MSMichael Silosky

Michael Silosky

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Michael Silosky has a total of 23 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    Pregnancy, cancer, and radiation—a modern refresher

    by Kelly Kisling, Sandra M Meyers, Rebecca Milman, Susan Richardson, Christina Chapman, Justine Dupere, Grace Eliason, Suzanne Evans, Titania Juang, Young Lee, et al.

    Cancer occurs in ∼1 per 1000 pregnancies; thousands of patients may require radiation procedures for diagnosis and treatment each year in the United States alone. This rare but high-risk scenario, coupled with fear of radiation, has created ambiguity in the ideal management of pregnant patients. Without a comprehensive guide for the use of radiation in imaging and treatment for pregnant cancer patients, it is difficult for providers to offer optimal patient-centered care without introducing disparities. The goal of this paper is to provide guidance on the use of radiation for screening, diagnosis, staging, and treatment of cancer, while highlighting gaps in existing knowledge and guidelines. The intention is that physicians, medical physicists, and patients could use this document as a resource for shared decision-making, ensuring safe and effective practice. A team of physicians and medical physicists with expertise in imaging, radiotherapy, and maternal fetal medicine was assembled, along with a patient advocate and lawyer. Existing guidelines and recent literature were reviewed. Authors also drew from their experience where published guidance was lacking. The resulting document discusses best practice to guide use of radiation for cancer, as well as patient-centered care management and legal considerations. Overall, it is possible to safely and effectively deliver radiation to pregnant patients in numerous circumstances. Use of radiation or other modalities should be discussed through shared decision-making with the physician and patient, contextualizing the maternal and fetal risk from treatments. Healthcare providers should support wide access to reproductive healthcare to allow equitable, evidence-based, patient-centered healthcare.


  • Open Access

    The impact of NaI(Tl) crystal hydration on gamma camera spectral response and image uniformity

    by Grace Eliason, Michael S Silosky

    Background: NaI(Tl) scintillators used in most gamma cameras are hermetically sealed to prevent the absorption of water molecules from the surrounding environment. If this seal is compromised, crystal hydration, a localized defect resulting in non‐uniform attenuation of scintillation photons, may occur. Purpose: The purpose of this study was to evaluate the effects of crystal hydration across multiple radionuclides, characterizing the impact on spectral response and image uniformity. Methods: Hydration was assessed using on‐peak and off‐peak imaging for several radionuclides. The progression of hydration was assessed by calculating the contrast‐to‐noise ratio in hydrated regions on monthly uniformity maps. Spectra were captured for both hydrated and non‐hydrated regions of the detector to determine photopeak energy and energy resolution. Results: The visual appearance of the hydration effect in off‐peak images was greatest for 133 Xe. The effect was more substantial in the 10% low off‐peak image than in the 10% high off‐peak image. Hydration was not observed in on‐peak images for either 133 Xe or 99m Tc and was only barely visible for 131 I. CNR measurements show a slow but notable progression of hydration in uniformity maps over time. The measured photopeak in hydrated regions was lower than that of non‐hydrated regions. Hydration also resulted in a degradation in energy resolution with the effect being more significant at lower energies. Conclusions: Hydrated regions of the detector demonstrated an energy‐dependent degradation in energy resolution which corresponded to an energy dependence in the detectability of hydration in off‐peak images. Monthly updates to the uniformity correction maps were able to adequately account for hydration. When hydration has been identified, routine evaluation of its impact on uniformity maps may provide a convenient method of tracking progression.

  • Learning Without Real Data Annotations to Detect Hepatic Lesions in PET Images

    by Xinyi Yang, Bennett B. Chin, Michael Silosky, Jonathan Wehrend, Daniel V. Litwiller, Debashis Ghosh, Fuyong Xing

    Objective: Deep neural networks have been recently applied to lesion identification in fluorodeoxyglucose (FDG) positron emission tomography (PET) images, but they typically rely on a large amount of well-annotated data for model training. This is extremely difficult to achieve for neuroendocrine tumors (NETs), because of low incidence of NETs and expensive lesion annotation in PET images. The objective of this study is to design a novel, adaptable deep learning method, which uses no real lesion annotations but instead low-cost, list mode-simulated data, for hepatic lesion detection in real-world clinical NET PET images. Methods: We first propose a region-guided generative adversarial network (RG-GAN) for lesion-preserved image-to-image translation. Then, we design a specific data augmentation module for our list-mode simulated data and incorporate this module into the RG-GAN to improve model training. Finally, we combine the RG-GAN, the data augmentation module and a lesion detection neural network into a unified framework for joint-task learning to adaptatively identify lesions in real-world PET data. Results: The proposed method outperforms recent state-of-the-art lesion detection methods in real clinical $^{68}$ Ga-DOTATATE PET images, and produces very competitive performance with the target model that is trained with real lesion annotations. Conclusion: With RG-GAN modeling and specific data augmentation, we can obtain good lesion detection performance without using any real data annotations. Significance: This study introduces an adaptable deep learning method for hepatic lesion identification in NETs, which can significantly reduce human effort for data annotation and improve model generalizability for lesion detection with PET imaging.


  • Location-Aware Encoding for Lesion Detection in $^{68}$Ga-DOTATATE Positron Emission Tomography Images

    by Fuyong Xing, Michael Silosky, Debashis Ghosh, Bennett B. Chin

    Objective: Lesion detection with positron emission tomography (PET) imaging is critical for tumor staging, treatment planning, and advancing novel therapies to improve patient outcomes, especially for neuroendocrine tumors (NETs). Current lesion detection methods often require manual cropping of regions/volumes of interest (ROIs/VOIs) a priori, or rely on multi-stage, cascaded models, or use multi-modality imaging to detect lesions in PET images. This leads to significant inefficiency, high variability and/or potential accumulative errors in lesion quantification. To tackle this issue, we propose a novel single-stage lesion detection method using only PET images. Methods: We design and incorporate a new, plug-and-play codebook learning module into a U-Net-like neural network and promote lesion location-specific feature learning at multiple scales. We explicitly regularize the codebook learning with direct supervision at the network's multi-level hidden layers and enforce the network to learn multi-scale discriminative features with respect to predicting lesion positions. The network automatically combines the predictions from the codebook learning module and other layers via a learnable fusion layer. Results: We evaluate the proposed method on a real-world clinical $^{68}$ Ga-DOTATATE PET image dataset, and our method produces significantly better lesion detection performance than recent state-of-the-art approaches. Conclusion: We present a novel deep learning method for single-stage lesion detection in PET imaging data, with no ROI/VOI cropping in advance, no multi-stage modeling and no multi-modality data. Significance: This study provides a new perspective for effective and efficient lesion identification in PET, potentially accelerating novel therapeutic regimen development for NETs and ultimately improving patient outcomes including survival.