A phase II trial of TGFβ type I receptor inhibitor, galunisertib, plus neoadjuvant chemoradiation in patients with locally advanced rectal cancer

3617 Background: Transforming growth factor beta (TGFβ) is an immunosuppressive cytokine upregulated in colorectal cancer. Preclinical data demonstrated improved response to chemoradiation with TGFβ blockade in colorectal adenocarcinoma. Here we report the results of our single arm Phase II study combining the TGFβ type I receptor kinase inhibitor, galunisertib, with neoadjuvant chemoradiation in patients with locally advanced rectal adenocarcinoma. Methods: Eligible patients had T3+ or N+ rectal adenocarcinoma planned for surgical resection. Enrolled patients completed a 14-day course of galunisertib, followed by chemoradiation with continuous infusion 5-fluorouracil or capecitabine with radiation to 50.4-54Gy in 28-30 fractions. On day 30, patients underwent another 14-day course of galunisertib concurrent with ongoing chemoradiation. Five to nine weeks after completing neoadjuvant therapy, patients underwent response assessment. Those with complete response by physical exam, proctoscopy, and MRI, could opt for non-operative management and proceed to mFOLFOX6. Those with less than a complete response underwent surgical resection. The primary endpoint was complete response rate (CR), which was a composite of pathologic complete responses in those patients who proceeded to surgery, and clinical complete responses maintained at 1 year after completion of therapy for those who chose non-operative management. Using a Simon Two-Stage approach with 90% power and α = 0.05, to detect a 20% improvement in complete response rate compared to historical control (15% vs 35%); we reject the null hypothesis if ≥10/38 patients have a CR. 38 patients were enrolled with 35 patients evaluable. Results: Median age was 51y, 68% of patients were male, 87% of patients were Stage III, 97% of patients were node-positive, and 97% were proficient in mismatch repair. 95% of patients completed study therapy. Toxicity attributed to galunisertib was Grade 1-2. 28 patients went to surgery, with a mean neoadjuvant rectal (NAR) score (NAR = 5pN-3(cT-pT)+12]2/9.61) of 11.29, and pCR in 7 patients. 7 patients underwent non-operative management, with 5 achieving cCR at 1 year. Therefore, complete responses were observed in 12 patients. Peripheral immune monitoring revealed percent change in CD3+CD4-CXCR3+ T cells and activated CD8EM T cells correlated with response to therapy. 2y PFS and OS were 81.5% and 97%. Conclusions: The addition of TGFβ inhibition to neoadjuvant chemoradiation in locally advanced rectal cancer markedly improved response to therapy, was well tolerated, and warrants further investigation. Clinical trial information: NCT02688712.