Diagnostic performance of prediction models for extraprostatic extension in prostate cancer: a systematic review and meta-analysis

Insights into Imaging

Table 4 Univariable meta-regression evaluating the effect of confounding factors on sensitivities and specificities of MRI-inclusive nomograms and clinical nomograms for EPE prediction

Parameter	Category	No. of cohorts	Sensitivity (95% CI)	p value	Specificity (95% CI)	p value
MRI-inclusive nomograms
EPE based on	Whole-gland	9	0.73 (0.64–0.82)	0.01	0.76 (0.68–0.85)	0.42
EPE based on	Side-specific	4	0.81 (0.71–0.92)	0.01	0.54 (0.37–0.71)	0.42
pEPE rate	≥ 0.4	5	0.74 (0.61–0.87)	0.09	0.81 (0.67–0.94)	0.82
pEPE rate	< 0.4	8	0.77 (0.68–0.86)	0.09	0.65 (0.52–0.78)	0.82
MRI time	Before biopsy	4	0.72 (0.59–0.85)	0.09	0.72 (0.49–0.95)	0.97
MRI time	After biopsy	4	0.77 (0.66–0.88)	0.09	0.68 (0.42–0.94)	0.97
Slice thickness	> 3 mm	4	0.68 (0.52–0.84)	0.07	0.70 (0.53–0.87)	0.84
Slice thickness	≤ 3 mm	3	0.82 (0.68–0.95)	0.07	0.61 (0.37–0.84)	0.84
AI/radiomics-based	Yes	6	0.69 (0.57–0.82)	0.01	0.73 (0.59–0.88)	0.60
AI/radiomics-based	No	7	0.80 (0.71–0.88)	0.01	0.69 (0.54–0.84)	0.60
Clinical nomograms
Model	Partin Table	8	0.73 (0.64–0.81)	0.26	0.60 (0.50–0.70)	< 0.001
Model	MSKCCn	5	0.68 (0.56–0.80)	0.26	0.77 (0.67–0.87)	< 0.001

EPE extraprostatic extension, pEPE pathological EPE, AI artificial intelligence, DCE dynamic contrast enhancement, MSKCCn Memorial Sloan Kettering Cancer Center nomogram