A feasibility study to predict 3D dose delivery accuracy for IMRT using DenseNet with log files.

OBJECTIVE: This study aims to explore the feasibility of DenseNet in the establishment of a three-dimensional (3D) gamma prediction model of IMRT based on the actual parameters recorded in the log files during delivery. METHODS: A total of 55 IMRT plans (including 367 fields) were randomly selected. The gamma analysis was performed using gamma criteria of 3% /3 mm (Dose Difference/Distance to Agreement), 3% /2 mm, 2% /3 mm, and 2% /2 mm with a 10% dose threshold. In addition, the log files that recorded the gantry angle, monitor units (MU), multi-leaf collimator (MLC), and jaws position during delivery were collected. These log files were then converted to MU-weighted fluence maps as the input of DenseNet, gamma passing rates (GPRs) under four different gamma criteria as the output, and mean square errors (MSEs) as the loss function of this model. RESULTS: Under different gamma criteria, the accuracy of a 3D GPR prediction model decreased with the implementation of stricter gamma criteria. In the test set, the mean absolute error (MAE) of the prediction model under the gamma criteria of 3% /3 mm, 2% /3 mm, 3% /2 mm, and 2% /2 mm was 1.41, 1.44, 3.29, and 3.54, respectively; the root mean square error (RMSE) was 1.91, 1.85, 4.27, and 4.40, respectively; the Sr was 0.487, 0.554, 0.573, and 0.506, respectively. There was a correlation between predicted and measured GPRs (P <  0.01). Additionally, there was no significant difference in the accuracy between the validation set and the test set. The accuracy in the high GPR group was high, and the MAE in the high GPR group was smaller than that in the low GPR group under four different gamma criteria. CONCLUSIONS: In this study, a 3D GPR prediction model of patient-specific QA using DenseNet was established based on log files. As an auxiliary tool for 3D dose verification in IMRT, this model is expected to improve the accuracy and efficiency of dose validation.

Deep-learning based segmentation of ultrasound adipose image for liposuction.

BACKGROUND: To develop an automatic and reliable ultrasonic visual system for robot- or computer-assisted liposuction, we examined the use of deep learning for the segmentation of adipose ultrasound images in clinical and educational settings. METHODS: To segment adipose layers, it is proposed to use an Attention Skip-Convolutions ResU-Net (Attention SCResU-Net) consisting of SC residual blocks, attention gates and U-Net architecture. Transfer learning is utilised to compensate for the deficiency of clinical data. The Bama pig and clinical human adipose ultrasound image datasets are utilized, respectively. RESULTS: The final model obtains a Dice of 99.06 ± 0.95% and an ASD of 0.19 ± 0.18 mm on clinical datasets, outperforming other methods. By fine-tuning the eight deepest layers, accurate and stable segmentation results are obtained. CONCLUSIONS: The new deep-learning method achieves the accurate and automatic segmentation of adipose ultrasound images in real-time, thereby enhancing the safety of liposuction and enabling novice surgeons to better control the cannula.

Three Glycosyltransferase Mutants in a One-Pot Multi-enzyme System with Enhanced Efficiency for Biosynthesis of Quercetin-3,4'-O-diglucoside.

Quercetin-3,4'-O-diglucoside (Q3,4'G), among the major dietary flavonoids, is superior to quercetin aglycone or quercetin monoglucoside in solubility. However, its low content in nature makes it hard to be prepared in large quantities by traditional extraction methods. In the present study, the F378S mutant of UGT78D2 (78D2_F378S) derived from Arabidopsis thaliana with improved regioselectivity and the V371A mutant of UGT73G1 (73G1_V371A) derived from Allium cepa were adopted to realize a two-step continuous glycosylation of quercetin to produce Q3,4'G. The mutation S31D was introduced to the sucrose synthase from Micractinium conductrix with enhanced activity, which was responsible for regenerating UDP-glucose by coupling with 78D2_F378S and 73G1_V371A. Using the aforementioned enzymes, prepared from the three-enzyme co-expression strain, 4.4 ± 0.03 g/L (7.0 ± 0.05 mM, yield 21.2%) Q3,4'G was produced from 10 g/L quercetin after reaction for 24 h at 45 °C.

Predictive effect of J waves on cardiac compression and clinical prognosis of esophageal tumors: a retrospective study.

Background: The J wave syndromes (JWS) could be observed in patients with mediastinal tumors, though few studies have verified the statistical correlation between J waves and cardiac compression by tumors. This study aimed to investigate the relationship between J waves and cardiac compression by esophageal tumor and to compare the prediction of J waves on clinical prognosis with that of cardiac compression by esophageal tumor. Methods: We enrolled 273 patients (228 males, 45 females; mean 63.8±7.5 years) with esophageal tumors admitted to Shanghai Chest Hospital between August 2016 and November 2020. The J wave was defined as a J-point elevation of ≥0.1 mV in a 12-lead electrocardiogram (ECG) and classified into multiple types. Chest computed tomography (CT) was reviewed to clarify the anatomical relationship between the heart and the esophageal tumor. The prognosis of severe cardiac events and survival status were followed up through medical history, examination records and telephone records. Results: J waves were present in 141 patients among all 273 cases. The sensitivity and specificity of cardiac compression by the tumor for J waves were 78.1% and 67.3%, respectively. The odds ratio (OR) of cardiac compression by the tumor to J waves was 7.33 [95% confidence interval (CI): 4.21-12.74; P<0.001]. The Kappa coefficient between J waves and cardiac compression was 0.44±0.05. The significance association between J waves and cardiac compression was independent from other clinical variables (P<0.001). Decreased J wave amplitude was correlated with the disappearance of cardiac compression during follow-up (P=0.03). Patients with J waves had a higher risk of severe cardiac events than those without J waves (OR =2.84, 95% CI: 1.22-6.63; P=0.01). During the follow-up period, we found that the presence of J waves [hazard ratio (HR) =2.28; 95% CI: 1.35-3.84; P=0.002] and cardiac compression by the tumor (HR =2.51; 95% CI: 1.51-4.17; P<0.001) were both negatively correlated with the survival time of patients. Conclusions: The presence of J waves could be used as an effective mean to predict the mechanical impact of esophageal tumor on the heart, and played an important role in predicting the survival of patients.

Virtual Patient-Specific Quality Assurance of IMRT Using UNet++: Classification, Gamma Passing Rates Prediction, and Dose Difference Prediction.

The dose verification in radiotherapy quality assurance (QA) is time-consuming and places a heavy workload on medical physicists. To provide a clinical tool to perform patient specific QA accurately, the UNet++ is investigated to classify failed or pass fields (the GPR lower than 85% is considered "failed" while the GPR higher than 85% is considered "pass"), predict gamma passing rates (GPR) for different gamma criteria, and predict dose difference from virtual patient-specific quality assurance in radiotherapy. UNet++ was trained and validated with 473 fields and tested with 95 fields. All plans used Portal Dosimetry for dose verification pre-treatment. Planar dose distribution of each field was used as the input for UNet++, with QA classification results, gamma passing rates of different gamma criteria, and dose difference were used as the output. In the test set, the accuracy of the classification model was 95.79%. The mean absolute error (MAE) were 0.82, 0.88, 2.11, 2.52, and the root mean squared error (RMSE) were 1.38, 1.57, 3.33, 3.72 for 3%/3mm, 3%/2 mm, 2%/3 mm, 2%/2 mm, respectively. The trend and position of the predicted dose difference were consistent with the measured dose difference. In conclusion, the Virtual QA based on UNet++ can be used to classify the field passed or not, predict gamma pass rate for different gamma criteria, and predict dose difference. The results show that UNet++ based Virtual QA is promising in quality assurance for radiotherapy.

Bioconversion of Stevioside to Rebaudioside E Using Glycosyltransferase UGTSL2.

Rebaudioside E, one of the minor components of steviol glycosides, was first isolated and identified from Stevia rebaudiana in 1977. It is a high-intensity sweetener that tastes about 150-200 times sweeter than sucrose and is also a precursor for biosynthesis of rebaudioside D and rebaudioside M, the next-generation Stevia sweeteners. In this work, new unknown steviol glycosides were enzymatically synthesized from stevioside by coupling UDP-glucosyltransferase UGTSL2 from Solanum lycopersicum and sucrose synthase StSUS1 from Solanum tuberosum. Rebaudioside E was speculated to be the main product of glucosylation of the Glc(ß1→C-19) residue of stevioside along with the formation of a (ß1→2) linkage based on the analysis of the regioselectivity and stereoselectivity of UGTSL2, and verified afterwards by LC-MS/MS with standard. In a 20-ml bioconversion reaction of 20 g/l stevioside by UGTSL2 and StSUS1, 15.92 g/l rebaudioside E was produced for 24 h.

Production of rebaudioside D from stevioside using a UGTSL2 Asn358Phe mutant in a multi-enzyme system.

Rebaudioside D is a sweetener from Stevia rebaudiana with superior sweetness and organoleptic properties, but its production is limited by its minute abundance in S. rebaudiana leaves. In this study, we established a multi-enzyme reaction system with S. rebaudiana UDP-glycosyltransferases UGT76G1, Solanum lycopersicum UGTSL2 and Solanum tuberosum sucrose synthase StSUS1, achieving a two-step glycosylation of stevioside to produce rebaudioside D. However, an increase in the accumulation of rebaudioside D required the optimization of UGTSL2 catalytic activity towards glucosylation of rebaudioside A and reducing the formation of the side-product rebaudioside M2. On the basis of homology modelling and structural analysis, Asn358 in UGTSL2 was subjected to saturating mutagenesis, and the Asn358Phe mutant was used instead of wild-type UGTSL2 for bioconversion. The established multi-enzyme reaction system employing the Asn358Phe mutant produced 14.4 g l-1 (1.6 times of wild-type UGTSL2) rebaudioside D from 20 g l-1 stevioside after reaction for 24 h. This system is useful for large-scale rebaudioside D production and expands our understanding of the pathways involved in its synthesis.

Natural Product Glycosylation: Biocatalytic Synthesis of Quercetin-3,4'-O-diglucoside.

Flavonoids have gained much attention for their proposed positive effects for human health. Glycosylation is a significant method for the structural modification of various flavanols, resulting in glycosides with increased solubility, stability, and bioavailability compared with the corresponding aglycone. Natural product glycosylation by using enzymes has emerged as a topic of interest as it offers a sustainable and economical alternative source so as to address supply scalability limitations associated with plant-based production. Quercetin-3,4'-O-diglucoside, as one of the major but trace bioactive flavonoids in onion (Allium cepa), is superior or at least equal to quercetin aglycone in its bioavailability. In the present study, the onion-derived enzyme, UGT73G1, coupled with sucrose synthase, StSUS1, from Solanum tuberosum formed a circulatory system to produce quercetin-3,4'-O-diglucoside from quercetin, which preferred sucrose as a sugar donor and quercetin as a sugar acceptor. The optimal conditions were determined in order to increase the production of quercetin-3,4'-O-diglucoside. The maximum concentration of quercetin-3,4'-O-diglucoside achieved in a 10-mL reaction was 427.11 mg/L, from the conversion of 1 g/L of quercetin for 16 h at 40 °C and pH 7.2.

Improved soluble bacterial expression and properties of the recombinant flavonoid glucosyltransferase UGT73G1 from Allium cepa.

Glycosylation of quercetin using flavonol-specific glycosyltransferases offers an alternate method for isoquercitrin production. Obtaining sufficient quantities of bioactive enzymes is an important prerequisite for highly effective biocatalysis and biotransformation. In this study, a codon-optimized gene for the flavonoid glucosyltransferase UGT73G1 from Allium cepa was heterologously expressed in the preferred prokaryotic expression host Escherichia coli. By combining expression as a fusion protein with 6-histidine tags with coexpression with molecular chaperones, increased soluble expression of UGT73G1 was achieved in E. coli. Two-terminal 6-histidine tags contributed more to the expression than molecular chaperones, as demonstrated by comparison of specific activities in crude extracts obtained from the recombinant E. coli strains. Studies of the catalytic properties of purified UGT73G1 indicated that its activity was significantly promoted by Mn2+ and Mg2+, while it was strongly inhibited by Cu2+. These expression strategies enhanced the solubility and activity of the overexpressed protein and enabled characterization of this plant-derived glucosyltransferase expressed in a prokaryotic host.