Development and validation of a machine-learning model for preoperative risk of gastric gastrointestinal stromal tumors.

BACKGROUND: Gastrointestinal stromal tumors (GISTs) have malignant potential, and treatment varies according to risk. However, no specific protocols exist to preoperatively assess the malignant potential of gastric stromal tumors (gGISTs). This study aimed to use machine learning (ML) to develop and validate clinically relevant preoperative models to predict the malignant potential of gGISTs. METHODS: We screened patients diagnosed with gGISTs at the Affiliated Hospital of North Sichuan Medical College. We employed the Least Absolute Shrinkage and Selection Operator (LASSO) and logistic regression to identify risk factors. Subsequently, an ensemble of ML models was deployed to determine the optimal classifier. Additionally, we harnessed SHapley Additive exPlanations (SHAP) for tailored risk profiling. RESULTS: We enrolled 318 patients with gGISTs. Utilizing LASSO regression and multifactorial logistic regression, we analyzed the training dataset, revealing that the presence of endoscopic ultrasound (EUS) high-risk features, tumor border clarity, tumor diameter, and monocyte-to-lymphocyte ratio (MLR) were significant predictors of high malignancy risk in gGIST. As determined by our ML approach, the logistic classification model demonstrated optimal performance, with an area under the receiver operating characteristic curve of 0.919 and 0.925 for the training and test sets, respectively. Furthermore, decision curve analysis substantiated the clinical relevance of the model. CONCLUSION: High-risk EUS features, ill-defined tumor margins, larger tumor diameters, and elevated MLR independently predicted heightened malignant potential in gGIST. We developed logistic regression models based on these factors, which were further interpreted using the SHAP methodology. This analytical approach facilitated personalized therapeutic decision-making for diverse patient populations.

A mesothelin microsensor based on an embedded thionine electronic medium within an imprinted polymer on an acupuncture needle electrode.

An electrochemical microsensor for mesothelin (MSLN) based on an acupuncture needle (AN) was constructed in this work. To prepare the microsensor, MSLN was self-assembled on 4-mercaptophenylboronic acid (4-MPBA) by an interaction force between the external cis-diol and phenylboronic acid. This was followed by the gradual electropolymerization of thionine (TH) and eriochrome black T (EBT) around the anchored protein. The thickness of the surface imprinted layers influenced the sensing performance and needed to be smaller than the height of the anchored protein. The polymerized EBT was not electrically active, but the polymerized TH provided a significant electrochemical signal. Therefore, electron transfer smoothly proceeded through the eluted nanocavities. The imprinted nanocavities were highly selective toward MSLN, and the rebinding of insulating proteins reduced the electrochemical signal of the embedded pTH. The functionalized interface was characterized by SEM and electrochemical methods, and the preparation conditions were studied. After optimization, the sensor showed a linear response in the range of 0.1 to 1000 ng mL-1 with a detection limit of 10 pg mL-1, indicating good performance compared with other reported methods. This microsensor also showed high sensitivity and stability, which can be attributed to the fine complementation of the imprinted organic nanocavities. The sensitivity of this sensor was related to the nanocavities used for electron transport around the AuNPs. In the future, microsensors that can directly provide electrochemical signals are expected to play important roles especially on AN matrices.

Machine learning models predict lymph node metastasis in patients with stage T1-T2 esophageal squamous cell carcinoma.

Background: For patients with stage T1-T2 esophageal squamous cell carcinoma (ESCC), accurately predicting lymph node metastasis (LNM) remains challenging. We aimed to investigate the performance of machine learning (ML) models for predicting LNM in patients with stage T1-T2 ESCC. Methods: Patients with T1-T2 ESCC at three centers between January 2014 and December 2019 were included in this retrospective study and divided into training and external test sets. All patients underwent esophagectomy and were pathologically examined to determine the LNM status. Thirty-six ML models were developed using six modeling algorithms and six feature selection techniques. The optimal model was determined by the bootstrap method. An external test set was used to further assess the model's generalizability and effectiveness. To evaluate prediction performance, the area under the receiver operating characteristic curve (AUC) was applied. Results: Of the 1097 included patients, 294 (26.8%) had LNM. The ML models based on clinical features showed good predictive performance for LNM status, with a median bootstrapped AUC of 0.659 (range: 0.592, 0.715). The optimal model using the naive Bayes algorithm with feature selection by determination coefficient had the highest AUC of 0.715 (95% CI: 0.671, 0.763). In the external test set, the optimal ML model achieved an AUC of 0.752 (95% CI: 0.674, 0.829), which was superior to that of T stage (0.624, 95% CI: 0.547, 0.701). Conclusions: ML models provide good LNM prediction value for stage T1-T2 ESCC patients, and the naive Bayes algorithm with feature selection by determination coefficient performed best.

First Organic-Inorganic Hybrid Compounds Formed by Ge-V-O Clusters and Transition Metal Complexes of Aromatic Organic Ligands.

Three compounds based on Ge-V-O clusters were hydrothermally synthesized and characterized by IR, UV-Vis, XRD, ESR, elemental analysis and X-ray crystal structural analysis. Both [Cd(phen)(en)]2[Cd2(phen)2V12O40Ge8(OH)8(H2O)]∙12.5H2O (1) and [Cd(DETA)]2[Cd(DETA)2]0.5[Cd2(phen)2V12O41Ge8(OH)7(0.5H2O)]∙7.5H2O (2) (1,10-phen = 1,10-phenanthroline, en = ethylenediamine, DETA = diethylenetriamine) are the first Ge-V-O cluster compounds containing aromatic organic ligands. Compound 1 is the first dimer of Ge-V-O clusters, which is linked by a double bridge of two [Cd(phen)(en)]2+. Compound 2 exhibits an unprecedented 1-D chain structure formed by Ge-V-O clusters and [Cd2(DETA)2]4+ transition metal complexes (TMCs). [Cd(en)3]{[Cd(η2-en)2]3[Cd(η2-en)(η2-µ2-en)(η2-en)Cd][Ge6V15O48(H2O)]}∙5.5H2O (3) is a novel 3-D structure which is constructed from [Ge6V15O48(H2O)]12- and four different types of TMCs. We also synthesized [Zn2(enMe)3][Zn(enMe)]2[Zn(enMe)2(H2O)]2[Ge6V15O48(H2O)]∙3H2O (4) and [Cd(en)2]2{H8[Cd(en)]2Ge8V12O48(H2O)}∙6H2O (5) (enMe = 1,2-propanediamine), which have been reported previously. In addition, the catalytic properties of these five compounds for styrene epoxidation have been assessed.

Development, validation and comparison of four methods for quantifying endogenous 25OH-D3 in human plasma.

To meet the increasing clinical needs for 25-hydroxyvitamin D3 (25OH-D3) detection, the development of an efficient and accurate high-performance liquid chromatography-mass spectrometry (HPLC-MS) method for plasma 25OH-D3 quantitation is important. Since 25OH-D3 is an endogenous compound, the lack of a plasma blank increases the difficulty of accurately quantifying 25OH-D3. Selection of a method suitable for clinical monitoring among various methods for endogenous compound quantification is necessary. Methyl tert butyl ether was chosen for the sample treatment in a liquid-liquid extraction protocol. Water as a blank matrix, 5% human serum albumin in water as a blank matrix, surrogate analyte and background subtraction were designed to address the problem of a deficiency of a plasma blank. Four liquid chromatography-tandem mass spectrometry methods were fully validated to verify the advantages and limitations owing to regulatory deficiencies for endogenous compound validation. All four methods met the criteria and could be used to monitor clinical samples. Overall 30 human plasma samples were quantified in parallel using the four methods. The difference between any two methods was <12.6% and the total relative standard deviation was <5.2%. Background subtraction and 5% human serum albumin in water as a blank matrix may be better choices considering data quality, matrix similarity, cost and practicality.

An unprecedented antimonato-polyoxovanadate (SbPOV) based on both α-[V14Sb8O42]4- and ß-[V14Sb8O42]4- isomers.

An unprecedented antimonato-polyoxovanadate (SbPOV) based on both α-[V14Sb8O42]4- and ß-[V14Sb8O42]4- isomers has been hydrothermally synthesized.

Tuning the structures based on polyoxometalates from 1-D to 2-D by using different secondary organic ligands.

Six new organic-inorganic hybrid compounds based on [XM12O40](4-) (X = heteroatom, M = metal atom), namely [Cu(pic)2][H2XM12O40]·2Hapy·2apy (X = Si, M = W for , X = Ge, M = W for and X = Si, M = Mo for ), [Cu(2,2'-bpy)2][Cu(2,2'-bpy)(H2O)][Cu(pic)2]0.5[XM12O40]·nH2O (X = Si, M = Mo, n = 0.5 for , X = Ge, M = W, n = 1 for ) and [Cu(phen)(H2O)]2[Cu(pic)2][GeW12O40]·2.5H2O () (pic = deprotonated picolinic acid, apy = 2-aminopyridine, 2,2'-bpy = 2,2'-bipyridine, phen = phenanthroline), have been synthesized and characterized by IR, UV-Vis, XRD, cyclic voltammetric measurements and single crystal X-ray diffraction analysis. Single crystal X-ray analysis reveals that compounds are isomorphous and isostructural, in which each is based on [H2XM12O40](2-) and [Cu(pic)2]. Compounds and are also isomorphous and isostructural, of which the structures are more interesting than those of compounds . Both structures are constructed from [XM12O40](4-) and metal mixed-organic-ligand complexes. Compound is also constructed from Keggin ions and metal mixed-organic-ligand complexes, which are, however, thoroughly different from those of compounds and . The photodegradation properties of compounds have been analyzed. Compounds also exhibit rapid absorption properties for RhB (Rhodamine B). Detailed analysis of the photodegradation properties of compounds reveals that the molybdate POM has stronger degradation ability for RhB than the tungstate one.

Polyoxometalate-based organic-inorganic hybrid compounds containing transition metal mixed-organic-ligand complexes of N-containing and pyridinecarboxylate ligands.

Five new organic­inorganic hybrid compounds based on the Keggin-type polyoxoanion [SiW12O40]4−, namely [Cu3(2,2'-bpy)3(inic)(µ2-OH)(H2O)][SiW12O40]·2H2O (1), [Cu6(phen)6(µ3-Cl)2(µ2-Cl)2Cl2(inic)2][SiW12O40]·6H2O (2), [Cu2(hnic)(2,2'-bpy)2Cl]2[H2SiW12O40] (3), [Cu2(nic)(phen)2Cl2]2[SiW12O40] (4) and [Cu2(pic)(2,2'-bpy)2Cl]2[SiW12O40] (5) (inic = isonicotinic acid, hnic = 2-hydroxy-nicotinic acid, nic = nicotinic acid, pic = picolinic acid, 2,2'-bpy = 2,2'-bipyridine, phen = 1,10-phenanthroline) have been synthesized and characterized by IR, UV-Vis, XPS, XRD, cyclic voltammetric measurements, photoluminescence analysis and single crystal X-ray diffraction analysis. Crystal analysis reveals that compound 1 exhibits a 2-D double layered framework structure constructed from [SiW12O40]4− and copper-aqua-2,2'-bipy-hydroxyl-isonicotinate complexes. Compound 2 is a 0-D discrete structure formed by [SiW12O40]4− and copper-chloro-isonicotinate-phenanthroline complexes. Compound 3 shows a 1-D single chain structure based on the linkage of copper-2,2-bpy-chloro-2-hydroxy-nicotinate complexes and [SiW12O40]4−. Compounds 4 and 5 both contain polyoxometalate supported transition metal complexes, one is a polyoxometalate supported copper-chloro-nicotinate-phenanthroline complex in 4, and the other is a polyoxometalate supported copper-2,2-bpy-chloro-nicotinate complex in 5. It should be noted that nicotinic, isonicotinic and picolinic acids are structural isomers and 2-hydroxy-nicotinic acid is an in situ hydroxylated product of nicotinic acid. In addition, photocatalytic degradation of Rhodamine B (RhB) by compounds 1­5 has been investigated in aqueous solutions.