The potential of large language models in medicine for education and decision making purposes has been demonstrated as they achieve decent scores on medical exams such as the United States Medical ...Licensing Exam (USMLE) and the MedQA exam. In this work, we evaluate the performance of ChatGPT-4 in the specialized field of radiation oncology using the 38th American College of Radiology (ACR) radiation oncology in-training (TXIT) exam and the 2022 Red Journal Gray Zone cases. For the TXIT exam, ChatGPT-3.5 and ChatGPT-4 have achieved the scores of 63.65% and 74.57%, respectively, highlighting the advantage of the latest ChatGPT-4 model. Based on the TXIT exam, ChatGPT-4's strong and weak areas in radiation oncology are identified to some extent. Specifically, ChatGPT-4 demonstrates better knowledge of statistics, CNS & eye, pediatrics, biology, and physics than knowledge of bone & soft tissue and gynecology, as per the ACR knowledge domain. Regarding clinical care paths, ChatGPT-4 performs better in diagnosis, prognosis, and toxicity than brachytherapy and dosimetry. It lacks proficiency in in-depth details of clinical trials. For the Gray Zone cases, ChatGPT-4 is able to suggest a personalized treatment approach to each case with high correctness and comprehensiveness. Importantly, it provides novel treatment aspects for many cases, which are not suggested by any human experts. Both evaluations demonstrate the potential of ChatGPT-4 in medical education for the general public and cancer patients, as well as the potential to aid clinical decision-making, while acknowledging its limitations in certain domains. Because of the risk of hallucination, facts provided by ChatGPT always need to be verified.
Brain metastases occur frequently in patients with metastatic cancer. Early and accurate detection of brain metastases is very essential for treatment planning and prognosis in radiation therapy. To ...improve brain metastasis detection performance with deep learning, a custom detection loss called volume-level sensitivity-specificity (VSS) is proposed, which rates individual metastasis detection sensitivity and specificity in (sub-)volume levels. As sensitivity and precision are always a trade-off in a metastasis level, either a high sensitivity or a high precision can be achieved by adjusting the weights in the VSS loss without decline in dice score coefficient for segmented metastases. To reduce metastasis-like structures being detected as false positive metastases, a temporal prior volume is proposed as an additional input of DeepMedic. The modified network is called DeepMedic+ for distinction. Our proposed VSS loss improves the sensitivity of brain metastasis detection for DeepMedic, increasing the sensitivity from 85.3% to 97.5%. Alternatively, it improves the precision from 69.1% to 98.7%. Comparing DeepMedic+ with DeepMedic with the same VSS loss, 44.4% of the false positive metastases are reduced in the high sensitivity model and the precision reaches 99.6% for the high specificity model. The mean dice coefficient for all metastases is about 0.81. With the ensemble of the high sensitivity and high specificity models, on average only 1.5 false positive metastases per patient needs further check, while the majority of true positive metastases are confirmed. The ensemble learning is able to distinguish high confidence true positive metastases from metastases candidates that require special expert review or further follow-up, being particularly well-fit to the requirements of expert support in real clinical practice.
Background: Tumor segmentation in MRI is crucial in radiotherapy (RT) treatment planning for brain tumor patients. Segment anything (SA), a novel promptable foundation model for autosegmentation, has ...shown high accuracy for multiple segmentation tasks but was not evaluated on medical datasets yet. Methods: SA was evaluated in a point-to-mask task for glioma brain tumor autosegmentation on 16744 transversal slices from 369 MRI datasets (BraTS 2020). Up to 9 point prompts were placed per slice. Tumor core (enhancing tumor + necrotic core) was segmented on contrast-enhanced T1w sequences. Out of the 3 masks predicted by SA, accuracy was evaluated for the mask with the highest calculated IoU (oracle mask) and with highest model predicted IoU (suggested mask). In addition to assessing SA on whole MRI slices, SA was also evaluated on images cropped to the tumor (max. 3D extent + 2 cm). Results: Mean best IoU (mbIoU) using oracle mask on full MRI slices was 0.762 (IQR 0.713-0.917). Best 2D mask was achieved after a mean of 6.6 point prompts (IQR 5-9). Segmentation accuracy was significantly better for high- compared to low-grade glioma cases (mbIoU 0.789 vs. 0.668). Accuracy was worse using MRI slices cropped to the tumor (mbIoU 0.759) and was much worse using suggested mask (full slices 0.572). For all experiments, accuracy was low on peripheral slices with few tumor voxels (mbIoU, <300: 0.537 vs. >=300: 0.841). Stacking best oracle segmentations from full axial MRI slices, mean 3D DSC for tumor core was 0.872, which was improved to 0.919 by combining axial, sagittal and coronal masks. Conclusions: The Segment Anything foundation model, while trained on photos, can achieve high zero-shot accuracy for glioma brain tumor segmentation on MRI slices. The results suggest that Segment Anything can accelerate and facilitate RT treatment planning, when properly integrated in a clinical application.
Background: Deep learning (DL)-based head and neck lymph node level (HN_LNL) autodelineation is of high relevance to radiotherapy research and clinical treatment planning but still underinvestigated ...in academic literature. Methods: An expert-delineated cohort of 35 planning CTs was used for training of an nnU-net 3D-fullres/2D-ensemble model for autosegmentation of 20 different HN_LNL. A second cohort acquired at the same institution later in time served as the test set (n=20). In a completely blinded evaluation, 3 clinical experts rated the quality of DL autosegmentations in a head-to-head comparison with expert-created contours. For a subgroup of 10 cases, intraobserver variability was compared to the average DL autosegmentation accuracy on the original and recontoured set of expert segmentations. A postprocessing step to adjust craniocaudal boundaries of level autosegmentations to the CT slice plane was introduced and the effect on geometric accuracy and expert rating was investigated. Results: Blinded expert ratings for DL segmentations and expert-created contours were not significantly different. DL segmentations with slice plane adjustment were rated numerically higher (mean, 81.0 vs. 79.6,p=0.185) and DL segmentations without slice plane adjustment were rated numerically lower (77.2 vs. 79.6,p=0.167) than manually drawn contours. DL segmentations with CT slice plane adjustment were rated significantly better than DL contours without slice plane adjustment (81.0 vs. 77.2,p=0.004). Geometric accuracy of DL segmentations was not different from intraobserver variability (mean, 0.76 vs. 0.77, p=0.307). Conclusions: We show that a nnU-net 3D-fullres/2D-ensemble model can be used for highly accurate autodelineation of HN_LNL using only a limited training dataset that is ideally suited for large-scale standardized autodelineation of HN_LNL in the research setting.
The aim of this exploratory study was to evaluate the influence of hepatic steatosis on the detection rate of metastases in gadoxetic acid-enhanced liver magnetic resonance imaging (MRI). A total of ...50 patients who underwent gadoxetic acid-enhanced MRI (unenhanced T1w in- and opposed-phase, T2w fat sat, unenhanced 3D-T1w fat sat and 3-phase dynamic contrast-enhanced (uDP), 3D-T1w fat sat hepatobiliary phase (HP)) were retrospectively included. Two blinded observers (O1/O2) independently assessed the images to determine the detection rate in uDP and HP. The hepatic signal fat fraction (HSFF) was determined as the relative signal intensity reduction in liver parenchyma from in- to opposed-phase images. A total of 451 liver metastases were detected (O1/O2, n = 447/411). O1/O2 detected 10.9%/9.3% of lesions exclusively in uDP and 20.2%/15.5% exclusively in HP. Lesions detected exclusively in uDP were significantly associated with a larger HSFF (area under curve (AUC) of receiver operating characteristic (ROC) analysis, 0.93; p < 0.001; cutoff, 41.5%). The exclusively HP-positive lesions were significantly associated with a smaller diameter (ROC-AUC, 0.82; p < 0.001; cutoff, 5 mm) and a smaller HSFF (ROC-AUC, 0.61; p < 0.001; cutoff, 13.3%). Gadoxetic acid imaging has the advantage of detecting small occult metastatic liver lesions in the HP. However, using non-optimized standard fat-saturated 3D-T1w protocols, severe steatosis (HSFF > 30%) is a potential pitfall for the detection of metastases in HP.
Triticale is adapted to a wide range of abiotic stress conditions, is an important high-quality feed stock and produces similar grain yield but more biomass compared to other crops. Modern genomic ...approaches aimed at enhancing breeding progress in cereals require high-quality genetic linkage maps. Consensus maps are genetic maps that are created by a joint analysis of the data from several segregating populations and different approaches are available for their construction. The phenomenon that alleles at a locus deviate from the Mendelian expectation has been defined as segregation distortion. The study of segregation distortion is of particular interest in doubled haploid (DH) populations due to the selection pressure exerted on the plants during the process of their establishment.
The final consensus map, constructed out of six segregating populations derived from nine parental lines, incorporated 2555 DArT markers mapped to 2602 loci (1929 unique). The map spanned 2309.9 cM with an average number of 123.9 loci per chromosome and an average marker density of one unique locus every 1.2 cM. The R genome showed the highest marker coverage followed by the B genome and the A genome. In general, locus order was well maintained between the consensus linkage map and the component maps. However, we observed several groups of loci for which the colinearity was slightly uneven. Among the 2602 loci mapped on the consensus map, 886 showed distorted segregation in at least one of the individual mapping populations. In several DH populations derived by androgenesis, we found chromosomes (2B, 3B, 1R, 2R, 4R and 7R) containing regions where markers exhibited a distorted segregation pattern. In addition, we observed evidence for segregation distortion between pairs of loci caused either by a predominance of parental or recombinant genotypes.
We have constructed a reliable, high-density DArT marker consensus genetic linkage map as a basis for genomic approaches in triticale research and breeding, for example for multiple-line cross QTL mapping experiments. The results of our study exemplify the tremendous impact of different DH production techniques on allele frequencies and segregation distortion covering whole chromosomes.
Hypoxia signaling plays a major role in non-malignant and malignant hyperproliferative diseases. Pulmonary hypertension (PH), a hypoxia-driven vascular disease, is characterized by a glycolytic ...switch similar to the Warburg effect in cancer. Ras association domain family 1A (RASSF1A) is a scaffold protein that acts as a tumour suppressor. Here we show that hypoxia promotes stabilization of RASSF1A through NOX-1- and protein kinase C- dependent phosphorylation. In parallel, hypoxia inducible factor-1 α (HIF-1α) activates RASSF1A transcription via HIF-binding sites in the RASSF1A promoter region. Vice versa, RASSF1A binds to HIF-1α, blocks its prolyl-hydroxylation and proteasomal degradation, and thus enhances the activation of the glycolytic switch. We find that this mechanism operates in experimental hypoxia-induced PH, which is blocked in RASSF1A knockout mice, in human primary PH vascular cells, and in a subset of human lung cancer cells. We conclude that RASSF1A-HIF-1α forms a feedforward loop driving hypoxia signaling in PH and cancer.
Amyotrophic lateral sclerosis (ALS) is a neurodegenerative disease primarily characterized by the progressive impairment of motor functions. However, a significant portion of affected patients ...develops severe cognitive dysfunction, developing a widespread white (WM) and gray matter (GM) microstructural impairment. The objective of this study is to determine if Gaussian and non-Gaussian diffusion models gathered by ultra-high field diffusion MRI (UHFD-MRI) are an appropriate tool to detect early structural changes in brain white and gray matter in a preclinical model of ALS. ALS brains (G93A-SOD1mice) were scanned in a 16.7 T magnet. Diffusion tensor imaging (DTI) and neurite orientation dispersion and density imaging (NODDI) have shown presymptomatic decrease in axonal organization by Fractional Anisotropy (FA) and neurite content by Intracellular Volume Fraction (ICVF) across deep WM (corpus callosum) as well as superficial (cortex) and deep (hippocampus) GM. Additional diffusion kurtosis imaging (DKI) analysis demonstrated broader and earlier GM reductions in mean kurtosis (MK), possibly related to the decrease in neuronal complexity. Histological validation was obtained by an ALS fluorescent mice reporter (YFP, G93A-SOD1 mice). The combination of DTI, NODDI, and DKI models have proved to provide a more complete assessment of the early microstructural changes in the ALS brain, particularly in areas associated with high cognitive functions. This comprehensive approach should be considered as a valuable tool for the early detection of neuroimaging markers.
Leaves of C4 crops usually have higher radiation, water and nitrogen use efficiencies compared to the C3 species. Engineering C4 traits into C3 crops has been proposed as one of the most promising ...ways to repeal the biomass yield ceiling. To better understand the function of C4 photosynthesis, and to identify candidate genes that are associated with the C4 pathways, a comparative transcription network analysis was conducted on leaf developmental gradients of three C4 species including maize, green foxtail and sorghum and one C3 species, rice. By combining the methods of gene co-expression and differentially co-expression networks, we identified a total of 128 C4 specific genes. Besides the classic C4 shuttle genes, a new set of genes associated with light reaction, starch and sucrose metabolism, metabolites transportation, as well as transcription regulation, were identified as involved in C4 photosynthesis. These findings will provide important insights into the differential gene regulation between C3 and C4 species, and a good genetic resource for establishing C4 pathways in C3 crops.
The angiogenic function of endothelial cells is regulated by numerous mechanisms, but the impact of long noncoding RNAs (lncRNAs) has hardly been studied. We set out to identify novel and ...functionally important endothelial lncRNAs.
Epigenetically controlled lncRNAs in human umbilical vein endothelial cells were searched by exon-array analysis after knockdown of the histone demethylase JARID1B. Molecular mechanisms were investigated by RNA pulldown and immunoprecipitation, mass spectrometry, microarray, several knockdown approaches, CRISPR-Cas9, assay for transposase-accessible chromatin sequencing, and chromatin immunoprecipitation in human umbilical vein endothelial cells. Patient samples from lung and tumors were studied for MANTIS expression.
A search for epigenetically controlled endothelial lncRNAs yielded lncRNA n342419, here termed MANTIS, as the most strongly regulated lncRNA. Controlled by the histone demethylase JARID1B, MANTIS was downregulated in patients with idiopathic pulmonary arterial hypertension and in rats treated with monocrotaline, whereas it was upregulated in carotid arteries of
subjected to atherosclerosis regression diet, and in endothelial cells isolated from human glioblastoma patients. CRISPR/Cas9-mediated deletion or silencing of MANTIS with small interfering RNAs or GapmeRs inhibited angiogenic sprouting and alignment of endothelial cells in response to shear stress. Mechanistically, the nuclear-localized MANTIS lncRNA interacted with BRG1, the catalytic subunit of the switch/sucrose nonfermentable chromatin-remodeling complex. This interaction was required for nucleosome remodeling by keeping the ATPase function of BRG1 active. Thereby, the transcription of key endothelial genes such as
,
, and
was regulated by ensuring efficient RNA polymerase II machinery binding.
MANTIS is a differentially regulated novel lncRNA facilitating endothelial angiogenic function.