One of the consequences of surveillance programs is the identification of smaller HCCs, as well as dysplastic nodules. The smaller the HCC lesion, the more difficult for us to distinguish malignant from benign nodules. This is true both radiologically and histologically. Recently, a distinction has been made between “very early HCC” [29, 30] and “small” or “progressed” HCC [31, 32]. Early HCC, as defined by Japanese pathologists, is generally hypovascular and has ill-defined margins [9]. Thus, it has a somewhat vague outline on ultrasound and may be hypovascular on CT scanning.

Histologically, there are few unpaired arteries, but the cells show varying grades of dysplasia [33]. There may be invasion of the portal space by hepatocytes, but vessel invasion is absent. These lesions have been called “very early HCC” in the Barcelona Clinic Liver Cancer (BCLC) staging scheme [34]. The pathology of these “very early HCC” lesions has been defined in resected specimens, and therefore, the natural history of these lesions is unclear. However, the presence of small foci of typical HCC within them has been noted, suggesting that these lesions are precursors of typical HCC lesions [29, 31, 32]. The frequency with which these lesions develop typical HCC is unknown either. In contrast, “small” or “progressed” HCC has well-defined margins on ultrasound and exhibits the typical features of well-differentiated HCC on CT and on histology [29, 31, 33]. These lesions often show microvascular invasion, despite their small size [33]. The presence of microvascular invasion suggests that the prognosis of these lesions is less good than for “early HCC” where vascular invasion is rare. However, this has not been proven in clinical studies.

The classification and description of dysplastic nodules and early HCC have been recently revised to harmonize the approaches taken by Western and Japanese pathologists [31]. These studies have been carried out on resected tissue, whereas samples from lesions detected on surveillance usually only have a needle biopsy to evaluate. It is important to understand that rather than being individual discrete states, there is a continuum between HGDN and HCC. This complicates the evaluation of biopsies from small nodules. Patients with liver nodules having a nonspecific vascular profile and negative biopsy should continue to undergo enhanced follow-up as well. There are no data to establish the best follow-up policy so far, but repeated biopsy or follow-up CT/MRI to detect further growth should be considered. There are emerging data indicating that the smaller the lesion, the less likely there is to be microscopic vascular invasion [9].

In addition, smaller lesions are more likely to be associated with treatment that will be curative [30, 35, 36]. Finally, decision analysis also confirms that ideally, for the best outcome, the lesion should be smaller than 2 cm at diagnosis [37]. However, it is equally important not to apply invasive treatment to lesions that do not have any malignant potential and may still regress. This is a fine distinction that is not always possible to make. Another concern about thin needle liver biopsy is the risk of bleeding and needle track seeding. Most studies that report needle track seeding do not specify the lesion size being biopsied [38]. Although the rate of needle track seeding after biopsy of small lesions (<2 cm) has not been accurately measured, it is probably uncommon. The current rate of bleeding from thin needle biopsy of small HCC has not been reported but is probably no different than for biopsy of the liver in general [38, 39].

The European Association for the Study of the Liver expert panel proposed the following surveillance recall and diagnostic strategy (Fig. 1.2) [40]. In nodules smaller than 1 cm, which are malignant in less than 50 % of cases, reliable HCC diagnosis is difficult. Thus, close follow-up is recommended. In nodules of 1–2 cm, HCC diagnosis requires positive cytohistology. However, there is a 30–40 % false-negative rate with fine-needle biopsy [41]. A negative result does not rule out malignant disease. Noninvasive diagnostic criteria to be applied solely in patients with cirrhosis and tumors larger than 2 cm were proposed. HCC diagnosis is established by the concomitant finding of two imaging techniques, showing a nodule larger than 2 cm with arterial hypervascularization, or by one positive imaging technique, showing hypervascularization associated with alpha-fetoprotein concentration higher than 400 ug/L [27]. The accuracy of imaging techniques is rapidly evolving. Ultrasonography plays a key part in the detection of HCC, but its sensitivity to detect additional small nodules is low [42]. New contrast agents might increase the accuracy of this technique and might be relevant to assess treatment response [43]. The best techniques are helical CT and MRI with contrast enhancement, which have an accuracy exceeding 80 % [44]. CT and MRI sensitivity decreases when assessing the extent of the disease. In a comparison of CT and MRI with pathological examination of explanted livers, MRI angiography was more precise in detecting nodules of 1–2 cm than CT scan [12]. However, 20–30 % of intrahepatic tumors, especially those smaller than 10 mm, are not diagnosed preoperatively with any technique. Preliminary data with positron emission tomography are not encouraging for detection of small nodules.