Callstrom, Matthew

Abstract The advent of high-throughput sequencing has led to an exponential increase in genomic data, highlighting the need for efficient and accurate models to analyze and interpret this information. In this study, we introduce a novel, exomic foundational model that leverages a combination of the human reference genome and multispecies data to improve variant detection and interpretation. Our model utilizes a short-range transformer architecture and is trained on a large dataset of human exomic sequences derived from the Tapestry study. Through a series of ablation studies and scaling experiments, we demonstrate the effectiveness of our model in predicting next token accuracy and identifying clinically pathogenic variants. We also show that our model outperforms existing models in a range of downstream tasks, including variant effect prediction and disease state identification. In fact, our largest sequence transformer for range adaptive nucleotide decoding variant (1B parameters) surpassed previous benchmarks, demonstrating a mean accuracy of 0.880 [an 8.2% improvement over the original nucleotide transformer (NT) and a 7% improvement over NT-v2]. Furthermore, we construct a unique exomic ClinVar dataset to evaluate the model’s performance on pathogenicity and disease states. Our results highlight the potential of this model to improve our understanding of the human exome and its role in disease. The model and its applications have significant implications for genomics-based diagnosis and personalized medicine, including tailored therapeutic development.

ObjectivesMetabolic dysfunction‐associated steatotic liver disease (MASLD) is the most prevalent liver disorder in Western countries, with approximately 20%–30% of the MASLD patients progressing to severe stages. There is an urgent need for noninvasive, cost‐effective, widely accessible, and precise biomarkers to evaluate liver steatosis. This study aims to assess and compare the diagnostic performance of a novel reference frequency method‐based ultrasound attenuation coefficient estimation (ACE) in both fundamental (RFM‐ACE‐FI) and harmonic (RFM‐ACE‐HI) imaging for detecting and grading liver steatosis.MethodsAn Institutional Review Board‐approved prospective study was carried out between December 2018 and October 2022. A total number of 130 subjects were enrolled in the study. The correlation between RFM‐ACE‐HI values and magnetic resonance imaging proton density fat fraction (MRI‐PDFF), as well as between RFM‐ACE‐FI values and MRI‐PDFF were calculated. The diagnostic performance of RFM‐ACE‐FI and RFM‐ACE‐HI was evaluated using receiver operating characteristic (ROC) curve analysis, as compared to MRI‐PDFF. The reproducibility of RFM‐ACE‐HI was assessed by interobserver agreement between two sonographers.ResultsA strong correlation was observed between RFM‐ACE‐HI and MRI‐PDFF, with R = 0.88 (95% confidence interval [CI]: 0.83–0.92; P < .001), while the correlation between RFM‐ACE‐FI and MRI‐PDFF was R = 0.65 (95% CI: 0.50–0.76; P < .001). The area under the ROC (AUROC) curve for RFM‐ACE‐HI in staging liver steatosis grades of S ≥ 1 and S ≥ 2 was 0.97 (95% CI: 0.91–0.99; P < .001) and 0.98 (95% CI: 0.93–1.00; P < .001), respectively, and 0.76 (95% CI: 0.65–0.85) and 0.80 (95% CI: 0.70–0.88) for RFM‐ACE‐FI, respectively. Great reproducibility was achieved for RFM‐ACE‐HI, with an interobserver agreement of R = 0.97 (95% CI: 0.94–0.99; P < .001).ConclusionsThe novel RFM‐ACE‐HI method offered high liver steatosis diagnostic accuracy and reproducibility, which has important clinical implications for early disease intervention and treatment evaluation.

AbstractBackgroundComputed tomography (CT) is routinely used to guide cryoablation procedures. Notably, CT‐guidance provides 3D localization of cryoprobes and can be used to delineate frozen tissue during ablation. However, metal‐induced artifacts from ablation probes can make accurate probe placement challenging and degrade the ice ball conspicuity, which in combination could lead to undertreatment of potentially curable lesions.PurposeIn this work, we propose an image‐based neural network (CNN) model for metal artifact reduction for CT‐guided interventional procedures.MethodsAn image domain metal artifact simulation framework was developed and validated for deep‐learning‐based metal artifact reduction for interventional oncology (MARIO). CT scans were acquired for 19 different cryoablation probe configurations. The probe configurations varied in the number of probes and the relative orientations. A combination of intensity thresholding and masking based on maximum intensity projections (MIPs) was used to segment both the probes only and probes + artifact in each phantom image. Each of the probe and probe + artifact images were then inserted into 19 unique patient exams, in the image domain, to simulate metal artifact appearance for CT‐guided interventional oncology procedures. The resulting 361 pairs of simulated image volumes were partitioned into disjoint training and test datasets of 304 and 57 volumes, respectively. From the training partition, 116 600 image patches with a shape of 128 × 128 × 5 pixels were randomly extracted to be used for training data. The input images consisted of a superposition of the patient and probe + artifact images. The target images consisted of a superposition of the patient and probe only images. This dataset was used to optimize a U‐Net type model. The trained model was then applied to 50 independent, previously unseen CT images obtained during renal cryoablations. Three board‐certified radiologists with experience in CT‐guided ablations performed a blinded review of the MARIO images. A total of 100 images (50 original, 50 MARIO processed) were assessed across different aspects of image quality on a 4‐point likert‐type item. Statistical analyses were performed using Wilcoxon signed‐rank test for paired samples.ResultsReader scores were significantly higher for MARIO processed images compared to the original images across all metrics (all p < 0.001). The average scores of the overall image quality, iceball conspicuity, overall metal artifact, needle tip visualization, target region confidence, and worst metal artifact, needle tip visualization, iceball conspicuity, and target region confidence improved by 34.91%, 36.29%, 39.94%, 34.17%, 35.13%, and 45.70%, respectively.ConclusionsThe proposed method of image‐based metal artifact simulation can be used to train a MARIO algorithm to effectively reduce probe‐related metal artifacts in CT‐guided cryoablation procedures.

Abstract Objective . Ultrasound attenuation coefficient estimation (ACE) has diagnostic potential for clinical applications such as quantifying fat content in the liver. Previously, we have proposed a system-independent ACE technique based on spectral normalization of different frequencies, called the reference frequency method (RFM). This technique does not require a well-calibrated reference phantom for normalization. However, this method may be vulnerable to severe reverberation clutter introduced by the body wall. The clutter superimposed on liver echoes may bias the estimation. Approach . We proposed to use robust principal component analysis, combined with wavelet-based sparsity promotion, to suppress the severe reverberation clutters. The capability to mitigate the reverberation clutters was validated through phantom and in vivo studies. Main Results . In the phantom studies with added reverberation clutters, higher normalized cross-correlation and smaller mean absolute errors were attained as compared to RFM results without the proposed method, demonstrating the capability to reconstruct tissue signals from reverberations. In a pilot patient study, the correlation between ACE and proton density fat fraction (PDFF), a measurement of liver fat by MRI as a reference standard, was investigated. The proposed method showed an improvement of the correlation (coefficient of determination, R = 0.82) as compared with the counterpart without the proposed method ( R = 0.69). Significance : The proposed method showed the feasibility of suppressing the reverberation clutters, providing an important basis for the development of a robust ACE with large reverberation clutters.

Metastatic pheochromocytoma and paraganglioma (PPGL) are incurable neuroendocrine tumors. The goals of treatment include palliating symptoms and reducing tumor burden. Little is known about the use of radiofrequency ablation (RFA), cryoablation (CRYO), and percutaneous ethanol injection (PEI) to treat metastatic PPGL. We performed a retrospective study of patients age 17 years and older with metastatic PPGL who were treated with ablative therapy at Mayo Clinic, USA, between June 14, 1999 and November 14, 2017. Our outcomes measures were radiographic response, procedure-related complications, and symptomatic improvement. Thirty-one patients with metastatic PPGL had 123 lesions treated during 42 RFA, 23 CRYO, and 4 PEI procedures. The median duration of follow-up was 60 months (range, 0–163 months) for non-deceased patients. Radiographic local control was achieved in 69/80 (86%) lesions. Improvement in metastasis-related pain or symptoms of catecholamine excess was achieved in 12/13 (92%) procedures. Thirty-three (67%) procedures had no known complications. Clavien-Dindo Grade I, II, IV, and V complications occurred after 7 (14%), 7 (14%), 1 (2%), and 1 (2%) of the procedures, respectively. In patients with metastatic PPGL, ablative therapy can effectively achieve local control and palliate symptoms.

Abstract Lessons Learned Percutaneous thermal ablation combined with in situ granulocyte-macrophage colony-stimulating factor cytokine therapy was technically feasible and well tolerated. No significant clinical or immunologic responses were seen. Background Melanoma tumor-derived heat-shock proteins (HSPs) and HSP-peptide complexes can elicit protective antitumor responses. The granulocyte-macrophage colony-stimulating factor (GM-CSF) chemokine can also promote uptake and processing by professional antigen presenting cells (APCs). On this basis, we designed a pilot study of percutaneous thermal ablation as a means to induce heat-shock protein vaccination plus GM-CSF to determine safety and preliminary antitumor activity of this combination. Materials and Methods This study was designed to assess overall safety of percutaneous ablation combined with GM-CSF for unresectable, metastatic melanoma including uveal and mucosal types. All patients received heat-shock therapy (42°C for 30 minutes), then received one of three treatments: (a) intralesional GM-CSF (500 mcg standard dose); (b) radiofrequency ablation (RFA) + GM-CSF; or (c) cryoablation plus GM-CSF. The primary endpoint of the study was the induction of endogenous HSP70 and melanoma-specific cytotoxic T lymphocytes (CTL). Results Nine patients (three per study arm) were enrolled. No dose-limiting toxicity was observed as specified per protocol. All patients developed progressive disease and went on to receive alternative therapy. Median overall survival (OS) was 8.2 months (95% confidence interval [CI] 2–17.2). The study was not powered to detect a difference in clinical outcome among treatment groups. Conclusion Percutaneous thermal ablation plus GM-CSF was well tolerated, technically feasible, and demonstrated an acceptable adverse event profile comparable to conventional RFA and cryoablation. While HSP70 was induced following therapy, the degree of HSP70 elevation was not associated with clinical outcome or induced CTL responses. While percutaneous thermal ablation plus GM-CSF combinations including checkpoint inhibitors could be considered in future studies, the use of GM-CSF remains experimental and for use in the context of clinical trials.