Tumor angiogenesis is vital to tumor growth, invasion and metastasis [14-15]. The results of angiogenesis lead to the formation of proliferous vasculature including the surrounding feeding vessels. A few studies have shown that the growth and progression of tumor are greatly associated with an increase in feeding vessels through the process of angiogenesis [16]. The feeding vessels of solid tumors usually originate from two sources. One source is the existing vascular network of the tumor body, morphologically mature and active blood vessels [17-18]. Another source is angiogenetic microvasculature formed through the process of tumor vascularization [19-20]. The angiogenetic microvasculature is characteristic with irregular vascular size, shape and branching pattern. HCC is typical hypervascular tumor with abundant network of feeding vessels, most of HCC demonstrated a homogeneous or heterogeneous hyperenhancement in AP of CEUS. A few studies have shown that microvascular density, vascular invasion are the strong predictors of tumor outcome and recurrence for HCC [21-23]. Therefore, the entire network of feeding vessels should be resected or ablated during surgical resection or interventional ablation therapy. Incomplete treatment may result in some survival vessels and residual tumors.
Conventional 2DUS is unable to image the vascular network of tumor, yet color Doppler flow imaging (CDFI) can only reveal the main branches of the network because of the poor sensitivity to low-speed blood flow. In small HCCs, the sensitivity of CDFI in showing abnormal arterial neovascularity is low, and flow can be demonstrated in less than 50% of the lesions [24-25].
CEUS is relatively new US imaging modality that can show the dynamic contrast perfusion of tumor body including capillary. In AP, CEUS is able to demonstrate an integrated wash-in hyperenhancement of intratumoral and peritumoral tumor vessels, while in LP the hyperenhancement quickly wash out and HCC becomes hypoenhanced [26]. This dynamic phase changes provide another two opportunities to measure HCC besides 2DUS.
In this study, we found that the HCC sizes of 2D, AP and LP are significantly different. The maximal longitudinal and transverse diameters of AP were 9.7% and 9.8% larger than that of 2D, and were 14.8% and 14.7% larger than that of LP. Likewise, the 2D longitudinal and transverse diameters were 5.6% and 5.4% larger than that of LP. The interobserver agreement between two readers was excellent, so there were no significant differences on size measurement between the readers. AP diameter of HCC appeared to be maximal followed by 2D diameter, and LP diameter was the minimal. Despite the order of three sizes was clearly seen in most of HCC cases in the scatter plot (Fig. 3), we also found the data dispersion of individual case vary greatly, indicating the heterogenicity of tumor angiogenesis.
We prefer to use routine 2DUS for HCC general screening high-risk subjects, because it is more convenient, time-saving and less costly. For HCC patients requiring surgical resection or interventional therapies, especially radiofrequency ablation, we could recommend AP as more precise measurement. It is obvious that AP diameter includes tumor vascular construction not only intratumoral but also peritumoral vascular invasion. So, it is understandable that AP diameter is larger than 2D diameter which does not contain peritumoral vascular construction. Considering the importance of peritumoral vascular invasion in HCC prognosis and recurrence [27-28], AP diameter should be selected for therapeutic reference.
In CEUS, LP usually provides the best contrast resolution for HCC imaging or size measurement. However, LP diameter was the minimal when compared to AP and 2D diameters in this study. LP diameter may underestimate HCC size probably because some well-differentiated HCC tend to iso-enhanced and incomplete washout in LP. The iso-enhancement may shrink the sizing range of HCC. The quick and significant washout is observed more frequently in HCC with poorer grades of differentiation than in well-differentiated HCC [29-31]. Thus, it may be inappropriate to use LP diameter for treatment reference.
There are some limitations in this study. Firstly, the study was designed and limited to only HCC population. Secondly, it is obvious that AP cannot substitute CT or MRI for tumor size measurement. The sections in ultrasound imaging are usually flexible and acoustic window dependent, not as standardized as in CT or MRI. The US imaging cannot match CT or MRI in section, therefor the comparison is not appropriate. So, we think that AP can only correct the bias of underestimation in ultrasound measurement. Thirdly, the HCC size did not compare with gross pathological size, which is the gold standard. This research is a preliminary, retrospective study not a randomized controlled trial. Further large sample study is needed to prove the results.