Establishment of a machine learning model for the diagnosis of clinically significant prostate cancer based on transrectal contrast-enhanced ultrasound parameters and clinical data / 中华超声影像学杂志

Objective:To establish a machine learning model for the diagnosis of clinically significant prostate cancer based on transrectal contrast-enhanced ultrasound parameters and clinically relevant data.Methods:A retrospective analysis was performed on 151 patients in Chongqing University Cancer Hospital who underwent transrectal contrast-enhanced ultrasonography and transrectal ultrasound-guided needle biopsy from November 2018 to September 2021. The time intensity curve was drawn using VueBox software and 12 parameters such as rise time, peak time, average transit time, peak intensity, and rising slope were quantitatively analyzed. Age, total prostate-specific antigen, free prostate-specific antigen, free prostate-specific antigen ratio, volume, prostate-specific antigen density, and transrectal contrast-enhanced ultrasonography parameters, a total of 18 characteristic parameters, were analyzed and screened through relevant attribute values and information gain attribute values. The screening features were trained and tested by the machine learning single algorithm and integrated algorithm, and then the model was evaluated by the F1 value and the area under the ROC curve(AUC).Results:Using the related attribute value and the information gain attribute value, 12 variables and 5 variables were screened out respectively to establish a machine learning model. The model established by the ensemble algorithm was better than the single algorithm. For the two variable selection methods, the AUC (0.810 vs 0.789) and F1 values (0.748 vs 0.742) of the Bagging ensemble algorithm model, which basic algorithm was decision tree, were the highest, followed by Logistic regression and support vector machine(SVM) in order of AUC and F1 values.Conclusions:Based on transrectal contrast-enhanced ultrasound parameters and clinical data, the Bagging ensemble model based on decision tree has the best performance in diagnosing clinically significant prostate cancer.

Risk factors and predictive effectiveness of PI-RADS after targeted combined system puncture for CsPCa patients with 3 points / 现代泌尿外科杂志

【Objective】 To investigate the risk factors and predictive effectiveness of prostate imaging reporting and data system (PI-RADS) score for patients with clinically significant prostate cancer (CsPCa) whose PI-RADS score was 3, so as to provide evidence for the diagnosis and treatment. 【Methods】 The clinical and multi-parameter magnetic resonance imaging (mpMRI) data of 153 CsPCa patients treated during Jan.2017 and Dec.2021 whose PI-RADS score was 3 were retrospectively analyzed. With PI-RADS score of 3 as the independent risk factor for CsPCa, the other relevant independent risk factors in predicting CsPCa were evaluated. 【Results】 Univariate and multivariate analyses showed that prostate-specific antigen (PSA) density and apparent dispersion coefficient (ADC) were independent risk factors for the diagnosis of CsPCa (P<0.05). Analysis of receiver operating characteristic (ROC) curve showed that combined PSA density and ADC were more effective than PSA density and ADC alone (P<0.05). 【Conclusion】 The combination of PSA density and ADC can guide clinicians to identify high-risk CsPCa patients from patients with PI-RADS score of 3 points.

Densidad de APE en pacientes PI-RADS 3: un parámetro clínico útil para su manejo / PSA density as a new approach for PIRADS 3 patient's management

Resumen: Objetivo: Analizar las biopsias realizadas en paciente categorizados PIRADS 3 en nuestra institución desde el segundo semestre del año 2016 al primer semestre del año 2018 y describir la correlación de la densidad de PSA con la incidencia de cáncer de próstata. Evaluar el rol de la densidad de PSA en la indicación de estudio histológico en pacientes PIRADS 3. Método: Trabajo autorizado por el comité de ética de nuestra institución. Se realizó búsqueda en el PACs, de todos los informes de RM multiparamétricas de próstata que incluyeran la categoría ¨PIRADS 3¨ en el periodo señalado. De ellos se calculó la densidad de PSA, con el último valor de PSA registrado en la ficha clínica previo a RM y volumen prostático en RM. Se procedió a buscar los pacientes con estudio histológico. Se correlacionó los resultados de biopsias con el valor de densidad de PSA. Realizamos análisis uni y multivariados, análisis estadísticos con sensibilidad, especificidad y uso de curva ROC. Resultados: De las 2416 RMmp de próstata realizadas en nuestra institución en las fechas ya descritas, se encontraron 424 informes catalogados con score PIRADS 3, y 267 de esos pacientes tenían estudio y seguimiento institucional, de los cuales 134 contaban con biopsia. La muestra tenía un promedio de edad de 60 años, y una mediana de densidad de PSA de 0,10 (RIC 0,07-0,14). Se encontraron 36 biopsias con cáncer clínicamente significativo (Gleason > 6), lo que corresponde a 26,8% de la muestra, valor similar al encontrado en la literuatua. En estos pacientes se obtuvo un punto de corte óptimo de densidad de PSA de 0,11, con una sensibilidad y especificidad de 67% y un AUC de 0,68. Una densidad de PSA de 0,11 presenta un OR de 4,1, con una probabilidad de 4 veces más de encontrar un cáncer de próstata por sobre este valor (IC 95% 1,3-9,8), lo cuál es estadísticamente significativo con un p igual a 0,01. Conclusión: La DAPE sobre 0,11 ng/ml/cc puede considerarse como una herramienta adicional para indicar biopsia en pacientes con RMmp PI-RADS 3, aumentando la precisión para la detección de cáncer de próstata clínicamente significativos ayudando a disminuir estudios histológicos innecesarios.

Abstract: Objective: To analyze the biopsies performed in patients categorized PIRADS 3 in our institution from the second half of 2016 to the first half of 2018 and describe the correlation of PSA density with the incidence of prostate cancer. To evaluate the role of PSA density in the indication of histological study in PIRADS 3 patients. Method: Work authorized by the ethics committee of our institution. The PACs were searched for all multiparameter prostate MRI reports that included the category "PIRADS 3" in the period indicated. The PSA density was calculated, with the last PSA value recorded in the clinical record before MRI and prostate volume in MRI. We proceeded to look for patients with the histological study. The biopsy results were correlated with the PSA density value. We perform uni and multivariate analyzes, statistical analyzes with sensitivity, specificity and use of the ROC curve. Results: Of the 2416 RMmp of the prostate performed in our institution on the dates already described, 424 reports catalogued with PIRADS 3 score were found, and 267 of those patients had study and institutional follow-up, of which 134 had a biopsy. The sample had an average age of 60 years and a median PSA density of 0.10 (RIC 0.075-0.146). We found 36 biopsies with clinically significant cancer (Gleason> 6), which corresponds to 26.8% of the sample, a value similar to that found in the literature. In these patients, an optimal cut-off point of PSA density of 0.11 was obtained, with a sensitivity and specificity of 67% and an AUC of 0.68. A PSA density of 0.11 has an OR of 4.1, with a 4-fold probability of finding prostate cancer above this value (95% CI 1.3-9.8), which It is statistically significant with a p equal to 0.01. Conclusion: DAPE over 0.11 ng/ml/cc can be considered as an additional tool to indicate biopsy in patients with RMmp PI-RADS 3, increasing the accuracy for the detection of clinically significant prostate cancer helping to reduce unnecessary histological studies.