Initial experience of preoperative gemcitabine and cisplatin plus durvalumab followed by resection for high-risk biliary tract cancer.

Abstract

e16329 Background: Althoughgemcitabine and cisplatin plus durvalumab (GC+D) are standard of care for metastatic biliary tract cancer (BTC), the regimen’s effectiveness in the neoadjuvant setting for resectable disease remains unkown. We sought to evaluate early outcomes for patients who received neoadjuvant therapy with GC+D followed by resection for high-risk BTC. Methods: Patients who received preoperative GC+D followed by attempted resection of BTC (2/2023-6/2024) were retrospectively identified from a prospective single-institution database. Examples of high-risk BTC include: a) tumor size > 5 cm, b) multifocality or satellitosis limited to the same lobe of the liver, c) vascular invasion, d) suspected or confirmed (via biopsy) regional lymph node metastases, and e) CA 19-9 > 200 U/mL. Response was graded by RECIST criteria. Results: Among 33 patients identified, the majority (76%) had intrahepatic cholangiocarcinoma (iCCA). Median tumor diameter was 4.8 cm (interquartile range [IQR] 3.5-5.6) and 40% of patients with iCCA had multifocal disease at baseline. Baseline CA19-9 was elevated in 55% of patients with a median of 37 U/mL (IQR 16.3-90.7). Vascular invasion and metastatic lymph nodes were radiographically detected preoperatively in 30% and 24% of cases, respectively. Cisplatin was replaced with other platinum-based drugs in two patients due to tinnitus. Durvalumab was held after cycle #3 due to pancreatitis in one patient. Partial response was achieved in 52% of patients and 6% had progression of disease after a median of 5 (3-7) cycles of treatment and with a median washout time from last dose to surgery of 40 days (IQR 29-44). Of 33 patients with attempted resection, a total of 82% underwent curative-intent resection with 56% undergoing major hepatectomy. Surgery was aborted in 18% patients. The median viability at final pathology was 60% (IQR 30-80%) with a pathologic complete response in 7%. Median follow-up was 5.9 months (IQR 4.2-8.1). Postoperatively, 53% of patients received some form of adjuvant therapy. Conclusions: Neoadjuvant therapy with GC+D led to a high curative-intent resection rate in patients with high-risk BTC, obtaining promising results in terms of pathologic response. Ongoing clinical trials of GC+D for patients with BTC are critical to evaluate these results prospectively.