TBCRC 039: A phase II study of preoperative ruxolitinib with or without paclitaxel for triple-negative inflammatory breast cancer

We conducted a randomized phase II study with nested window-of-opportunity in patients with TN-IBC. Patients were randomized to receive a 7-day run-in of ruxolitinib (RUX) alone or ruxolitinib plus paclitaxel (RUX+PAC) in a 1:2 ratio, followed by 12 weeks of neoadjuvant paclitaxel alone or in combination with ruxolitinib in a 1:2 ratio. All patients subsequently received 4 cycles of doxorubicin plus cyclophosphamide, followed by modified radical mastectomy. Patients provided tumor tissue from research biopsies performed at baseline (pre-run-in) and after the completion of run-in therapy. Tumor tissue was evaluated for phosphorylated STAT3 (pSTAT3) using immunohistochemistry (IHC) and immunofluorescence, and a subset of cases were also analyze by RNA-seq. The primary objective was to evaluate the percent of patients whose tumors were pSTAT3-positive at baseline, and which subsequently became pSTAT3-negative after run-in therapy. Secondary endpoints included pathologic complete response (pCR), residual tumor burden (RCB) and event-free survival (EFS) from the date of surgery. There were 23 patients enrolled, 11 in RUX and 12 in RUX+PAC combination. The most common adverse events were fatigue and nausea. Two patients achieved pCR and 21 completed pre-operative therapy. We detected a decrease in pSTAT3 and IL6/JAK/STAT3 signaling pathways in post-run-in biopsies of RUX-treated samples, but this did not correlate with pCR, while RUX+PAC upregulated IL6/JAK/STAT3 signaling. Both treatments decreased GZM+ T cells implying immune suppression, which might have contributed to lack of efficacy. We performed RNA-seq on paired pre- and post-run-in tumor biopsy samples from 5 patients who received run-in ruxolitinib (only pre-run-in sample was available for one patient) and 4 patients who received run-in ruxolitinib plus paclitaxel. The eight patients with paried specimens were subjected into downstream analysis.