In the article ‘Tumor Radiomic Features on Pretreatment MRI to Predict Response to Lenvatinib plus an Anti-PD-1 Antibody in Advanced Hepatocellular Carcinoma: A Multicenter Study’ [Liver Cancer. 2023;12(3):262–276. https://doi.org/10.1159/000528034] by Xu et al., after publication, the authors found an error in Table 3, which represents the performance of the clinicopathologic-radiomic model on the validation cohort. The Area Under the Curve (AUC) with its 95% Confidence Interval (CI) was incorrectly listed in Table 3 as ‘0.876 (0.750–1.000)’, while the correct value should be listed as ‘0.884 (0.762–1.000).’
The corrected Table 3 is shown here.
Table 3.
ROC analyses of prediction models
| AUC (95% CI)* | Cut-off | Numerator/denominator (%) | |||||
|---|---|---|---|---|---|---|---|
| Acc | Sens | Spec | PPV | NPV | |||
| Clinicopathologic model | |||||||
| Training cohort | 0.748 (0.656–0.840) | 0.614 | 105/124 (84.7) | 18/37 (48.6) | 87/87 (100) | 18/18 (100) | 87/106 (82.1) |
| Validation cohort | 0.702 (0.547–0.884) | 33/46 (71.7) | 4/13 (30.8) | 29/33 (87.9) | 4/8 (50.0) | 29/38 (76.3) | |
| Radiomic model | |||||||
| Training cohort | 0.886 (0.815–0.957) | 0.504 | 114/124 (91.9) | 30/37 (81.1) | 84/87 (96.6) | 30/33 (90.9) | 84/91 (92.3) |
| Validation cohort | 0.820 (0.648–0.984) | 40/46 (87.0) | 8/13 (61.5) | 32/33 (97.0) | 8/9 (88.9) | 32/37 (86.5) | |
| Clinicopathologic-radiomic model | |||||||
| Training cohort | 0.987 (0.968–1.000) | 0.443 | 121/124 (97.6) | 37/37 (100) | 84/87 (96.6) | 37/40 (92.5) | 84/84 (100) |
| Validation cohort | 0.884 (0.762–1.000) | 39/46 (84.8) | 9/13 (69.2) | 30/33 (90.9) | 9/12 (75.0) | 30/34 (88.2) | |
Acc, accuracy; AUC, area under the receiver operating characteristic curve; CI, confidence interval; NPV, negative predictive value; PPV, positive predictive value; Sens, sensitivity; Spec, specificity.
*Calculated with bootstrap resampling (n = 1,000).