Comparison of effective radiation doses from X-ray, CT, and PET/CT in pediatric patients with neuroblastoma using a dose monitoring program

CoVSeverity-Net: an efficient deep learning model for COVID-19 severity estimation from Chest X-Ray images

Purpose

COVID-19 is not going anywhere and is slowly becoming a part of our life. The World Health Organization declared it a pandemic in 2020, and it has affected all of us in many ways. Several deep learning techniques have been developed to detect COVID-19 from Chest X-Ray images. COVID-19 infection severity scoring can aid in establishing the optimum course of treatment and care for a positive patient, as all COVID-19 positive patients do not require special medical attention. Still, very few works are reported to estimate the severity of the disease from the Chest X-Ray images. The unavailability of the large-scale dataset might be a reason.

Methods

We aim to propose CoVSeverity-Net, a deep learning-based architecture for predicting the severity of COVID-19 from Chest X-ray images. CoVSeverity-Net is trained on a public COVID-19 dataset, curated by experienced radiologists for severity estimation. For that, a large publicly available dataset is collected and divided into three levels of severity, namely Mild, Moderate, and Severe.

Results

An accuracy of 85.71% is reported. Conducting 5-fold cross-validation, we have obtained an accuracy of 87.82 ± 6.25%. Similarly, conducting 10-fold cross-validation we obtained accuracy of 91.26 ± 3.42. The results were better when compared with other state-of-the-art architectures.

Conclusion

We strongly believe that this study has a high chance of reducing the workload of overworked front-line radiologists, speeding up patient diagnosis and treatment, and easing pandemic control. Future work would be to train a novel deep learning-based architecture on a larger dataset for severity estimation.

COVID detection from Chest X-Ray Images using multi-scale attention

Deep learning based methods have shown great promise in achieving accurate automatic detection of Coronavirus Disease (COVID) - 19 from Chest X-Ray (CXR) images. However, incorporating explainability in these solutions remains relatively less explored. We present a hierarchical classification approach for separating normal, non-COVID pneumonia (NCP) and COVID cases using CXR images. We demonstrate that the proposed method achieves clinically consistent explanations. We achieve this using a novel multi-scale attention architecture called Multi-scale Attention Residual Learning (MARL) and a new loss function based on conicity for training the proposed architecture. The proposed classification strategy has two stages. The first stage uses a model derived from DenseNet to separate pneumonia cases from normal cases while the second stage uses the MARL architecture to discriminate between COVID and NCP cases. With a five-fold cross validation, the proposed method achieves 93%, 96.28%, and 84.51% accuracy respectively over three public datasets for normal vs. NCP vs. COVID classification. This is competitive to the state-of-the-art methods. We also provide explanations in the form of GradCAM attributions, which are well aligned with expert annotations. The attributions are also seen to clearly indicate that MARL deems the peripheral regions of the lungs to be more important in the case of COVID cases while central regions are seen as more important in NCP cases. This observation matches the criteria described by radiologists in clinical literature, thereby attesting to the utility of the derived explanations.

Deep-Precognitive Diagnosis: Preventing Future Pandemics by Novel Disease Detection With Biologically-Inspired Conv-Fuzzy Network

Deep learning-based Computer-Aided Diagnosis has gained immense attention in recent years due to its capability to enhance diagnostic performance and elucidate complex clinical tasks. However, conventional supervised deep learning models are incapable of recognizing novel diseases that do not exist in the training dataset. Automated early-stage detection of novel infectious diseases can be vital in controlling their rapid spread. Moreover, the development of a conventional CAD model is only possible after disease outbreaks and datasets become available for training (viz. COVID-19 outbreak). Since novel diseases are unknown and cannot be included in training data, it is challenging to recognize them through existing supervised deep learning models. Even after data becomes available, recognizing new classes with conventional models requires a complete extensive re-training. The present study is the first to report this problem and propose a novel solution to it. In this study, we propose a new class of CAD models, i.e., Deep-Precognitive Diagnosis, wherein artificial agents are enabled to identify unknown diseases that have the potential to cause a pandemic in the future. A de novo biologically-inspired Conv-Fuzzy network is developed. Experimental results show that the model trained to classify Chest X-Ray (CXR) scans into normal and bacterial pneumonia detected a novel disease during testing, unseen by it in the training sample and confirmed to be COVID-19 later. The model is also tested on SARS-CoV-1 and MERS-CoV samples as unseen diseases and achieved state-of-the-art accuracy. The proposed model eliminates the need for model re-training by creating a new class in real-time for the detected novel disease, thus classifying it on all subsequent occurrences. Second, the model addresses the challenge of limited labeled data availability, which renders most supervised learning techniques ineffective and establishes that modified fuzzy classifiers can achieve high accuracy on image classification tasks.

The cumulative radiation dose paradigm in pediatric imaging

Medical imaging professionals have an accountability for both quality and safety in the care of patients that have unexpected or anticipated repeated imaging examinations that use ionizing radiation. One measure in the safety realm for repeated imaging is cumulative effective dose (CED). CED has been increasingly scrutinized in patient populations, including adults and children. Recognizing the challenges with effective dose, including the cumulative nature, effective dose is still the most prevalent exposure currency for recurrent imaging examinations. While the responsibility for dose monitoring incorporates an element of tracking an individual patient cumulative radiation record, a more complex aspect is what should be done with this information. This challenge also differs between the pediatric and adult population, including the fact that high cumulative doses (e.g.,>100 mSv) are reported to occur much less frequently in children than in the adult population. It is worthwhile, then, to review the general construct of CED, including the comparison between the relative percentage occurrence in adult and pediatric populations, the relevant pediatric medical settings in which high CED occurs, the advances in medical care that may affect CED determinations in the future, and offer proposals for the application of the CED paradigm, considering the unique aspects of pediatric care.

A multi model ensemble based deep convolution neural network structure for detection of COVID19

The year 2020 will certainly be remembered for the COVID-19 outbreak. First reported in Wuhan city of China back in December 2019, the number of people getting affected by this contagious virus has grown exponentially. Given the population density of India, the implementation of the mantra of the test, track, and isolate is not obtaining satisfactory results. A shortage of testing kits and an increasing number of fresh cases encouraged us to come up with a model that can aid radiologists in detecting COVID19 using chest Xray images. In the proposed framework the low level features from the Chest X-ray images are extracted using an ensemble of four pre-trained Deep Convolutional Neural Network (DCNN) architectures, namely VGGNet, GoogleNet, DenseNet, and NASNet and later on are fed to a fully connected layer for classification. The proposed multi model ensemble architecture is validated on two publicly available datasets and one private dataset. We have shown that our multi model ensemble architecture performs better than single classifier. On the publicly available dataset we have obtained an accuracy of 88.98% for three class classification and for binary class classification we report an accuracy of 98.58%. Validating the performance on private dataset we obtained an accuracy of 93.48%. The source code and the dataset are made available in the github linkhttps://github.com/sagardeepdeb/ensemble-model-for-COVID-detection.

Cumulative radiation doses due to nuclear medicine examinations: a systematic review

OBJECTIVES

To systematically review the published data regarding the cumulative exposure to radiation in selected cohorts of adults or paediatric patients due to diagnostic nuclear medicine examinations.

METHODS

We conducted PubMed/Medline searches of peer-reviewed papers on cumulated effective dose (CED) from diagnostic nuclear medicine procedures published between 01 January 2010 until 31 January 2021. Studies were considered eligible if the contribution of nuclear medicine examinations to total CED was >10%. Studies reporting cumulative doses in a single episode of care or in a limited time (≤1 year) were excluded. The main outcomes for which data were sought were the CED accrued by patients, the period in which the CED was accrued, the percentage of patients with CED > 100 mSv and the percentage contribution due to nuclear medicine procedures to the overall CED.

RESULTS

The studies included in the synthesis were 18 which enrolled a total of 1,76,371 patients. Eleven (1,757 patients), three (1,74,079 patients) and four (535 patients) were related to oncological, cardiologic and transplanted patients, respectively. All the studies were retrospective; some of the source materials referred to small number of patients and some of the patients were followed for a short time. Not many studies accurately quantified the contribution of nuclear medicine procedures to the overall radiation exposure due to medical imaging. Finally, most of the studies covered an observation period which extended mainly in the 2000-2010 decade.

CONCLUSIONS

There is a need of prospective, multicentric studies enrolling a greater number of patients, followed for longer period in selected groups of patients to fully capture the cumulative exposure to radiation in these settings.

ADVANCES IN KNOWLEDGE

This systematic review allows to identify selected group of patients with a specific health status in which the cumulated exposure to radiation may be of concern and where the contribution of nuclear medicine procedures to the total CED is significant.

Automatically discriminating and localizing COVID-19 from community-acquired pneumonia on chest X-rays

The COVID-19 outbreak continues to threaten the health and life of people worldwide. It is an immediate priority to develop and test a computer-aided detection (CAD) scheme based on deep learning (DL) to automatically localize and differentiate COVID-19 from community-acquired pneumonia (CAP) on chest X-rays. Therefore, this study aims to develop and test an efficient and accurate deep learning scheme that assists radiologists in automatically recognizing and localizing COVID-19. A retrospective chest X-ray image dataset was collected from open image data and the Xiangya Hospital, which was divided into a training group and a testing group. The proposed CAD framework is composed of two steps with DLs: the Discrimination-DL and the Localization-DL. The first DL was developed to extract lung features from chest X-ray radiographs for COVID-19 discrimination and trained using 3548 chest X-ray radiographs. The second DL was trained with 406-pixel patches and applied to the recognized X-ray radiographs to localize and assign them into the left lung, right lung or bipulmonary. X-ray radiographs of CAP and healthy controls were enrolled to evaluate the robustness of the model. Compared to the radiologists' discrimination and localization results, the accuracy of COVID-19 discrimination using the Discrimination-DL yielded 98.71%, while the accuracy of localization using the Localization-DL was 93.03%. This work represents the feasibility of using a novel deep learning-based CAD scheme to efficiently and accurately distinguish COVID-19 from CAP and detect localization with high accuracy and agreement with radiologists.

Role of Machine Learning Techniques to Tackle the COVID-19 Crisis: Systematic Review

BACKGROUND

SARS-CoV-2, the novel coronavirus responsible for COVID-19, has caused havoc worldwide, with patients presenting a spectrum of complications that have pushed health care experts to explore new technological solutions and treatment plans. Artificial Intelligence (AI)-based technologies have played a substantial role in solving complex problems, and several organizations have been swift to adopt and customize these technologies in response to the challenges posed by the COVID-19 pandemic.

OBJECTIVE

The objective of this study was to conduct a systematic review of the literature on the role of AI as a comprehensive and decisive technology to fight the COVID-19 crisis in the fields of epidemiology, diagnosis, and disease progression.

METHODS

A systematic search of PubMed, Web of Science, and CINAHL databases was performed according to PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analysis) guidelines to identify all potentially relevant studies published and made available online between December 1, 2019, and June 27, 2020. The search syntax was built using keywords specific to COVID-19 and AI.

RESULTS

The search strategy resulted in 419 articles published and made available online during the aforementioned period. Of these, 130 publications were selected for further analyses. These publications were classified into 3 themes based on AI applications employed to combat the COVID-19 crisis: Computational Epidemiology, Early Detection and Diagnosis, and Disease Progression. Of the 130 studies, 71 (54.6%) focused on predicting the COVID-19 outbreak, the impact of containment policies, and potential drug discoveries, which were classified under the Computational Epidemiology theme. Next, 40 of 130 (30.8%) studies that applied AI techniques to detect COVID-19 by using patients' radiological images or laboratory test results were classified under the Early Detection and Diagnosis theme. Finally, 19 of the 130 studies (14.6%) that focused on predicting disease progression, outcomes (ie, recovery and mortality), length of hospital stay, and number of days spent in the intensive care unit for patients with COVID-19 were classified under the Disease Progression theme.

CONCLUSIONS

In this systematic review, we assembled studies in the current COVID-19 literature that utilized AI-based methods to provide insights into different COVID-19 themes. Our findings highlight important variables, data types, and available COVID-19 resources that can assist in facilitating clinical and translational research.

Feasibility and possible value of quantitative semi-automated diffusion weighted imaging volumetry of neuroblastic tumors

Background

To assess the feasibility and possible value of semi-automated diffusion weighted imaging (DWI) volumetry of whole neuroblastic tumors with apparent diffusion coefficient (ADC) map evaluation after neoadjuvant chemotherapy.

Methods

Pediatric patients who underwent surgical resection of neuroblastic tumors at our institution from 2013 to 2019 and who received a preoperative MRI scan with DWI after chemotherapy were included. Tumor volume was assessed with a semi-automated approach in DWI using a dedicated software prototype. Quantitative ADC values were calculated automatically of the total tumor volume after manual exclusion of necrosis. Manual segmentation in T1 weighted and T2 weighted sequences was used as reference standard for tumor volume comparison. The Student’s t test was used for parametric data while the Wilcoxon rank sum test and the Kruskal-Wallis test were applied for non-parametric data.

Results

Twenty seven patients with 28 lesions (neuroblastoma (NB): n = 19, ganglioneuroblastoma (GNB): n = 7, ganglioneuroma (GN): n = 2) could be evaluated. Mean patient age was 4.5 ± 3.2 years. Median volume of standard volumetry (T1w or T2w) was 50.2 ml (interquartile range (IQR): 91.9 ml) vs. 45.1 ml (IQR: 98.4 ml) of DWI (p = 0.145). Mean ADC values (× 10^− 6 mm²/s) of the total tumor volume (without necrosis) were 1187 ± 301 in NB vs. 1552 ± 114 in GNB/GN (p = 0.037). The 5th percentile of ADC values of NB (614 ± 275) and GNB/GN (1053 ± 362) provided the most significant difference (p = 0.007) with an area under the curve of 0.848 (p < 0.001).

Conclusions

Quantitative semi-automated DWI volumetry is feasible in neuroblastic tumors with integrated analysis of tissue characteristics by providing automatically calculated ADC values of the whole tumor as well as an ADC heatmap. The 5th percentile of the ADC values of the whole tumor volume proved to be the most significant parameter for differentiation of the histopathological subtypes in our patient cohort and further investigation seems to be worthwhile.

Supplementary Information

The online version contains supplementary material available at 10.1186/s40644-020-00366-3.

Efficient photoacoustic imaging using indocyanine green (ICG) loaded functionalized mesoporous silica nanoparticles

Photoacoustic (PA) imaging is gaining momentum due to its greater depth of field, low background, and 3D imaging capabilities. However, traditional PA imaging agents (e.g. dyes, quantum dots, etc.) are usually unstable in plasma and bind to serum proteins, and thus cleared rapidly. Because of this, the nanoparticle encapsulation of PA imaging agents is becoming increasingly popular. Therefore, the rational design of carrier nanoparticles for this purpose is necessary for strong imaging signal intensity, high biosafety, and precise targeting. Herein, we systematically evaluate the influence of the chemical and physical surface functionalization of mesoporous silica nanoparticles (MSNs) on the photo-stability, loading, release, and photoacoustic (PA) signal strength of the FDA approved small molecule contrast agent, indocyanine green (ICG). Chemical functionalization involved the modification of MSNs with silanes having amine (NH2) or phosphonate (PO3) terminal groups, whereas physical modifications were performed by capping the ICG loaded MSNs with lipid bilayer (LB) or layer-by-layer (LBL) polyelectrolyte coatings. The NH2-MSNs display the highest ICG mass loading capacity (16.5 wt%) with a limited release of ICG (5%) in PBS over 48 h, while PO3-MSNs only loaded ICG around 3.5 wt%. The physically modified MSNs (i.e. LBMSNs and LBLMSNs) were vacuum loaded resulting in approximately 9 wt% loading and less than 10% ICG release in 48 h. Pure ICG was highly photo-unstable and showed 20% reduction in photoluminescence (PL) within 3 h of exposure to 800 nm, while the ICG loaded onto functionalized MSNs did not photo-degrade. Among the tested formulations, NH2-MSNs and LBLMSNs presented 4-fold in vitro PA signal intensity enhancement at a 200 μg mL-1 equivalent ICG dose. Similar to the in vitro PA imaging, NH2-MSNs and LBLMSNs performed the best when subcutaneously injected into mouse cadavers with 1.29- and 1.43-fold PA signal enhancement in comparison to the pure ICG, respectively.

Patient-adapted organ absorbed dose and effective dose estimates in pediatric 18F-FDG positron emission tomography/computed tomography studies

BACKGROUND

Organ absorbed doses and effective doses can be used to compare radiation exposure among medical imaging procedures, compare alternative imaging options, and guide dose optimization efforts. Individual dose estimates are important for relatively radiosensitive patient populations such as children and for radiosensitive organs such as the eye lens. Software-based dose calculation methods conveniently calculate organ dose using patient-adjusted and examination-specific inputs.

METHODS

Organ absorbed doses and effective doses were calculated for 429 pediatric 18F-FDG PET-CT patients. Patient-adjusted and scan-specific information was extracted from the electronic medical record and scanner dose-monitoring software. The VirtualDose and OLINDA/EXM (version 2.0) programs, respectively, were used to calculate the CT and the radiopharmaceutical organ absorbed doses and effective doses. Patients were grouped according to age at the time of the scan as follows: less than 1 year old, 1 to 5 years old, 6 to 10 years old, 11 to 15 years old, and 16 to 17 years old.

RESULTS

The mean (+/- standard deviation, range) total PET plus CT effective dose was 14.5 (1.9, 11.2-22.3) mSv. The mean (+/- standard deviation, range) PET effective dose was 8.1 (1.2, 5.7-16.5) mSv. The mean (+/- standard deviation, range) CT effective dose was 6.4 (1.8, 2.9-14.7) mSv. The five organs with highest PET dose were: Urinary bladder, heart, liver, lungs, and brain. The five organs with highest CT dose were: Thymus, thyroid, kidneys, eye lens, and gonads.

CONCLUSIONS

Organ and effective dose for both the CT and PET components can be estimated with actual patient and scan data using commercial software. Doses calculated using software generally agree with those calculated using dose conversion factors, although some organ doses were found to be appreciably different. Software-based dose calculation methods allow patient-adjusted dose factors. The effort to gather the needed patient data is justified by the resulting value of the characterization of patient-adjusted dosimetry.

Quantification of DNA damage in patients undergoing non-contrast and contrast enhanced whole body PET/CT investigations using comet assay and micronucleus assay

Objective: To quantify DNA damage in patients undergoing non-contrast and contrast-enhanced ¹⁸F-FDG PET/CT whole body positron emission tomography/computed tomography (WB PET/CT) investigations using comet assay technique and micronucleus assay, and to study the effect of other baseline parameters of patients on DNA damage. Methodology: Eighty-four patients referred for ¹⁸F-FDG PET/CT investigation were included in the study of which 44 patients underwent contrast-enhanced WB PET/CT and 40 patients underwent non-contrast WB PET/CT investigations. The investigations were performed on Discovery 690 PET/CT. For contrast-enhanced investigation, Omnipaque300 was injected intravenously based on the patient body weight. Absorbed dose resulting from the intravenous administration of ¹⁸F-FDG was estimated using the ICRP 106 dose coefficients. Radiation dose from the acquisition of CT scans was estimated using CT dose index and dose-length product. Blood samples were collected from the patients for DNA damage analysis. Comet assay and MN assay was used to assess the DNA damage. The Differences in the comet TM (Tail Moment) and MNBC % in both groups were calculated. Result: The radiation dose received by the study population during ¹⁸F-FDG WB PET/CT examination was 27.03 ± 2.33 mSv. Comet TM and percentage frequency of MNBC % was 65.22 ± 35.42 and 18.55 ± 10.14, respectively in the patients injected with contrast and 42.49 ± 28.52 and 13.76 ± 7.52 for non-contrast group. Significant increase in DNA damage was observed in the contrast group as compared to non-contrast group. Significant association was observed between patient weight, contrast volume and TM and MNBC%. Baseline parameters of the patients did not show significant correlation with TM and MNBC%. Conclusion: The patients undergoing contrast-enhanced WB PET/CT investigations have demonstrated higher DNA damage. The DNA damage was also observed to be more in heavier patients. The other baseline parameters of patients like age, sex, CBG, serum creatinine did not show any correlation with DNA damage.

Acquired resistance to tyrosine kinase inhibitors may be linked with the decreased sensitivity to X-ray irradiation

Acquired resistance to chemotherapy and radiation therapy is one of the major obstacles decreasing efficiency of treatment of the oncologic diseases. In this study, on the two cell lines (ovarian carcinoma SKOV-3 and neuroblastoma NGP-127), we modeled acquired resistance to five target anticancer drugs. The cells were grown on gradually increasing concentrations of the clinically relevant tyrosine kinase inhibitors (TKIs) Sorafenib, Pazopanib and Sunitinib, and rapalogs Everolimus and Temsirolimus, for 20 weeks. After 20 weeks of culturing, the half-inhibitory concentrations (IC₅₀) increased by 25 – 186% for the particular combinations of the drugs and cell types. We next subjected cells to 10 Gy irradiation, a dose frequently used in clinical radiation therapy. For the SKOV-3, but not NGP-127 cells, for the TKIs Sorafenib, Pazopanib and Sunitinib, we noticed statistically significant increase in capacity to repair radiation-induced DNA double strand breaks compared to naïve control cells not previously treated with TKIs. These peculiarities were linked with the increased activation of ATM DNA repair pathway in the TKI-treated SKOV-3, but not NGP-127 cells. Our results provide a new cell culture model for studying anti-cancer therapy efficiency and evidence that there may be a tissue-specific radioresistance emerging as a side effect of treatment with TKIs.