The 2022 Gastrointestinal Cancers Symposium

This year's GI Cancers Symposium was conducted virtually and in person January 20–22 in San Francisco. NEJM Journal Watch Oncology and Hematology Associate Editor David H. Ilson, MD, PhD, reports on a number of important and potentially practice-changing studies presented there. Abstracts can be viewed in the symposium's meeting library.

Researchers assessed the effect of early treatment discontinuation (ETD) in a pooled analysis of 11 trials in the ACCENT/IDEA database that evaluated 6 months of either adjuvant capecitabine/oxaliplatin or FOLFOX in patients with stage III colon cancer (abstract 11). ETD, defined as discontinuation of treatment before 75% of planned chemotherapy cycles, occurred in 20.9% of 10,444 patients.

ETD was associated with decreased 3-year disease-free survival (DFS; 69.0% vs. 78.8% in patients without ETD; hazard ratio, 1.61; P<0.001) and 5-year overall survival (OS; 64.7% vs. 74.7%; HR, 1.73; P<0.001); detriments were similar with FOLFOX and capecitabine/oxaliplatin. However, no detriment was associated with ETD in patients with low-risk stage III disease treated with capecitabine/oxaliplatin (HR, 1.03; P= 0.9 for DFS; HR, 1.02; P= 0.9 for OS).

In contrast, early discontinuation of oxaliplatin only (EOD), which occurred in 18.8% of 7243 patients, was not associated with a detriment either in 3-year DFS (74.9% vs. 77.2% in patients without EOD; HR, 1.07; P=0.2758) or 5-year OS (83% vs. 84%; HR, 1.13; P=0.0804) There were no differences by regimen or risk status. However, 3-year DFS was decreased with EOD in the subset of patients treated with less than 50% of the planned oxaliplatin cycles.

Previous reports from the IDEA analysis indicated that in patients with low-risk stage III disease 3 months of either FOLFOX or capecitabine/oxaliplatin are acceptable therapy. The current analysis now indicates that in high-risk patients, extension of oxaliplatin beyond 3 months may offer no additional benefit, regardless of regimen used or risk status, and that continuation of infusional 5FU or capecitabine beyond 3 months, if opted for, may convey a DFS benefit. This analysis may change practice to limit oxaliplatin treatment in patients with high-risk and low-risk stage III disease to 3 months and to individualize the need to continue fluorinated pyrimidine therapy out to 6 months in patients with high-risk disease.

Investigators reported results from the JCOG1018 trial, in which elderly patients with advanced colorectal cancer were randomized to first-line chemotherapy with bevacizumab plus mFOLFOX7 or capecitabine/oxaliplatin or those treatments without oxaliplatin (abstract 10).

Of 251 patients, median age was 79 and most had ECOG performance status 0–1 (92%). Median progression-free survival (PFS) was similar in patients treated with and without oxaliplatin (10.0 and. 9.4 months; HR, 0.837; P=0.086), as was overall survival (19.7 and 21.3 months; HR, 1.054). Response rate was higher with oxaliplatin (47.7% vs. 29.5%).

This trial adds to the literature questioning a benefit from the addition of oxaliplatin to chemotherapy treatment in elderly patients with advanced colorectal cancer and supports de-escalation of therapy.

One of the largest prospective databases monitoring circulating tumor DNA (ctDNA) as a biomarker in colorectal cancer was reported by investigators for the observational GALAXY study (abstract 9). Patients with colon cancer had individual tumor-informed ctDNA tested prior to planned surgery and at regular intervals from 4 up to 96 weeks afterwards. Of 1040 patients, 38% had stage III disease and 30% had stage II disease; 18% had positive ctDNA after surgery, including 23% of patients with stage III disease and 31% with stage IV disease.

Across all disease stages, positive ctDNA status by 4 weeks postsurgery was significantly prognostic for recurrence compared with negative status (HR, 10.9); 12-month DFS was inferior for those with positive status (47.5% vs. 92.7%). Evaluation of serial ctDNA dynamics in patients with and without chemotherapy treatment indicated improved 6-month DFS in patients with ctDNA conversion from positive to negative status compared with those with persistent positive-positive status (HR, 15.8). The 6-month DFS was similar in patients with conversion from positive to negative status compared with those with persistent negative-negative status, with and without treatment (100% and 98%). In patients who were ctDNA-positive postsurgery, adjuvant chemotherapy improved DFS across disease stages compared to no adjuvant chemotherapy (HRs, 2.4–9.4). However, in patients with high-risk stage II–III disease who were ctDNA negative postsurgery, 12-month DFS was similar with and without adjuvant chemotherapy (96.2% and 94.7%; HR, 1.3).

These provocative, hypothesis-generating results underscore the strong prognostic implications of positive ctDNA status after surgery. The results indicate that ctDNA-negative patients fare better than ctDNA-positive patients, with or without adjuvant chemotherapy, and that patients who convert from positive to negative status have better outcomes with chemotherapy than without. Ongoing trials are evaluating the use of postoperative ctDNA status to risk stratify patients in adjuvant trials, with escalation of intensity of treatment in positive patients and de-escalation of therapy in negative patients.

Practice-changing results for hepatocellular cancer were reported by investigators from the open-label, industry-sponsored, phase 3 HIMALAYA trial in which patients with advanced Child-Pugh class A hepatocellular cancer were randomized to sorafenib, durvalumab, or a single-dose of tremelimumab followed by durvalumab (abstract 379).

Of 1171 patients, 31% had hepatitis B and 27% hepatitis C, 62% had ECOG performance status of 0, and 38% were PDL-1 positive. At a median follow up of 32 months, the primary endpoint of OS with tremelimumab/durvalumab versus sorafenib indicated superiority for the immunotherapy combination (median, 16.4 vs. 13.8 months; HR, 0.78; P=0.0035); 3-year OS was also superior with the combination (30.7% vs. 20.2%). PFS trended better for the combination (HR, 0.90), and response rate was higher (20% vs. 5.1%). Single-agent durvalumab was noninferior to sorafenib for OS (median, 16.6 vs. 13.8 months; P=0.86), with a higher response rate (17.0% vs. 5.1%). OS, PFS, and 3-year OS trended similar with single-agent durvalumab and durvalumab/tremelimumab.

The results suggest that the combination of durvalumab/tremelimumab and single-agent durvalumab may represent new first-line treatment options in patients with hepatocellular cancer.

Potentially practice changing results were also reported for advanced biliary tract cancer. The industry-sponsored, double-blind placebo-controlled TOPAZ-1 trial compared chemotherapy with gemcitabine and cisplatin with or without the addition of durvalumab (abstract 378). Of 685 patients, 56% were Asian, 56% had intrahepatic and 25% had gallbladder primaries, 86% had metastatic disease, and 59% were PDL-1 positive.

At a median follow up of 16 months, the primary endpoint of OS was superior for durvalumab compared to placebo (median,12.8 vs. 11.5 months; HR, 0.80; P=0.021) and 24-month OS was improved from 10.4% to 24.9%. PFS was also improved with durvalumab (median, 7.2 vs. 5.7 months; HR, 0.75; P<0.001) as was response rate (26.7% vs. 18.7%). Rates of grade 3/4 treatment-related adverse events were similar (62.7% vs. 64.9%).

The addition of durvalumab to first-line chemotherapy may become a new treatment option in biliary tract cancer.

Evidence supporting a benefit for adjuvant chemotherapy in resected biliary cancer was reported by investigators for the open-label phase 3 JCOG1202 ASCOT trial (abstract 382). Patients with curatively resected biliary or gallbladder cancer were randomized to observation or treatment with 6 months of adjuvant S-1. Of 440 patients, most were male (87%) and had ECOG performance status of 0 (87%), extrahepatic primaries (57%), R0 resection (85%) of stage II disease (58%), and node negative status (60%).

The primary endpoint of 3-year OS was superior for S-1 (77.1% vs. 67.6%; HR, 0.797; P=0.008) as was relapse-free survival (62.4% vs. 50.9%). No new safety signals were observed. The ASCOT trial provides additional evidence for a survival benefit for oral adjuvant fluorinated pyrimidine chemotherapy after resection of biliary cancer.

An important trial from Japan comparing different neoadjuvant strategies for esophageal squamous cancer was reported from the JCOG1109 NExT trial (abstract 238). Patients with operable stage IB–III esophageal squamous cancer were randomized to treatment with two cycles of cisplatin and infusional 5-FU (CF) every 3 weeks, two cycles of CF with 41.4 Gy radiation therapy every 4 weeks, or three cycles of chemotherapy with docetaxel, cisplatin, and infusional 5-FU (DCF) every 3 weeks, followed by surgery.

Of 601 patients, most were male (80%) with cT3 (68%) and cN+ (80%) disease. The primary endpoint of 3-year OS was superior for DCF compared with CF (72.1% vs. 62.6%; HR, 0.68; P=0.006), but not for CF with radiotherapy compared with CF (68.3% vs. 62.6%; HR, 0.84; P=0.12). Median PFS also favored DCF compared with CF (not reached vs. 2.7 years; HR, 0.67), and trended superior for CF with radiotherapy compared to CF (5.3 vs. 2.7 years; HR, 0.77). Rates of curative resection in the 92% of patients sent to surgery were high in all treatment arms (95%–99%) and much higher than reported in Western trials. Rates of pathologic complete response were highest with chemoradiotherapy (36.7%) followed by DCF (18.6%) and CF (2.2%).

This important trial suggests that for patients with inoperable esophageal squamous cancer, DCF is superior to CF when chemotherapy alone is used as preoperative therapy. There was limited benefit from the addition of radiotherapy to CF in patients for whom surgery was planned. However, given the potential for chemoradiotherapy as primary therapy to avoid surgery, the optimal strategy in these patients will depend on the chosen treatment approach (surgery or no surgery), the expertise of the surgeon, and potential tolerance of therapy.

Lastly, two provocative trials were reported in patients with microsatellite instability (MSI)–high cancer treated with neoadjuvant immune checkpoint inhibitor therapy rather than conventional chemotherapy or chemoradiotherapy.

In a single-arm phase 2 trial, patients with MSI-high/DNA mismatch repair protein-deficient clinical stage II or III rectal cancer were treated with the anti-PD-1 agent dostarlimab for a total of 6 months (abstract 16). Of 16 patients enrolled, 75% had stage T3/4 disease and 94% were node positive. Of 11 patients completing treatment, 100% achieved a clinical complete response on endoscopy and rectal magnetic resonance imaging and none required chemoradiotherapy or surgery. This exciting signal of substantial response in patients with MSI-high rectal cancer may offer a new treatment paradigm; further follow-up and other ongoing trials are awaited.

In another phase 2 trial, patients with MSI-high/DNA mismatch repair protein-deficient esophagogastric adenocarcinoma received a combination of ipilimumab and nivolumab prior to planned surgical resection (abstract 244). Patients received six cycles over 12 weeks of nivolumab combined with four doses of ipilimumab followed by surgery.

Of the 32 patients enrolled, half had gastric primaries, 69% had cT3Nx disease, and 31% were not endoscopic ultrasound staged. One patient developed metastatic disease during therapy and two achieved clinical complete response without surgery. Of the 29 patients who underwent surgery, 17 (59%) had pathologic complete response. At 12 months' follow-up, 30 of 32 patients were alive without relapse This innovative trial of checkpoint inhibitor therapy indicates a high degree of activity and warrants further study in larger-scale trials. It may also open the possibility of nonoperative management in these patients.