Correlations of blast cell Pgp features with achievement of complete remission (CR and CRp) or clearance of marrow blasts as an indication of in vivo drug susceptibility.

(A) The median baseline DiOC₂ efflux ratio (indicated by the horizontal line) was significantly higher among the group of samples from patients who failed to respond to gemtuzumab ozogamicin compared with samples from patients who achieved a CR or CRp (1.58 vs 1.24; *P = .002). In addition, samples with CSA-sensitive DiOC₂ efflux (black circles) were more common in the group without a response (52%) compared with responders (24%; P = .003). (B) The median levels of Pgp surface expression as indicated by 4E3.16 staining intensity ratios were not significantly different between samples from patients with CR or CRp and samples from nonresponders. (C) In comparison, Pgp surface expression was significantly higher among samples from patients who failed to clear marrow blasts compared with patients who achieved 5% or fewer blasts after 2 doses of gemtuzumab ozogamicin (1.11 vs 1.03; *P = .03). (D) The median baseline DiOC₂efflux ratio among samples from patients with residual leukemia was significantly higher than the median value among samples from patients who cleared marrow blasts (1.73 vs 1.27; *P = 0.0002), and nonresponder patient samples more frequently exhibited inhibition of dye efflux by CSA (72%) compared with samples from patients with 5% or fewer marrow blasts after 2 doses of gemtuzumab ozogamicin (29%;P < .001).