Use of Artificial Intelligence for Duodenal Biopsy for Celiac Disease: Developing a Prediction Model for Histopathology Diagnosis.

DIAGNOSIS (patient must be on a gluten-containing diet). Serologic testing consists of TTG-IgA plus, if IgA deficiency has not previously been excluded, concurrent total IgA; if IgA deficiency is present, measure an IgG-based serology (DGP-IgG and/or TTG-IgG). This algorithm, previously restricted to those >=2 years, now applies to adults and children at ANY age. An elevated TTG-IgA should proceed to EGD with duodenal biopsy. A negative TTG-IgA in a non-IgA-deficient patient adequately rules out CD only when pretest probability is low or moderate; in symptomatic patients whose pretest suspicion is high (>5%), EGD with duodenal biopsy should be performed irrespective of serologic result. ACG recommends EGD with multiple duodenal biopsies to confirm the diagnosis in both children and adults with suspicion of CD (STRONG recommendation, moderate quality; 1 dissent); sampling should be 1 or 2 biopsies from the bulb (9-o'clock or 12-o'clock position) and at least 4 from the distal duodenum. Lymphocytic duodenosis (>=25 IELs/100 epithelial cells) without villous atrophy is NOT specific for CD. NONBIOPSY PATHWAY (CONDITIONAL recommendation, moderate quality): a combination of high-level TTG-IgA (>10x ULN) with a positive EMA in a SECOND blood sample is suggested as reliable for diagnosis in CHILDREN, and only if the family agrees with the no-biopsy strategy; in SYMPTOMATIC ADULTS unwilling or unable to undergo upper GI endoscopy the same criteria may be considered AFTER THE FACT, as a diagnosis of 'likely CD', ACG does not endorse a prospective no-biopsy pathway in adults, noting a >=10-fold TTG-IgA elevation carries a PPV of only ~95% in adults, which may be unacceptably low given lifelong treatment implications. HLA-DQ2/DQ8 is not required for diagnosis but is useful for serology-histology discrepancy or in patients already on a GFD; if negative, CD is ruled out. TREATMENT AND MONITORING. Strict lifelong GFD. A visit with a dietitian after diagnosis is MANDATORY, with subsequent visits as needed to reinforce education and adherence encouraged; interview with a dietitian experienced in the GFD is the standard of care for assessing adherence. ACG RECOMMENDS consumption of gluten-free oats (STRONG recommendation, moderate quality), with the qualifier that gluten contamination of oats, variable toxicity across oat varieties, and a small risk of immune reaction to avenin require monitoring for oat tolerance (monitoring intervals are not known). ACG suggests AGAINST routine use of gluten detection devices in food or biospecimens (conditional, low quality; 1 dissent), and finds INSUFFICIENT EVIDENCE to recommend for or against probiotics (evidence gap). ACG suggests setting a goal of intestinal healing as an end point of GFD therapy, advocating individualized discussion of goals beyond clinical and serological remission (CONDITIONAL recommendation, low quality). Follow-up may require multiple visits in the first year (e.g. 3, 6, and 12 months) and regular visits (e.g. twice a year or yearly) thereafter, tracking symptoms and TTG or DGP serology; other tests may include CBC, ALT, AST, vitamins A/D/E/B12, copper, zinc, folic acid, ferritin, and iron, with follow-up bloodwork individualized to verify correction of values abnormal at baseline. Preventive care includes vaccines and DXA. Symptoms improve within days of strict adherence (diarrhea improved in 80% within 60 days), but mucosal healing lags: median time from GFD onset to mucosal healing in adults is 3 years, so it is reasonable to consider a follow-up biopsy in adults after 2 years of starting a GFD, in the ABSENCE of symptoms, after shared decision-making; children heal faster (95% within 2 years), altering the risk-benefit calculus. Repeat biopsy is the only reliable method to document mucosal healing, seroconversion correlates poorly. Upper endoscopy with biopsies is also indicated for lack of clinical response or relapse of symptoms despite a GFD. VACCINATION: ACG suggests vaccination to prevent pneumococcal disease in patients with CD (CONDITIONAL recommendation, low quality), because increased pneumococcal risk is attributed to hyposplenism (frequently subclinical) present in roughly one-third of patients. The regimen follows CDC recommendations by age, immunization record, and comorbidity: for an adult with CD and functional asplenia not previously vaccinated or with unknown history, 1 dose of PCV15 (followed by PPSV23 at least 1 year later) or 1 dose of PCV20 (no PPSV23 indicated); for an adult with CD and no other condition carrying a specific CDC recommendation, ACG suggests 1 dose of PCV20 alone, or 1 dose of PCV15 first and then consider PPSV23 at least 1 year later. NONRESPONSIVE CD (NRCD): defined as persistent symptoms, signs, or laboratory abnormalities typical of CD despite 6-12 MONTHS of dietary gluten avoidance. Evaluation should review and confirm the initial diagnosis of CD, assess for inadvertent gluten exposure (via expert dietitian assessment and serology, positive serologies despite 12 months of GFD suggest ongoing gluten ingestion), assess for a coexisting functional disorder, and selectively test based on clinical suspicion for food intolerances (e.g. lactose, fructose), pancreatic insufficiency, microscopic colitis, and small intestinal bacterial overgrowth, among others. REFRACTORY CD (RCD) is a separate, rarer entity (<1% of patients outside referral centers): ongoing malabsorptive symptoms and signs with villous atrophy despite strict adherence to a GFD for MORE THAN 12 months and absence of other disorders including overt lymphoma; initial assessment centers on distinguishing RCD type 1 from type 2 by CD3/CD8 immunostains, T-cell receptor clonality by PCR, and/or flow cytometry of duodenal biopsies.