{
  "abstract": "Background Hepatic inflammation grade is regarded as the key indicator for antiviral therapy in multiple guidelines. Liver biopsy remains the gold standard for assessing hepatic inflammation, yet its clinical utility is limited by drawbacks such as delayed reporting and potential complications. Methods Between May 2020 and April 2024, we prospectively enrolled 100 consecutive CHB patients who underwent MRI examination within 6 hours prior to liver biopsy on a GE 3.0T Signa HDxt MRI scanner. The imaging protocols were routine non-enhancement MRI protocols. ADC (Apparent diffusion coefficient) values derived from DWI and SII (signal intensity index) values derived from T2WI which were defined as the ratio of hepatic signal intensity adjacent to the liver biopsy site in the right liver lobe to that of the right erector spinal muscle at the same level were measured using ROI (region-of-interest) delineation by two observers. Meanwhile, clinical characteristics of patients within 3 days before biopsy including age, age range, gender, ALT, HBV DNA exponents, HBsAg, TBIL, DBIL, ALB and PLT were documented. Using pathological hepatic inflammation grade as the gold standard, patients were classified into a case group (N=50, inflammation grade ≥G2) and a control group (N=50, inflammation grade ≤G1). Univariate and multivariable logistic regression were used to identify the significant independent risk factors and establish a predictive model.ROC (Receiver operating characteristic) analysis was performed to evaluate the diagnostic performance of the radiological model, clinical model and combined model for predicting hepatic inflammation. Results The radiological model was constructed by a combination of ADC and SII. For the clinical model, age, ALT, and HBV DNA exponents were identified as independent predictors and thus included in its development. The combined model was established by merging the radiological and clinical predictors.AUC (The area under the curve) of the radiological model in predicting hepatic inflammation grade≥G2 was 0.904, the AUC of the clinical model was 0.862, and the AUC of the combined model was 0.947 ( IDDF2026-ABS-0063 Figure 1. Comparison of ROC Curves of Radiological models, Clinical models, and Combined models for Predicting Liver Inflammation Grade ≥ G2 in Patients with Chronic Hepatitis B (CHB), IDDF2026-ABS-0063 Figure 2. The nomogram for predicting liver inflammation ≥ G2 in CHB patients generated based on the combined model (both ADC and SII values are values ×10−2)). Calibration plot and DCA curve of the combined model demonstrated good predictive performance (IDDF2026-ABS-0063 Figure 3. Calibration Plot and Decision Curve Analysis Plot for Predicting Liver Inflammation grade ≥ G2 in Patients with CHB using the Combined Model).Conclusions Multiple parameters MRI combined with clinical characteristics achieved superior accuracy in predicting hepatic inflammation grade ≥G2. This approach provides a non-invasive method, serving as a potential biomarker for early clinical assessment of pathological outcomes.Abstract IDDF2026-ABS-0063 Figure 2Abstract IDDF2026-ABS-0063 Figure 1Abstract IDDF2026-ABS-0063 Figure 3",
  "authors": [
    {
      "affiliations": [
        "Fifth Medical Centre of Chinese PLA General Hospital, China"
      ],
      "name": "Yuan Liu"
    },
    {
      "affiliations": [
        "Fifth Medical Centre of Chinese PLA General Hospital, China"
      ],
      "name": "JInghui Dong"
    },
    {
      "affiliations": [
        "Fifth Medical Centre of Chinese PLA General Hospital, China"
      ],
      "name": "Hongwei Ren"
    },
    {
      "affiliations": [
        "Third Medical Centre of Chinese PLA General Hospital, China"
      ],
      "name": "Yanan Zhang"
    },
    {
      "affiliations": [
        "Fifth Medical Centre of Chinese PLA General Hospital, China"
      ],
      "name": "Changchun Liu"
    },
    {
      "affiliations": [
        "Fifth Medical Centre of Chinese PLA General Hospital, China"
      ],
      "name": "Mengmeng Zhang"
    },
    {
      "affiliations": [
        "Fifth Medical Centre of Chinese PLA General Hospital, China"
      ],
      "name": "Meng Zheng"
    },
    {
      "affiliations": [
        "Medical School of Chinese PLA, China"
      ],
      "name": "Baobao Li"
    },
    {
      "affiliations": [
        "Medical School of Chinese PLA, China"
      ],
      "name": "Xiangrong Du"
    },
    {
      "affiliations": [
        "Fifth Medical Centre of Chinese PLA General Hospital, China"
      ],
      "name": "Jianming Cai"
    }
  ],
  "title": "IDDF2026-ABS-0063 Non-invasive prediction of hepatic inflammation grade in chronic hepatitis B patients based on multiple parameters non-enhancement MRI combined with clinical characteristics",
  "uid": "84386484-4dd4-5cac-9b52-2052ba7092c2"
}