was higher when considering the biomarkers on a binary scale using the median as the cutoff than on a numerical scale. Spatially-informed radiomics models for tumor progression vs treatment effect are possible and can play an instrumental role in navigating confounding imaging changed common during treatment progression.

Abstract BACKGROUND An important challenge in radiomics research is reproducibility. Images are collected on different image scanners and protocols, which introduces significant variability even for the same type of image across institutions. In the present proof-of-concept study, we address the reproducibility issue by using domain adaptation – an algorithm that transforms the radiomic features of each new patient to align with the distribution of features formed by the patient samples in a training set. METHOD Our dataset included 18 patients in training with a total of 82 biopsy sample. The pathological tumor cell density was available for each sample. Radiomic (statistical + texture) features were extracted from the region of six image contrasts locally matched with each biopsy sample. A Gaussian Process (GP) classifier was built to predict tumor cell density using radiomic features. Another 6 patients were used to test the training model. These patients had a total of 31 biopsy samples. The images of each test patient were purposely normalized using a different approach, i.e., using the CSF instead of the whole brain as the reference. This was to mimic the practical scenario of image source discrepancy between different patients. Domain adaptation was applied to each test patient. RESULTS Among the 18 training patients, the leave-one-patient-out cross validation accuracy is 0.81 AUC, 0.78 sensitivity, and 0.83 specificity. When the trained model was applied to the 6 test patients (purposely normalized using a different approach than that of the training data), the accuracy dramatically reduced to 0.39 AUC, 0.08 sensitivity, and 0.61 specificity. After using domain adaption, the accuracy improved to 0.68 AUC, 0.62 sensitivity, and 0.72 specificity. CONCLUSION We provide candidate enabling tools to address reproducibility in radiomics models by using domain adaption algorithms to account for discrepancy of the images between different patients.

Abstract INTRODUCTION The quantification of intratumoral heterogeneity – through radiomics-based approaches - can help resolve the regionally distinct genetic drug targets that may co-exist within a single Glioblastoma (GBM) tumor. While this offers potential diagnostic value under the paradigm of individualized oncology, clinical decision-making must also consider the degree of uncertainty associated with each model. In this study, we evaluate the performance of a novel machine-learning (ML) algorithm, called Gaussian Process (GP) modeling, that can quantify the impact of multiple sources of uncertainty in ML model development and prediction accuracy, including variabilities in the copy number measurement, radiomics features, training sample characteristics, and training sample size. METHOD We collected 95 image-localized biopsies from 25 primary GBM patients. We coregistered stereotactic locations with preoperative multi-parametric MRI features (conventional MRI, DSC perfusion, Diffusion Tensor Imaging) to generate spatially matched pairs of MRI and copy number variants (CNV) for for each biopsy. We developed a Gaussian Process (GP) model to predict CNV for Epidermal Growth Factor Receptor (EGFR) based on MRI radiomic features in each patient. We used leave-one-patient-out cross validation to quantify prediction accuracy and model uncertainty. Spatial prediction and uncertainty (p-value) maps were overlaid to help visualize regional genetic variation of EGFR and uncertainty of the radiomic predictions. RESULT: The initial GP radiomics model for EGFR amplification (CNV > 3.5) produced a sensitivity of 0.8 and specificity of 0.8. Samples/regions associated with high uncertainty (p-value >0.05) correlated with either 1) extrapolation of radiomic features from the training set-defined feature space or 2) insufficient training samples in the feature space. CONCLUSION We present a ML-based model that quantifies spatial genetic heterogeneity in GBM, while also estimating model uncertainties that result from multi-source data variabilities. This approach lays the groundwork for prospective clinical integration of modeling-based diagnostic approaches in the paradigm of individualized medicine.

e15626 Background: FGFR2 fusions are found in 10-15% of intrahepatic cholangiocarcinomas. Individuals with FGFR2 fusions have a better prognosis and more indolent natural history. More recently FGFR2 fusions have been described as a predictive marker for tyrosine kinase inhibitors targeting the FGFR2 pathway. Texture analysis from CT imaging has been used in the past to estimate degree of fibrosis and angiogenesis in various tumors types; however, texture analysis has not been used in the past to detect FGFR2 rearrangements. Methods: This is a retrospective study of 33 patients with intrahepatic cholangiocarcinoma who were previously tested for FGFR2 fusions with either a FISH break apart assay or with clinically available next generation sequencing assays. All patients had a contrast enhanced CT performed. To address physiological and pathological variations, matched-size regions of interest (ROIs) over areas of non-necrotic tumor and unaffected liver during the portal venous phase were drawn. The paired ROIs for each patient were then processed via an imaging analytic pipeline for (1) ROI normalization, (2) texture panel creation, and (3) machine learning predictive model development. Results: Fifteen patients with FGFR2 fusions and 18 patients without FGFR2 fusions were included in the study. Using the developed predictive model, accurate FGFR2 fusion classification in 90% of cases was achieved (sensitivity = 87%, specificity = 94%). Conclusions: Utilizing advanced analytics with machine learning techniques, multi-textural data obtained from conventional CT images has the capability to detect genetic aberrations in intrahepatic cholangiocarcinoma, including FGFR2 gene fusions. Radiogenomic evaluation has several attractive benefits including the cost effective and expeditious detection of actionable mutations such as FGFR2 fusions as well as avoidance of invasive procedures such as percutaneous core biopsies. [Table: see text]

ObjectiveTo determine whether machine learning can accurately classify human papillomavirus (HPV) status of oropharyngeal squamous cell carcinoma (OPSCC) using computed tomography (CT)-based texture analysis.MethodsTexture analyses were retrospectively applied to regions of interest from OPSCC primary tumors on contrast-enhanced neck CT, and machine learning was used to create a model that classified HPV status with the highest accuracy. Results were compared against the blinded review of 2 neuroradiologists.ResultsThe HPV-positive (n = 92) and -negative (n = 15) cohorts were well matched clinically. Neuroradiologist classification accuracies for HPV status (44.9%, 55.1%) were not significantly different (P= 0.13), and there was a lack of agreement between the 2 neuroradiologists (κ = −0.145). The best machine learning model had an accuracy of 75.7%, which was greater than either neuroradiologist (P< 0.001,P= 0.002).ConclusionsUseful diagnostic information regarding HPV infection can be extracted from the CT appearance of OPSCC beyond what is apparent to the trained human eye.