[Tree-Augmented NaÏve Bayesian network model for predicting prostate cancer].

OBJECTIVE: To evaluate the integrated performance of age, serum PSA, and transrectal ultrasound images in the prediction of prostate cancer using a Tree-Augmented NaÏve (TAN) Bayesian network model. METHODS: We collected such data as age, serum PSA, transrectal ultrasound findings, and pathological diagnoses from 941 male patients who underwent prostate biopsy from January 2008 to September 2011. Using a TAN Bayesian network model, we analyzed the data for predicting prostate cancer, and compared them with the gold standards of pathological diagnosis. RESULTS: The accuracy, sensitivity, specificity, positive prediction rate, and negative prediction rate of the TAN Bayesian network model were 85.11%, 88.37%, 83.67%, 70.37%, and 94.25%, respectively. CONCLUSIONS: Based on age, serum PSA, and transrectal ultrasound images, the TAN Bayesian network model has a high value for the prediction of prostate cancer, and can help improve the clinical screening and diagnosis of the disease.

[Diagnosis Values of Back Propagation Neural Network Integrating Age, Transrectal Ultrasound Characteristics and Serum PSA for Prostate Cancer].

OBJECTIVE: To explore the diagnosis value of back propagation (BP) neural network integrating age, transrectal ultrasound characteristics and serum prostate specific antigen (PSA) for prostate cancer. METHODS: The data of age, PSA, and transrectal ultrasound characteristics were collected from 941 patients who received color doppler transrectal ultrasound scan and systemic biopsies of prostates. A prostate cancer diagnosis system of BP neural network with age, transrectal ultrasound characteristics and serum PSA was developed in MATLAB software, and its diagnostic value for prostate cancer was analyzed based on the pathological results of prostatic biopsy. RESULTS: The biopsy results confirmed 358 cases of prostate cancer (38.04%) and 583 cases noncancerous prostate diseases (61.96%). The sensitivity, specificity, accuracy, positive value and negative predictive value of BP neural networks for prostate cancer diagnosis were 78.57%, 92.94%, 87.23%, 88.00% and 86.81% respectively. CONCLUSION: Back propagation neural network with age, transrectal ultrasound characteristics and PSA shows good diagnosis value for prostate cancer.

Geographic variations in the incidence of congenital hypothyroidism in China: a retrospective study based on 92 million newborns screened in 2013-2018.

BACKGROUND: Although congenital hypothyroidism (CH) has been widely studied in Western countries, CH incidence at different administrative levels in China during the past decade remains unknown. This study aimed to update the incidence and revealed the spatial pattern of CH incidence in the mainland of China, which could be helpful in the planning and implementation of preventative measures. METHODS: The data used in our study were derived from 245 newborns screening centers that cover 30 provinces of the Chinese Newborn Screening Information System. Spatial auto-correlation was analyzed by Global Moran I and Getis-Ord Gi statistics at the provincial level. Kriging interpolation methods were applied to estimate a further detailed spatial distribution of CH incidence at city level throughout the mainland of China, and Kulldorff space scanning statistical methods were used to identify the spatial clusters of CH cases at the city level. RESULTS: A total of 91,921,334 neonates were screened from 2013 to 2018 and 42,861 cases of primary CH were identified, yielding an incidence of 4.66 per 10,000 newborns screened (95% confidence interval [CI]: 4.62-4.71). Neonates in central (risk ratio [RR] = 0.84, 95% CI: 0.82-0.85) and western districts (RR = 0.71, 95% CI: 0.69-0.73) had lower probability of CH cases compared with the eastern region. The CH incidence indicated a moderate positive global spatial autocorrelation (Global Moran I value = 0.394, P  < 0.05), and the CH cases were significantly clustered in spatial distribution. A most likely city-cluster (log-likelihood ratio [LLR] = 588.82, RR = 2.36, P  < 0.01) and 25 secondary city-clusters of high incidence were scanned. The incidence of each province and each city in the mainland of China was estimated by kriging interpolation, revealing the most affected province and city to be Zhejiang Province and Hangzhou city, respectively. CONCLUSION: This study offers an insight into the space clustering of CH incidence at provincial and city scales. Future work on environmental factors need to focus on the effects of CH occurrence.

Prostate cancer prediction using the random forest algorithm that takes into account transrectal ultrasound findings, age, and serum levels of prostate-specific antigen.

The aim of this study is to evaluate the ability of the random forest algorithm that combines data on transrectal ultrasound findings, age, and serum levels of prostate-specific antigen to predict prostate carcinoma. Clinico-demographic data were analyzed for 941 patients with prostate diseases treated at our hospital, including age, serum prostate-specific antigen levels, transrectal ultrasound findings, and pathology diagnosis based on ultrasound-guided needle biopsy of the prostate. These data were compared between patients with and without prostate cancer using the Chi-square test, and then entered into the random forest model to predict diagnosis. Patients with and without prostate cancer differed significantly in age and serum prostate-specific antigen levels (P < 0.001), as well as in all transrectal ultrasound characteristics (P < 0.05) except uneven echo (P = 0.609). The random forest model based on age, prostate-specific antigen and ultrasound predicted prostate cancer with an accuracy of 83.10%, sensitivity of 65.64%, and specificity of 93.83%. Positive predictive value was 86.72%, and negative predictive value was 81.64%. By integrating age, prostate-specific antigen levels and transrectal ultrasound findings, the random forest algorithm shows better diagnostic performance for prostate cancer than either diagnostic indicator on its own. This algorithm may help improve diagnosis of the disease by identifying patients at high risk for biopsy.