December 10, 2020 — Journal of Hepatology
Refers to: Contrast-enhanced ultrasound can guide the therapeutic strategy by improving the detection of colorectal liver metastases, Journal of Hepatology, October 14, 2020
Contrast-enhanced ultrasound (CEUS) was introduced around 20 years ago and hepatological applications have been the mainstay of its use since then. Several studies have demonstrated its high accuracy in characterizing focal liver lesions, not only to distinguish benign from malignant lesions,1 but also to accurately diagnose either hepatocellular carcinoma in the setting of background cirrhosis2,3 or benign entities, such as hemangioma, focal nodular hyperplasia or adenoma.4 Extensive work has been carried out in this field, with hundreds of publications dealing with CEUS of focal liver lesions. Since the early works, CEUS has also been shown to detect liver metastases missed by radiologic techniques.5 However, the fact that CEUS can detect some lesions missed by other radiological techniques does not confirm whether it is worth recommending in all or in specified categories of patients. To this end, International Guidelines are unspecific,6 when they state that “CEUS can be used for liver metastases detection as part of a multimodality imaging approach”, as a weak recommendation. Indeed, not enough evidence has been produced in almost 20 years to draw firm conclusions on this topic. Additionally, these guidelines state that “CEUS is recommended in patients with inconclusive findings at CT or MR imaging” as a strong recommendation.6
In this issue of the Journal of Hepatology, Sawatzki and colleagues7 produced a study finally answering some of these questions. They performed CEUS in all patients with a new diagnosis of colorectal cancer in which the staging CT showed no clear metastasis (in a total of 296 patients), regardless of whether no lesion at all, presumably benign lesions or lesions of uncertain dignity had been identified. One strength of the study, among others, is the prospective design, which allowed for the analysis of consecutive patients. Interestingly, a significant portion of this population showed either lesions of unclear dignity at CT (20.9%) or presumably benign lesions (percentage not reported), probably a higher rate than we would have assumed. In this patient population, CEUS was able to definitively characterize the vast majority (98%) of lesions of uncertain dignity at CT, avoiding more expensive and time demanding MRI. In 229 cases, CT and CEUS were consistent in either ruling out any focal liver lesion or in correctly diagnosing specific benign entities (assuming these were cysts or hemangiomas or focal fatty sparing or fatty infiltration). Among the remaining cases, CEUS was able to detect 8 additional metastases not diagnosed by CT. These 8 additional lesions were found in patients that were judged as free from focal liver lesions or affected by lesions considered benign (hemangiomas or cyst) or of unclear dignity at CT. In 6 of these 8 patients the finding of a liver metastasis modified the treatment strategy. These 8 cases were all found in patients with T3/T4 tumor stage, but independent of the presence/absence of nodal involvement or CEA levels above or below 5 ng/ml. Therefore, considering this patient population (T3/T4 tumor stage without definite liver metastasis at CT), the number of CEUS exams needed to detect 1 additional metastasis would be 24.5. Conversely no additional metastasis was found in patients with T1/T2 tumor stage. If CEUS was targeted to only T3/T4 colorectal tumor stage patients with at least 1 lesion detected at CT, despite apparently not being malignant, the number of CEUS exams needed to detect 1 additional metastasis would be even lower than 24.5, maximizing the cost-effectiveness. In fact, the role of CEUS in detecting metastases in patients in whom staging CT did not find any focal lesion was not specifically addressed by the study. However, trying to dissect this information from the study results, it seemed that all the 8 additional metastases identified by CEUS were already visible at CT (2 classified as hemangiomas, 1 as a cyst, and 5 as lesions of unclear dignity). According to these data, it appears possible that CEUS has no strategic role in patients in whom CT did not find any lesion at all, but no firm conclusion can be drawn at present, as the number of such patients was not specified in the work.
Altogether, this study provides important new information, that definitively confirms the statement that “CEUS is recommended in patients with inconclusive findings at CT or MR imaging”6 but also suggests that the next update of the CEUS guidelines could upgrade the slightly vague statement “CEUS can be used for liver metastases detection as part of a multimodality imaging approach” to a stronger recommendation for staging purposes, at least in the type of patient population analyzed in this study.
There are some additional points worthy of comment. The ultrasound explorability was good or moderate in 89% of patients and explorability correlated with BMI. Mean BMI was in fact 24.5 kg/m2 in patients with a good explorability, 27.7 kg/m2 in those with moderate and 30.6 kg/m2 in those with poor explorability.7 As expected, these data confirm that CEUS is probably not particularly useful in obese patients; if technically possible, MRI might be preferable in these patients.
Whether patients with liver metastases, but with a potential for liver surgery, might benefit from preoperative CEUS to maximize the sensitivity of metastasis detection remains an open question.
One single operator performed all CEUS exams with high-end equipment. This abolished interobserver variability and maximized performance, which makes sense for study design. However, the message that is derived is that it is probably better to refer this type of patient to expert centers, which is not impossible in view of their limited numbers. A key message for our readers is that great expertise and high quality equipment can also be provided by hepatology/gastroenterology centers, as CEUS can be performed not only in radiological but also in clinical units. Moreover, the need for expertise stands valid for all techniques. The CT images of 3 of the 8 missed metastasis were reclassified as metastasis when re-evaluated by another radiologist,7 who was, however, not blinded to the clinical information. This suggests that the level of expertise might also be highly relevant for CT. However, it must be acknowledged that there is a possibility for a second opinion on the same images with CT, which is hardly possible with CEUS.
Author: Fabio Piscaglia, Vito Sansone, Francesco Tovoli
Division of Internal Medicine, IRCCS Azienda Ospedaliero Universitaria di Bologna, Italy