A Risk Prediction Model For Bone Metastasis In Prostate Cancer

Objective:By analyzing the possible predictors of bone metastasis in prostate cancer and establishing a prediction model,the risk of bone metastasis of prostate cancer can be predicted early and the necessity of bone scintigraphy can be evaluated,so that low-risk patients with prostate cancer can safely omit bone scintigraphy by Single photon emission computed tomography(SPECT).Methods:Retrospectively analyze the clinical data of 298 patients with prostate cancer(PCa),perform univariate and multivariate logistic regression analysis of the age,total prostate specific antigen(tPSA),free prostate specific antigen(fPSA),fPSA/tPSA,prostatic volume(PV),alkaline phosphatase(ALP),serum calcium and phosphorus,clinical tumor stage(cTx),Gleason score(GS).And then the predictors of bone metastasis in prostate cancer were defiined and a prediction model was established.The prediction model was validated with 33 data.Results:Univariate analysis showed that there were significant differences in tPSA,fPSA,ALP,cTx and GS between the positive group and the negative group.Multivariate logistic regression analysis showed that there were significant differences in GS,cTx,tPSA and ALP between the two groups,which could be used as independent predictors of bone metastasis of prostate cancer.Among the four independent predictors,the area under the curve(AUC)of tPSA is the largest(0.783),which had the best prediction efficiency for bone metastasis of prostate cancer.When four factors were used to build the prediction model,the AUC(0.844)was larger than all the individual indicators,showing the best prediction efficiency.When the four predictors were all negative according to the optimal threshold values(tPSA < 80.77ng/ml,ALP< 155.5u /L,cTx < 3,GS < grade 4 /8),102 patients were involved,among which 6were positive for bone metastasis and 96 were negative for bone metastasis,which means the false negative rate was 5.9%(6/102),but unnecessary bone scintigraphy could be avoided for 32.2%(96/298)of patients.33 external data was used to validate The nomogram of the predictive model,the sensitivity of the prediction model was72.7%(8/11)and the specificity was 81.8%(18/22),which were basically consistent with the results of 72.9% and 80.6% when the model was established,proving that the model had high accuracy and stability.Conclusion:1.GS,cTx,tPSA and ALP were independent predictors of bone metastasis in prostate cancer.2.This model has a good prediction efficiency,which is helpful to predict the risk of bone metastasis of prostate cancer at the first diagnosis and to evaluate the necessity of bone imaging in clinical application;3.The external data prove that the prediction model has high accuracy and stability.