Diagnostic Accuracy of Ultrasound in Cholestatic Infants with Biliary Atresia

Abstract

Objective:

Biliary atresia (BA) is an obstructive cholangiopathy that involves the intrahepatic and extrahepatic bile ducts. Ultrasound (US) can aid in evaluation of the biliary system and be efficiently used in daily practice. However, most studies on US for diagnosing BA have been conducted in developed countries. Therefore, we have aimed to evaluate the diagnostic accuracy of US in BA in infants with cholestasis from a developing country.

Materials and Methods:

This retrospective study used data collected from our hospital medical records. The US findings were compared with the gold standard intraoperative or cholangiography findings.

Results:

Thirty-five BA patients (19 males and 16 females) and 36 controls (20 males and 16 females) were included in the study. Most of the patients (85.7%) were ≤ 6 months old. The absence of a gallbladder demonstrated 71.42% sensitivity (Sn), 91.67% specificity (Sp), 89.29% positive predictive value (PPV), 76.74% negative predictive value (NPV), 8.57 positive likelihood ratio (LR+), and 0.31 negative likelihood ratio (LR−) for diagnosing BA. The triangular cord sign demonstrated 14.28% Sn, 100% Sp, 100% PPV, 76.74% NPV, ∞ LR+, and 0.86 LR- for diagnosing BA. The combination of gallbladder absence and a positive triangular cord sign demonstrated 82.85% Sn, 91.67% Sp, 90.63% PPV, 84.61% NPV, 9.95 LR+, and 0.19 LR− for diagnosing BA.

Conclusion:

The diagnostic accuracy of US in BA is high, indicating that it can be the imaging tool of choice in infants with cholestasis. Ultrasound is safe and can be easily used in daily practice without the risk of radiation exposure.

What is already known on this topic?

• Biliary atresia is a significant cause of neonatal jaundice and liver failure.Biliary atresia is a significant cause of neonatal jaundice and liver failure.

• Early diagnosis and timely intervention, usually involving surgical correction, are crucial for better outcomes in infants with biliary atresia.

• Ultrasound imaging is commonly used as an initial diagnostic tool due to its non-invasive nature and accessibility.

What this study adds on this topic?

• Evaluation of the accuracy of ultrasound for specifically diagnosing biliary atresia in infants and its ability to differentiate biliary atresia from other causes of neonatal jaundice.

• Assessment of the sensitivity and specificity of ultrasound in distinguishing biliary atresia from the other causes of neonatal jaundice support the use of ultrasound as the diagnostic tool of choice in infants with cholestasis.

• Identify potential advantages and challenges in using ultrasound for diagnosing biliary atresia in infants in clinical practice.

Introduction

Biliary atresia (BA) is an obstructive cholangiopathy that affects the intrahepatic and extrahepatic bile ducts and obstructs the flow of bile.¹ If left untreated, BA can cause liver fibrosis, which can be fatal within 2 years of life. Two in 3 children with BA will require liver transplantation at some point.²

Biliary atresia is diagnosed on the basis of clinical manifestations, radiological findings, and intraoperative cholangiography (IOC) findings. Intraoperative cholangiography is the gold standard for diagnosing BA.³ However, it is an invasive method and cannot be used in primary health centers or district hospitals. Ultrasound (US) is an imaging technology that can be used to evaluate the biliary system and identify any abnormalities.^4-7 Furthermore, US is frequently used in daily practice.⁸

Several US findings have been described as valuable predictors of BA, including the triangular cord sign and gallbladder absence.^1,4-7,9 The triangular cord sign is an echogenic tube of fibrous tissue that is found on the hepatic portal vein. It is a characteristic finding of BA that can help distinguish BA from other neonatal cholestatic diseases.⁹ Previous studies have demonstrated varied sensitivities and specificities of US for detecting BA, which depend on the abnormalities detected. The sensitivity and specificity of gallbladder absence on US for diagnosing BA are 67.5% and 65%, respectively.¹ However, the sensitivity and specificity of the combination of gallbladder absence and a positive triangular cord sign for diagnosing BA are 100% and 94.4%, respectively.¹⁰ A recent review demonstrated that US is essential for the diagnosis of BA.¹¹ However, most of the studies on US for diagnosing BA have been conducted in developed countries.⁷ Therefore, in this study, we aimed to evaluate the accuracy of US for diagnosing BA in infants with cholestasis from a developing country.

Materials and Methods

Patients

In our methodological study, data were collected from Dr. Sardjito Hospital medical records. The inclusion criteria were patients with BA who presented to our institution between January 2014 and December 2016. Cholestatic patients without BA were included as controls. Patients with incomplete medical records were excluded from our study. The study was approved by the Institutional Review Board of the Faculty of Medicine, Public Health, and Nursing, Universitas Gadjah Mada/Dr. Sardjito Hospital (number: KE/FK/0708/EC/2017, date: 21 June 2017).

Based on the ICD-10 codes for cholestasis (Q44.2, K71.0, B25.8), 122 patients were enrolled in the study. Of these patients, 71 were confirmed to have BA on the basis of the findings of the gold standard diagnostic tool (intraoperative/IOC findings) and US (Figure 1).

Figure 1.

Flowchart of the inclusion and exclusion criteria for this study. Gold standard, intraoperative, or cholangiography findings.

Ultrasound Findings

We used the absence of the gallbladder and the presence of the triangular cord sign (Figure 2) as US findings to be tested against the gold-standard findings (Figure 3). The infants were kept fasting for at least 3 hours before the US examination. All examinations were performed by an experienced consultant pediatric radiologist using the Logiq S9 series scanner (GE Healthcare, Boston, USA) with convex (3.5-5.5 MHz) and linear (7.5-10 MHz) transducers. The triangular cord sign was defined as the presence of a > 4 mm-thick echogenic anterior wall of the right portal vein on the longitudinal view of a US. The experienced consultant pediatric radiologist identified the absence of a gallbladder and a positive triangular cord sign on the captured images immediately after the initial US scan.

Figure 2.

Ultrasound showing (1) the absence of gallbladder, (2) triangular cord sign. TCA, triangular cord; v porta, portal vein.

Figure 3.

Intraoperative cholangiography of biliary atresia type 1, 2a, 2b, and 3.

Statistical Analysis

Data are presented as numbers and percentages. The McNemar or chi-square test was used to compare nonparametric data. The null hypothesis is that paired samples (gold standard and US finding) are equal, and no significant change has occurred. A two-sided P-value of < .05 was considered statistically significant. Sensitivity (Sn), specificity (Sp), positive predictive value (PPV), negative predictive value (NPV), positive likelihood ratio (LR+), and negative likelihood ratio (LR–) were determined using a standard formula based on a 2 × 2 table. Sensitivity was determined using the following formula: number of true positives [TP)/(number of TP + number of false negatives (FN)] × 100. Specificity was determined using the following formula: number of true negatives [TN)/(number of TN + number of false positives (FP)] × 100. PPV was calculated using the following formula: TP/(TP + FP) × 100. Negative predictive value was calculated using the following formula: TN/(TN + FN) × 100. Positive likelihood ratio was calculated using the following formula: Sn/(1 − Sp). Negative likelihood ratio was calculated using the following formula: (1 − Sn)/Sp.

Results

Patient Characteristics

In this study, 35 patients with BA (19 men and 16 women) and 36 controls (patients without BA; 20 men and 16 women) were included (Table 1). All patients simultaneously underwent surgical intervention and liver biopsy. Most of the included patients (65.7%) exhibited type III BA (Table 1).

Table 1.

Baseline Characteristics of BA Patients in Our Study

Characteristic	BA, n (%)	Control, n (%)	P*
Sex  Male  Female	19 (54.3) 16 (46.7)	20 (55.6) 16 (44.4)	.914
Age, months  < 3  3–6  > 6	15 (42.9) 15 (42.9) 5 (14.2)	20 (55.6) 9 (25) 7 (20.4)	.282
BA classification  Type I  Type IIa  Type IIb  Type III	1 (2.9) 8 (22.9) 3 (8.6) 23 (65.7)	—	—

BA, biliary atresia.

*Chi-square test was used to define the statistical differences among groups and P < .05 was considered statistically significant.

Ultrasound Findings

Of the 71 enrolled patients, 35 were diagnosed with BA. The US of 80% (n = 28) of the patients with BA revealed the absence of a gallbladder (Figure 2). Furthermore, the US of 14.3% (n = 5) of the patients with BA demonstrated a positive triangular cord sign (Figure 2).

Diagnostic Accuracy of Ultrasound in Biliary Atresia

We used the following 2 US findings to determine the diagnostic accuracy of US in BA: absence of the gallbladder and a positive triangular cord sign. We found that the absence of a gallbladder demonstrated 71.42% Sn, 91.67% Sp, 89.29% PPV, 76.74% NPV, 8.57 LR+, and 0.31 LR−. The triangular cord sign exhibited 14.28% Sn, 100% Sp, 100% PPV, 76.74% NPV, ~ LR+, and 0.86 LR− (Table 2). The triangular cord sign or absence of gallbladder combination exhibited 82.85% Sn, 91.67% Sp, 90.63% PPV, 84.61% NPV, 9.95 LR+, and 0.19 LR− (Table 2).

Table 2.

Diagnostic Value of the Absence of the Gallbladder and Triangular Cord Sign for the Diagnosis of BA

US Finding	Intraoperative/IOC Finding of BA		P
Absence of gallbladder	Yes, n (%)	No, n (%)
Yes	25 (71.4)	3 (8.3)	.096
No	10 (28.6)	33 (91.7)	–
Triangular cord sign
Yes	5 (14.3)	0	.0001*
No	30 (85.7)	36 (100)	–
Triangular cord sign or absence of gallbladder combination
Yes	29 (82.9)	3 (8.3)	.505
No	6 (17.1)	33 (91.7)	–

BA, biliary atresia; IOC, intraoperative cholangiography.

*McNemar test was used to determine the statistical differences among groups and P < .05 was considered statistically significant.

Discussion

In this study, we demonstrated that a combination of US findings (gallbladder absence and the triangular cord sign) can diagnose BA with relatively high accuracy. Ultrasound is the first choice of imaging modality to screen infants with cholestasis. It is a safe and affordable diagnostic tool that can be easily used in daily practice without the risk of radiation exposure. Several US findings have been described in patients with BA, including the triangular cord sign and the absence of a gallbladder.^12,13 These findings were also seen in our patients.

In our study, US demonstrated an 82.85% sensitivity for detecting BA, which is higher than that in previous studies (67.6%-78% sensitivity).^12,14 In our study, US exhibited a 91.67% specificity for detecting BA, which is higher than that in previous studies (65%-81%; Table 3).^14,12 The differences in sensitivity and specificity may be attributable to the different sample sizes. For an expected sensitivity of 80% when compared to the gold standard, the estimated minimal sample size was 61. Thus, we included 71 patients in this study. Furthermore, our study provides new evidence of the high accuracy of US for diagnosing BA, particularly in a developing country. Most studies to date have been conducted in developed countries.⁷

Table 3.

Comparison of Accuracy Value of US for Diagnosis of Biliary Atresia Between Our Study and Previous Reports.^7,12-14

Accuracy	Our study (%)	[7] (%)	[12] (%)	[13] (%)	[14] (%)
Sn	82.85	95	78	92.6	67.6
Sp	91.67	89	81	51.9	65
PPV	90.63	N/A	88	65.8	76.7
NPV	84.61	N/A	63	87.5	54.2

NPV, negative predictive value; PPV, positive predictive value; Sn, sensitivity; Sp, specificity; US, ultrasound.

The PPV and NPV in this study were 90.63% and 84.61%, respectively, which are higher than that in previous studies (Table 3).^12,14 The PPV and NPV are affected by disease prevalence.¹⁵ A higher prevalence is associated with an increase in PPV and a decrease in NPV. In this study, the prevalence of BA was 49.3% (Table 1). The prevalence of BA in the Asia-Pacific region and Europe is higher than that in the Middle East.¹⁶ Moreover, the frequency of BA as the etiology of cholestasis in infants is 4.7% in Saudi Arabia, which is 20%-25% lower than the frequencies in Western countries.¹⁶ Indonesia is one of the countries in the Asia-Pacific region; therefore, it is expected that the PPV will be higher than that in previous studies from the Middle East region.^13,14,16

Ultrasound is an operator-dependent tool, and the radiologist’s ability to interpret the imaging findings depends on their experience. Several recent studies have reported a scoring system for diagnosing BA.^4,5 This scoring system includes US findings, including the presence of the triangular cord sign.^4,5

A recent study demonstrated a delay in the diagnosis and treatment of BA. This may be attributed to the fact that most medical students and primary health care physicians are unable to recognize the signs of prolonged jaundice and acholic feces in infants.¹⁷

This study has some limitations. First, we only determined the diagnostic accuracy of US in BA using the overall means of only 2 findings. Other US findings, including gallbladder abnormalities such as a < 1.5 cm-long gallbladder and a gallbladder with wall abnormalities, amount of tissue stiffness (shear wave elastography), hepatic artery diameter, and the presence of a subcapsular hepatic flow, were not considered.^4-6,13,18-20 Second, due to the study’s retrospective design and the fact that the US was performed only once by an experienced consultant pediatric radiologist, we were unable to determine the intrarater and interrater agreement. Third, the study was conducted at a single institution. Thus, multicenter prospective studies with larger sample sizes are required in the future to validate our study findings.

The diagnostic accuracy of US in BA is high, indicating that it can be a diagnostic tool of choice in infants with cholestasis. Furthermore, US is safe and can be easily used in daily practice without the risk of radiation exposure.

Data Availability Statement:

All data generated or analyzed during this study are included in the submission. The raw data are available from the corresponding author upon reasonable request.

Funding Statement

This study received no funding.