NSABP B-38: Definitive analysis of a randomized adjuvant trial comparing dose-dense (DD) ACpaclitaxel (P) plus gemcitabine (G) with DD ACP and with docetaxel, doxorubicin, and cyclophosphamide (TAC) in women with operable, node-positive breast cancer

Background: The primary aims were to determine whether adjuvant DD ACPG will be superior to DD ACP as well as to TAC in DFS and to compare the relative DFS of TAC and DD ACP. Secondary endpoints include survival and toxicity. Methods: From Nov 3, 2004 to May 3, 2007, 4894 women were randomized; 1630 to TAC (docetaxel [T] 75 mg/m2, doxorubicin [A] 50 mg/m2, cyclophosphamide [C] 500 mg/m2 q3 wks x 6), 1634 to DD ACP (A 60 mg/m2 and C 600 mg/m2 q2 wks x 4 followed by P 175 mg/m2 q2 wks x 4), and 1630 to DD ACPG (A 60 mg/m2 and C 600 mg/m2 q2 wks x 4 P 175 mg/m2 + G 2000 mg/m2q2 wks x 4). Primary G-CSF support was required and erythropoiesis-stimulating agents (ESA) were used at investigator discretion. 52% were postmenopausal, 65% had 1 - 3 positive nodes, and 80% had HR+ breast cancer. Log-rank tests were used for pair-wise comparisons of the primary (DFS) and secondary (OS) endpoints among the three treatment arms. Results: With 64 months median follow-up, 5-year DFS in DD ACPG group was 80.6% compared with 82.2% in DD ACP group (HR=1.1; p=0.27) and 80.1 % (HR=0.97; p=0.71) in TAC group. 5-year OS was 90.8% in DD ACPG group as compared with 89.1% (HR=.89; p=0.25) in DD ACP group and 89.6 % (HR=0.90; p=0.32) in TAC group. HR for DFS and OS of DD ACP vs. TAC were 0.89 (p=0.14) and 1.01 (p=0.92) respectively. Toxicities for TAC, DD ACP, DD ACPG, respectively: febrile neutropenia (Gr 3/4: 9%, 4%, 4% [p<0.001]), sensory neuropathy (Gr 3/4: <1%, 7%, 6% [p<0.001]), diarrhea (Gr 3/4: 8 %, 2%, 2% [p<0.001]). Hgb was <10 in 12%, 26%, 33% with ESA use in 35.2%, 47%, 51.6% and transfusions in 3.7%, 6.3%, 9.4%. Deaths on treatment: N=13, 5, 7 (p=0.2). AML/MDS: N=5, 8, 11. All cycles completed in 91% for TAC and 88% for DD regimens. Conclusions: Addition of G to DD ACP did not improve outcomes. No significant differences in efficacy endpoints were identified between DD ACP and TAC, though toxicity profiles differed.