RESUMEN
OBJECTIVE: To investigate the feasibility of standardizing RT simulation CT scanner protocols between vendors using target-based image quality (IQ) metrics. METHOD AND MATERIALS: A systematic assessment process in phantom was developed to standardize clinical scan protocols for scanners from different vendors following these steps: (a) images were acquired by varying CTDIvol and using an iterative reconstruction (IR) method (IR: iDose and model-based iterative reconstruction [IMR] of CTp-Philips Big Bore scanner, SAFIRE of CTs-Siemens biograph PETCT scanner), (b) CT exams were classified into body and brain protocols, (c) the rescaled noise power spectrum (NPS) was calculated, (d) quantified the IQ change due to varied CTDIvol and IR, and (e) matched the IR strength level. IQ metrics included noise and texture from NPS, contrast, and contrast-to-noise ratio (CNR), low contrast detectability (d'). Area under curve (AUC) of the receiver operation characteristic curve of d' was calculated and compared. RESULTS: The level of change in the IQ ratio was significant (>0.6) when using IMR. The IQ ratio change was relatively low to moderate when using either iDose in CTp (0.1-0.5) or SAFIRE in CTs (0.1-0.6). SAFIRE-2 in CTs showed a closer match to the reference body protocol when compared to iDose-3 in CTp. In the brain protocol, iDose-3 in CTp could be matched to the low to moderate level of SAFIRE in CTs. The AUC of d' was highest when using IMR in CTp with lower CTDIvol, and SAFIRE in CTs performed better than iDose in CTp CONCLUSION: It is possible to use target-based IQ metrics to evaluate the performance of the system and operations across various scanners in a phantom. This can serve as an initial reference to convert clinical scanned protocols from one CT simulation scanner to another.
RESUMEN
Innovation in medical imaging artificial intelligence (AI)/machine learning (ML) demands extensive data collection, algorithmic advancements, and rigorous performance assessments encompassing aspects such as generalizability, uncertainty, bias, fairness, trustworthiness, and interpretability. Achieving widespread integration of AI/ML algorithms into diverse clinical tasks will demand a steadfast commitment to overcoming issues in model design, development, and performance assessment. The complexities of AI/ML clinical translation present substantial challenges, requiring engagement with relevant stakeholders, assessment of cost-effectiveness for user and patient benefit, timely dissemination of information relevant to robust functioning throughout the AI/ML lifecycle, consideration of regulatory compliance, and feedback loops for real-world performance evidence. This commentary addresses several hurdles for the development and adoption of AI/ML technologies in medical imaging. Comprehensive attention to these underlying and often subtle factors is critical not only for tackling the challenges but also for exploring novel opportunities for the advancement of AI in radiology.
RESUMEN
BACKGROUND AND OBJECTIVE: Robotic-assisted bronchoscopy (RAB) is an emerging modality to sample pulmonary lesions. Cone-beam computed tomography (CBCT) can be incorporated into RAB. We investigated the magnitude and predictors of patient and staff radiation exposure during mobile CBCT-guided shape-sensing RAB. METHODS: Patient radiation dose was estimated by cumulative dose area product (cDAP) and cumulative reference air kerma (cRAK). Staff equivalent dose was calculated based on isokerma maps and a phantom simulation. Patient, lesion and procedure-related factors associated with higher radiation doses were identified by logistic regression models. RESULTS: A total of 198 RAB cases were included in the analysis. The median patient cDAP and cRAK were 10.86 Gy cm2 (IQR: 4.62-20.84) and 76.20 mGy (IQR: 38.96-148.38), respectively. Among staff members, the bronchoscopist was exposed to the highest median equivalent dose of 1.48 µSv (IQR: 0.85-2.69). Both patient and staff radiation doses increased with the number of CBCT spins and targeted lesions (p < 0.001 for all comparisons). Patient obesity, negative bronchus sign, lesion size <2.0 cm and inadequate sampling by on-site evaluation were associated with a higher patient dose, while patient obesity and inadequate sampling by on-site evaluation were associated with a higher bronchoscopist equivalent dose. CONCLUSION: The magnitude of patient and staff radiation exposure during CBCT-RAB is aligned with safety thresholds recommended by regulatory authorities. Factors associated with a higher radiation exposure during CBCT-RAB can be identified pre-operatively and solicit procedural optimization by reinforcing radiation protective measures. Future studies are needed to confirm these findings across multiple institutions and practices.
Asunto(s)
Broncoscopía , Tomografía Computarizada de Haz Cónico , Exposición a la Radiación , Procedimientos Quirúrgicos Robotizados , Humanos , Tomografía Computarizada de Haz Cónico/métodos , Broncoscopía/métodos , Broncoscopía/efectos adversos , Masculino , Femenino , Exposición a la Radiación/efectos adversos , Persona de Mediana Edad , Procedimientos Quirúrgicos Robotizados/efectos adversos , Exposición Profesional/efectos adversos , Anciano , Dosis de Radiación , Fantasmas de Imagen , Adulto , Neoplasias Pulmonares/diagnóstico por imagen , Neoplasias Pulmonares/radioterapiaRESUMEN
BACKGROUND: Global shortages of iodinated contrast media (ICM) during COVID-19 pandemic forced the imaging community to use ICM more strategically in CT exams. PURPOSE: The purpose of this work is to provide a quantitative framework for preserving iodine CNR while reducing ICM dosage by either lowering kV in single-energy CT (SECT) or using lower energy virtual monochromatic images (VMI) from dual-energy CT (DECT) in a phantom study. MATERIALS AND METHODS: In SECT study, phantoms with effective diameters of 9.7, 15.9, 21.1, and 28.5 cm were scanned on SECT scanners of two different manufacturers at a range of tube voltages. Statistical based iterative reconstruction and deep learning reconstruction were used. In DECT study, phantoms with effective diameters of 20, 29.5, 34.6, and 39.7 cm were scanned on DECT scanners from three different manufacturers. VMIs were created from 40 to 140 keV. ICM reduction by lowering kV levels for SECT or switching from SECT to DECT was calculated based on the linear relationship between iodine CNR and its concentration under different scanning conditions. RESULTS: On SECT scanner A, while matching CNR at 120 kV, ICM reductions of 21%, 58%, and 72% were achieved at 100, 80, and 70 kV, respectively. On SECT scanner B, 27% and 80% ICM reduction was obtained at 80 and 100 kV. On the Fast-kV switch DECT, with CNR matched at 120 kV, ICM reductions were 35%, 30%, 23%, and 15% with VMIs at 40, 50, 60, and 68 keV, respectively. On the dual-source DECT, ICM reductions were 52%, 48%, 42%, 33%, and 22% with VMIs at 40, 50, 60, 70, and 80 keV. On the dual-layer DECT, ICM reductions were 74%, 62%, 45%, and 22% with VMIs at 40, 50, 60, and 70 keV. CONCLUSIONS: Our work provided a quantitative baseline for other institutions to further optimize their scanning protocols to reduce the use of ICM.
Asunto(s)
COVID-19 , Medios de Contraste , Fantasmas de Imagen , Tomografía Computarizada por Rayos X , Humanos , Medios de Contraste/química , Tomografía Computarizada por Rayos X/métodos , Tomografía Computarizada por Rayos X/instrumentación , SARS-CoV-2 , Adulto , Niño , Relación Señal-Ruido , Dosis de Radiación , Procesamiento de Imagen Asistido por Computador/métodos , Imagen Radiográfica por Emisión de Doble Fotón/métodosRESUMEN
The adoption of artificial intelligence (AI) tools in medicine poses challenges to existing clinical workflows. This commentary discusses the necessity of context-specific quality assurance (QA), emphasizing the need for robust QA measures with quality control (QC) procedures that encompass (1) acceptance testing (AT) before clinical use, (2) continuous QC monitoring, and (3) adequate user training. The discussion also covers essential components of AT and QA, illustrated with real-world examples. We also highlight what we see as the shared responsibility of manufacturers or vendors, regulators, healthcare systems, medical physicists, and clinicians to enact appropriate testing and oversight to ensure a safe and equitable transformation of medicine through AI.
RESUMEN
Automated assessment of noise level in clinical computed tomography (CT) images is a crucial technique for evaluating and ensuring the quality of these images. There are various factors that can impact CT image noise, such as statistical noise, electronic noise, structure noise, texture noise, artifact noise, etc. In this study, a method was developed to measure the global noise index (GNI) in clinical CT scans due to the fluctuation of x-ray quanta. Initially, a noise map is generated by sliding a 10 × 10 pixel for calculating Hounsfield unit (HU) standard deviation and the noise map is further combined with the gradient magnitude map. By employing Boolean operation, pixels with high gradients are excluded from the noise histogram generated with the noise map. By comparing the shape of the noise histogram from this method with Christianson's tissue-type global noise measurement algorithm, it was observed that the noise histogram computed in anthropomorphic phantoms had a similar shape with a close GNI value. In patient CT images, excluding the HU deviation due the structure change demonstrated to have consistent GNI values across the entire CT scan range with high heterogeneous tissue compared to the GNI values using Christianson's tissue-type method. The proposed GNI was evaluated in phantom scans and was found to be capable of comparing scan protocols between different scanners. The variation of GNI when using different reconstruction kernels in clinical CT images demonstrated a similar relationship between noise level and kernel sharpness as observed in uniform phantom: sharper kernel resulted in noisier images. This indicated that GNI was a suitable index for estimating the noise level in clinical CT images with either a smooth or grainy appearance. The study's results suggested that the algorithm can be effectively utilized to screen the noise level for a better CT image quality control.
Asunto(s)
Algoritmos , Tomografía Computarizada por Rayos X , Humanos , Tomografía Computarizada por Rayos X/métodos , Fantasmas de Imagen , Control de Calidad , Artefactos , Dosis de Radiación , Procesamiento de Imagen Asistido por Computador/métodosRESUMEN
Deep learning models can perform as well or better than humans in many tasks, especially vision related. Almost exclusively, these models are used to perform classification or prediction. However, deep learning models are usually of black-box nature, and it is often difficult to interpret the model or the features. The lack of interpretability causes a restrain from applying deep learning to fields such as neuroimaging, where the results must be transparent, and interpretable. Therefore, we present a 'glass-box' deep learning model and apply it to the field of neuroimaging. Our model mixes spatial and temporal dimensions in succession to estimate dynamic connectivity between the brain's intrinsic networks. The interpretable connectivity matrices produced by our model result in beating state-of-the-art models on many tasks using multiple functional MRI datasets. More importantly, our model estimates task-based flexible connectivity matrices, unlike static methods such as Pearson's correlation coefficients.
RESUMEN
Cancer care increasingly relies on imaging for patient management. The two most common cross-sectional imaging modalities in oncology are computed tomography (CT) and magnetic resonance imaging (MRI), which provide high-resolution anatomic and physiological imaging. Herewith is a summary of recent applications of rapidly advancing artificial intelligence (AI) in CT and MRI oncological imaging that addresses the benefits and challenges of the resultant opportunities with examples. Major challenges remain, such as how best to integrate AI developments into clinical radiology practice, the vigorous assessment of quantitative CT and MR imaging data accuracy, and reliability for clinical utility and research integrity in oncology. Such challenges necessitate an evaluation of the robustness of imaging biomarkers to be included in AI developments, a culture of data sharing, and the cooperation of knowledgeable academics with vendor scientists and companies operating in radiology and oncology fields. Herein, we will illustrate a few challenges and solutions of these efforts using novel methods for synthesizing different contrast modality images, auto-segmentation, and image reconstruction with examples from lung CT as well as abdome, pelvis, and head and neck MRI. The imaging community must embrace the need for quantitative CT and MRI metrics beyond lesion size measurement. AI methods for the extraction and longitudinal tracking of imaging metrics from registered lesions and understanding the tumor environment will be invaluable for interpreting disease status and treatment efficacy. This is an exciting time to work together to move the imaging field forward with narrow AI-specific tasks. New AI developments using CT and MRI datasets will be used to improve the personalized management of cancer patients.
RESUMEN
PURPOSE: To compare radiation dose index distributions for fluoroscopically guided interventions in interventional radiology from the American College of Radiology (ACR) Fluoroscopy Dose Index Registry (DIR-Fluoro) pilot to those from the Radiation Doses in Interventional Radiology (RAD-IR) study. MATERIALS AND METHODS: Individual and grouped ACR Common identification numbers (procedure types) from the DIR-Fluoro pilot were matched to procedure types in the RAD-IR study. Fifteen comparisons were made. Distribution parameters, including the 10th, 25th, 50th, 75th, and 95th percentiles, were compared for fluoroscopy time (FT), cumulative air kerma (Ka,r), and kerma area product (PKA). Two derived indices were computed using median dose indices. The procedure-averaged reference air kerma rate (Ka,r¯) was computed as Ka,r / FT. The procedure-averaged x-ray field size at the reference point (Ar) was computed as PKA / (Ka,r × 1,000). RESULTS: The median FT was equally likely to be higher or lower in the DIR-Fluoro pilot as it was in the RAD-IR study, whereas the maximum FT was almost twice as likely to be higher in the DIR-Fluoro pilot than it was in the RAD-IR study. The median Ka,r was lower in the DIR-Fluoro pilot for all procedures, as was median PKA. The maximum Ka,r and PKA were more often higher in the DIR-Fluoro pilot than in the RAD-IR study. Ka,r¯ followed the same pattern as Ka,r, whereas Ar was often greater in DIR-Fluoro. CONCLUSIONS: The median dose indices have decreased since the RAD-IR study. The typical Ka,r rates are lower, a result of the use of lower default dose rates. However, opportunities for quality improvement exist, including renewed focus on tight collimation of the imaging field of view.
Asunto(s)
Radiografía Intervencional , Radiología Intervencionista , Humanos , Radiología Intervencionista/métodos , Dosis de Radiación , Fluoroscopía , Radiografía Intervencional/efectos adversos , Sistema de RegistrosRESUMEN
PURPOSE: To update normative data on fluoroscopy dose indices in the United States for the first time since the Radiation Doses in Interventional Radiology study in the late 1990s. MATERIALS AND METHODS: The Dose Index Registry-Fluoroscopy pilot study collected data from March 2018 through December 2019, with 50 fluoroscopes from 10 sites submitting data. Primary radiation dose indices including fluoroscopy time (FT), cumulative air kerma (Ka,r), and kerma area product (PKA) were collected for interventional radiology fluoroscopically guided interventional (FGI) procedures. Clinical facility procedure names were mapped to the American College of Radiology (ACR) common procedure lexicon. Distribution parameters including the 10th, 25th, 50th, 75th, 95th, and 99th percentiles were computed. RESULTS: Dose indices were collected for 70,377 FGI procedures, with 50,501 ultimately eligible for analysis. Distribution parameters are reported for 100 ACR Common IDs. FT in minutes, Ka,r in mGy, and PKA in Gy-cm2 are reported in this study as (n; median) for select ACR Common IDs: inferior vena cava filter insertion (1,726; FT: 2.9; Ka,r: 55.8; PKA: 14.19); inferior vena cava filter removal (464; FT: 5.7; Ka,r: 178.6; PKA: 34.73); nephrostomy placement (2,037; FT: 4.1; Ka,r: 39.2; PKA: 6.61); percutaneous biliary drainage (952; FT: 12.4; Ka,r: 160.5; PKA: 21.32); gastrostomy placement (1,643; FT: 3.2; Ka,r: 29.1; PKA: 7.29); and transjugular intrahepatic portosystemic shunt placement (327; FT: 34.8; Ka,r: 813.0; PKA: 181.47). CONCLUSIONS: The ACR DIR-Fluoro pilot has provided state-of-the-practice statistics for radiation dose indices from IR FGI procedures. These data can be used to prioritize procedures for radiation optimization, as demonstrated in this work.
Asunto(s)
Radiografía Intervencional , Radiología Intervencionista , Humanos , Dosis de Radiación , Proyectos Piloto , Fluoroscopía , Radiología Intervencionista/métodos , Sistema de Registros , Radiografía Intervencional/efectos adversosRESUMEN
BACKGROUND: Pre-training peritonitis (PTP), defined as peritonitis that occurred after catheter insertion and before peritoneal dialysis (PD) training, is increasingly recognized as a risk factor for adverse patient outcomes, yet poorly understood with limited studies conducted to date. This study was conducted to identify the associations, microbiologic profiles and outcomes of PTP compared to post-training peritonitis. METHODS: This single-centre, case-control study involved patients with kidney failure who had PD as their first kidney replacement therapy and had experienced an episode of PD peritonitis between 1 January 2005 and 31 December 2015. Individuals experiencing their first episode of peritonitis were included in the study and categorized according to whether it occurred pre- or post-training. The primary outcome was peritonitis cure rates and composite outcome of hemodialysis (HD) transfer for ≥30 days or death. The secondary outcomes included catheter removal and refractory peritonitis rates. RESULTS: Among 683 patients who received PD for the first time, 121 (17.7%) had PTP while 265 (38.8%) had post-training peritonitis. PTP patients were more likely to have had exit-site infection (ESI) prior to peritonitis (24.8% compared to 17% in the post-training peritonitis group, p = 0.2). Culture-negative peritonitis was significantly more common in the PTP patients (53.7%) than in the post-training group (27.3%, p < 0.001). The cure was achieved in 68.9% of cases and was not significantly different between the PTP and post-training peritonitis groups (66.1% vs. 70.2%; OR 0.83, 95% CI 0.51-1.35). Lower odds of cure were associated with peritonitis caused by moderate and high severity organisms (OR 0.49, 95% CI 0.29-0.85; OR 0.18, 95% CI 0.08-0.43, respectively). Composite outcome of HD transfer or death was more commonly observed among patients with PTP (87.5% vs. 75.8%; OR 2.2, 95% CI 1.20-4.48) in whom significantly shorter median time to HD transfer was observed (PTP 10.7 months vs. post-training peritonitis 21.9 months, p < 0.0001). CONCLUSIONS: PTP is a common condition that is highly associated with preceding ESI, is frequently culture-negative and is associated with worse composite outcome of HD transfer or death. PTP rates should be routinely monitored and reported by PD units for continuous quality improvement.
Asunto(s)
Diálisis Peritoneal , Peritonitis , Humanos , Diálisis Peritoneal/efectos adversos , Estudios de Casos y Controles , Diálisis Renal/efectos adversos , Cateterismo/efectos adversos , Peritonitis/etiología , Peritonitis/microbiologíaRESUMEN
Brain network interactions are commonly assessed via functional (network) connectivity, captured as an undirected matrix of Pearson correlation coefficients. Functional connectivity can represent static and dynamic relations, but often these are modeled using a fixed choice for the data window Alternatively, deep learning models may flexibly learn various representations from the same data based on the model architecture and the training task. However, the representations produced by deep learning models are often difficult to interpret and require additional posthoc methods, e.g., saliency maps. In this work, we integrate the strengths of deep learning and functional connectivity methods while also mitigating their weaknesses. With interpretability in mind, we present a deep learning architecture that exposes a directed graph layer that represents what the model has learned about relevant brain connectivity. A surprising benefit of this architectural interpretability is significantly improved accuracy in discriminating controls and patients with schizophrenia, autism, and dementia, as well as age and gender prediction from functional MRI data. We also resolve the window size selection problem for dynamic directed connectivity estimation as we estimate windowing functions from the data, capturing what is needed to estimate the graph at each time-point. We demonstrate efficacy of our method in comparison with multiple existing models that focus on classification accuracy, unlike our interpretability-focused architecture. Using the same data but training different models on their own discriminative tasks we are able to estimate task-specific directed connectivity matrices for each subject. Results show that the proposed approach is also more robust to confounding factors compared to standard dynamic functional connectivity models. The dynamic patterns captured by our model are naturally interpretable since they highlight the intervals in the signal that are most important for the prediction. The proposed approach reveals that differences in connectivity among sensorimotor networks relative to default-mode networks are an important indicator of dementia and gender. Dysconnectivity between networks, specially sensorimotor and visual, is linked with schizophrenic patients, however schizophrenic patients show increased intra-network default-mode connectivity compared to healthy controls. Sensorimotor connectivity was important for both dementia and schizophrenia prediction, but schizophrenia is more related to dysconnectivity between networks whereas, dementia bio-markers were mostly intra-network connectivity.
Asunto(s)
Demencia , Esquizofrenia , Humanos , Imagen por Resonancia Magnética/métodos , Mapeo Encefálico/métodos , Encéfalo/diagnóstico por imagen , Esquizofrenia/diagnóstico por imagenRESUMEN
Brain dynamics are highly complex and yet hold the key to understanding brain function and dysfunction. The dynamics captured by resting-state functional magnetic resonance imaging data are noisy, high-dimensional, and not readily interpretable. The typical approach of reducing this data to low-dimensional features and focusing on the most predictive features comes with strong assumptions and can miss essential aspects of the underlying dynamics. In contrast, introspection of discriminatively trained deep learning models may uncover disorder-relevant elements of the signal at the level of individual time points and spatial locations. Yet, the difficulty of reliable training on high-dimensional low sample size datasets and the unclear relevance of the resulting predictive markers prevent the widespread use of deep learning in functional neuroimaging. In this work, we introduce a deep learning framework to learn from high-dimensional dynamical data while maintaining stable, ecologically valid interpretations. Results successfully demonstrate that the proposed framework enables learning the dynamics of resting-state fMRI directly from small data and capturing compact, stable interpretations of features predictive of function and dysfunction.
Asunto(s)
Encéfalo , Imagen por Resonancia Magnética , Encéfalo/diagnóstico por imagen , Mapeo Encefálico/métodos , Neuroimagen Funcional , Imagen por Resonancia Magnética/métodosRESUMEN
We map single energy CT (SECT) scans to synthetic dual-energy CT (synth-DECT) material density iodine (MDI) scans using deep learning (DL) and demonstrate their value for liver segmentation. A 2D pix2pix (P2P) network was trained on 100 abdominal DECT scans to infer synth-DECT MDI scans from SECT scans. The source and target domain were paired with DECT monochromatic 70 keV and MDI scans. The trained P2P algorithm then transformed 140 public SECT scans to synth-DECT scans. We split 131 scans into 60% train, 20% tune, and 20% held-out test to train four existing liver segmentation frameworks. The remaining nine low-dose SECT scans tested system generalization. Segmentation accuracy was measured with the dice coefficient (DSC). The DSC per slice was computed to identify sources of error. With synth-DECT (and SECT) scans, an average DSC score of 0.93±0.06 (0.89±0.01) and 0.89±0.01 (0.81±0.02) was achieved on the held-out and generalization test sets. Synth-DECT-trained systems required less data to perform as well as SECT-trained systems. Low DSC scores were primarily observed around the scan margin or due to non-liver tissue or distortions within ground-truth annotations. In general, training with synth-DECT scans resulted in improved segmentation performance with less data.
RESUMEN
Modern fluoroscopes used for image guidance have become quite complex. Adding to this complexity are the many regulatory and accreditation requirements that must be fulfilled during acceptance testing of a new unit. Further, some of these acceptance tests have pass/fail criteria, whereas others do not, making acceptance testing a subjective and time-consuming task. The AAPM Task Group 272 Report spells out the details of tests that are required and gives visibility to some of the tests that while not yet required are recommended as good practice. The organization of the report begins with the most complicated fluoroscopes used in interventional radiology or cardiology and continues with general fluoroscopy and mobile C-arms. Finally, the appendices of the report provide useful information, an example report form and topics that needed their own section due to the level of detail.
Asunto(s)
Cardiología , Radiología Intervencionista , Fluoroscopía/métodos , Dosis de Radiación , Radiología Intervencionista/métodos , Informe de InvestigaciónRESUMEN
Artificial intelligence (AI) has been successful at solving numerous problems in machine perception. In radiology, AI systems are rapidly evolving and show progress in guiding treatment decisions, diagnosing, localizing disease on medical images, and improving radiologists' efficiency. A critical component to deploying AI in radiology is to gain confidence in a developed system's efficacy and safety. The current gold standard approach is to conduct an analytical validation of performance on a generalization dataset from one or more institutions, followed by a clinical validation study of the system's efficacy during deployment. Clinical validation studies are time-consuming, and best practices dictate limited re-use of analytical validation data, so it is ideal to know ahead of time if a system is likely to fail analytical or clinical validation. In this paper, we describe a series of sanity tests to identify when a system performs well on development data for the wrong reasons. We illustrate the sanity tests' value by designing a deep learning system to classify pancreatic cancer seen in computed tomography scans.
RESUMEN
Purpose: The lack of standardization in quantitative radiomic measures of tumors seen on computed tomography (CT) scans is generally recognized as an unresolved issue. To develop reliable clinical applications, radiomics must be robust across different CT scan modes, protocols, software, and systems. We demonstrate how custom-designed phantoms, imprinted with human-derived patterns, can provide a straightforward approach to validating longitudinally stable radiomic signature values in a clinical setting. Approach: Described herein is a prototype process to design an anatomically informed 3D-printed radiomic phantom. We used a multimaterial, ultra-high-resolution 3D printer with voxel printing capabilities. Multiple tissue regions of interest (ROIs), from four pancreas tumors, one lung tumor, and a liver background, were extracted from digital imaging and communication in medicine (DICOM) CT exam files and were merged together to develop a multipurpose, circular radiomic phantom (18 cm diameter and 4 cm width). The phantom was scanned 30 times using standard clinical CT protocols to test repeatability. Features that have been found to be prognostic for various diseases were then investigated for their repeatability and reproducibility across different CT scan modes. Results: The structural similarity index between the segment used from the patients' DICOM image and the phantom CT scan was 0.71. The coefficient variation for all assessed radiomic features was < 1.0 % across 30 repeat scans of the phantom. The percent deviation (pDV) from the baseline value, which was the mean feature value determined from repeat scans, increased with the application of the lung convolution kernel, changes to the voxel size, and increases in the image noise. Gray level co-occurrence features, contrast, dissimilarity, and entropy were particularly affected by different scan modes, presenting with pDV > ± 15 % . Conclusions: Previously discovered prognostic and popular radiomic features are variable in practice and need to be interpreted with caution or excluded from clinical implementation. Voxel-based 3D printing can reproduce tissue morphology seen on CT exams. We believe that this is a flexible, yet practical, way to design custom phantoms to validate and compare radiomic metrics longitudinally, over time, and across systems.
RESUMEN
Purpose: The goal of this study is to develop innovative methods for identifying radiomic features that are reproducible over varying image acquisition settings. Approach: We propose a regularized partial correlation network to identify reliable and reproducible radiomic features. This approach was tested on two radiomic feature sets generated using two different reconstruction methods on computed tomography (CT) scans from a cohort of 47 lung cancer patients. The largest common network component between the two networks was tested on phantom data consisting of five cancer samples. To further investigate whether radiomic features found can identify phenotypes, we propose a k -means clustering algorithm coupled with the optimal mass transport theory. This approach following the regularized partial correlation network analysis was tested on CT scans from 77 head and neck squamous cell carcinoma (HNSCC) patients in the Cancer Imaging Archive (TCIA) and validated using an independent dataset. Results: A set of common radiomic features was found in relatively large network components between the resultant two partial correlation networks resulting from a cohort of lung cancer patients. The reliability and reproducibility of those radiomic features were further validated on phantom data using the Wasserstein distance. Further analysis using the network-based Wasserstein k -means algorithm on the TCIA HNSCC data showed that the resulting clusters separate tumor subsites as well as HPV status, and this was validated on an independent dataset. Conclusion: We showed that a network-based analysis enables identifying reproducible radiomic features and use of the selected set of features can enhance clustering results.
RESUMEN
INTRODUCTION: Oncologic patients who develop chemotherapy-associated liver injury (CALI) secondary to chemotherapy treatment tend to have worse outcomes. Biopsy remains the gold standard for the diagnosis of hepatic steatosis. The purpose of this article is to compare 2 alternatives: Proton-Density-Fat-Fraction (PDFF) MRI and MultiMaterial-Decomposition (MMD) DECT. MATERIALS AND METHODS: 49 consecutive oncologic patients treated with Chemotherapy underwent abdominal DECT and abdominal MRI within 2 weeks of each other. Two radiologists tracked Regions of Interest independently both in the PDFF fat maps and in the MMD DECT fat maps. Non-parametric exact Wilcoxon signed rank test and Cohen's K were used to compare the 2 sequences and to evaluate the agreement. RESULTS: There was no statistically significant difference in the fat fraction measured as a continuous value between PDFF and DECT between 2 readers. Within the same imaging method (PDFF) the degree of agreement based on the k coefficient between reader 1 and reader 2 is 0.88 (p-value < 0.05). Similarly, for single-source DECT(ssDECT) the degree of agreement based on the k coefficient between reader 1 and reader 2 is 0.97 (p-value < 0.05). CONCLUSIONS: The results of this study demonstrate that the hepatic fat fraction of ssDECT with MMD are not significantly different from PDFF. This could be an advantage in an oncological population that undergoes serial CT scans for follow up of chemotherapy response.
RESUMEN
BACKGROUND: To evaluate quantitative iodine parameters from the arterial phase dual-energy computed tomography (DECT) scans as an imaging biomarker for tumor grade (TG), mitotic index (MI), and Ki-67 proliferation index of hepatic metastases from neuroendocrine tumors (NETs) of the gastrointestinal (GI) tract. Imaging biomarkers have the potential to provide relevant clinical information about pathologic processes beyond lesion morphology. NETs are a group of rare, heterogeneous neoplasms classified by World Health Organization (WHO) TG, which is derived from MI and Ki-67 proliferation index. Imaging biomarkers for these pathologic features and TG may be useful. METHODS: Between January 2014 and April 2019, 73 unique patients with hepatic metastases from NET of the GI tract underwent DECT of the abdomen with an arterial phase were analyzed after exclusions. Using GSIViewer software (GE Healthcare, Madison, Wisconsin), elliptical regions of interest (ROIs) were placed over selected hepatic metastases by a fellowship trained abdominal radiologist. Quantitative iodine concentration (IC) data was extracted from the lesion ROIs, and the normalized IC (lesion IC/aorta IC) and relative IC (lesion IC/liver IC) for each liver were calculated. Spearman correlation was calculated for lesion mean IC, normalized IC, and relative IC to both Ki-67 proliferation and mitotic indices. Student's t-test was performed to compare lesion mean IC, normalized IC and relative IC between WHO TGs. RESULTS: There was very weak correlation between both normalized IC and relative IC for both Ki-67 proliferation and mitotic indices. A significant difference was not observed between normalized IC and relative IC to distinguish metastases from G1 and G2/3 tumors. CONCLUSIONS: Our study finds limited potential for quantitative parameters from DECT to distinguish neuroendocrine hepatic metastases by WHO TG, as well as limited potential as an imaging biomarker for Ki-67 proliferation and mitotic indices in this setting. Our findings of a lack of correlation between Ki-67 and quantitative iodine parameters stands in contrast to existing literature that reports positive correlations for these parameters in the rectum and stomach.