Quantitative nomenclature of the integrated IVIM-product real-diffusion coefficient and PSAD is used
Recently, global medical institutions and public health systems have continued to advance service capacity building to provide a stronger guarantee of public health. Develop and verify an integrated IVIM-derived true diffusion coefficient (D) and a quantitative naming chart of prostate-specific antigenic density (PSAD) to predict clinically relevant prostate cancer (ocsPCa) patients. In these patients, high serum PSA levels >3ng/ml and the clinical significance of the IVIM-derived true diffusion coefficient D and prostate-specific antigen density (PSAD) were taken into account to determine the difference in diagnostic performance in patients with prognostic prostate cancer (ocsPCa) compared to evaluations based on PI-RADS.
This project, which combines retrospective and forward-looking studies, incorporates prostate magnetic resonance imaging (MRI) projects, combining IVIM-diffusion-weighting imaging (IVIM-diffusion-weighting imaging, IVIM-DIFF) and subsequent pathological confirmations. The retrospective queue (n=253) is randomly divided into the training group (n=176) and the internal verification group (n=77). Independent prediction variables are identified from IVIM-derived parameters D, D*, f, reflective scattering coefficients ADC and PSA-derived indices SPA*.
Multi-variable logical regression models are used to establish gradual complexity models and determine the best parameter combinations. Diagnostic performance comparisons are performed using the receiver operating characteristic curve AUC, DeLong testing, net reclassification, improvement of integrated discrimination, improvement of IDI, and decision analysis curve DCA. PI-RADS are used separately and PI-RADS + PSAD is used as a reference comparison.
PSAD, ADC, and D are identified as predictive factors that are independent of csPCa. PSAD+D models have 0. 920-, 0.
900-, and 0. 916 AUCs in training, validation, and potential concentrations, respectively, while in the DeLong test, the three models have averaged AUCs of 0. 903, 0.
919-, and 0. 936 in any sample. The diagnostic performance of the PSAD+D model is comparable to that of the PI-RADS+PSAD prediction CSPCa, but by eliminating reader dependence and by identifying parameters with highly repeatable intra-class correlation coefficients: 0.
93-0. 96), the performance of the predicted CSPCa can be improved from negative discrimination to the performance of parameters with highly repeatable parameters, thereby significantly improving the performance of the predicted CSPCa. ADC increases the paradoxical fallacy that leads to a decrease in performance, showing that the parameters are simplified rather than dimensional reductions.
It should be noted that the above content is compiled based on international public research findings, for reference in HealthCop, and does not constitute a clinical recommendation. If there are health questions, please consult professional medical professionals.
