Recently, Shim et al conducted a retrospective analysis to determine the optimum radiologic response criteria to precisely predict the outcome of 332 patients with intermediate stage HCC after TACE. Of note, they selected patients with intermediate stage HCC which were classified as BCLC stage B according to BCLC system as a control for tumor burden and Child-Pugh grade A to control for liver function. All the included patients were treated similarly with conventional TACE and followed-up with CT scans which were evaluated using WHO, RECIST, EASL, and mRECIST criteria. They reported that patients with a complete response by enhancement-based criteria (EASL or mRECIST) had longer survival than those who had any other type of response. Moreover, a follow-up assessment according to EASL (p<0.001) and mRECIST (p<0.001) criteria was found to be an independent predictor of survival. Of note, these findings could not be achieved for WHO and RECIST criteria which are size-based.19

In addition, Kim et al performed a retrospective analysis which included 314 patients with intermediate stage HCC (BCLC B and Child-Pugh grade A) to assess predictors of treatment outcome. Noteworthy, they standardized the treatment protocol and follow-up regimen using conventional TACE and contrast enhanced imaging. They evaluated the responses only by mRECIST criteria. In multivariate analysis, they reported that both the initial response according to mRECIST criteria (p<0.001; defined as the response at the first follow-up scan) and the best response by mRECIST (p<0.001; defined as response after on-demand retreatment when necessary) were independent predictors of overall survival.20


Continue Reading

Actual survival fit the predicted survival in many studies better when patients received the recommended treatment according to BCLC 0/A compared to other stages such as BCLC B, C or D. This was found to be better after subclassification of BCLC B according to Bolondi or Kinki criteria,11Kudo, 2016 #321,.22 However, this was still not the optimum prognosis achieved for this intermediate stage. According to the current prognostic scores and the evolution of many TACE-specific scores which are tailored to intermediate stage HCC, there is strong growing evidence that the optimum score should include tumor number, size, vascular invasion, etiology, and TACE response in addition to liver function parameters such as serum albumin and bilirubin. Some of these were already proposed in HAP score and its modification.15,23–26

In general, many recently developed scoring systems such as ALBI, PALBI, ALBI-T, and modified ALBI-T have shown better performance compared to conventional Child-Pugh score and BCLC staging system. Also, indocyanine green retention test has been reported by several studies as a beneficial tool to predict liver failure. However, the performance of most of these scores was not optimum in patients with intermediate stage HCC, especially those undergoing TACE in comparison to other TACE-specific scores such as HAP score and its modified versions.15,26

Recently, Campani et al conducted a comparative study that included 1,058 patients with intermediate stage HCC where they compared TACE-specific scores such as HAP, mHAP II, and mHAP III to other general grading systems such as ALBI and PALBI grades. They concluded that mHAP III was superior to all of these scores including ALBI and PALBI grades. Noteworthy, even HAP and mHAP II were superior to ALBI and PALBI grades. So, the future direction is toward TACE-specific scores which would offer more precise patient selection and improve response to therapy.26

Among the cohort classified as “ideal candidates” for TACE, an expected median survival in the order of 30 months was quoted, but even within this patient group there was a wide variation in survival.27–29 However, in practice, many patients receive TACE outside the guideline criteria. For example, because TACE is apparently “safe” even in the presence of vascular invasion, the latter is not always considered a contraindication.30 In this expanded population, variation in survival may be even greater. This wide variability in survival has led to attempts to define the prognostic features and combine these into scores (or “models”) that can be applied to assess prognosis at a sub-group or individual patient level. One frequently quoted aim is to identify the sub-group that responds poorly to TACE, which may be more appropriately treated with less aggressive therapies.31,32

For years, many scores have been proposed and despite the scientific criticism of the Child-Pugh score, guidelines still adopt it in the assessment of liver functions. It is clear that a variable percentage of patients with HCC do not have cirrhosis, but they have a range of liver pathology varying from mild pathological changes to advanced fibrosis, and the resultant liver dysfunction relies on the state of the non-tumorous liver and extent of the tumor. Child-Pugh score’s weak points appear in the subjective assessment of encephalopathy and amount of ascites together with the inclusion of inter-related variables such as ascites and serum albumin level. In addition, the individual parameters are scored based on arbitrarily defined and predetermined cutoff points.33–36

Of note, one of the reasons for the concrete adherence to Child-Pugh score is that it is easily calculated and uses daily routine parameters. Nowadays, smart phones are abundant and hospital information systems can include any complex calculator. This was shown in MELD score which is widely used regardless of having a complex equation. Despite the lack of validation of any of these scores to fit clinical decision-making, there was an agreement on the need for a substitute for the Child-Pugh score to assess liver functions in patients with HCC, in addition to a reliable specific score to predict response to treatment.


One of the first TACE-specific scores was the HAP score, which is based on baseline pre-TACE serum bilirubin >17 mmol/L, albumin <36 g/dL which are, respectively, the upper and lower limits of the normal range; AFP >400 ng/mL and tumor size >7 cm in order to guide initial TACE treatment. Kadalayil et al proposed this score after reviewing 114 sequential patients with HCC who received TACE/TAE and the minimum follow-up was 6 months. The patients received one point for each, where HAP A stage has 0 points and HAP D stage has >2 points (Table 4). They concluded that HAP stages C and D had bad prognosis and benefited poorly from TACE.37

Of note, Park et al proposed a modification of the previously published HAP score (Table 5) by removing the bilirubin parameter and incorporating portal vein involvement and mRECIST criteria response. Further validation of the score showed no superiority compared to HAP score.38Nevertheless, Cappelli et al also modified the HAP score by incorporating tumor number, where single tumor receives 0 points while two or more lesions receive 1 point. The proposed model showed a Harrell’s c‐statistic of 0.649 (95% CI: 0.610–0.688), which was more significant than that of the original HAP score (0.589; 95% CI: 0.552–0.626; P=0.001), in addition to better performance in comparison to the modified HAP‐II score (0.611; 95% CI: 0.572–0.650; P=0.005).39

In addition, Ogasawara et al proposed CHIP score (Table 6) which consists of CP score of the patients together with number of tumors and presence of HCV as an etiology. CHIP score also included presence of hepatitis C viremia as a factor which might be good for patients receiving direct-acting antivirals. However, we come again to the weak points due to the presence of CP score and the predetermined cut-off points.

In late 2017, Op den Winkel et al proposed Munich TACE (Table 7) score which included AFP, serum bilirubin, prothrombin concentration, creatinine, CRP, and tumor extension. The score gives 0–6 points for each element, resulting in three stages. In early 2018, they also did further validation of the proposed score and it showed AUROC of 0.71 which was superior to TACE-tailored CLIP, HAP, JIS, GETCH, BCLC, CP, Okuda, and STATE scores.40,41

(To view a larger version of Table 7, click here.)