Background
Background: Point-of-care ultrasound (POCUS) has emerged globally as inexpensive tool used by various healthcare professionals. We evaluated the accuracy of a novel point-of-care liver ultrasound protocol for hepatitis (PUSH) in detecting significant fibrosis, cirrhosis, and suspicious liver lesions in Zambia.
Methods
Method: we performed a real-world validation of the PUSH at University Teaching hospital in Lusaka, Zambia, from March to June 2024. Four nurses and 6 physicians with no experience in ultrasound (US) were trained over 3 days to visualize 3 liver windows (epigastric, subcostal, and right transverse), identify 4 cirrhosis-suggestive features (enlarged caudate lobe, nodular liver surface, coarse echotexture, tortuous vascularity), and liver lesions suspicious for hepatocellular carcinoma. Post-training, consecutive adult participants (PTs) with chronic hepatitis B (CHB) alone or HBV/HIV underwent PUSH. We excluded PTs with currently/recently pregnancy or known hepatitis C. Paticipants underwent reference standard tests (RST, transient elastography, fibroscan) for significant fibrosis or cirrhosis: liver stiffness, LS>7.0 kPa and LS>12.5 kPa, respectively. Reference standard for lesions was abdominal US by an experienced radiographer. Nonparametric analysis of ROC curve for PUSH was adjusted for operator type (doctor vs. nurse), sex, and HIV status.
Results
Results: 202 PTs were included in the study including 57 (28%) with HIV (see Table 1 for characteristics). According to RSTs, 30% of PTs had significant fibrosis, 12% had cirrhosis, and 4% had liver lesions. PUSH had low accuracy in detecting significant fibrosis, with sensitivity of 28.3% (17.5 – 41.4), specificity of 99.3% (96.1 – 100), and an overall ROC area under the curve (ROC-AUC) of 0.64 (0.57 – 0.70), and moderate accuracy for cirrhosis, with sensitivity of 56% (34.9 – 75.6), specificity of 97.7% (94.3 – 99.4), and ROC-AUC of 0.77(0.70 – 0.82). Higher accuracy was seen for liver lesions, with 85.7% (42.1–99.6) sensitivity, 93.5% (88.9 – 96.6) specificity, ROC-AUC of 0.90 (0.84 – 0.94).
Conclusions
Conclusion: Newly-trained nurses and physicians in Zambia were able to use PUSH to diagnose cirrhosis with moderate accuracy and liver lesions with high accuracy in people with CHB with and without HIV coinfection. PUSH could be a useful, inexpensive tool to rapidly assess advanced liver disease and eligibility for HBV treatment in low-income settings where laboratory systems are lacking.