Endoscopy, in full.

WORKUP (NCCN ESOPH-1): H&P; EGD with biopsy; chest/abdomen CT with oral and IV contrast; pelvis CT with contrast ONLY as clinically indicated; FDG-PET/CT (skull base to mid-thigh) if no evidence of M1 disease; CBC and comprehensive chemistry; EUS if no evidence of M1 or unresectable disease; endoscopic resection (ER) is recommended for accurate staging of early-stage cancers (Tis, T1a, or T1b) and may also be therapeutic; bronchoscopy if the tumor is at or above the carina with no evidence of M1; biopsy of metastatic disease as clinically indicated; assign Siewert category; nutritional assessment and counseling; smoking-cessation counseling; screen for family history. BIOMARKERS: universal MSI (PCR/NGS) or MMR (IHC) testing AND universal PD-L1 testing in ALL newly diagnosed patients; HER2 and CLDN18.2 testing if advanced/metastatic adenocarcinoma is documented or suspected; NGS should be considered. Multidisciplinary evaluation is recommended for stage I-IVA (locoregional) disease (ESOPH-E). ENDOSCOPIC THERAPY (ESOPH-A): the goal of endoscopic therapy (EMR, ESD, and/or ablation) is complete removal or eradication of early-stage disease, pTis, pT1a, and SELECTED superficial pT1b without LVI, plus the pre-neoplastic Barrett's segment. Tis/HGD must first be fully characterized for nodularity, lateral spread and multifocality, with EUS to rule out nodal metastases in select higher-risk cases. Areas of nodularity or ulceration must be RESECTED, not ablated. Completely flat lesions ≤2 cm (squamous HGD/Tis, and BE with flat HGD) should be treated by ER because it gives more accurate histologic assessment; flat lesions >2 cm can be treated by ER but with greater complication risk, and may be treated by ablation alone (data for ablation-alone in squamous HGD are very limited). Lesions pathologically limited to lamina propria/muscularis mucosae (pT1a) or superficial submucosa (pT1b), in the absence of nodal metastases, LVI, or poor differentiation, can be treated with full ER, HOWEVER a thorough patient-and-surgeon discussion of esophagectomy versus the risk of concurrent nodal disease should be undertaken, especially for larger tumors or deeper invasion. Ablation of residual Barrett's should follow ER; complete BE eradication may also be achieved by widefield EMR or ESD at the initial intervention when needed to resect superficial tumor or nodularity ≤2 cm. For SCC, the level of evidence for ablation after ER is LOW; additional ablation may be needed only for multifocal HGD/CIS elsewhere, and may not be needed for completely excised lesions. Endoscopic therapy is 'PREFERRED' for limited early-stage disease: Tis and T1a, ≤2 cm, well or moderately differentiated. Esophagectomy is indicated for extensive carcinoma in situ (pTis/HGD), pT1a, or superficial pT1b, especially nodular disease not adequately controlled endoscopically. PRIMARY TREATMENT, MEDICALLY FIT PATIENTS, ADENOCARCINOMA (ESOPH-13): cT1b-cT2, N0 LOW-RISK (<3 cm, well differentiated) → esophagectomy. cT2 N0 HIGH-RISK (LVI, ≥3 cm, or poorly differentiated), cT1b-cT2 N+, or cT3-cT4a any N → perioperative systemic therapy is PREFERRED (FLOT: 5-FU 2600 mg/m2 IV over 24 h day 1, leucovorin 200 mg/m2, oxaliplatin 85 mg/m2, docetaxel 50 mg/m2, every 14 days, 4 cycles pre- and 4 cycles postoperatively; or FLOT + durvalumab 1500 mg for PD-L1 CPS ≥1 or TAP ≥1%, category 1 for EGJ, category 2A for esophageal adenocarcinoma; note MATTERHORN showed no EFS advantage for durvalumab in diffuse-type disease), QUALIFIER: perioperative therapy is preferred for patients medically fit and with access to frequent toxicity evaluation; preoperative chemoradiation for planned esophagectomy remains an option and may be considered for borderline-resectable patients or patients who are not FLOT candidates (perioperative FOLFOX/CAPOX is another alternative). Consider neoadjuvant or perioperative immune checkpoint inhibitor if the tumor is MSI-H/dMMR (in multidisciplinary consultation; the role of surgery after complete response is unclear). Definitive chemoradiation for patients who are medically unfit/inoperable for surgery or who decline surgery. cT4b → definitive chemoradiation, or chemotherapy alone with invasion of trachea, great vessels, vertebral body, or heart. SQUAMOUS CELL CARCINOMA (ESOPH-2): cT1b-cT2, N0 LOW-RISK (<3 cm, well differentiated) → esophagectomy (non-cervical esophagus). cT2 N0 high-risk (LVI, ≥3 cm, poorly differentiated), cT1b-cT2 N+, or cT3-cT4a any N → PREOPERATIVE CHEMORADIATION is preferred (or definitive chemoradiation). cT4b → definitive chemoradiation, or chemotherapy alone for trachea/great-vessel/vertebral/cardiac invasion. NCCN states explicitly: 'Preoperative chemoradiation is preferred for esophageal SCC.' Preoperative RT dose 41.4-50.4 Gy at 1.8-2.0 Gy/day (23-28 fractions); preferred concurrent regimens include weekly paclitaxel 50 mg/m2 + carboplatin AUC 2 (CROSS, category 1) and fluorouracil + oxaliplatin (category 1). POSTOPERATIVE (ESOPH-F): nivolumab is the preferred postoperative regimen ONLY after PREOPERATIVE CHEMORADIATION with R0 resection and residual pathologic disease (category 1), 240 mg IV every 14 days for 16 weeks, then 480 mg IV every 28 days, maximum treatment duration 1 year. PALLIATION (ESOPH-A/H): esophageal dilation with balloons or bougies for temporary relief of malignant or treatment-related obstruction (avoid overdilation, perforation risk); long-term dysphagia palliation by endoscopic tumor ablation (Nd:YAG laser, PDT, cryoablation) or expandable metal/plastic stents; feeding gastrostomy or jejunostomy for anorexia/dysphagia/malnutrition, but placement of a GASTROSTOMY preoperatively may compromise the gastric vasculature and interfere with gastric-conduit reconstruction and SHOULD BE AVOIDED (feeding jejunostomy is generally preferred for postoperative nutritional support; multidisciplinary expertise is recommended before percutaneous gastrostomy).

Match the pre-sedation assessment, the drug regimen and the monitoring to the patient's risk and the complexity of the procedure. Before every sedated procedure, check ASA class, Mallampati score, BMI, OSA risk and fasting status. Propofol and midazolam-with-opiate are both strongly recommended. Using propofol as first line depends on national legislation and staffing, not on proof that it works better. Remimazolam is an ultra-short-acting benzodiazepine, FDA-approved for procedural sedation in July 2020, with fast onset, short duration, organ-independent metabolism and flumazenil reversibility. It is an accepted alternative to midazolam and propofol. Consider it especially in high-risk patients: the elderly and those with cardiorespiratory comorbidity. ESGE/ESGENA 2026 says this, and the ASGE 2026 clinical practice update now covers it, the first ASGE document to address remimazolam (the ASGE 2018 sedation guideline covered only midazolam and propofol). Whoever gives the sedation must be able to rescue a patient one level deeper than intended. Involve an anesthesiology specialist for emergency endoscopy with raised aspiration risk, hemodynamic instability or significant comorbidities. Do the same electively for ASA >=3, Mallampati >=3, severe OSA or an expected difficult airway. Manage GLP-1 receptor agonists case by case rather than stopping them by default. Discharge the patient against a scoring system, with an adult accompanying them and 24-hour restrictions.

Surveillance practice is now defined by the AGA 2025 surveillance guideline (Wani et al., Gastroenterology 2025;169:1184-1231; endorsed by the Canadian Association of Gastroenterology). Its 8 GRADE recommendations are: (1) endoscopic surveillance for nondysplastic BE rather than none (conditional, low certainty); (2) NO endoscopic surveillance for columnar-lined esophagus <1 cm, even with intestinal metaplasia, when there is no neoplasia (conditional); (3) high-definition white-light endoscopy PLUS chromoendoscopy over white light alone (STRONG); (4) no recommendation for or against wide-area transepithelial sampling (WATS-3D) as an adjunct to structured biopsy; (5) no recommendation on p53 IHC and (6) no recommendation on TissueCypher as adjuncts to histopathology in NDBE, indefinite for dysplasia (IND), or LGD; (7) daily PPI over no PPI to prevent neoplastic progression (conditional); (8) PPI over antireflux surgery to prevent progression to HGD/EAC (conditional). Implementation statements require a high-quality exam, a structured (Seattle-type) biopsy protocol, confirmation of any BE-related dysplasia by an expert GI pathologist, intervals tailored to segment length and dysplasia grade, and shared decision-making. Surveillance should stop when a patient could not tolerate endoscopic therapy, surgery, or oncologic treatment for EAC, at any age. The AGA 2025 interval table could not be read directly (full text paywalled). Until it is, the intervals below are the ACG 2022 intervals, which remain consistent with AGA's segment-length framing: NDBE <3 cm every 5 years; NDBE >=3 cm every 3 years; IND, after a second-pathologist read, optimize to twice-daily PPI, repeat EGD within 6 months, then annually if IND persists; confirmed LGD on surveillance, EGD at 6 and 12 months, then annually. Diagnosis (IM plus >=1 cm columnar mucosa, no biopsy of a normal or <1 cm irregular Z line, >=8 biopsies), endoscopic eradication therapy (EET over esophagectomy for HGD/T1a (strong); EET suggested for confirmed LGD with surveillance an acceptable alternative; resect visible lesions before RFA; sm1 low-risk T1b may be considered for EET) and post-EET intervals (baseline LGD: 1 year, 3 years, then every 2 years; baseline HGD/T1a: 3, 6, 12 months, then annually) are still governed by ACG 2022. The AGA surveillance guideline does not address them, and AGA has said a separate screening guideline will follow in 2026.

Diagnosis is unchanged and still follows the separate ESGE diagnostic guideline (Endoscopy 2025;57:166-185). Start with MRI/MRCP plus laboratory tests, read CA19-9 only after decompression, and check IgG4 when IgG4-related sclerosing cholangitis is suspected. For distal extrahepatic strictures with no pancreatic mass, do combined same-session EUS-guided tissue acquisition plus ERCP brushing and biopsy. For perihilar strictures, sample by ERCP first and move to cholangioscopy-guided biopsy if the result is indeterminate. Do not place an uncovered SEMS (USEMS) until malignancy is confirmed. Management now follows the 2026 ESGE biliary stricture management guideline, which replaces the 2018 stenting guideline and brings it into line with the 2022 therapeutic EUS guideline. Benign: post-cholecystectomy strictures get temporary multiple plastic stents (MPSs), exchanged every 3-4 months with as many stents as possible, for about 12 months. A fully covered SEMS (FCSEMS) is an option only if the stricture is more than 2 cm below the hepatic confluence or the patient will not comply with repeated ERCPs. Anastomotic strictures after liver transplant get MPSs, or a temporary FCSEMS if more than 2 cm below the confluence. Nonanastomotic post-transplant strictures that can be treated endoscopically get MPSs plus balloon dilation if needed. Chronic pancreatitis strictures get a temporary 8-10 mm FCSEMS for 6-12 months. High-grade PSC strictures get balloon dilation alone, with no stent, because stenting adds no benefit and may add adverse events. Endotherapy is indicated for a benign stricture with liver tests raised for more than 1 month and/or cholangitis. Malignant distal: drain before surgery only selectively (cholangitis, severe jaundice or pruritus, delayed surgery). When drainage is needed before surgery or neoadjuvant therapy, or for palliation, place a 10-mm SEMS endoscopically. Endoscopic drainage is preferred over surgery or PTBD for inoperable obstruction. Malignant hilar: manage in high-volume centres with a hepatobiliary multidisciplinary team (MDT). ERCP is preferred over PTBD for preoperative drainage. Palliate Bismuth I-II with ERCP. For Bismuth III-IV, ERCP or ERCP combined with EUS-guided drainage (EUS-BD) is preferred over primary PTBD, with PTBD as a complement if needed. For proven Bismuth II-IV, aim to drain more than 50%, ideally 70% or more, of the viable nonatrophic liver volume using several routes as needed. Use plastic stents when the drainage strategy is undefined or the centre is not expert. USEMS are for expert centres when the diagnosis is proven, the tumour is unresectable, and complete drainage is predicted. After failed or unfeasible ERCP, EUS-BD is recommended over PTBD. Use hepaticogastrostomy (EUS-HGS) if there is a duodenal stricture, and EUS-HGS or choledochoduodenostomy (EUS-CDS) if there is not. PTBD is used if EUS-BD fails. In surgically altered anatomy, benign disease gets enteroscopy-assisted ERCP first; malignant disease gets enteroscopy-assisted ERCP or EUS-BD. Give antibiotics before ERCP in selected cases: predicted incomplete drainage, hilar strictures, immunocompromised patients, PSC, cholangioscopy. This follows the ESGE ERCP-adverse-events guideline. Etiology-specific care (Strasberg injuries, IgG4-SC, Mirizzi, choledochal cysts) is unchanged.