Page 115 - Read Online

Page 4 of 11 Muñoz-Martínez et al. Hepatoma Res 2022;8:30 https://dx.doi.org/10.20517/2394-5079.2022.22

Selecting which patients should be enrolled in the surveillance program and the frequency of the tests is a
crucial decision to develop a cost-effective surveillance strategy. Imaging tests with high sensitivity may
increase false-positive results and consequently increase exposure to invasive tests like endoscopic
retrograde cholangiopancreatography (ERCP). On the contrary, selecting a test imbalanced to a higher
specificity would increase the number of false-negative cases, missing early CCA diagnosis .
[23]
All these considerations thicken the plot, where the clinician will have to decide if the findings justify
exposing the patient to invasive procedures for histopathological confirmation.

Imaging techniques for CCA surveillance in PSC
The ideal imaging test for surveillance should offer an adequate diagnostic accuracy balancing a very high
sensitivity to detect an early-stage CCA on asymptomatic patients with an acceptable specificity for avoiding
unacceptable false-positive results. On top of that, the test should be acceptable by the population and
widely available. More importantly, an appropriate and validated recall strategy in the case of a positive
surveillance test should be available.

There are three non-invasive imagine techniques that could be considered for surveillance: ultrasonography
(US), computerized tomography (CT), and magnetic resonance imaging (MRI) with
cholangiopancreatography technique (MRCP). US is widely available and relatively cheap, but the
sensitivity of US for early-stage CCA detection and the ability to provide a full virtual an reproducible map
of the biliary tree is far from optimal and it is inferior to MRI for early-stage CCA detection . CT is
[22]
associated with radiation exposure and limited image quality of the biliary tree when the biliary ducts are
not dilated.

[24]
MRI/MRCP is considered the imaging standard for diagnosis and follow-up in patients with PSC . The
MRI/MRCP is non-invasive without radiation exposure, and is the best image technique to explore the
biliary tree with a pooled sensitivity of 98.9% (95%CI: 98.6-99.3) and specificity close to 100% as reported in
a recent meta-analysis, increasing CCA detection sensitivity by combining with contrast agents, without
affecting the specificity [23,25] . MRI should be considered as the study of choice as it is superior against US in
the detection of early-stage perihilar CCA in patients with PSC, showing better area under the curve (AUC)
in the entire cohort (0.87 vs. 0.70) and also in asymptomatic patients (0.81 vs. 0.59) .
[22]

There are different findings on imaging tests which are considered to be suspicious for CCA [3,26] [Table 1].
The most frequent finding is the presence of an obstructive biliary stricture, but this finding is not a specific
as patients with PSC usually have inflammatory/fibrotic obstructive biliary strictures on imaging studies
known as dominant strictures (DS), which may mimic malignant strictures. Early CCA has different
presentations in imaging tests: dilatation and/or thickening of the biliary duct, tumor infiltration along the
biliary tree resulting in ductal narrowing, beading irregularities of the central hepatic ducts, diffuse
strictures, or a discrete mass-forming lesion, making the diagnosis between malignant and benign strictures
extremely difficult [3,23] .

The presence of a dominant stricture (DS) is the imaging hallmark for CCA diagnosis. Chapman et al.
found a correlation between DS and the development of CCA in a retrospective 25-year study of PSC
patients . This relationship has been described in further retrospective and prospective studies, showing an
[27]
increase of risk of 6.2% to 26.3% in patients with PSC with a DS in comparison of patients with PSC without
a DS, with CCA being diagnosed over a 6.2- to 9.8-year follow-up period [28,29] . However, the presence of a DS
for CCA diagnosis is not mandatory since patients without DS are diagnosed with CCA [29,30] . In addition, the