Targeted Strategies, Systemic Therapies, and AI Advances Emerge as Central Themes From ASCO GI 2026

The 2026 Gastrointestinal (GI) Cancers Symposium (ASCO GI) concluded on January 10 following 3 days of data-driven discussions highlighting the continued refinement of biomarker-directed strategies, evolving roles for immunotherapy, and growing interest in systemic approaches across GI malignancies.

Against this backdrop, OncLive® spoke with the following experts to identify the most practice-informing studies and cross-cutting themes to emerge from the meeting:

• Kanwal P.S. Raghav, MBBS, MD, an associate professor in the Department of Gastrointestinal Medical Oncology; associate vice president in the Department of Ambulatory Medical Operations, Division of Cancer Medicine; and executive medical director of the Department of Ambulatory Treatment Centers at The University of Texas MD Anderson Cancer Center in Houston
• Stephen Lam Chan, MBBS, MD (CUHK), FRCP, FHKCP, FHKAM (Medicine), clinical professor in the Department of Clinical Oncology at the Chinese University of Hong Kong
• Caroline Chung, MD, MSc, FRCPC, CIP, vice president and chief data officer; chief data and analytics officer and co-director of the Institute for Data Science in Oncology; and a professor in the Department of CNS Radiation Oncology, Division of Radiation Oncology at the University of Texas MD Anderson Cancer Center
• Daniela Molena, MD, a thoracic surgeon at Memorial Sloan Kettering Cancer Center in New York, New York
• William A. Hall, MD, a professor and chair of Radiation Oncology, the medical director of Froedtert Radiation Oncology, and Bob Uecker Endowed Chair in the Department of Surgery at the Medical College of Wisconsin’s School of Graduate Studies in Wauwatosa

Key meeting takeaways included the expansion of targeted therapies into earlier disease settings, challenges to long-standing locoregional treatment paradigms, and increasing reliance on molecular and data-driven tools to individualize care.

Read on to learn how these findings may reshape clinical decision-making across colorectal cancer (CRC), hepatocellular carcinoma (HCC), and esophageal and gastric cancers.

Why are data from BREAKWATER considered potentially paradigm-shifting for the management of BRAF V600E–mutated metastatic CRC?

Raghav: One of the key presentations [at the meeting was the readout of data from] the FOLFIRI [folinic acid, fluorouracil, and irinotecan] cohort of the [phase 3] BREAKWATER study [NCT04607421].¹ [The data] are showing really promising efficacy [with the combination of encorafenib (Braftovi), cetuximab (Erbitux), and FOLFIRI), as was seen previously with mFOLFOX6 [(oxaliplatin, leucovorin, 5-fluorouracil) plus encorafenib and cetuximab in the same study]. That is an important finding, because that means that the chemotherapy backbone [used with encorafenib and cetuximab] really doesn't matter, as long as you attach it to a BRAF-targeted therapy.

What do interim data from the phase 3b ABC-HCC trial (NCT04803994) suggest about the potential benefit of systemic therapy over transarterial chemoembolization (TACE) in intermediate-stage HCC?

Chan: There were also data from [an interim analysis of the phase 3b] IKF-035/ABC-HCC trial evaluating [atezolizumab (Tecentriq) plus bevacizumab (Avastin)] compared with TACE in [patients with] intermediate-stage HCC.²

These results will also be interesting to challenge the concept of whether patients with intermediate-stage [disease] should be treated with TACE. Some of them could potentially benefit from systemic therapy alone. [Doing so] could also potentially improve liver function and quality of life for these patients.

What strides have been made with biomarker-directed therapies in esophageal and gastric cancers?

Molena: I was very excited to hear about some of the new combinations for biomarker-directed therapies. We do not have many biomarkers in esophageal and gastric cancer yet, but there are now trials using targeted medications for HER2 that have shown really good results, as well as for Claudin 18.2 [CLDN18.2]. Although HER2 [positivity] is not very common in our patients, CLDN18.2 [expression] is very common; we see that biomarker often in our disease. I hope that more patients can be treated effectively. The way to increase the [rate of] complete response is by having more effective and personalized treatments targeted directly to the [patients’] specific histology and disease biomarkers.

New ways to use immunotherapy are also coming [down the pike]—not just for PD-L1 [inhibitors], but other agents that involve different [aspects] of the immune system to help cure the disease.

How might the role of artificial intelligence (AI) tools evolve and expand in the GI oncology space, according to data from the meeting?

Chung: Although AI tools are continuing to evolve and are very exciting—and the entire panel [at the session on Data in the Era of AI] definitely encouraged people to explore them—we must also take a step back to ensure that, if we are utilizing these tools, we think about how they may influence our work moving forward. One of the studies presented [showed that] that colonoscopists who were very seasoned and experienced actually started to lose their skill set and attention span after 3 months of using an AI-based tool. There was a difference in measurements before and after, so we need to think about what it means to utilize and integrate this into our workflow. [To use an analogy]: pilots have been using autopilot support for many years. I am curious to know what their training requirements are so they know when to intervene; I think it will be a similar [trajectory] for how we must figure out how to use these tools while remaining critical in our thinking so that we do not become too reliant on the technology.

Hall: I find AI-based tools that help guide precision oncology to be very interesting and exciting. [There are several] tools that currently help oncologists and researchers accelerate the more fundamental and traditional biomarker development that has been done for many years. These tools are very conversational and simplistic to interact with and do not require comprehensive programming experience. They do not require deep statistical knowledge or an understanding of how to use sophisticated statistical software packages like SPSS and SAS; instead, they can interact with oncologists in a conversational way.

I could ask these tools questions such as, ‘What is the association between P53 mutations and the stage of rectal cancer?’ or ‘What is the association [between] KRAS mutational status and responses with FOLFOX [fluorouracil, leucovorin, and oxaliplatin]?’ in the same way you would interact with ChatGPT or Gemini. I find that very exciting because it will open up the opportunity for more of the oncologic community to ask critical questions and innovate. As these tools deploy, we are going to see a rapid acceleration in the discovery of novel associations and biomarkers simply because more people will be using them. More people will be able to investigate and ask critical questions from their clinical practice, even if they do not have access to a team of bioinformatics-skilled statisticians or programmers.

References

1. Kopetz S, Wasan HS, Yoshino T, et al. BREAKWATER: Primary analysis of first-line (1L) encorafenib + cetuximab (EC) + FOLFIRI in BRAF V600E-mutant metastatic colorectal cancer (mCRC). J Clin Oncol. 2026;44(suppl 2):13. doi:10.1200/JCO.2026.44.2_suppl.13
2. Galle PR, Bruix J, Kloeckner R, et al. IKF-035/ABC-HCC: a phase IIIb, randomized, multicenter, open-label trial of atezolizumab plus bevacizumab versus transarterial chemoembolization (TACE) in intermediate-stage hepatocellular carcinoma. J Clin Oncol. 2026;44(suppl 2):478. doi:10.1200/JCO.2026.44.2_suppl.478