Table 2.
Development and performance characteristics of prediction models
| Study | Predictive factors | Development methodology |
Gender | Age (Years) | Sample size | Validation Methodology |
Model presentation | |
|---|---|---|---|---|---|---|---|---|
| Modeling (Positive Events) |
Validation (Positive Events) |
|||||||
| Li 2021 [22] | 4 factors: Intravertebral vacuum cleft, posterior fascia oedema, paraspinal muscle degeneration, and bone cement distribution | Logistic regression |
Male: 53 patients Female: 215 patients |
< 60: 14 patients 60∼70 years: 100 patients 70∼80 years: 108 patients > 80 years: 46 patients |
268 (37) | – | Internal validation with 1000 bootstrap samples | Nomogram |
| Lin 2023 [23] |
4 factors: Preoperative bone mineral density, thoracolumbar fascia injury, facet joint injury, and incomplete cementing of the fracture line |
Logistic regression |
Male: 69 patients Female: 338 patients |
Training set: 74.00 (Mean) Validation set: 74.50 (Mean) |
281 (47) | 162 (22) |
Internal validation with 1000 bootstrap samples; External validation |
Nomogram |
| Liu 2022 [24] | 3 factors: fracture segment, the number of surgical vertebrae, and smoking | Logistic regression |
Male: 49 patients Female: 247 patients |
208 (64) | 88 (19) |
Internal validation (Random split Validation: 7:3) |
Nomogram | |
| Tu 2024 [25] | 5 factors: Posterior fascia oedema, intravertebral vacuum cleft, time from fracture to surgery, sarcopenia, and interspinous ligament degeneration | Logistic regression |
Male: 46 patients Female: 221 patients |
Training set: 71.5 (Mean) Validation set: 71.0 (Mean) |
186 | 81 |
Internal validation (Random split Validation: 7:3) |
Nomogram |
| Yu 2023 [26] | 5 factors: Depression, intravertebral vacuum cleft, no anti-osteoporosis treatment, cement volume, and bone cement distribution | Logistic regression |
Male: 43 patients Female: 193 patients |
Training set: ≤ 75: 157 patients > 75: 49 patients Validation set: ≤ 75: 20 patients > 75: 10 patients |
236 (30) | – | Internal validation with 1000 bootstrap samples | Nomogram |
| Cheng 2023 [27] | 4 factors: Intravertebral vacuum cleft, posterior fascia oedema, paravertebral muscle degeneration, and bone cement distribution | Logistic regression |
Male: 178 patients Female: 124 patients |
No residual pain group: 68.5 ± 6.3 (Mean ± SD) Residual pain group: 70.2 ± 7.4 (Mean ± SD) |
302 (43) | – | – | Nomogram |
| Liao 2023 [28] | 4 factors: Intravertebral vacuum cleft, thoracolumbar fascia injury, bone mineral density, and incomplete cementing of the fracture line | Logistic regression |
Male: 48 patients Female: 66 patients |
No residual pain group: 64.95 ± 4.27 (Mean ± SD) Residual pain group: 65.76 ± 4.32 (Mean ± SD) |
114 (41) | – | Internal validation with bootstrap samples | Nomogram |
| Lin 2022 [29] | 5 factors: Thoracolumbar fascia injury, bone mineral density, minor joint injury, lumbar disc herniation, and incomplete cementing of the fracture line | Logistic regression |
Male: 52 patients Female: 325 patients |
No residual pain group: 75.00 (Mean) Residual pain group: 74.00 (Mean) |
377 (64) | – | Internal validation with 200 bootstrap samples | Nomogram |
| Qiu 2023 [30] | 8 factors: Age, bone mineral density, educational level, smoking, psychological disorders, ASA grading, amount of bone cement injected, and bone cement leakage | Logistic regression |
Male: 62 patients Female: 74 patients |
No residual pain group: 63.85 ± 1.71 (Mean ± SD) Residual pain group: 66.17 ± 2.56 (Mean ± SD) |
136 (58) | – | – | Nomogram |
| Tian 2023 [31] | 7 factors: Previous history of low back injury, fracture severity, cortical rupture, percentage of bone cement vertebral body, recovery of anterior edge height of vertebral body, bone cement leakage, and bone cement distribution | Logistic regression |
Male: 94 patients Female: 134 patients |
No residual pain group: 66.17 ± 10.74 (Mean ± SD) Residual pain group: 65.45 ± 11.29 (Mean ± SD) |
228 (35) | – | Internal validation with 1000 bootstrap samples | Nomogram |
| Wu 2023 [32] | 4 factors: intervertebral foramen reduction rate (%), fracture severity, thoracolumbar intervertebral injury, and fracture type | Logistic regression | Training set: 51 male and 206 female |
Training set: 1) no residual pain group: 72.64 ± 12.38 (Mean ± SD); 2) residual pain group: 73.71 ± 11.59 (Mean ± SD) Validation set: – |
257 (48) | 85 (17) |
Internal validation (Random split Validation: 3:1) |
Nomogram |
| Ge 2022 [33] | 4 factors: Bone mineral density, intravertebral cleft, thoracolumbar fascia injury, and radiomics score | Logistic regression | – | All patients > 55 years | 548 (59) | 183 (22) |
Internal validation (Random split Validation: 3:1) |
Nomogram |
| Deng 2023 [34] | 5 factors: Diabetes, preoperative T value, number of fractured vertebrae, lumbar compression rate, and bone cement leakage | Logistic regression |
Male: 238 patients Female: 338 patients |
< 60: 165 patients ≥ 60: 391 patients |
556 (–) | 236 (–) |
Internal validation (Random split Validation: 7:3) |
Nomogram |
| Xu 2024 [35] | 7 factors: Age, bone mineral density, smoking, ASA grading, thoracolumbar fascia injury, amount of bone cement injected, and bone cement leakage | Logistic regression |
Male: 61 patients Female: 75 patients |
No residual pain group: 62.63 ± 5.71 (Mean ± SD) Residual pain group: 67.79 ± 7.56 (Mean ± SD) |
136 (58) | – | – | Nomogram |
| Zhou 2024 [36] | 6 factors: Thoracolumbar fascia injury, bone mineral density, bone cement distribution, short term complications, implementing early rehabilitation interventions, and compliance with rehabilitation interventions | Logistic regression |
Male: 29 patients Female: 141 patients |
No residual pain group: 68.20 ± 6.86 (Mean ± SD) Residual pain group: 72.75 ± 7.85 (Mean ± SD) |
170 (85) | – | – | Formula |
SD Standard Deviation