Performance Characteristics of 5 Methods of Early Tumor Response Evaluation in Predicting Response (RECIST 1.1) to ICI Therapy at 4 Months
| Method no. | Tumor response evaluation method description | SCAN-1 to SCAN-2 optimal percentage change cutoff | Sensitivity (%) | Specificity (%) | PPV (%) | NPV (%) | Accuracy (%) |
| 1 | Change in sum of RECIST 1.1–based target lesion diameters | ≤0 | 80.0 (28.8–96.7) | 86.7 (59.5–98.0) | 66.7 (22.7–94.7) | 92.9 (66.1–98.8) | 85.0 |
| 2 | Change in sum of the products of the 2 largest perpendicular diameters of irRC-based index lesions | ≤ −14.7 | 60.0 (15.4–93.5) | 93.3 (68.0–98.9) | 75.0 (20.3–95.9) | 87.5 (61.6–98.1) | 85.0 |
| 3 | Change in SULpeak of the hottest lesion | >15.5 | 80.0 (28.8–96.7) | 73.3 (44.9–92.0) | 50.0 (16.0–84.0) | 91.7 (61.5–98.6) | 75.0 |
| 4 | Change in sum of SUVmax of all 18F-FDG–avid metastatic lesions | >14.7 | 80.0 (28.8–96.7) | 66.7 (38.4–88.1) | 44.4 (14.0–78.6) | 90.9 (58.7–98.5) | 70.0 |
| Methods 1 and 3, above, combined (PECRIT) | 100.0 (48.0–100) | 93.3 (68.0–98.9) | 83.3 (36.1–97.2) | 100.0 (76.7–100.0) | 95.0 |
PPV = positive predictive value; NPV = negative predictive value; method 1 = change in sum of target lesion diameters, selected based on RECIST 1.1; method 2 = change in sum of the products of the 2 largest perpendicular diameters of index lesions, selected based on irRC criteria; method 3 = change in peak SUV, normalized by lean body mass, of the hottest lesion (SULpeak) seen on PET scan (PERCIST 1.0); method 4 = change in the SUVmax of all 18F-FDG–avid metastatic lesions; PECRIT = PET/CT Criteria for early prediction of Response to Immune checkpoint inhibitor Therapy.
Changes in tumor burden seen on PET/CT scans from baseline (SCAN-1) to 3–4 wk (SCAN-2) were calculated using 4 methods, each based on standard response criteria. Optimal cutoff percentage changes to predict response to ICI therapy based on RECIST 1.1 at 4 mo were determined from ROC analysis. Data in parentheses are 95% confidence intervals.