Multi-modality deep learning model reaches high prediction accuracy in the diagnosis of ovarian cancer.

We evaluated the diagnostic performance of a multimodal deep-learning (DL) model for ovarian mass differential diagnosis. This single-center retrospective study included 1,054 ultrasound (US)-detected ovarian tumors (699 benign and 355 malignant). Patients were randomly divided into training (n = 675), validation (n = 169), and testing (n = 210) sets. The model was developed using ResNet-50. Three DL-based models were proposed for benign-malignant classification of these lesions: single-modality model that only utilized US images; dual-modality model that used US images and menopausal status as inputs; and multi-modality model that integrated US images, menopausal status, and serum indicators. After 5-fold cross-validation, 210 lesions were tested. We evaluated the three models using the area under the curve (AUC), accuracy, sensitivity, and specificity. The multimodal model outperformed the single- and dual-modality models with 93.80% accuracy and 0.983 AUC. The Multimodal ResNet-50 DL model outperformed the single- and dual-modality models in identifying benign and malignant ovarian tumors.

Evaluation of left ventricular myocardial work in patients with hyperthyroidism with different heart rates with noninvasive pressure-strain loop based on two-dimensional speck tracking imaging.

Background: We investigated the application value of no-invasive myocardial work in evaluating left ventricular (LV) function in patients with hyperthyroidism. Methods: Sixty-five patients with an initial hyperthyroidism diagnosis were sorted into tachycardia (group TH1, n=31) and without tachycardia (group TH2, n=34) groups. Thirty healthy participants served as the control group (group CON). LV strain parameters and LV myocardial work parameters were evaluated at rest. Each parameter's value in identifying myocardial damage was analyzed using receiver operating characteristic curves. The correlation of myocardial work parameters with global longitudinal strain (GLS), longitudinal peak strain dispersion (normalized by heart rate, PSDN), and systolic blood pressure (SBP) was analyzed. Results: There was no difference in classic echocardiographic parameters between the groups. Compared with that in group CON, GLS decreased in groups TH1 and TH2 (TH1 17.99%±2.21% and TH2: 19.00%±2.85% vs. 20.27%±1.49%; both P<0.05); there was no significant difference between groups TH1 and TH2. PSDN increased in groups TH1 and TH2 (TH1 73.13±19.51 ms and TH2 55.06±17.03 vs. 44.13±8.65 ms; both P<0.05); it was higher in group TH1 than in group TH2 (P<0.05). Myocardial global work efficiency (GWE) decreased in groups TH1 and TH2 {TH1 95% [interquartile range (IQR), 94-95%] and TH2 96% (IQR, 95-97%) vs. 97% (IQR, 96-97%); both P<0.05}; it was lower in group TH1 than in group TH2 (P<0.05). Global constructive work (GCW) decreased in group TH1 (1,865.29±284.13 vs. 2,030.33±252.52 mmHg%; P<0.05), but was not different from that in group TH2; there was no difference between groups TH2 and CON. Global wasted work (GWW) increased in groups TH1 and TH2 [TH1 83.00 (IQR, 74.00-97.00) mmHg% and TH2 69.50 (IQR, 51.25-84.25) vs. 50.50 (IQR, 40.75-65.25) mmHg%; both P<0.05]; it was higher in group TH1 than in group TH2 (P<0.05). The area under the GWE curve was the largest (area under the curve =0.835), and the optimal cutoff point was 96.5%, with a sensitivity of 0.83 and a specificity of 0.70. GWE and GCW were positively correlated with GLS and negatively correlated with PSDN. GWW was negatively correlated with GLS and positively correlated with PSDN. In group CON, GCW and GWW were positively correlated with SBP; GWE was not correlated with SBP. In groups TH1 and TH2, GCW was positively correlated with SBP, but not with GWW or GWE. Conclusions: Hyperthyroidism can significantly decrease the GWE and increase GWW of the left ventricle. This change is more pronounced in patients with tachycardia. Myocardial work could be a novel method for the evaluation of LV myocardial function in patients with hyperthyroidism.

Prediction of Surgical Approach in Mitral Valve Disease by XGBoost Algorithm Based on Echocardiographic Features.

In this study, we aimed to develop a prediction model to assist surgeons in choosing an appropriate surgical approach for mitral valve disease patients. We retrospectively analyzed a total of 143 patients who underwent surgery for mitral valve disease. The XGBoost algorithm was used to establish a predictive model to decide a surgical approach (mitral valve repair or replacement) based on the echocardiographic features of the mitral valve apparatus, such as leaflets, the annulus, and sub-valvular structures. The results showed that the accuracy of the predictive model was 81.09% in predicting the appropriate surgical approach based on the patient's preoperative echocardiography. The result of the predictive model was superior to the traditional complexity score (81.09% vs. 75%). Additionally, the predictive model showed that the three main factors affecting the choice of surgical approach were leaflet restriction, calcification of the leaflet, and perforation or cleft of the leaflet. We developed a novel predictive model using the XGBoost algorithm based on echocardiographic features to assist surgeons in choosing an appropriate surgical approach for patients with mitral valve disease.

Development of PI3Kα inhibitors for tumor therapy.

The PI3K/AKT/mTOR signaling pathway is well known to be involved in cell growth, proliferation, metabolism and other cellular physiological processes. Abnormal activation of this pathway is closely related to tumorigenesis and metastasis. As the starting node of the pathway, PI3K is known to contain 4 isoforms, including PI3Kα, a heterodimer composed of the catalytic subunit p110α and the regulatory subunit p85. PIK3CA, which encodes p110α, is frequently mutated in cancer, especially breast cancer. Abnormal activation of PI3Kα promotes cancer cell proliferation, migration, invasion, and angiogenesis; therefore, PI3Kα has become a key target for the development of anticancer drugs. The hinge region and the region of the mutation site in the PI3Kα protein are important for designing PI3Kα-specific inhibitors. As the group shared by the most PI3Kα-specific inhibitors reported thus far, carboxamide can produce hydrogen bonds with Gln859 and Ser854. Gln859 is specific to the p110α protein in producing hydrogen bond interactions with PI3Kα-specific inhibitors and this is a key point for designing PI3Kα inhibitors. To date, alpelisib is the only PI3Kα inhibitor approved for the treatment of breast cancer. Several other PI3Kα inhibitors are under evaluation in clinical trials. In this review, we briefly describe PI3Kα and its role in tumorigenesis, summarize the clinical trial results of some PI3Kα inhibitors as well as the synthetic routes of alpelisib, and finally give our proposal for the development of novel PI3Kα inhibitors for tumor therapy. HighlightsWe summarize the progress of PI3Kα and PI3Kα inhibitors in cancer from the second half of the 20th century to the present.We describe the clinical trial results of PI3Kα inhibitors as well as the synthetic routes of the only approved PI3Kα inhibitor alpelisib.Crystal structure of alpelisib bound to the PI3Kα receptor binding domain.This review gives proposal for the development of novel PI3Kα inhibitors and will serve as a complementary summary to other reviews in the research field of PI3K inhibitors.Communicated by Ramaswamy H. Sarma.

Discovery of Novel PI3Kδ Inhibitors Based on the p110δ Crystal Structure.

PI3Kδ is a key mediator of B-cell receptor signaling and plays an important role in the pathogenesis of certain hematological malignancies, such as chronic lymphocytic leukemia. Idelalisib, which targets PI3Kδ specifically, is the first approved PI3K inhibitor for cancer therapy. Recently, we carried out virtual screening, cell-based assays, adapta kinase assays, and molecular dynamic analysis to discover novel PI3Kδ inhibitors and identified NSC348884 as a lead PI3Kδ inhibitor. NSC348884 had an excellent docking score, potent PI3Kδ-inhibitory activity, antitumor effects on various cancer cell lines, and a favorable binding mode with the active site of PI3Kδ. Moreover, through the structural modification of NSC348884, we further discovered comp#1, which forms H-bonds with both Val828 and Lys779 in the ATP binding pocket of PI3Kδ, with a more favorable conformation binding to PI3Kδ. In addition, we found that N1, N1, N2-trimethyl-N2-((6-methyl-1H-benzo[d]imidazol-2-yl) methyl) ethane-1,2-diamine might be a potential scaffold structure. Thus, the result of this study provides a far more efficient approach for discovering novel inhibitors targeting PI3Kδ.

Effect of Lemon Essential Oil Microemulsion on the Cariogenic Virulence Factor of Streptococcus mutans via the Glycolytic Pathway.

PURPOSE: To investigate the effects and mechanisms of lemon essential oil products on dental caries prevention. MATERIALS AND METHODS: Lemon essential oil microemulsions (LEOM) with concentrations of 1/8 minimum inhibitory concentration (MIC), 1/4 MIC, and 1/2 MIC were applied to S. mutans at concentrations of 0.2%, 1%, and 5% glucose, respectively. Changes in acid production capacity of S. mutans were measured based on changes in pH. The effect of the reductive coenzyme I oxidation method on LDH activity was examined. The effect of lemon essential oil microemulsion on the expression of the lactate dehydrogenase gene (ldh) was detected by a quantitative real-time polymerase chain reaction. RESULTS: Lemon essential oil microemulsion at 1/2 MIC concentration reduced the environmental pH value at different glucose concentrations, compared to those observed in the control group (p < 0.05). LDH activity of S. mutans was decreased at three subinhibitory concentrations of lemon essential oil microemulsions (p < 0.05). The effect of lemon essential oil microemulsions on S. mutans LDH activity and bacterial acid production were positively correlated (r = 0.825, p < 0.05). Lemon essential oil microemulsion at 1/2 MIC concentration downregulated the expression of the ldh gene of S. mutans at different glucose concentrations (p < 0.05). In different glucose environments, lemon essential oil microemulsions at subminimum inhibitory concentrations can inhibit the acid production of S. mutans by reducing ldh expression and LDH activity in the glycolytic pathway, proving its anti-caries potential. CONCLUSIONS: LEOM can effectively prevent dental caries and maintain the microecological balance of the oral environment.

Growth Inhibitory and Pro-Apoptotic Effects of Hirsuteine in Chronic Myeloid Leukemia Cells through Targeting Sphingosine Kinase 1.

Chronic myeloid leukemia (CML) is a slowly progressing hematopoietic cell disorder. Sphingosine kinase 1 (SPHK1) plays established roles in tumor initiation, progression, and chemotherapy resistance in a wide range of cancers, including leukemia. However, small-molecule inhibitors targeting SPHK1 in CML still need to be developed. This study revealed the role of SPHK1 in CML and investigated the potential anti-leukemic activity of hirsuteine (HST), an indole alkaloid obtained from the oriental plant Uncaria rhynchophylla, in CML cells. These results suggest that SPHK1 is highly expressed in CML cells and that overexpression of SPHK1 represents poor clinical outcomes in CML patients. HST exposure led to G2/M phase arrest, cellular apoptosis, and downregulation of Cyclin B1 and CDC2 and cleavage of Caspase 3 and PARP in CML cells. HST shifted sphingolipid rheostat from sphingosine 1-phosphate (S1P) towards the ceramide coupled with a marked inhibition of SPHK1. Mechanistically, HST significantly blocked SPHK1/S1P/S1PR1 and BCR-ABL/PI3K/Akt pathways. In addition, HST can be docked with residues of SPHK1 and shifts the SPHK1 melting curve, indicating the potential protein-ligand interactions between SPHK1 and HST in both CML cells. SPHK1 overexpression impaired apoptosis and proliferation of CML cells induced by HST alone. These results suggest that HST, which may serve as a novel and specific SPHK1 inhibitor, exerts anti-leukemic activity by inhibiting the SPHK1/S1P/S1PR1 and BCR-ABL/PI3K/Akt pathways in CML cells, thus conferring HST as a promising anti-leukemic drug for CML therapy in the future.

Prospective Study About the Relationship Between CEUS of Carotid Intraplaque Neovascularization and Ischemic Stroke in TIA Patients.

Objective: To evaluate the relationship between contrast-enhanced ultrasonography (CEUS) of carotid intraplaque neovascularization and ischemic stroke in transient ischemic attack (TIA) patients. Methods: A total of 112 TIA patients were selected for the study. Routine carotid ultrasonic examination was performed for all the patients. CEUS was carried out for consecutive patients with plaque thicker than 2.5 mm in carotid bifurcation and follow-up for at least 24 months. The number of patients with incurrence of ischemic stroke or recurrence of TIA was obtained during the follow-up period. To detect the risk factors for incurrence of ischemic stroke or recurrence of TIA in 24 months, multivariate logistic regression analyses were performed for all the risk factors in all the selected patients. Results: Ninety-one patients underwent CEUS and were followed up at least 24 months. There were statistical differences between recurrent and non-recurrent groups about hypertension, diabetes, hyperlipemia, smoking history, family history of stroke, medication compliance, two-dimensional ultrasound, and CEUS (P < 0.05). The higher CEUS intensity in the carotid plaque was, the higher was the possibility of ischemic stroke or recurrent TIA. Multivariate logistic regression analysis showed that the CEUS characteristics of carotid plaque such as linear enhancement or diffuse enhancement were independent risk factors for ischemic stroke or recurrent TIA in TIA patients (P < 0.05). Conclusion: For carotid plaques, CEUS could evaluate the infusion mode, which could reflect the neovascularization in plaques. CEUS could predict the incurrence of ischemic stroke or recurrence of TIA in TIA patients, which is useful information when making a clinical decision.

Mutant monocyte chemoattractant protein-1 protein (7ND) inhibits osteoclast differentiation and reduces oral squamous carcinoma cell bone invasion.

The seven-amino acid truncated (7ND) protein is an N-terminal deletion mutant of monocyte chemoattractant protein-1 (MCP-1) and it functions as a dominant-negative inhibitor. 7ND and wild-type MCP-1 form a heterodimer, which binds to MCP-1 receptors and inhibits monocyte chemotaxis. In the present study, the 7ND protein was cloned, expressed and purified. An MTT assay revealed that the proliferation of oral squamous cell carcinoma (OSCC) SCC25 cells was not affected following 3 days of treatment with synthetic 7ND protein. Serial dilutions of the 7ND protein were tested for monocyte migration and osteoclast differentiation, and tartrate-resistant acid phosphatase staining demonstrated that significantly fewer osteoclasts were differentiated from cluster of differentiation 14+ (CD14+) monocytes using magnetic activated cell sorting. Immunofluorescence confirmed these results and significantly less F-actin staining was observed in 7ND-treated osteoclasts. Furthermore, bone invasion was examined by subcutaneously injecting SCC25 cells into the area overlaying the calvariae of nude mice. The results demonstrated that the average tumor volume of SCC25 cells with 7ND protein was similar to the average volume of tumors formed by untreated SCC25 cells. Flow cytometric analysis suggested that the CD14+ subpopulation in the bone marrow of 7ND-treated mice was reduced compared with that of untreated mice. Micro-computed tomography imaging revealed significantly less bone resorption in the calvariae injected with SCC25 cells plus the 7ND protein. Taken together, the results of the present study demonstrated the potential therapeutic value of the 7ND protein. The 7ND MCP-1 variant not only functions in vitro to inhibit osteoclast differentiation, but also reduces the progression of bone invasion by OSCC cells in vivo.

[The effect of GDNF on matrix-degrading and cell-adhesion during perineural invasion of salivary adenoid cystic carcinoma].

PURPOSE: To investigate the possible contribution of GDNF in matrix-degrading, cell-adhesion during perineural invasion (PNI) of salivary adenoid cystic carcinoma (ACC). METHODS: Totally 42 ACCs and 5 normal salivary tissues were included in the present study．Immunohistochemical staining SP method was used to detect the expression of GDNF,MMP-9,NF-κB,integrin ß1 in ACC specimens and normal salivary tissues. Statistical analysis was performed using SPSS16.0 software package. RESULTS：GDNF was strongly expressed in ACC tumor cells and nerve fibers adjacent to ACC tumor cells. NF-κB, MMP-9, integrin ß1 were positively expressed in ACC cell cytoplasm, integrin ß1 was also found in ACC cell membrane, and NF-κB in nuclei occasionally. The positive expression rate was 69.05%（29/42）,66.67%（28/42）,61.90%（26/42）, respectively. The differences between the expression of NF-κB, MMP-9, integrinß1 in PNI group and non-PNI group were significant (P=0.005,P=0.011,P=0.001, respectively). Expression of NF-κB, MMP-9, integrin ß1 was correlated to that of GDNF(r=0.443, P=0.003; r=0.401, P=0.009; r=0.535, P=0.000, respectively). Expression of MMP-9 and integrin ß1 was positively correlated to that of NF-κB(r=0.501, P=0.001; r=0.429, P=0.005). Expression of MMP-9 was correlated positively to that of integrin ß1 (r=0.381, P=0.013). CONCLUSIONS：GDNF may increase the matrix-degrading and cell-adhesion of ACC in the process of PNI. NF-κB, MMP-9 and integrin ß1 involve in ACC cells invading nerves.

[Effect of nerve growth factor on adhesion and motional abilities of tumor cells and their correlation].

OBJECTIVE: To investigate the effect on adhesion and motion capbilities of adenoid cystic carcinoma (ACC), by detecting the expression of nerve growth factor (NGF), E-cadherin (E-cad) and S100A4 and the clinical significance in ACC tissues and analyzing the relationships between them with perineural invasion (PNI). METHODS: The expression of NGF, E-cad and S100A4 in ACC was detected with the way of immunohistochemistry SP, and then analyzing the expression level of them in different pathological types and histological regions in statistical ways on the basis of their relation with clinical and pathological parameters. RESULTS: The expression of NGF and S100A4 in PNI group [88% (23/26) and 77% (20/26)], was higher than that in NPNI group (8/16 and 7/16, P < 0.05), and a positive correlation between them was identified in PNI group (r = 0.316, P < 0.05). However the E-cad expression was lower in PNI group [31% (8/26), P < 0.05]. On the other hand it suggested a negative correlation with NGF (r = 0.385, P < 0.05) as well as that with S100A4 (r = -0.612, P = 0.000). The expression level of NGF in fasciculus [83% (25/30)] has significant deviation compared with it in distant tumor tissues [47% (14/30), P < 0.05]. CONCLUSIONS: In the PNI process of ACC, NGF plays important parts but not the only factor. It can increase the expression and activity of S100A4 but decrease E-cad expression through binding with its receptor. Thus, adhesion abilities of tumor cells was weakened and motional abilities was strengthen.