A new predictor is comparable to the updated nomogram in predicting the intermediate- and high-risk prostate cancer but outperforms nomogram in reducing the overtreatment for the low-risk Pca

Figures & data

Table 1 The clinical characteristics, demographics and pathological findings of all men

Characteristic	Training cohort (N=1,078)				Validation cohort (N=96)
	Total	N-LPca	IH-Pca	P-value	L-Pca	IH-Pca
n=640	n=438	n=19	n=77
Mean(SD)
Total score	13.24 (4.29)	10.52 (3.32)	16.13 (3.11)	<0.01	13.67 (1.53)	17.18 (2.6)
Nomogram risk	0.41 (0.32)	0.2 (0.2)	0.63 (0.28)	<0.01	0.44 (0.19)	0.72 (0.24)
Age, years	68.90 (8.28)	67.06 (8.26)	70.91 (7.84)	<0.01	72.33 (8.51)	71.9 (7.23)
tPSA level, ng/mL	30.95 (30.90)	17.24 (17.68)	45.88 (35.04)	<0.01	7.32 (2.68)	53.91 (36.23)
fPSA, ng/mL	6.59 (11.49)	3.36 (6.57)	10.1 (14.33)	<0.01	0.89 (0.17)	13.81 (16.63)
%fPSA	0.22 (1.52)	0.17 (0.13)	0.27 (2.2)	0.29	0.13 (0.05)	0.37 (1.61)
PSAD, ng/mL/mL	0.74 (0.90)	0.39 (0.5)	1.11 (1.07)	<0.01	0.22 (0.11)	1.32 (1.37)
PV, mL	53.87 (36.22)	57.79 (35.77)	49.66 (36.21)	0.003	34.86 (6.16)	47.81 (30.73)
Erythrocyte	4.46 (0.59)	4.52 (0.56)	4.39 (0.62)	0.002	4.41 (0.28)	4.36 (0.66)
Platelets	196.48 (65.02)	200.5 (68.59)	192.03 (60.6)	0.073	155.67 (28.57)	188.87 (66.01)
Neutrophil	4.06 (3.41)	4.07 (1.78)	4.04 (4.59)	0.886	3.15 (0.88)	3.9 (1.34)
Lymphocyte	1.73 (2.3)	1.67 (0.73)	1.79 (3.25)	0.483	1.54 (0.13)	1.53 (0.55)
NLR	2.88 (2.58)	2.86 (2.18)	2.91 (2.96)	0.776	2.08 (0.71)	2.89 (1.6)
BMI	23.35 (3.37)	23.48 (3.07)	23.22 (3.66)	0.283	22.43 (2.95)	22.89 (3.23)
N (%)
DRE
Normal	604 (56)	502 (83)	102 (17)	<0.01	11 (58)	8 (42)
Abnormal	323 (30)	65 (20)	258 (80)	<0.01	8 (11)	64 (89)
Unclear	151 (14)	73 (48)	78 (52)	<0.01	0	3 (100)
PI-RADS V2
1–2	224 (21)	211 (94)	13 (6)	<0.01	2 (40)	3 (60)
3	331(30)	249 (75)	82 (25)	<0.01	5 (23)	13 (72)
4	306 (29)	146 (47)	160 (63)	<0.01	9 (17)	39 (83)
5	217 (20)	34 (16)	183 (84)	<0.01	3 (12)	22 (88)

Abbreviations: N-Pca (Pca), non-Pca (prostate cancer); N-LPca, normal men and low-risk Pca; IH-Pca, intermediate- and high-risk Pca; t(f)PSA, total (free) prostate-specific antigen; %fPSA, the ratio between tPSA and fPSA; PSAD, PSA density; PV, prostate volume; NLR, neutrophil to lymphocyte ratio; BMI, body mass index; DRE, digital rectal examination; PI-RADS V2, prostate imaging reporting and data system version 2.

Table 2 Distribution of demographics and biopsy cores in different Gleason score, clinical stage and D’Amico risk

Characteristic	Number of patients (%)	Mean (SD)
Biopsy cores	Positive cores
Total	1078	12.15 (1.31)	3.39 (4.25)
N-Pca	534 (49)	12.38 (0.87)	-
Pca	544 (51)	11.92 (1.60)	6.59 (3.75)
Gleason score
≤6	107 (20)	12.40 (0.66)	3.2 (2.46)
7	160 (29)	12.01 (1.57)	5.91 (3.47)
≥8	277 (51)	11.69 (1.82)	8.26 (3.30)
Clinical stage
≤T2a	146 (27)	12.25 (1.25)	2.18 (1.57)
T2b	155 (29)	11.43 (2.22)	5.41 (1.65)
≥T2c	243 (44)	12.03 (1.21)	9.98 (2.05)
D’Amico risk
Low	28 (5)	12.25 (0.75)	2.36 (1.93)
Intermediate	113 (21)	12.50 (0.94)	4.10 (2.81)
High	403 (74)	11.73 (1.75)	7.58 (3.55)

Abbreviation: N-Pca (Pca), non-Pca (prostate cancer).

Table 3 The regression probability of PI-RADS in predicting Pca and the upper bound of 95% CI for mean under the regression probability of PI-RADS

Indicators	0 point	1 point	2 points	3 points
Unlikely (normal)	Equivocally	Likely	Highly likely
PI-RADS v2	1–2	3	4	5
Regression probability	≤0.15	0.15–0.40	0.40–0.65	>0.65
Grouping boundary
Main indicators
PSA (ng/mL)	0–4	4–10	10–26	>26
PSAD (ng/mL^2)	≤0.15	0.16–0.20	0.20–0.60	>0.60
PI-RADS*	1–2	3	4	5
Auxiliary indicators
Age (years)	<50	50–58	59–70	>70
PV (mL)	>454	232–454	42–232	<42

Note: *This variable has 3 times the point of the other variables.

Abbreviations PSA, prostate-specific antigen; PSAD, PSA density; PV, prostate volume; PI-RADS v2, prostate imaging reporting and data system version.

Figure 1 The thresholds for predicting Pca and IH-Pca and the updated nomogram. (A) The new predictor and screening thresholds. (B) The updated nomogram.

Abbreviations: Pca, prostate cancer; IH-Pca, intermediate- and high-risk Pca; PSA, prostate-specific antigen; PSAD, PSA density; PV, prostate volume; PI-RADS, prostate imaging reporting and data system.

Table 4 Performance of PSA, PI-RADS, nomogram and TS in predicting the IH-Pca

Evaluation indexes	PSA	PI-RADS	Nomogram	TS
Youden’s index	0.32	0.45	0.56	0.57
Likelihood ratio	170.97	285.36	388.13	398.75
PPV, %	59	64	70	71
OR (95%CI)	7.82 (5.50–11.12)	14.42 (9.89–21.02)	16.19 (11.76–22.30)	20.11 (14.03–28.82)
Overall accuracy,%	66	72	77	78
AUC (95%CI) for all cancer	0.75 (0.72–0.78)	0.80 (0.78–0.83)	0.87 (0.85–0.89)	0.87 (0.85–0.89)
AUC (95%CI) for GS >6	0.77 (0.74–0.80)	0.81 (0.79–0.84)	0.87 (0.85–0.89)	0.88 (0.85–0.89)
AUC (95%CI) for Cs ≥T2b	0.79 (0.76–0.82)	0.80 (0.77–0.83)	0.87 (0.84–0.89)	0.87 (0.85–0.89)

Abbreviations: IH-Pca, intermediate- and high-risk Pca; PSA, prostate-specific antigen; PSAD, PSA density; PI-RADS, prostate imaging reporting and data system; TS, total scores; PPV, positive predictive value; AUC, area under the curve; GS, Gleason score; CS, clinical stage.

Table 5 The value of NRI, OR and AUC of predictive methods in predicting the IH-Pca

Predictive methods	Compared to TS			Logistic regression		For IH-Pca
NRI %	Z value	P-value	OR (95%CI)	P-value	AUC (95%CI)	P-value (compared to TS)
PSA	24.13	4.58	<0.001	1.01 (1.00–1.02)	0.024	0.79 (0.76–0.82)	<0.001
PI-RADS	13.19	6.97	<0.001	0.98 (0.73–1.32)	0.91	0.80 (0.79–0.84)	<0.001
Nomogram	1.31	0.60	0.55	15.68 (3.43–71.73)	<0.001	0.87 (0.86–0.90)	0.89
TS	-	-	-	16.87 (4.25–67.00)	<0.001	0.88 (0.87–0.90)	-

Abbreviations: NRI, net reclassification index; TS, total score; IH-Pca, intermediate- and high-risk Pca; PSA, prostate-specific antigen; PSAD, PSA density; PI-RADS, prostate imaging reporting and data system; AUC, area under the curve.

Figure 2 The predictive performance of different methods. (A) The area under the curve (AUC) is used to compare the predictive power in predicting IH-Pca. (B) The AUC is performed to show the predictive power in predicting IH-Pca in population with radical prostatectomy. (C) The decision curve analysis is showing the predictive power in predicting IH-Pca. (D) The AUC is utilized to illustrate the performance of predicting significant Pca in cohort with radical prostatectomy. (E) Using AUC is to show the ability in forecasting the men following the criteria of PRIAS.

Abbreviations: Pca, prostate cancer; IH-Pca, intermediate- and high-risk Pca; PRIAS, Prostate Cancer Research International Active Surveillance.

Figure 3 The distribution of men with N-LPca and with IH-Pca in the different screening methods. (A) The distribution of TS for men with N-LPca or with IH-Pca in positive tests of nomogram, PSA and PI-RADS. (B) The distribution of TS groups in different screening methods for men with N-LPca.

Abbreviations: N-LPca, normal men or low-risk prostate cancer; IH-Pca, intermediate- and high-risk Pca;TS, total score; PSA, prostate-specific antigen; PI-RADS, prostate imaging reporting and data system.

Figure 4 The value and mean of total scores and nomogram. (A) The value and mean of total scores in D’Amico risks. (B) The value and mean of nomogram in D’Amico risks.

Figure 5 Calibration curves for the prediction models. (A) Total scores. (B) Nomogram. (C) PSA and (D) PI-RADS.

Abbreviations: PSA, prostate-specific antigen; PI-RADS, prostate imaging reporting and data system.