Table 3.
Disease entity | Clinical features | Training cohort (n = 55) |
P | European validation cohort (n = 80) |
P | United States validation cohort (n = 100) |
P | Pooled validation (n = 180) |
P | ||||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|
High (n = 32) |
Low (n = 23) |
High (n = 42) |
Low (n = 38) |
High (n = 41) |
Low (n = 59) |
High (n = 83) |
Low (n = 97) |
||||||
Severe thrombocytopenia | Platelet count <50 g/L | 28 (88) | 12 (52) | .006 | 37 (88) | 15 (39) | <.001 | 35 (85) | 18 (31) | <.001 | 72 (87) | 33 (34) | <.001 |
Anemia | Hb <8 g/dL or requiring transfusion | 29 (91) | 14 (61) | .02 | 40 (95) | 14 (37) | <.001 | 40 (98) | 25 (42) | <.001 | 80 (96) | 39 (40) | <.001 |
Neutropenia | |||||||||||||
Severe | ANC ≤500 cells per µL | 32 (100) | 22 (96) | .4 | 42 (100) | 27 (71) | <.001 | 40 (98) | 50 (85) | .04 | 82 (99) | 77 (79) | <.001 |
Profound | ANC ≤100 cells per µL | 32 (100) | 15 (65) | <.001 | 37 (88) | 17 (45) | <.001 | 37 (90) | 31 (53) | <.001 | 74 (89) | 48 (49) | <.001 |
Protracted, severe | ANC ≤500 cells per µL for ≥7 days | 31 (97) | 11 (48) | <.001 | 37 (88) | 14 (37) | <.001 | 36 (88) | 31 (53) | <.001 | 73 (88) | 45 (46) | <.001 |
Protracted, profound | ANC ≤100 cells per µL for ≥7 days | 16 (50) | 0 (0) | <.001 | 21 (50) | 4 (11) | <.001 | 18 (44) | 1 (2) | <.001 | 39 (47) | 5 (5) | <.001 |
Prolonged | ANC ≤1000 cells per µL measured at ≥21 days after CAR transfusion | 29 (91) | 14 (61) | .02 | 35 (83) | 23 (61) | .003 | 32 (78) | 18 (31) | <.001 | 67 (81) | 41 (42) | <.001 |
Primary end point: duration of severe neutropenia from day 0 to day 60 (95% CI) | 16.5 (13-43) |
7 (6-10) |
<.001 | 17 (13-19) |
5 (3-8) |
<.001 | 10 (9-12) |
7 (5-8) |
<.001 | 12 (10-16) |
5.5 (5-8) |
<.001 | |
Binary end point: severe neutropenia ≥14 days | 21 (66) | 1 (4) | <.001 | 28 (67) | 5 (13) | <.001 | 12 (29) | 0 (0) | <.001 | 40 (48) | 5 (5) | <.001 |
Comparison of the incidence of clinically significant thrombocytopenia, anemia, and neutropenia in patients with a high risk for hematotoxicity (CAR-HEMATOTOX score ≥2) compared with patients with a low risk (CAR-HEMATOTOX score <2). Statistical significance (P < .05) between groups was determined by Fisher’s exact 2-sided t test for incidence rates and Mann-Whitney U test for the primary end point.