Search

Aim: This systematic review aims to present the available evidence on the use of radiomic features (RFs) extracted from PET imaging in patients with prostate cancer (PCa). Materials and methods: A comprehensive literature search of studies on the utility of PET-derived RFs in patients with PCa was performed in the PubMed/MEDLINE database through February 24th, 2021 using the following search string: [“positron-emission tomography” (MeSh terms) OR “positron emission tomography computed tomography” (MeSh terms) OR “positron-emission tomography” (all fields) OR “positron emission tomography computed tomography” (all fields) OR “PET” (all fields)] AND [“radiomics” (all fields) OR “radiomic” (all fields) OR “radiogenomics” (all fields) OR “radiogenomic”(all fields) OR “machine learning”(all fields) OR “machine learning”(MeSh terms) OR “artificial intelligence”(MeSh terms) OR “artificial intelligence”(all fields)] AND [“prostatic neoplasms” (MeSh terms) OR “prostate cancer”(all fields) OR “prostatic carcinoma” (all fields) OR “prostate carcinoma” (all fields) OR “prostatic tumor” (all fields) OR “prostatic tumour” (all fields)]. The Google scholar database was interrogated to find additional studies. Results: Seven studies were ultimately included in the systematic review and summarized in two relevant clinical sections: (1) primary staging and (2) restaging. In primary staging, RFs, extracted from 68Ga-prostate-specific membrane antigen (PSMA) PET may characterize intraprostatic radiotracer hotspots in patients with high- and intermediate-risk, discriminate between Gleason Score (GS) 7 and ≥ 8 and between pN1 and pN0 disease, and suggest presence of intraprostatic lesions missed at visual PET examination. Machine learning (ML) may help selecting RFs able to predict risk classification (low vs. high), lymph node involvement, presence of nodal or distant metastasis, GS and extracapsular extension. At restaging, PET_Kurtosis may correlate with OS in patients with advanced PCa scheduled for 177Lu-PSMA treatment, whereas ML may assist discrimination of malignant lesions from physiologic/unspecific tracer accumulation, and predict disease progression. Conclusion: To date, although PET literature appears still too narrow to draw definitive conclusions, PET-derived RFs appear promising in PCa. ML seems an important tool that may contribute to the widespread use of radiomics and subsequent implementation in the clinical setting.

Ingestion of caustic substances, whether accidental or for the purpose of suicide, can cause severe lesions of the lips, oral cavity, pharynx, upper gastrointestinal system, and upper airways. In particular, caustic agents could be responsible for severe esophageal injuries resulting in short- and long-term complications. Because of these important clinical implications, timely diagnosis and appropriate management are crucial. In the evaluation of esophageal injuries, thoraco-abdominal computed tomography (CT) is preferable to endoscopy as it avoids the risk of esophageal perforation and allows the evaluation of esophageal injuries as well as of the surrounding tissue. In this review, we report CT findings of esophageal injuries and possible related thoracic complications caused by caustic ingestion.

Congenital cystic lesions of the bile ducts represent a spectrum of liver and biliary system lesions, resulting from abnormal embryologic development of the ductal plate. These disorders include Caroli disease, choledochal cysts, autosomal dominant polycystic liver disease, congenital hepatic fibrosis, and biliary hamartomas. Each disorder carries a peculiar clinical presentation, prognosis, and risk of complications. Knowledge of radiological findings of fibropolycystic liver diseases is crucial for their appropriate detection and for differential diagnosis with other similar hepatic cystic lesions, in order to avoid relevant misdiagnosis. The aim of this review is to provide an illustrative summary of the most relevant imaging findings of these conditions as encountered on ultrasound, computed tomography, and magnetic resonance imaging, and provide pearls for imaging-based differential diagnosis. (c) 2019 Elsevier Inc. All rights reserved.

Magnetic Resonance Imaging-based prostate segmentation is an essential task for adaptive radiotherapy and for radiomics studies whose purpose is to identify associations between imaging features and patient outcomes. Because manual delineation is a time-consuming task, we present three deep-learning (DL) approaches, namely UNet, efficient neural network (ENet), and efficient residual factorized convNet (ERFNet), whose aim is to tackle the fully-automated, real-time, and 3D delineation process of the prostate gland on T2-weighted MRI. While UNet is used in many biomedical image delineation applications, ENet and ERFNet are mainly applied in self-driving cars to compensate for limited hardware availability while still achieving accurate segmentation. We apply these models to a limited set of 85 manual prostate segmentations using the k-fold validation strategy and the Tversky loss function and we compare their results. We find that ENet and UNet are more accurate than ERFNet, with ENet much faster than UNet. Specifically, ENet obtains a dice similarity coefficient of 90.89% and a segmentation time of about 6 s using central processing unit (CPU) hardware to simulate real clinical conditions where graphics processing unit (GPU) is not always available. In conclusion, ENet could be efficiently applied for prostate delineation even in small image training datasets with potential benefit for patient management personalization.