Table 1.
Summary of studies conducted with nonimaging data.
| Publication reference | Task | Data set (nonimaging data) | Performance | |
|---|---|---|---|---|
| With nonimaging data | Without nonimaging data | |||
| Lim et al. [9] | Predict presence of OA | 5749 subjects with 24 features including demographics, personal characteristics, lifestyle variables, and health status (3795 training (30% validation), 1955 testing) | AUC: 76.8% | — |
| Accuracy: 71.97% | ||||
| SN: 66.67% | ||||
| SP: 73.35% | ||||
| Positive predictive value: 39.53% | ||||
|
| ||||
| Christodoulou et al. [4] | Predict progression of OA | OAI: 4796 subjects with 141 features including joint symptoms, disability, functionality, lifestyle, and general health status | Overall accuracy: 79.39% | — |
| Guan et al. [11] | Predicting progression of radiographic medial joint space loss | 7 features of demographic data and radiographic risk factors | AUC: 0.863; SN: 80.5%; SP: 80.5% | AUC: 0.799; SN: 78.0%; SP: 75.5% |
| Kim et al. [17] | Predict knee OA severity based on KL-grade | 6 features including demographics, alignment, and metabolic data | AUC: 0.97 (KL0), 0.85 (KL1), 0.75 (KL2), 0.86 (KL3), and 0.95 (KL4) | AUC: 0.91 (KL0), 0.80 (KL1), 0.69 (KL2), 0.86 (KL3), and 0.96 (KL4) |
| Martinez et al. [18] | Detect OA and predict future onset OA | 3 features including demographic data | Detecting OA: SN: 81.03%; SP: 79.01% | Detecting OA: SN: 79.0%; SP: 77.1% |
| Predicting onset: SN: 76.77%; SP: 62.5% | Predicting onset: SN: 76.8%; SP: 57.5% | |||
|
| ||||
| Nunes et al. [19] | Stage severity of cartilage lesion | 3 features including demographic data | Accuracy: 86.7% | Accuracy: 82.79% |
| Pedoia et al. [20] | Detect and stage severity of meniscus and patellofemoral cartilage lesions | 2 features including demographic data | Accuracy: 80.74% (normal), 78.02% (mild-moderate), 75% (severe) | Accuracy: 87.55% (normal), 71.43% (mild-moderate), 66.7% (severe) |
| Tolpadi et al. [21] | Predict total knee replacement | 27 features including demographic data, health status, disability, pain scores | AUC± SD: 0.890 ± 0.021 (X-ray), 0.834 ± 0.036 (MRI) | AUC ± SD: 0.848 ± 0.039 (X-ray), 0.886 ± 0.020 (MRI) |
| Guan et al. [22] | Predict knee pain | 7 features including demographic and radiographic risk factors | AUC: 0.804; SN: 75.2%; SP: 76.2% | AUC: 0.753; SN: 65.77%; SP: 73.51% |
Note. Task: osteoarthritis (OA); performance: magnetic resonance imaging (MRI), specificity (SP), sensitivity (SN), and area under receiver operating characteristics curve (AUC).