ESMO GI Congress 2026, Munich, Germany

Physical activity is moving from general lifestyle advice to an evidence-based component of cancer treatment and survivorship. The phase III CHALLENGE trial showed that a structured exercise programme started after adjuvant chemotherapy improved disease-free and overall survival in patients with colon cancer. Exercise also improves physical fitness, fatigue, muscle strength, mobility and emotional well-being, helping patients recover after surgery and systemic treatment. Exercise should be prescribed individually, usually combining aerobic activity with resistance training and gradual progression according to ability. Assessment should consider neuropathy, anaemia, heart or lung disease, bone metastases, fall risk, stomas, recent surgery and treatment-related fatigue. Wearable devices and activity data may help monitor progress and adherence, but they should support—not replace—clinical assessment. Integration of physical activity data was featured as an ESMO GI keynote topic. Lifestyle care should also address smoking cessation, healthy body weight, balanced nutrition, sleep and avoidance of excessive alcohol. Patients with weight loss or cancer-related muscle wasting need dietitian-led nutrition and supervised exercise rather than restrictive “anticancer” diets. GastroAGI take-home message: Exercise should increasingly be prescribed like a treatment—personalized, monitored and integrated into routine GI cancer care.

Supportive care should begin with treatment planning, because modern drugs produce new and sometimes unfamiliar toxicities. Immune-checkpoint inhibitors can cause inflammation in almost any organ, including the bowel, liver, lungs, thyroid and skin, and reactions may occur even after treatment has stopped. Patient education and early reporting of symptoms are essential. Treatment should be withheld and anti-inflammatory therapy started according to severity and the affected organ. Restarting immunotherapy after toxicity should be individualized according to recovery, previous severity, available alternatives and expected treatment benefit. With trastuzumab deruxtecan, new cough, breathlessness or fever requires urgent assessment for treatment-related lung inflammation and immediate interruption until evaluated. Zolbetuximab frequently requires proactive prevention of nausea and vomiting, hydration support and careful infusion management. FGFR inhibitors require monitoring for phosphate disturbance, eye problems, mouth ulcers and nail or skin changes. These practical toxicity issues formed a dedicated ESMO GI educational session. Nutrition, pain relief, pancreatic enzyme replacement, thrombosis prevention, psychological support and early palliative care remain as important as drug-specific toxicity management. GastroAGI take-home message: The success of newer GI cancer therapies depends not only on tumour response, but also on anticipating toxicity, recognizing it early and maintaining quality of life.

Precision surgery means selecting the correct operation for the correct patient according to anatomy, tumour biology and response to treatment—not simply performing more extensive surgery. Robotic platforms provide improved visualization and instrument movement, particularly in narrow or technically difficult areas such as the lower pelvis. The randomized REAL trial supports potential advantages of robotic rectal surgery in experienced centres, although costs, training and access remain important limitations. In rectal cancer, total neoadjuvant therapy followed by structured watch-and-wait surveillance can preserve the rectum in selected patients with a complete clinical response. Long-term OPRA results support this approach. Immunotherapy is further expanding organ preservation for MSI-H/dMMR rectal cancer, although careful and prolonged surveillance remains mandatory. The SANO trial showed that selected oesophageal-cancer patients with a complete response after chemoradiotherapy may undergo active surveillance rather than immediate surgery, provided salvage surgery is available when needed. Three-dimensional planning, fluorescence guidance, minimally invasive surgery and AI-supported imaging may further improve operative precision. These themes formed part of the ESMO GI surgical keynote and innovation symposium. GastroAGI take-home message: The future of GI cancer surgery is not surgery for everyone, but response-guided surgery, safer minimally invasive procedures and organ preservation whenever oncologically appropriate.

Neuroendocrine neoplasms must be separated into well-differentiated NETs and poorly differentiated neuroendocrine carcinomas , because their behaviour and treatment are fundamentally different. Well-differentiated NETs are graded according to growth activity, while neuroendocrine carcinomas are aggressive high-grade cancers. Morphology and Ki-67 should always be interpreted together. Expert pathology review is especially important in grade 3 disease, where a well-differentiated NET may be mistaken for neuroendocrine carcinoma. Somatostatin-receptor imaging has become central for staging, treatment selection and determining suitability for radionuclide therapy. Management of small localized NETs depends on the organ of origin, tumour size, depth and grade; the same approach cannot be applied to gastric, rectal, appendiceal and pancreatic NETs. Somatostatin analogues remain a treatment backbone for receptor-positive, slower-growing disease and for controlling hormone-related symptoms. NETTER-2 moved lutetium-177–DOTATATE earlier in the treatment pathway for selected advanced, well-differentiated gastroenteropancreatic NETs. The phase III CABINET trial established cabozantinib as an important later-line treatment for progressive pancreatic and extra-pancreatic NETs. GastroAGI take-home message: Modern NET care depends on correct classification, receptor-based imaging and thoughtful sequencing of somatostatin analogues, radionuclide therapy, targeted agents and chemotherapy.

Colorectal and several other GI cancers are increasingly being diagnosed in adults younger than 50 years, although the reasons remain incompletely understood. Early-onset cancers are clinically different because younger patients often experience delayed diagnosis, major disruption of work and family life, fertility concerns and long-term treatment consequences. Early-onset colorectal cancer more commonly involves the distal colon and rectum, but it should not be considered one uniform biological disease. Molecular studies show mixed findings: some tumours have distinct hereditary, microbial or molecular features, while many share the same major pathways as cancers occurring later in life. A major international study identified a DNA-damage pattern linked to the bacterial toxin colibactin , suggesting that early-life microbial exposure may contribute to some early-onset colorectal cancers. Younger age alone should not justify more aggressive treatment. Therapy should remain guided by stage, tumour location, molecular profile and patient goals. All patients with early-onset colorectal cancer should undergo germline multigene testing, with genetic counselling and family testing when an inherited alteration is identified. Fertility preservation, sexual health and psychosocial support should be discussed before treatment. These issues were central to the ESMO GI early-onset cancer symposium. GastroAGI take-home message: Early-onset GI cancers are clinically distinct but not yet a separate therapeutic category. Better understanding of genetics, microbiome and early-life exposures is urgently needed.

HER2 is now an actionable target not only in gastric cancer, but also in selected colorectal and biliary tract cancers. ESMO GI specifically highlighted HER2 treatment across upper and lower GI malignancies. In metastatic colorectal cancer, HER2 amplification is most relevant in RAS wild-type disease and may contribute to resistance to anti-EGFR therapy. The MOUNTAINEER trial established tucatinib plus trastuzumab as an active chemotherapy-free option for previously treated HER2-positive, RAS wild-type metastatic colorectal cancer. Trastuzumab deruxtecan has also demonstrated meaningful activity in previously treated HER2-positive colorectal cancer, including selected patients with prior HER2 therapy. In biliary tract cancer, zanidatamab provides an important treatment option for previously treated HER2-amplified disease. HER2-directed treatment in pancreatic, ampullary and other rare GI cancers remains investigational, although tumour-wide studies show encouraging activity in selected HER2-altered cancers. HER2 testing must use tumour-specific methods, as expression may vary between tumour sites and may change after treatment. GastroAGI take-home message: HER2 should no longer be viewed as a gastric-cancer-only biomarker. Its role is expanding across GI oncology, but accurate testing and disease-specific treatment selection remain essential.

Oligometastatic colorectal cancer represents limited metastatic disease in which complete local treatment may still offer prolonged disease control or cure. Selection should consider tumour biology, response to treatment, disease location and the possibility of treating every visible lesion—not simply the number of metastases. Management requires joint assessment by medical oncology, colorectal and liver surgery, interventional radiology and radiation oncology. The congress highlighted both multidisciplinary management and local treatment of colorectal liver metastases. Surgery remains central for resectable liver or lung metastases. However, the COLLISION trial established thermal ablation as a valid alternative for appropriately selected small liver metastases, with faster recovery and fewer complications. Stereotactic radiotherapy is useful when surgery or ablation is unsuitable. ESMO GI also presented the LAVA-CRLM comparison of microwave ablation and stereotactic radiotherapy. Systemic therapy may test tumour behaviour, shrink initially unresectable disease or treat microscopic spread, but should not unnecessarily delay potentially curative local treatment. The TransMet trial showed that liver transplantation may benefit highly selected patients with permanently unresectable, liver-only disease, but this remains restricted to expert programmes. GastroAGI take-home message: Oligometastatic CRC requires biology-guided, multidisciplinary treatment aimed at complete disease clearance using surgery, ablation, radiotherapy or, rarely, transplantation.

Artificial intelligence is moving from experimental image analysis toward practical support for pathology, endoscopy, radiology, surgery and clinical-trial selection. In digital pathology, AI can examine routine tumour slides and identify patterns associated with MSI/MMR status and other molecular changes. Large multicentre studies have shown that AI can accurately screen colorectal-cancer slides for MSI, but positive or uncertain findings should still undergo standard confirmatory testing. In endoscopy, AI may improve lesion detection, define tumour extent and support more consistent reporting. In radiology, AI can assist with staging, treatment-response assessment and identification of patients who may benefit from surgery or local treatment. AI-assisted platforms may also match patients to suitable clinical trials and combine pathology, imaging and molecular information into a single treatment-support system. ESMO GI included AI pathology, automated trial matching and AI-based prognostic studies. Major concerns remain: biased training data, poor performance outside the original hospital, lack of transparency, privacy and over-reliance on automated recommendations. Clinical responsibility must remain with the treating multidisciplinary team. GastroAGI take-home message: AI should be viewed as a clinical co-pilot—not an independent decision-maker. Its real value will come from validated tools that improve speed, consistency and access without compromising patient safety.

Therapeutic cancer vaccines aim to train the immune system to recognize proteins that are specific to an individual patient’s tumour. The greatest progress has come from personalized mRNA vaccines , created after sequencing the patient’s tumour and identifying its unique mutations. In resected pancreatic cancer, autogene cevumeran combined with immunotherapy and chemotherapy generated durable tumour-specific immune cells in selected patients and was associated with delayed recurrence. These findings are encouraging because pancreatic cancer generally has limited natural immune activity. Personalized vaccines combined with pembrolizumab have also shown early immune and clinical activity in advanced hepatocellular carcinoma. However, current evidence comes mainly from small early-phase studies; therapeutic cancer vaccines are not yet standard treatment in GI oncology. Important challenges include manufacturing time, cost, selecting the correct tumour targets and preventing the cancer from escaping immune recognition. Vaccines will probably work best in patients with minimal residual disease and in combination with immunotherapy rather than as stand-alone treatment. ESMO GI featured this subject as a keynote lecture, reflecting its importance as a future treatment platform. GastroAGI take-home message: Personalized cancer vaccines are scientifically convincing and clinically promising, but randomized trials must confirm that strong immune responses translate into longer survival.

Circulating tumour DNA, or ctDNA , consists of small fragments of cancer DNA detectable in blood and can provide information without another tissue biopsy. In metastatic disease, liquid biopsy is useful when tissue is inadequate and can identify emerging resistance mutations during treatment. After colorectal-cancer surgery, detectable ctDNA strongly identifies patients at high risk of recurrence. The GALAXY study confirmed its powerful prognostic value. The randomized DYNAMIC trial showed that ctDNA-guided treatment could reduce unnecessary chemotherapy in stage II colon cancer without worsening long-term outcomes. However, the DYNAMIC-III trial did not fully establish that chemotherapy can safely be reduced in all ctDNA-negative stage III patients. A negative result does not always exclude residual cancer because some tumours release very little DNA into the bloodstream. ctDNA should not replace high-quality imaging, endoscopy, pathology or clinical assessment. ESMO GI appropriately presented this as a “yes versus no” debate, reflecting that routine use remains dependent on the clinical setting. GastroAGI take-home message: ctDNA is ready for selected molecular and risk-assessment uses, but universal treatment escalation or de-escalation based only on ctDNA remains premature.

Advanced gastric and GE-junction cancer should now be tested for HER2, CLDN18.2, PD-L1 and MSI/MMR before first-line treatment. In HER2-positive disease, trastuzumab with chemotherapy remains the treatment backbone, with pembrolizumab added for suitable patients according to biomarker status and local approval. Final KEYNOTE-811 results confirmed the importance of this approach. Trastuzumab deruxtecan is an important later-line option after trastuzumab-based treatment and has shown superiority over conventional second-line therapy. In HER2-negative, CLDN18.2-positive disease, zolbetuximab plus chemotherapy improved disease control and survival in the SPOTLIGHT and GLOW trials. Nausea and vomiting are important zolbetuximab-related problems and require proactive prevention and monitoring. HER2 and CLDN18.2 expression may vary between tumour areas and may change after treatment; repeat biopsy can be useful when treatment decisions depend on the target. New CLDN18.2 antibody-drug conjugates, bispecific antibodies and cell therapies remain investigational. ESMO GI highlighted both targets as central parts of the emerging gastric-cancer pathway. GastroAGI take-home message: Gastric cancer treatment is no longer selected by stage alone; early, comprehensive biomarker testing now determines the first treatment and future sequencing.

Treatment choice should begin with tumour location, histology, stage, patient fitness and surgical plan . The ESMO GI session specifically addressed this question for distal oesophageal adenocarcinoma. For resectable oesophageal adenocarcinoma, the phase III ESOPEC trial showed better long-term outcomes with perioperative FLOT than with preoperative CROSS chemoradiotherapy. FLOT is therefore increasingly preferred for fit patients with resectable distal oesophageal or junctional adenocarcinoma. CROSS remains a valid option when local tumour control is the major concern, when FLOT is unsuitable, or when the multidisciplinary team favours preoperative chemoradiotherapy. MATTERHORN applies mainly to resectable gastric and GE-junction adenocarcinoma treated with a FLOT-based approach; it should not automatically be extended to all oesophageal cancers. For oesophageal squamous-cell carcinoma, chemoradiotherapy—with or without surgery—remains the central approach; FLOT and MATTERHORN are not standard strategies. Accurate distinction between oesophageal, junctional and gastric origin is therefore essential. GastroAGI take-home message: FLOT is emerging as the preferred strategy for fit patients with resectable oesophageal adenocarcinoma, while CROSS and MATTERHORN should be selected according to anatomy, histology and multidisciplinary judgement.

Surgery remains the only potentially curative treatment, but many patients relapse because microscopic metastatic disease is already present at diagnosis. Upfront surgery followed by adjuvant modified FOLFIRINOX remains a reasonable approach for fit patients with clearly resectable disease and no high-risk clinical features. Neoadjuvant treatment ensures that systemic therapy is delivered early, tests tumour biology and may prevent unnecessary surgery in patients who develop rapid metastatic progression. It is generally preferred for borderline-resectable disease and should be considered when CA19-9 remains markedly elevated, suspicious lymph nodes are present or the tumour appears biologically aggressive. NORPACT-1 did not show a clear advantage for a short course of neoadjuvant FOLFIRINOX over upfront surgery in clearly resectable pancreatic head cancer. PREOPANC supported a preoperative approach, although its chemotherapy and radiotherapy regimen differs from many current practices. PACT-21/CASSANDRA showed that preoperative PAXG improved outcomes compared with preoperative modified FOLFIRINOX, but it did not directly compare neoadjuvant therapy with upfront surgery. Ongoing phase III trials, including Alliance A021806 and PREOPANC-3, should clarify the best strategy for clearly resectable disease. GastroAGI take-home message: Neoadjuvant therapy is preferred for borderline-resectable pancreatic cancer, while clearly resectable disease still requires individualized multidisciplinary decision-making.

Antibody–drug conjugates combine a targeted antibody with a powerful anticancer drug, allowing treatment to be delivered more directly to tumour cells. Trastuzumab deruxtecan has established the value of this approach in previously treated HER2-positive gastric and GE-junction cancer. More recent evidence has confirmed its survival advantage over conventional later-line treatment in HER2-positive metastatic gastric cancer. ADC development is now expanding beyond HER2 to targets such as CLDN18.2, TROP2, c-MET, CDH17 and Nectin-4 . ESMO GI 2026 featured emerging ADC studies across colorectal, pancreatic, oesophageal and gastric cancers, including anti-CDH17, c-MET, CLDN18.2 and topoisomerase-based ADCs. These agents may help patients whose cancers no longer respond to chemotherapy, immunotherapy or earlier targeted treatment. However, most newer GI cancer ADCs remain in early clinical development and should not yet be considered routine treatment. Accurate biomarker testing is essential because target expression may vary between the primary tumour and metastases or change after treatment. Important toxicities include low blood counts, nausea, fatigue and, with some agents, potentially serious lung inflammation. GastroAGI take-home message: ADCs are becoming a major treatment platform in GI oncology, but success will depend on selecting the correct target, patient and treatment sequence.

Biliary tract cancer is not one disease; intrahepatic, extrahepatic and gallbladder cancers have different molecular profiles and treatment opportunities. Gemcitabine–cisplatin combined with durvalumab or pembrolizumab has established chemoimmunotherapy as the first-line approach for many patients with advanced disease. Comprehensive molecular profiling should be performed early, preferably before first-line treatment is completed. Important actionable alterations include FGFR2 fusions, IDH1 mutations, HER2 alterations, BRAF V600E, NTRK and RET fusions, and MSI-H/dMMR . FGFR inhibitors can provide meaningful benefit in FGFR2-altered intrahepatic cholangiocarcinoma, while ivosidenib is an established option for IDH1-mutated disease. HER2-directed treatment, including zanidatamab, has expanded options for HER2-positive biliary tract cancer. ESMO GI 2026 highlighted fanregratinib, FGFR-based combinations, long-term benefit with ivosidenib and real-world HER2-targeted therapy. Tissue testing remains preferred, although liquid biopsy may help when adequate tissue cannot be obtained. The major challenges are delayed testing, limited access, acquired resistance and uncertainty regarding treatment sequencing. GastroAGI take-home message: Molecular profiling is now a core component of biliary tract cancer care and should be planned at diagnosis—not after all standard treatments have failed.

MSI-H/dMMR colorectal cancers are exceptionally sensitive to immune-checkpoint therapy, making MSI/MMR testing essential before treatment. In locally advanced rectal cancer , dostarlimab produced complete clinical responses in a large proportion of patients, allowing many to avoid chemotherapy, radiotherapy and surgery under strict surveillance. This represents a major advance in organ preservation and may prevent permanent stomas, pelvic complications and impaired quality of life. In colon cancer , the NICHE-2 study showed remarkable tumour regression with short-course nivolumab plus ipilimumab before surgery. However, patients in NICHE-2 still underwent surgery; therefore, current evidence does not support routinely omitting colectomy. Non-operative management requires expert endoscopy, high-quality imaging, repeated biopsies and long-term follow-up because clinical assessment may miss residual disease. Lynch syndrome evaluation and genetic counselling should also be considered. ESMO GI 2026 highlighted both organ preservation in dMMR rectal cancer and whether dMMR colon cancer still requires surgery. GastroAGI take-home message: Immunotherapy may replace surgery in carefully selected dMMR rectal cancer patients, but omission of surgery in colon cancer remains investigational.

Metastatic colorectal cancer is no longer treated as one disease; therapy increasingly depends on molecular subtype, tumour location, disease burden and treatment goals. Comprehensive testing should include RAS, BRAF, MSI/MMR and HER2 , with broader sequencing used to identify rare actionable alterations. MSI-H/dMMR tumours are highly sensitive to immunotherapy and should be identified before starting first-line treatment. HER2-positive disease has become an important treatment subgroup, particularly in RAS wild-type cancers, with several HER2-directed options now available. Circulating tumour DNA can identify emerging resistance mutations and help select patients for anti-EGFR rechallenge after previous cetuximab or panitumumab treatment. ctDNA is also valuable for molecular profiling when tissue is unavailable, although routine treatment decisions based on minimal residual disease remain under evaluation. The phase III KRYSTAL-10 trial showed that adagrasib plus cetuximab did not outperform chemotherapy in second-line KRAS G12C-mutated disease, despite producing more tumour responses. Antibody-drug conjugates, bispecific antibodies and new immunotherapy combinations are being explored, particularly for microsatellite-stable disease. GastroAGI take-home message: Precision treatment in metastatic CRC now extends beyond RAS and BRAF. MSI/MMR, HER2, ctDNA, tumour sidedness and resistance profiling should increasingly guide treatment sequencing.

HCC treatment is moving from rigid stage-based therapy toward individualized combinations of surgery, ablation, embolization, radiotherapy and systemic immunotherapy. TACE remains appropriate for selected patients with embolization-eligible disease, while immunotherapy-based systemic treatment is preferred for advanced, high-burden or TACE-unsuitable disease. The phase III EMERALD-1 trial showed that improved disease control from adding durvalumab-based therapy to TACE did not translate into longer overall survival. This finding warns against adopting every TACE–immunotherapy combination based only on tumour response or delayed progression. EMERALD-3 provided more encouraging results with durvalumab–tremelimumab plus TACE, with improved disease control and an emerging survival signal; however, longer follow-up and careful patient selection remain important. Yttrium-90 radioembolization, stereotactic radiotherapy and neoadjuvant immunotherapy before resection or transplantation are also being actively studied. Repeated ineffective TACE may damage liver function and delay effective systemic therapy. Decisions should consider tumour burden, vascular invasion, liver function, bleeding risk and transplant eligibility. GastroAGI take-home message: Combining immunotherapy with locoregional treatment is promising but not yet a universal standard. Correct timing and patient selection are more important than simply adding more treatments.

KRAS has historically been considered “undruggable,” despite being the main driver of pancreatic ductal adenocarcinoma. The phase III RASolute 302 trial showed that daraxonrasib improved survival and disease control compared with standard chemotherapy in previously treated RAS-mutated metastatic pancreatic cancer. This provides the first strong phase III validation of direct RAS targeting in pancreatic cancer. At ESMO GI 2026, zoldonrasib , a KRAS G12D-selective inhibitor, showed encouraging early activity when combined with chemotherapy in untreated metastatic disease. Zoldonrasib combined with daraxonrasib also showed promising activity in heavily pretreated KRAS G12D-mutated disease. These findings remain early for zoldonrasib and require confirmation in larger randomized trials before routine use. Tumour molecular profiling is becoming essential to identify the exact KRAS variant and guide clinical-trial eligibility. Germline testing must continue, as BRCA, PALB2 and other DNA-repair alterations may support platinum-based or targeted strategies. Resistance, optimal sequencing and combination with chemotherapy remain important unanswered questions. The programme highlighted first-line and adjuvant phase III daraxonrasib trials. GastroAGI take-home message: KRAS is no longer an untreatable target. Daraxonrasib represents a major breakthrough, while G12D-selective combinations may further transform pancreatic cancer treatment.

Perioperative FLOT chemotherapy followed by surgery has been the standard treatment for fit patients with resectable gastric and gastro-oesophageal junction adenocarcinoma. Despite curative-intent treatment, recurrence remains common, creating a need for more effective perioperative therapy. The phase III MATTERHORN trial evaluated durvalumab combined with perioperative FLOT. Adding durvalumab produced deeper tumour regression, reduced recurrence risk and improved survival. Importantly, the addition of immunotherapy did not significantly reduce the likelihood of patients proceeding to surgery. Earlier studies such as KEYNOTE-585 improved pathological response but did not clearly establish a survival benefit, suggesting that results cannot be generalized to every checkpoint inhibitor or chemotherapy combination. The best candidates are fit patients with resectable gastric or GE-junction adenocarcinoma who can tolerate FLOT and undergo curative surgery. Accurate staging, nutritional assessment and multidisciplinary discussion remain essential before starting treatment. MSI/MMR testing should be performed in all patients , as MSI-H/dMMR tumours are particularly sensitive to immunotherapy. In selected MSI-H/dMMR patients, future strategies may reduce the need for chemotherapy or even surgery, but this remains investigational. GastroAGI take-home message: Durvalumab plus FLOT is a practice-changing option for suitable patients, while the future will increasingly depend on biomarker-guided perioperative treatment.