Citations

820 Background: The number of FDA-approved anticancer drugs have increased in recent years with a more use of surrogate endpoints. However, it is unclear whether the same trend applies to GI cancers. Therefore, we aimed to evaluate the trend in FDA-approved GI cancer drugs over the past two decades and the associated clinical benefit. Methods: The FDA’s online database was searched for drugs approved for GI cancers from 2006 until 2024. Primary and secondary endpoints along with trial characteristics were extracted. ESMO Magnitude of Clinical Benefit Scale (ESMO-MCBS) was assessed by two blinded investigators with a third resolving disagreements. We divided the study into period A (01/2006-12/2014) and period B (01/2015-08/2024). Results: Since 2006, FDA approved 62 regimens for GI cancer based on 70 trials, mostly (n= 45;72.6%) in period B. Most trials (n=44) were phase III with only 26 phase II: 22 single arm, and 4 randomized. All the 22 single-arm phase II trials were approved in period B with two randomized phase II trials approved in each period. Approvals based on phase III trials dropped by 33.5% (from n=15 to n=29) between the two periods. Half (n=34) of the approvals were based on overall survival (OS) benefit, while the use of objective response rate (ORR) as an endpoint has increased from 11.8% (n=2) in period A to 45.3% (n=24) in period B. Quality of life (QoL) was assessed only in 27% of the studies and found benefit only in 22.7%. The mean gain in OS was similar across both periods; 2.3 months [1.2-4.7]. Out of the 19 accelerated approvals, 17 were in period B, of which 23% (4/17) were based on secondary endpoint, 18% (3/17) were withdrawn, and 59% (10/17) received full approval. Substantial clinical benefit (ESMO score 4-5) was seen in 11.8% of the trials in period A and 19.6% in period B, with 17.4% overall from 2006-2024. Conclusions: Despite rising FDA approvals for GI cancer drugs since 2006, only 17.4% demonstrated substantial clinical benefit based on ESMO-MCBS. Recent approvals rely more on single-arm phase II studies and ORR, often without OS or QoL benefit. These trends underscore the need for more rigorous clinical trials that prioritize meaningful endpoint to ensure that new therapies provide real clinical value to the patient.

451 Background: Perioperative chemotherapy with surgery is the standard of care for patients (pts) with resectable gastric or gastroesophageal junction adenocarcinoma (GEA). However, recent studies have shown more favorable responses after neoadjuvant immunotherapy (NIO) in patients with mismatch repair deficient (dMMR) cancers. Herein, we present the outcomes of 10 cases with dMMR GEA. Methods: We retrospectively reviewed patients with GEA at Mayo Clinic Arizona who met the following criteria: have dMMR disease by IHC, nonmetastatic operable disease, and received NIO +/- chemotherapy. Patient demographics, disease characteristics, and clinical outcomes were extracted. The primary endpoint was clinical complete response (cCR) in patients treated with non-operative management (NOM) defined as complete radiological response and no residual disease on endoscopic ultrasound (EUS) and biopsy. Results: 10 pts with dMMR GEA with a median age of 70 (43-85) were identified. 6 pts had T2N0, 1 had T4N0, 3 had N+ disease. 8 pts received NIO monotherapy upfront, 1 received NIO after progressing on chemotherapy, and 1 concurrently with chemotherapy. No grade 3/4 toxicities were reported, however, 2 pts discontinued therapy due to arthralgia and colitis. After NIO, 9/10 pts had significant or complete resolution of FGD avidity on PET/CT, while 1/10 had stable disease. 3/10 pts underwent surgical resection, with 2 achieving pathological complete response (pCR) and the third showing T1N0 disease with no recurrence 1 year after surgery. Notably, one of the pts with pCR passed away due to surgical complications. 7/10 patients had NOM, of whom, 4 had cCR and 3 had radiological response only, as no EUS data was available. 4/7 NOM pts had follow up data at 6 months with no evidence of progression, of whom one pt remained progression-free at 33 months. ctDNA testing was performed in 6/10 pts: 2 had undetectable ctDNA at baseline, while 3 became MRD-negative after NIO, consistent with radiological and/or pathological response, and 1 had persistent ctDNA both before and after NIO, despite cCR. Conclusions: This study demonstrates exceptional and prolonged response, along with good tolerability to NIO in pts with dMMR GEA, supporting the feasibility of non-operative management in such patients.

732 Background: Ampullary carcinoma (AC) is a rare malignancy arising from the ampulla of Vater which has a more favorable prognosis compared to other pancreatic malignancies. Current guidelines favor surgical resection followed by adjuvant therapy for average risk patients (pts), however the type of regimen as well as other perioperative treatment modalities are still being explored. The objective of this retrospective study aims to provide insight into the management of locoregional disease. Methods: Pathology records from Mayo Clinic (AZ, FL, MN) denoting AC between 2010 to 2024 were searched using Mayo Data Explorer and selected for retrospective review. Pt demographics, treatment courses, and next generation sequencing (NGS) data were collected. Statistical analysis was conducted using SAS version 9.04. Results: A total of 137 pts (62% male, n = 85) were identified with median age 66 years old. Histologic subtypes were 54% pancreatobiliary (44/81), 40% intestinal (32/81), 6% mixed (5/81), with 56 pts unknown. Stages at diagnosis were 81% resectable (111/137), 11% locally advanced (15/137), and 8% metastatic (11/137). Nearly all (93%, n = 103/111) pts with localized disease had resection with a 97% R0 resection rate (100/103). Majority (68%, n = 75/111) of resectable pts had perioperative chemotherapy with 5-FU or gemcitabine-based regimens (24%, n = 18/75 neoadjuvant vs. 76%, n = 57/75 adjuvant). Perioperative chemoradiation was given in 17% (19/111) of pts. At median follow up of 30.5 months, recurrence rates were 40% (41/103), primarily to the liver (18/41). Median recurrence free survival (RFS) was 29.3 months (95% CI 24.7 – NE). Median overall survival (OS) has not been reached. The 5-year RFS and OS rates were 0.37 (95% CI 0.25 - 0.55) and 0.69 (95% CI 0.58 - 0.82), respectively. Univariate analysis showed no difference in OS with the addition of neoadjuvant or adjuvant chemotherapy or chemoradiation. Somatic NGS testing showed pathogenic mutations in 90% of pts (63/70): 57% KRAS (36/63; 44% G12D, 19% G12V, 8% G12C, 8% G12D, 19% others); 13% homologous recombination (HR) (8/63); 6% ERRB2 /Her2 amplification (4/63), and 5% mismatch repair (MMR) (3/63). Conclusions: AC has a favorable prognosis in resectable pts. However, high recurrence rates indicate the need for better systemic therapies and improved selection of pts who would benefit from these treatments. Future research should focus on refining perioperative strategies to enhance outcomes and reduce recurrence.