Establishing a prediction model for prostate cancer bone metastasis

Abstract

We collected clinical data from 308 prostate cancer (PCa) patients to investigate the clinical characteristics and independent risk factors of bone metastasis (BM) and to establish a prediction model for BM of PCa and determine the necessity of bone scans. Univariate and multivariate analyses were performed based on age, biopsy Gleason score (BGS), clinical tumor stage (cTx), total prostate specific antigen (tPSA), free prostate specific antigen (fPSA), fPSA/tPSA, prostate volume, alkaline phosphatase (ALP), serum calcium and serum phosphorus. Moreover, 80 of the 308 PCa patients had a PI-RADS v2 score and were analysed retrospectively. The univariate analysis showed that the BGS, cTx, tPSA, fPSA, prostate volume and ALP were significant. The multivariate logistic regression analysis showed significant differences among the BGS, cTx, tPSA and ALP. Four cases should be highly suspected with BM: (i) cTl-cT2, BGS ≤7, ALP >120 U/L and tPSA >90.64 ng/ml; (ii) cTl-cT2, BGS ≥8, and ALP >120 U/L; (iii) cT3-cT4, BGS ≤7, and ALP >120 U/L; and (iv) cT3-cT4 and BGS ≥8. After the PI-RADS v2 score was included in the model, the AUC of the prediction model rose from 0.884 (95% CI: 0.813-0.996) to 0.934 (95% CI: 0.883-0.986). This model may help determine the necessity of bone scans to diagnose BM for PCa patients.

Keywords: ALP; BGS; PI-RADS v2; Prediction analysis model; bone metastasis; cTx; prostate cancer; tPSA.

Figure 1

The distribution ratio of each variable for patients with and without BM. (A) tPSA level, (B) fPSA/tPSA, (C) ALP level, (D) prostate volume, (E) clinical tumour stage, (F) Gleason score.

Figure 2

The level of each variable for patients with and without BM. (A) tPSA level, (B) fPSA/tPSA, (C) ALP level, (D) prostate volume, (E) fPSA (* p<0.05, ** p<0.01, *** p<0.001).

Figure 3

ROC curves of variables for the prediction of bone metastasis in patients with prostate cancer. (A) Comparison of the ROC curve for each single variable, (B) tPSA, AUC: 0.820 (0.772-0.867), (C) fPSA, AUC: 0.829 (0.781-0.876), (D) Gleason score, AUC: 0.713 (0.655-0.771), (E) clinical tumour stage, AUC: 0.826 (0.777-0.876), (F) ALP, AUC: 0.726 (0.665-0.787).

Figure 4

ROC curves of the multivariate regression logistic analysis models. (A) Comparison of ROC curves for models A-E. (B) Model A, AUC: 0.899 (0.863-0.935), (C) model B, AUC: 0.870 (0.829-0.911), (D) model C, AUC: 0.902 (0.868-0.937), (E) model D, AUC: 0.877 (0.838-0.916), (F) model E, AUC: 0.910 (0.878-0.942).

Figure 5

The nomogram and calibration curve developed for model E. (A) Nomogram. To estimate the risk of BM, the points for each variable were calculated by drawing a straight line from a patient's variable value to the axis labelled “Points”. The score sum is converted to a probability in the lowest axis. (B) Calibration curve. The nomogram-predicted probability is plotted on the x-axis, and the actual probability is plotted on the y-axis. Mean absolute error=0.03.

Figure 6

Appraisal of the value of the PI-RADS v2 score in predicting PCa BM. (A) i. Tc 99m MDP imaging of an 84-year-old BM-negative PCa patient. ii. Tc 99m MDP imaging of a 61-year-old BM-positive PCa patient. (B) i. The white arrows indicate a clear prostate tumour lesion in the 84-year-old BM-negative PCa patient. ii. The white arrows indicate no obvious bone tissue destruction in the 84-year-old BM-negative PCa patient. iii. The white arrows show an unambiguous prostate tumour lesion in the 61-year-old BM-positive PCa patient. iv. The white arrows show obvious bone tissue destruction in the 61-year-old BM-positive PCa patient. (C) i. The distribution ratio of the PI-RADS v2 score. ii. The average PI-RADS v2 score. (D) i. Comparison of ROC curves for models F and G, ii. Model F, AUC: 0.884 (0.813-0.996), iii. Model G, AUC: 0.934 (0.883-0.986).