Correlation between 18F-FDG PET CT SUV and symptomatic or asymptomatic pulmonary tuberculosis

Revolutionizing tumor detection and classification in multimodality imaging based on deep learning approaches: methods, applications and limitations

BACKGROUND

The emergence of deep learning (DL) techniques has revolutionized tumor detection and classification in medical imaging, with multimodal medical imaging (MMI) gaining recognition for its precision in diagnosis, treatment, and progression tracking.

OBJECTIVE

This review comprehensively examines DL methods in transforming tumor detection and classification across MMI modalities, aiming to provide insights into advancements, limitations, and key challenges for further progress.

METHODS

Systematic literature analysis identifies DL studies for tumor detection and classification, outlining methodologies including convolutional neural networks (CNNs), recurrent neural networks (RNNs), and their variants. Integration of multimodality imaging enhances accuracy and robustness.

RESULTS

Recent advancements in DL-based MMI evaluation methods are surveyed, focusing on tumor detection and classification tasks. Various DL approaches, including CNNs, YOLO, Siamese Networks, Fusion-Based Models, Attention-Based Models, and Generative Adversarial Networks, are discussed with emphasis on PET-MRI, PET-CT, and SPECT-CT.

FUTURE DIRECTIONS

The review outlines emerging trends and future directions in DL-based tumor analysis, aiming to guide researchers and clinicians toward more effective diagnosis and prognosis. Continued innovation and collaboration are stressed in this rapidly evolving domain.

CONCLUSION

Conclusions drawn from literature analysis underscore the efficacy of DL approaches in tumor detection and classification, highlighting their potential to address challenges in MMI analysis and their implications for clinical practice.

Estimation of the probability of daily fluctuations of incidence of COVID-19 according to official data

When studying the dynamics of morbidity and mortality, one should not limit ourselves to analyzing general trends. Interesting information can be obtained from the analysis of deviations in morbidity and mortality from the general dynamics. Comparison of the cases of morbidity or death for adjacent time intervals allows us to find out whether the changes in conditions were for short periods of time and whether the cases of morbidity or death were independent. The article consists of two parts: Study of the probability distribution (CDF) of the difference between two independent observations of the Poisson distribution; Application of the results to analyze the morbidity and mortality trends by day for the new coronavirus infection. For the distribution function of the module of difference between two independent observations of the Poisson distribution, an analytical expression has been obtained that allows to get an exact solution. A program has been created, whose software can be downloaded at http://1mgmu.com/nau/DeltaPoisson/DeltaPoisson.zip. An approximate solution that does not require complex calculations has also been obtained, which can be used for an average of more than 20. If real difference is greater than expected, it may be in the following cases: morbidity or mortality varies considerably during the day. That could happen, for example, if the registered number of morbidity on Saturday and Sunday is less than on weekdays due to the management model of the health system, or if the cases are not independent; for example, due to the active identification of infected people among those who have come into contact with the patient. If the difference is less than expected, it may be due to external limiting factors, such as a shortage of test systems for making a diagnosis, a limited number of pathologists to determine the cause of death, and so on. In the analysis of the actual data for COVID-19 it was found that for Poland and Russia, excluding Moscow, the difference in the number of cases and deaths is greater than expected, while for Moscow-less than expected. This may be due to the information policy-the effort to somehow reassure Moscow's population, which in the spring of 2020 had a high incidence rate of the new coronavirus infection.

State of the art of 18F-FDG PET/CT application in inflammation and infection: a guide for image acquisition and interpretation

AIM

The diagnosis, severity and extent of a sterile inflammation or a septic infection could be challenging since there is not one single test able to achieve an accurate diagnosis. The clinical use of 18F-fluorodeoxyglucose ([¹⁸F]FDG) positron emission tomography/computed tomography (PET/CT) imaging in the assessment of inflammation and infection is increasing worldwide. The purpose of this paper is to achieve an Italian consensus document on [¹⁸F]FDG PET/CT or PET/MRI in inflammatory and infectious diseases, such as osteomyelitis (OM), prosthetic joint infections (PJI), infective endocarditis (IE), prosthetic valve endocarditis (PVE), cardiac implantable electronic device infections (CIEDI), systemic and cardiac sarcoidosis (SS/CS), diabetic foot (DF), fungal infections (FI), tuberculosis (TBC), fever and inflammation of unknown origin (FUO/IUO), pediatric infections (PI), inflammatory bowel diseases (IBD), spine infections (SI), vascular graft infections (VGI), large vessel vasculitis (LVV), retroperitoneal fibrosis (RF) and COVID-19 infections.

METHODS

In September 2020, the inflammatory and infectious diseases focus group (IIFG) of the Italian Association of Nuclear Medicine (AIMN) proposed to realize a procedural paper about the clinical applications of [¹⁸F]FDG PET/CT or PET/MRI in inflammatory and infectious diseases. The project was carried out thanks to the collaboration of 13 Italian nuclear medicine centers, with a consolidate experience in this field. With the endorsement of AIMN, IIFG contacted each center, and the pediatric diseases focus group (PDFC). IIFG provided for each team involved, a draft with essential information regarding the execution of [¹⁸F]FDG PET/CT or PET/MRI scan (i.e., indications, patient preparation, standard or specific acquisition modalities, interpretation criteria, reporting methods, pitfalls and artifacts), by limiting the literature research to the last 20 years. Moreover, some clinical cases were required from each center, to underline the teaching points. Time for the collection of each report was from October to December 2020.

RESULTS

Overall, we summarized 291 scientific papers and guidelines published between 1998 and 2021. Papers were divided in several sub-topics and summarized in the following paragraphs: clinical indications, image interpretation criteria, future perspectivess and new trends (for each single disease), while patient preparation, image acquisition, possible pitfalls and reporting modalities were described afterwards. Moreover, a specific section was dedicated to pediatric and PET/MRI indications. A collection of images was described for each indication.

CONCLUSIONS

Currently, [¹⁸F]FDG PET/CT in oncology is globally accepted and standardized in main diagnostic algorithms for neoplasms. In recent years, the ever-closer collaboration among different European associations has tried to overcome the absence of a standardization also in the field of inflammation and infections. The collaboration of several nuclear medicine centers with a long experience in this field, as well as among different AIMN focus groups represents a further attempt in this direction. We hope that this document will be the basis for a "common nuclear physicians' language" throughout all the country.

SUPPLEMENTARY INFORMATION

The online version contains supplementary material available at 10.1007/s40336-021-00445-w.