Intraoperative ultrasound for prediction of hepatocellular carcinoma biological behaviour: Prospective comparison with pathology

Preoperative prediction of both microinvasive hepatocellular carcinoma and histological grade of hepatocellular carcinoma is pivotal to treatment planning and prognostication. The aim of this study was to evaluate whether some intraoperative ultrasound features correlate with both the presence of same histological patterns and differentiation grade of hepatocellular carcinoma on the histological features of the primary resected tumour.


| INTRODUCTION
Disease recurrence after hepatic resection (HR) for hepatocellular carcinoma (HCC) might be predicted by pathological findings such as differentiation degree, presence of satellites and the existence of vascular microinfiltration. These features could allow patients to be shortlisted for liver transplantation without proven malignant disease. [1][2][3] The presence of microinvasion (MI) defined as portal venous, hepatic vein, bile duct infiltration and/or intrahepatic metastasis (satellite nodules) is the best indicator of poor prognosis after HR and transplantation; in fact, in this study of Yamashita et al, 4 the recurrence-free survival of the MI-positive group after HR was significantly worse than that of the MI-negative group. Furthermore, the histological grade of HCC is a significant prognostic factor after surgery, and high-grade HCC was found to be an independent predictor of vascular microinfiltration. [5][6][7][8][9] Therefore, preoperative prediction of both microinvasive hepatocellular carcinoma (MI-HCC) and histological grade of HCC is pivotal for treatment planning and prognostication. In this high-risk group of patients, prevention and treatment of recurrence could improve the data of survival. 10,11 However, histological grade of HCC and MI patterns are difficult to detect preoperatively with the current advanced imaging modality. Recent studies have shown that positron emission tomography, 12 diffusion-weighted magnetic resonance imaging (MRI) 13 and contrast-enhanced ultrasound 14 can be used to predict the histological grade and/or the presence of MI. Furthermore, one study 15 showed that the use of noninvasive radiological criteria can predict biology of HCC accurately and correlated radiological patterns directly to outcome. Our recent study 15 suggested that HCC recurrence is associated with HCC nodules presenting more aggressive biological behaviour and these patterns could be identified by the intraoperative ultrasound (IOUS) evaluation: IOUS remains as the best real-time imaging method for evaluation of HCC nodules during laparoscopic and open liver surgery.
In this prospective, double-blind study, we prospectively classified, for the first time to our knowledge, pattern combinations for each primary HCC nodule obtained by IOUS during HR. Then, we investigated whether these IOUS patterns correlate with both the presence of same histological patterns and differentiation grade of HCC on the histological features of the primary resected tumour.  20 Differing from the BCLC treatment protocol, 19 in the time interval of this study we did not consider nodule size and number as absolute exclusion criteria from surgical treatment.

| Patients' characteristics
In this series, the residual liver function was evaluated using the Child-Pugh classification 21 and MELD score. 22 Upon referral, laboratory tests including complete blood cell count, coagulation profile, liver functions, plasma levels of alpha-foetoprotein (AFP) and a chest x-ray were performed. This study was approved by Institutional Review Board at Milan University. The patient records were anonymized and de-identified prior to analysis, and informed consent to storage the clinical data of patients with HCC was obtained from each participant.  According to the histological definition proposed by Yamashita, 4 MI-HCC was defined as a tumour with vascular/biliary infiltration and/or the presence of satellites.

| Histopathological evaluation
The pathology reports collected prospectively after surgery were used as a reference of standard for this study. All lesions were confirmed

| Statistical analysis
The primary objective was to evaluate the ability of IOUS to depict the same pathological characteristics of HCC. The secondary objective was to determine whether IOUS patterns could be used to differentiate histopathological grade of HCC. Data following a normal distribution were expressed as mean ± standard deviation. If data were nonparametric, median and interquartile range values were reported.

| RESULTS
We enrolled patients who underwent curative HR for HCC at our uni-

| DISCUSSION
Several studies demonstrated that the presence of vascular microinfiltration, satellites and/or high-grade HCC affects disease-free survival of patients with solitary HCC after curative HR. 1,4,5,7,10,11 Therefore, it is important to predict these features prior to treatment (in the absence of a surgical specimen).
The use of preoperative tumour biopsy is controversial both for the risk of complications (which is because of associated coagulopathy and the possibility of tumour spreading) and for the interobserver variability to define the histological grade. In fact, few studies have evaluated, with conflicting results, the accuracy of percutaneous biopsy in comparison with the surgical specimen: HCC tumour differentiation is not homogeneous 8,26,27 and vascular microinfiltration is difficult to detect at the preoperative biopsy. 7,9 Also because of these reasons, many factors and scoring systems evaluated by preoperative noninvasive procedures have been proposed to predict MI-HCC and/or HCC tumour differentiation. 7 needs a prospective validation, while its single-centre nature suggests substantial agreement among radiologists with a wide range of experience: in fact, other studies confirmed that there is no standard for MRI evaluation and the reproducibility of MRI features is still being investigated. [32][33][34] Other studies demonstrate that preoperative 18 F-FDG PET-TC is useful for predicting vascular microinfiltration in small HCC. 12 However, the low sensitivity of 18 F-FDG PET-CT for detection of primary HCC was a major concern in the prestudy planning. In particular, well-differentiated and early HCC may not be detected using FDG-PET. 12,35 Yamashita et al 4 demonstrated that not only the presence of microinvasion such as portal venous, hepatic vein, bile duct infiltration, but also the presence of intrahepatic metastasis (satellites) is the best indicator of a poor prognosis after HR and transplantation.
Therefore, MI-HCC may be considered an important entity (superior to the only presence of vascular microinfiltration) to determine the prognosis of patients submitted not only to HR or liver transplantation 4,7,36 but also to ablation therapies. [37][38][39] Our study showed that IOUS definition of MI-HCC had a high agreement with the histological findings, such as the satellites detection and the correct measurement of HCC diameter (see Table 3) and the type of IOUS examination (through either a laparoscopic or an open approach) did not influence these results (see Table 4). This analysis is also valid in the subgroup of patients with HCC diameter ≤2 cm (see Table 5). It is important to outline that vascular microinfiltration showed weaker correlation with histological features: the main reason is intrinsic to IOUS definition of "vascular microinfiltration" that is either presence of vascular encasement or strict contact between vascular wall and nodule margins. In this last case, it is possible that the contact with vascular wall does not represent a histological encasement ( Figure 2). Furthermore, multivariate analysis showed that vascular microinfiltration and infiltrative HCC aspect are related to tumour differentiation. With recent improvements in technology, IOUS has now be-