Table 2 Summary demographic characteristics of patients included in the studies selected for narrative synthesis
Study | Year | Country | No. of patients | Age, years | PSA, ng/mL | Patient population | Bx | MRI vs Bx | Time MRI to Bx | No. of centres /vendors | No. of readers | Reader experience, years |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
Studies using deep learning-based fully-automated AI methods | ||||||||||||
Wang [20] | 2020 | Netherlands | 346 | 66 (48–83) | 13 (1–56) | Clinically suspected | TB | Pre-Bx | NR | 1/1 | 1 | 20 |
Fernandez-Quilez [21] | 2021 | Netherlands | 200 | 66 (48–83) | 13 (1–56) | Clinically suspected | TB | Pre-Bx | NR | 1/1 | 4 | NR |
Schelb [22] | 2019 | Germany | 312 | Training: 64 [58–71] Test: 64 [60–69] | Training: 7.0 [5.0–10.2] Test: 6.9 [5.1–8.9] | Clinically suspected | TB | Pre-Bx | NR | 1/1 | 2 | 0.5, 10 |
Deniffel [23] | 2020 | Canada | 499 | Training: 63.8 ± 8.1 Test: 64.4 ± 8.4 | Training: 7.6 [5.0–10.8]a Test: 7.2 [5.2–11.2] | Clinically suspected | TB | Pre-Bx | NR | 1/1 | 2 | 15, 3 |
Seetharaman [24] | 2021 | USA | 424 | Training: 63.8 (49–76) Test: 65 (38–82) | Training: 6.8 (3.3–28.6) Test: 7.1 (0.9–63.0) | Clinically suspected | RP or TB | Pre-Bx or Pre-Op | NR | 1/1 | Unclear | Unclear |
Studies using traditional machine learning-based semi-automated AI methods | ||||||||||||
Bonekamp [25] | 2018 | Germany | 316 | 64 [58–71] | Training: 6.6 [4.9–9.5] Test: 7.5 [5.4–11.0] | Clinically suspected | TB | Pre-Bx | NR | 1/1 | 2 | 0.5, 8 |
Min [26] | 2019 | China | 280 | Training,csPCa: 68.8 ± 8.3Training, iPCa: 71.5 ± 8.4 Test, csPCa: 70.3 ± 7.8 Test, iPCa: 71.6 ± 5.7 | NRb | Clinically suspected | TB | Pre-Bx | NR | 1/1 | 2 | NR, 20 |
Kwon [27] | 2018 | Netherlands | 344 | 66 (48–83) | 13 (1–56) | Clinically suspected | TB | Pre-Bx | NR | 1/1 | 2 |  > 25 |
Castillo [28] | 2021 | Netherlands | 107 | C1: 64 ± 7 C2: N/A C3: N/A | C1: 12 ± 10 C2: 9 ± 5 C3: 10 ± 8 | Clinically suspected | RP | Pre-Op | NR | 3/3 | 1 | NR |
Bleker [29] | 2019 | Netherlands | 206 | 66 (48–83) | 13 (1–56) | Clinically suspected | TB | Pre-Bx | NR | 1/1 | Unclear | Unclear |
Li [30] | 2020 | China | 381 | csPCa:75 [68–81] iPCa: 69 [63–75] | csPCa:49.3 [21.1–83.4 iPCa:9.9 [6.7–15.9] | Clinically suspected | TB | Pre-Bx | NR | 1/1 | 2 | 3, 9 |
Woźnicki [31] | 2020 | Germany | 191 | Training: 68 [63–74] Test: 69 [63–72] | Training:7.6 [5.7–11.0] Test: 8.2 [6.8–11.9] | Clinically suspected | TB | Pre-Bx | Bx 3 months before MRI | 1/2 | 2 | 7, 7 |
Bevilacqua [32] | 2021 | Italy | 76 | csPCa: 66 ± 6.8 iPCa: 65 ± 8.8 | csPCa: 7.8 ± 7.5 iPCa: 5.3 ± 3.0 | Biopsy-proven | TB | Post-Bx | Bx 6 weeks before MRI | 1/1 | 2 | 7, 25 |
Toivonen [33] | 2019 | Finland | 62 | 65 (45–73) | 9.3 (1.3–30) | Biopsy-proven | RP | Pre-Bx | NR | 1/1 | 2 | NR |
Antonelli [34] | 2019 | UK | 164 | 64 (43–83) | 7.4 (2.5–30.3) | Clinically suspected | TB | Pre-Bx | NR | 1/1 | 1 | 3 |
Yoo [35] | 2019 | Canada | 427 | NR | NR | Clinically suspected | NR | Pre-Bx | NR | 1/1 | NR | NR |
Hiremath [36] | 2021 | USA, Netherlands | 592 | C1: 65.5 (59–72) C2: 63 (59–68) C3: 62 (56–66) C4: 65.5 (62–73) | C1: 6.6 (0.25–88.2) C2: 6.7 (5–10) C3: 5.7 (4.54–9.58) C4: 7.7 (4.8–11.3) | Clinically suspected | RP or SB or TB | Pre-Bx | NR | 5/3 | 5 |  > 15,  > 15,  > 15,  > 10,  > 10 |