We aimed to elucidate the diagnostic potential of the C-X-C motif chemokine receptor 4 (CXCR4)-directed positron emission tomography (PET) tracer \(^{68}\)Ga-Pentixafor in patients with poorly differentiated neuroendocrine carcinomas (NEC), relative to the established reference standard \(^{18}\)F-FDG PET/computed tomography (CT). In our database, we retrospectively identified 11 treatment-naïve patients with histologically proven NEC, who underwent \(^{18}\)F-FDG and CXCR4-directed PET/CT for staging and therapy planning. The images were analyzed on a per-patient and per-lesion basis and compared to immunohistochemical staining (IHC) of CXCR4 from PET-guided biopsies. \(^{68}\)Ga-Pentixafor visualized tumor lesions in 10/11 subjects, while \(^{18}\)F-FDG revealed sites of disease in all 11 patients. Although weak to moderate CXCR4 expression could be corroborated by IHC in 10/11 cases, \(^{18}\)F-FDG PET/CT detected significantly more tumor lesions (102 vs. 42; total lesions, n = 107; p < 0.001). Semi-quantitative analysis revealed markedly higher 18F-FDG uptake as compared to \(^{68}\)Ga-Pentixafor (maximum and mean standardized uptake values (SUV) and tumor-to-background ratios (TBR) of cancerous lesions, SUVmax: 12.8 ± 9.8 vs. 5.2 ± 3.7; SUVmean: 7.4 ± 5.4 vs. 3.1 ± 3.2, p < 0.001; and, TBR 7.2 ± 7.9 vs. 3.4 ± 3.0, p < 0.001). Non-invasive imaging of CXCR4 expression in NEC is inferior to the reference standard \(^{18}\)F-FDG PET/CT.

Background The efficiency of artificial intelligence as computer-aided detection (CADe) systems for colorectal polyps has been demonstrated in several randomized trials. However, CADe systems generate many distracting detections, especially during interventions such as polypectomies. Those distracting CADe detections are often induced by the introduction of snares or biopsy forceps as the systems have not been trained for such situations. In addition, there are a significant number of non-false but not relevant detections, since the polyp has already been previously detected. All these detections have the potential to disturb the examiner's work. Objectives Development and evaluation of a convolutional neuronal network that recognizes instruments in the endoscopic image, suppresses distracting CADe detections, and reliably detects endoscopic interventions. Methods A total of 580 different examination videos from 9 different centers using 4 different processor types were screened for instruments and represented the training dataset (519,856 images in total, 144,217 contained a visible instrument). The test dataset included 10 full-colonoscopy videos that were analyzed for the recognition of visible instruments and detections by a commercially available CADe system (GI Genius, Medtronic). Results The test dataset contained 153,623 images, 8.84% of those presented visible instruments (12 interventions, 19 instruments used). The convolutional neuronal network reached an overall accuracy in the detection of visible instruments of 98.59%. Sensitivity and specificity were 98.55% and 98.92%, respectively. A mean of 462.8 frames containing distracting CADe detections per colonoscopy were avoided using the convolutional neuronal network. This accounted for 95.6% of all distracting CADe detections. Conclusions Detection of endoscopic instruments in colonoscopy using artificial intelligence technology is reliable and achieves high sensitivity and specificity. Accordingly, the new convolutional neuronal network could be used to reduce distracting CADe detections during endoscopic procedures. Thus, our study demonstrates the great potential of artificial intelligence technology beyond mucosal assessment.

Introduction Malnutrition in cancer patients often remains undetected and underestimated in clinical practice despite studies revealing prevalences from 20 to 70%. Therefore, this study aimed to identify patient groups exposed to an increased nutritional risk in a university oncological outpatient center. Methods Between May 2017 and January 2018 we screened oncological patients there using the malnutrition universal screening tool (MUST). Qualitative data were collected by a questionnaire to learn about patients’ individual information needs and changes in patients’ diets and stressful personal nutrition restrictions. Results We included 311 patients with various cancers. 20.3% (n = 63) were found to be at high risk of malnutrition, 16.4% (n = 51) at moderate risk despite a mean body mass index (BMI) of 26.5 ± 4.7 kg/m2. The average age was 62.7 (± 11.8) with equal gender distribution (52% women, n = 162). In 94.8% (n = 295) unintended weight loss led to MUST scoring. Patients with gastrointestinal tumors (25%, n = 78) and patients >65 years (22%, n = 68) were at higher risk. Furthermore, there was a significant association between surgery or chemotherapy within six months before survey and a MUST score ≥2 (OR = 3.6). Taste changes, dysphagia, and appetite loss were also particular risk factors (OR = 2.3–3.2). Young, female and normal-weight patients showed most interest in nutrition in cancer. However, only 38% (n = 118) had a nutritional counseling. Conclusion This study confirms that using the MUST score is a valid screening procedure to identify outpatients at risk of developing malnutrition. Here one in five was at high risk, but only 1% would have been detected by BMI alone. Therefore, an ongoing screening procedure with meaningful parameters should be urgently implemented into the clinical routine of cancer outpatients as recommended in international guidelines.