Citations

Abstract Introduction Tumefactive multiple sclerosis (MS) and glioblastoma (GBM) can be difficult to differentiate. Clinical presentations are nonspecific and may occur in patients with both diseases. Magnetic resonance imaging (MRI) scans are routinely used to generate an initial differential diagnosis. Although imaging features can help distinguish these entities, there are no pathognomonic signs, neither on traditional nor advanced MRI. Moreover, the ability to differentiate these entities depends on the radiologists’ experience. Ensuring the correct diagnosis has important therapeutic and prognostic implications, as the management of these diseases is very different. To date, a brain biopsy often represents the only option to obtain a diagnosis, but it may be inconclusive and could add unnecessary morbidity. We hypothesize that a machine-learning tool based on routinely performed MRI scans could guarantee a more reliable, fast, and reproducible diagnosis. Methods Only patients with a histological diagnosis were included in the study. Patients that underwent any form of radiation therapy before the surgery were excluded to avoid confounding factors due to treatment-related effects. A total of 99 MS and 257 GBMs were included, and preoperative T2-weighted images were collected. For each patient, the lesions' segmentation was performed, including both the contrast-enhancing and non-enhancing areas of the lesions. The segmentation masks were used to extract radiomic features from the T2-weighted sequences. All MRI images were standardized to the same mean and standard deviation before feature extraction and resampled to a voxel size of 1.0 × 1.0 × 1.0 mm. Features extraction was performed using Pyradiomics (v. 3.0.1) and included first order, shape, and texture features, totaling 100 features. The classifier consisted of a support vector machine (SVM) model. Patients were randomly assigned to the training and test sets with an 80-20% proportion for each class. To select the best hyperparameters for the SVM classifier, a grid-search analysis was performed on the training set using a nested cross-validation scheme (k=5). All experiments were performed using Python (v. 3.5) and the Scikit-learn library (v. 0.23.1). Results The mean accuracy and area under the ROC curve for the best set of hyperparameters on the validation set were 0.83 and 0.86, respectively. The weighted-average precision, recall, and F1 scores on the test set were 0.78, 0.76, and 0.77, respectively. Conclusions Radiomic features extracted from a routinely acquired MRI sequence showed promising results in differentiating tumefactive multiple sclerosis from glioblastoma, opening the possibility to obtain a reliable diagnosis without the need for a brain biopsy. Future experiments Future experiments will focus on using post-contrast T1-weighted images for the radiomic features extractions, testing multiple machine learning models and features selection algorithms, and including more diagnostic classes (i.e., metastasis, lymphomas). Citation Format: Gian Marco Conte, Jeanette E. Eckel-Passow, W. Oliver Tobin, Paul Decker, Daniel H. Lachance, Robert B. Jenkins, Bradley J. Erickson. Differentiation of tumefactive multiple sclerosis and glioblastoma using radiomics features extracted from magnetic resonance imaging and machine learning [abstract]. In: Proceedings of the AACR Virtual Special Conference on Artificial Intelligence, Diagnosis, and Imaging; 2021 Jan 13-14. Philadelphia (PA): AACR; Clin Cancer Res 2021;27(5_Suppl):Abstract nr PO-023.

431 Background: Patients (pts) with localized PC do not routinely undergo comprehensive genomic profiling (CGP) unless they develop recurrent or metastatic disease. KRAS is the most frequently mutated gene in PC, however, the impact of different KRAS mutations in localized PC has not been well characterized. We interrogated our genomic database to analyze the KRAS status in PC pts who presented with localized disease at diagnosis (Dx). Methods: We identified PC pts at our institution who underwent CGP utilizing the Foundation One CDx assay and had localized disease at initial Dx; these pts were categorized into resectable/borderline resectable PC (LPC) and locally advanced PC (LAPC). All pts with LPC and LAPC underwent neoadjuvant chemotherapy and chemoradiation prior to possible surgery (all intended therapy - AIT). Tissue from metastatic sites was used for CGP in pts who developed recurrent/metastatic disease before or after completion of AIT. The primary tumor was used for CGP in pts who completed AIT without subsequent relapse or in the absence of adequate metastatic tissue. Effect of each gene on response and survival outcomes was estimated using proportional odds and Cox regression analysis, respectively, adjusting for stage. Results: 75 pts were identified, median age at Dx was 65 years, 59% were male, 65% had a primary tumor in the pancreatic head. 38 (86%) pts with LPC completed AIT compared to 21 (68%) pts with LAPC (p<0.001). KRAS mutation was detected in 95% (71/75) of pts– 94% (67/71) in codon 12 and 6% (4/71) in codon 61. The various KRAS mutations and their association with completion of AIT is summarized in the table. The likelihood of completing AIT did not differ based on KRAS wildtype (WT) vs mutated status (p =1.00), the mutated codon (codon 12 vs. codon 61; p =1.00) or the individual KRAS point mutations (p = 0.7); however, all patients with G12A (N= 1), G12C (N=1), G12L (N=1) and G12R (N=11) mutations completed AIT. KRAS status (mutated vs. WT) and the individual KRAS mutations were not associated with overall survival (OS) after adjusting for stage (p= 0.13 and p = 0.26 respectively). Median OS for patients with LPC and LAPC, was 39 months (mos) and 29 mos respectively. Conclusions: KRAS status and individual KRAS mutations did not have an impact on completing AIT or mOS; however, these findings need to be interpreted with caution due to the inherent biases involved in such analyses. The clinical significance and functional relevance of KRAS G12A, G12C, G12L and G12R mutations, though relatively rare, needs further characterization as well as mechanistic elucidation. [Table: see text]

Abstract Objective Determine if information regarding hypertrophic cardiomyopathy (HCM) can be accurately retrieved from cardiac magnetic resonance (CMR) reports using natural language processing (NLP). Background CMR imaging is used for diagnosis and risk stratification of HCM. Manual annotation of information from CMR is time-consuming. NLP is an artificial intelligence method for automating extraction of information from narrative text. Methods We identified 200 HCM patients who had CMR reports from 1998 to 2018. These patients were randomly allocated into training (100 patients with 185 CMR reports) and testing sets (100 patients with 206 reports). An NLP system with 2 tiers was developed; the first extracted information regarding HCM diagnosis while second extracted categorical or numeric concepts for HCM classification. NLP performance was compared with gold-standard manual annotation. Results NLP algorithms achieved very high performance across all concepts with mean positive predictive value (PPV) = 0.96. An outlier was the performance for abstracting the presence of an apical pouch from CMR reports, which had noticeably lower PPV= 0.78 which be attributed to the low number of cases with this finding. Conclusions The algorithms developed can be translated to clinical decision support systems to increase efficiency and contribute to improved quality of care. Funding Acknowledgement Type of funding source: Other. Main funding source(s): Study supported by the National Heart, Lung and Blood Institute of National Institutes of Health (K01HL124045), the Mayo Clinic Center for Clinical and Translational Science (CCaTS), and the Mayo Clinic K2R award. Content is solely the responsibility of authors and does not necessarily represent official views of the National Institutes of Health.

Abstract Background Autosomal dominant polycystic kidney disease (ADPKD) is one of the most common monogenetic disorders in humans and is characterized by numerous fluid-filled cysts that grow slowly, resulting in end-stage renal disease in the majority of patients. Preclinical studies have indicated that treatment with low-dose thiazolidinediones, such as pioglitazone, decrease cyst growth in rodent models of PKD. Methods This Phase 1b cross-over study compared the safety of treatment with a low dose (15 mg) of the peroxisome proliferator-activated receptor-γ (PPAR-γ) agonist pioglitazone or placebo in PKD patients, with each treatment given for 1 year. The study monitored known side effects of PPAR-γ agonist treatment, including fluid retention and edema. Liver enzymes and risk of hypoglycemia were assessed throughout the study. As a secondary objective, the efficacy of low-dose pioglitazone was followed using a primary assessment of total kidney volume (TKV), blood pressure (BP) and kidney function. Results Eighteen patients were randomized and 15 completed both arms. Compared with placebo, allocation to pioglitazone resulted in a significant decrease in total body water as assessed by bioimpedance analysis {mean difference 0.16 Ω [95% confidence interval (CI) 0.24–2.96], P = 0.024} and no differences in episodes of heart failure, clinical edema or change in echocardiography. Allocation to pioglitazone led to no difference in the percent change in TKV of −3.5% (95% CI −8.4–1.4, P = 0.14), diastolic BP and microalbumin:creatinine ratio. Conclusions In this small pilot trial in people with ADPKD but without diabetes, pioglitazone 15 mg was found to be as safe as placebo. Larger and longer-term randomized trials powered to assess effects on TKV are needed.

PurposeOrgan segmentation of computed tomography (CT) imaging is essential for radiotherapy treatment planning. Treatment planning requires segmentation not only of the affected tissue, but nearby healthy organs‐at‐risk, which is laborious and time‐consuming. We present a fully automated segmentation method based on the three‐dimensional (3D) U‐Net convolutional neural network (CNN) capable of whole abdomen and pelvis segmentation into 33 unique organ and tissue structures, including tissues that may be overlooked by other automated segmentation approaches such as adipose tissue, skeletal muscle, and connective tissue and vessels. Whole abdomen segmentation is capable of quantifying exposure beyond a handful of organs‐at‐risk to all tissues within the abdomen.MethodsSixty‐six (66) CT examinations of 64 individuals were included in the training and validation sets and 18 CT examinations from 16 individuals were included in the test set. All pixels in each examination were segmented by image analysts (with physician correction) and assigned one of 33 labels. Segmentation was performed with a 3D U‐Net variant architecture which included residual blocks, and model performance was quantified on 18 test cases. Human interobserver variability (using semiautomated segmentation) was also reported on two scans, and manual interobserver variability of three individuals was reported on one scan. Model performance was also compared to several of the best models reported in the literature for multiple organ segmentation.ResultsThe accuracy of the 3D U‐Net model ranges from a Dice coefficient of 0.95 in the liver, 0.93 in the kidneys, 0.79 in the pancreas, 0.69 in the adrenals, and 0.51 in the renal arteries. Model accuracy is within 5% of human segmentation in eight of 19 organs and within 10% accuracy in 13 of 19 organs.ConclusionsThe CNN approaches the accuracy of human tracers and on certain complex organs displays more consistent prediction than human tracers. Fully automated deep learning‐based segmentation of CT abdomen has the potential to improve both the speed and accuracy of radiotherapy dose prediction for organs‐at‐risk.

Abstract Despite the widespread clinical use of dynamic susceptibility contrast (DSC) MRI, DSC-MRI methodology has not been standardized, hindering its utilization for response assessment in multicenter trials. Recently, the DSC-MRI Standardization Subcommittee of the Jumpstarting Brain Tumor Drug Development Coalition issued an updated consensus DSC-MRI protocol compatible with the standardized brain tumor imaging protocol (BTIP) for high-grade gliomas that is increasingly used in the clinical setting and is the default MRI protocol for the National Clinical Trials Network. After reviewing the basis for controversy over DSC-MRI protocols, this paper provides evidence-based best practices for clinical DSC-MRI as determined by the Committee, including pulse sequence (gradient echo vs spin echo), BTIP-compliant contrast agent dosing (preload and bolus), flip angle (FA), echo time (TE), and post-processing leakage correction. In summary, full-dose preload, full-dose bolus dosing using intermediate (60°) FA and field strength-dependent TE (40–50 ms at 1.5 T, 20–35 ms at 3 T) provides overall best accuracy and precision for cerebral blood volume estimates. When single-dose contrast agent usage is desired, no-preload, full-dose bolus dosing using low FA (30°) and field strength-dependent TE provides excellent performance, with reduced contrast agent usage and elimination of potential systematic errors introduced by variations in preload dose and incubation time.