TABLE 3.
Margin-level and specimen-level assessment results and performance metrics with 95% Clopper–Pearson confidence intervals for all 100 initial resections
| Parameter | Value |
|||
|---|---|---|---|---|
| Margin-level results | Path+, Image+ (TP) | Path−, Image− (TN) | Path−, Image+ (FP) | Path+, Image− (FN) |
| SIA | 7 | 555 | 24 | 14 |
| Micro-CT Reader 1 | 6 | 514 | 65 | 15 |
| Micro-CT Reader 2 | 5 | 520 | 59 | 16 |
| Micro-CT Reader 3 | 9 | 508 | 71 | 12 |
|
| ||||
| Margin-level performance, n (95% CI) | Sensitivity (%) | Specificity (%) | PPV (%) | NPV (%) |
|
| ||||
| SIA | 33.3 (14.6–57.0) | 95.9 (93.9–97.3) | 22.6 (12.4–37.5) | 97.5 (96.7–98.2) |
| Micro-CT Reader 1 | 28.6 (11.3–52.2) | 88.8 (85.9–91.2) | 8.5 (4.3–15.9) | 97.2 (96.3–97.8) |
| Micro-CT Reader 2 | 23.8 (8.2–47.2) | 89.8 (87.0–92.1) | 7.8 (3.7–15.9) | 97.0 (96.2–97.6) |
| Micro-CT Reader 3 | 42.9 (21.8–66.0) | 87.7 (84.8–90.3) | 11.3 (6.9–17.9) | 97.7 (96.7–98.4) |
|
| ||||
| Specimen-level results | Path+, Image+ (TP) | Path−, Image− (TN) | Path−, Image+ (FP) | Path+, Image− (FN) |
|
| ||||
| SIA | 6 | 66 | 20 | 8 |
| Micro-CT Reader 1 | 6 | 54 | 32 | 8 |
| Micro-CT Reader 2 | 5 | 59 | 27 | 9 |
| Micro-CT Reader 3 | 7 | 48 | 38 | 7 |
|
| ||||
| Specimen-level performance, n (95% CI) | Sensitivity (%) | Specificity (%) | PPV (%) | NPV (%) |
|
| ||||
| SIA | 42.9 (17.7–71.1) | 76.7 (66.4–85.2) | 23.1 (12.8–38.1) | 89.2 (83.8–93.0) |
| Micro-CT Reader 1 | 42.9 (17.7–71.1) | 62.8 (51.7–73.0) | 15.8 (8.8–26.7) | 87.1 (80.7–91.6) |
| Micro-CT Reader 2 | 35.7 (12.8–64.9) | 68.6 (57.7–78.2) | 15.6 (7.9–28.6) | 86.8 (81.2–90.9) |
| Micro-CT Reader 3 | 50.0 (23.0–77.0) | 55.8 (44.7–66.5) | 15.6 (9.4–24.7) | 87.3 (79.7–92.3) |
Path pathology; Image intraoperative margin assessment technique; TP true positive; TN true negative; FP false positive; FN false negative; SIA standard-of-care intraoperative margin assessment; CI confidence interval; PPV positive predictive value; NPV negative predictive value