Table 5 Results of the ROC analysis.

From: The grade of individual prostate cancer lesions predicted by magnetic resonance imaging and positron emission tomography

Modalities

AUC

95% CI

Youden thresholds

Sensitivity (%)

Specificity (%)

\(\widetilde{{{\mbox{T}}}2}\)

0.48 (0.49)

38–58 (41–57)

0.90 (0.90)

78 (72)

35 (35)

\(\widetilde{{{\mbox{ADC}}}}\)

0.60 (0.58)

50–71 (49–66)

0.85 (0.85)

67 (62)

54 (55)

ACE

0.61 (0.59)

51–71 (51–67)

3.4 (4.3)

67 (44)

57 (75)

\(\widetilde{{{\mbox{K}}}}\)trans

0.68 (0.70)

58–78 (63–78)

4.7 (4.7)

53 (51)

82 (84)

PSMA

0.72 (0.79)

63–82 (72–86)

6.5 (6.5)

63 (68)

82 (87)

MRI2

0.66 (0.61)

57–76 (53–69)

–

–

–

MRI3

0.70 (0.70)

61–80 (63–78)

–

–

–

PSMA + MRI3

0.79 (0.82)

71–87 (76–88)

–

–

–

  1. Figures in parentheses refer to IGG ≥ 3 vs. IGG ≤ 2, as opposed to IGG 3 vs. IGG 2.
  2. PSMA, ACE, \(\widetilde{{{\mbox{K}}}}\)trans, \(\widetilde{{{\mbox{ADC}}}}\) and \(\widetilde{{{\mbox{T}}}2}\) as explained in Table 3; MRI2 = fitting a logistic regression model combining \(\widetilde{{{\mbox{ADC}}}}\) and \(\widetilde{{{\mbox{T}}}2}\) as individual variables; MRI3 = combining \(\widetilde{{{\mbox{K}}}}\)trans, \(\widetilde{{{\mbox{ADC}}}}\) and \(\widetilde{{{\mbox{T}}}2}\) as individual variables; PSMA + MRI3 = combining PSMA and the components of MRI3; AUC = area under the receiver operating characteristic (ROC) curve; CI = confidence interval for the AUCs; n = 194 independent lesions for IGG ≥ 3 vs. IGG ≤ 2 and n = 123 for IGG 3 vs. IGG 2; IGG International Society of Urological Pathology grade group; Youden thresholds = cut-off values resulting in the largest Youden index (sensitivity + specificity − 1). The reported sensitivities and specificities correspond to the Youden thresholds. See Supplementary Table 1 for the complete list of combinations.
Back to article page