Artificial intelligence vs. radiologist: accuracy of wrist fracture detection on radiographs.

OBJECTIVE: To compare the performances of artificial intelligence (AI) to those of radiologists in wrist fracture detection on radiographs. METHODS: This retrospective study included 637 patients (1917 radiographs) with wrist trauma between January 2017 and December 2019. The AI software used was a deep neuronal network algorithm. Ground truth was established by three senior musculoskeletal radiologists who compared the initial radiology reports (IRR) made by non-specialized radiologists, the results of AI, and the combination of AI and IRR (IR+AI) RESULTS: A total of 318 fractures were reported by the senior radiologists in 247 patients. Sensitivity of AI (83%; 95% CI: 78-87%) was significantly greater than that of IRR (76%; 95% CI: 70-81%) (p < 0.001). Specificities were similar for AI (96%; 95% CI: 93-97%) and for IRR (96%; 95% CI: 94-98%) (p = 0.80). The combination of AI+IRR had a significantly greater sensitivity (88%; 95% CI: 84-92%) compared to AI and IRR (p < 0.001) and a lower specificity (92%; 95% CI: 89-95%) (p < 0.001). The sensitivity for scaphoid fracture detection was acceptable for AI (84%) and IRR (80%) but poor for the detection of other carpal bones fracture (41% for AI and 26% for IRR). CONCLUSIONS: Performance of AI in wrist fracture detection on radiographs is better than that of non-specialized radiologists. The combination of AI and radiologist's analysis yields best performances. KEY POINTS: • Artificial intelligence has better performances for wrist fracture detection compared to non-expert radiologists in daily practice. • Performance of artificial intelligence greatly differs depending on the anatomical area. • Sensitivity of artificial intelligence for the detection of carpal bones fractures is 56%.

Body packing, body stuffing and body pushing: Characteristics and pitfalls on low-dose CT.

Because of availability and low radiation dose level, low dose computed tomography (CT) is now commonly used to identify illicit in corpore drug transportation. This review illustrates the most common CT findings of in corpore drug transportation and describes complications due to in corpore drug transportation, with a special emphasis on low dose CT. Major information such as number of packets, exact location and aspect of packets must be assessed. Radiologist must be aware of the imaging characteristics of "in corpore" illicit drug transportation, and should know situations that may alter drug smugglers management.

Imaging for evaluation of endometriosis and adenomyosis.

Endometriosis and adenomyosis are two frequent diseases that impair women's quality of life by causing pain and infertility. Both endometriosis and adenomyosis are heterogeneous diseases that manifest as different forms. Adenomyosis may be described as diffuse adenomyosis, focal adenomyosis especially of the outer myometrium and cystic adenomyoma. Endometriosis has three phenotypes: superficial peritoneal endometriosis (SUP), ovarian endometrioma (OMA), and deep infiltrating endometriosis (DIE). These two diseases are closely linked, and it is now clear that adenomyosis can either arise on its own or coexist with endometriosis. There is a strong clinical relationship between endometriosis and adenomyosis according to their respective phenotypes. Various classifications are available to describe both diseases. Transvaginal ultrasonography (TVUS) and/or pelvic magnetic resonance imaging (MRI) are the first examination performed when endometriosis or adenomyosis are suspected. These two imaging techniques, used in a combination manner, allow accurate description of both endometriosis and adenomyosis, to assess the diagnosis and to improve clinical and surgical care. In this review, we described the different imaging aspects of endometriosis and adenomyosis to help the less experienced radiologist or gynecologist in the diagnosis and evaluation of those diseases.

Historadiological correlations in high-grade glioma with the histone 3.3 G34R mutation.

BACKGROUND AND PURPOSE: Molecular alterations were recently added to the World Health Organization (WHO) 2016 classification of central nervous system (CNS) tumors. We correlated the histological and radiological features of G34R mutant high-grade gliomas, a recently described hemispheric and supratentorial glioma of children and young adults. MATERIALS AND METHODS: We performed a retrospective multicenter study on the histopathological and MRI results of 12 patients. RESULTS: All tumors were supratentorial. Several radiological aspects were observed. Height over 12 were bulky and well delineated tumors, without visible peritumoral infiltration on MRI and pathologically characterized by highly cellular tissue associated with a moderate peritumoral infiltrative component. Two tumors were ill-defined and hyperintense on T2 sequences and pathologically characterized by diffuse tumoral infiltration. Two tumors were bulky and well delineated with an infiltrative component, both radiologically and histopathologically. CONCLUSIONS: These different patterns may correspond to different pathological mechanisms and a potential link with prognosis should be assessed in further studies.

Evaluation of a novel antibody to define histone 3.3 G34R mutant brain tumours.

Missense somatic mutations affecting histone H3.1 and H3.3 proteins are now accepted as the hallmark of paediatric diffuse intrinsic pontine gliomas (DIPG), non-brain stem paediatric high grade gliomas (pHGG) as well as a subset of adult glioblastoma multiforme (GBM). Different mutations give rise to one of three amino acid substitutions at two critical positions within the histone tails, K27M, G34R/V. Several studies have highlighted gene expression and epigenetic changes associated with histone H3 mutations; however their precise roles in tumourigenesis remain incompletely understood. Determining how such amino acid substitutions in a protein affect its properties can be challenging because of difficulties in detecting and tracking mutant proteins within cells and tissues. Here we describe a strategy for the generation of antibodies to discriminate G34R and G34V mutant histone H3 proteins from their wild-type counterparts. Antibodies were validated by western blotting and immunocytochemistry, using recombinant H3.3 proteins and paediatric GBM cell lines. The H3-G34R antibody demonstrated a high degree of selectivity towards its target sequence. Accordingly, immunostaining on a cohort of 22 formalin-fixed paraffin embedded tumours with a previously known H3.3 G34R mutation status, detected successfully the corresponding mutant protein in 11/11 G34R cases. Since there was a high concordance between genotype and immunohistochemical analysis of G34R mutant tumour samples, we analysed a series of tissue microarrays (TMAs) to assess the specificity of the antibody in a range of paediatric brain tumours, and noted immunoreactivity in 2/634 cases. Importantly, we describe the generation and validation of highly specific antibodies for G34 mutations. Overall our work adds to an extremely valuable portfolio of antibodies, not only for histopathologic detection of tumour-associated mutant histone sequences, but also facilitating the study of spatial/anatomical aspects of tumour formation and the identification of downstream targets and pathways in malignant glioma progression.