HCC is one of the most common malignancies globally, with incidence increasing yearly by 3–9%, posing a huge global health risk. HCC patients typically have a low 5-year overall survival rate and poor clinical prognoses, due to the delay in diagnosing the disease. As a result, effective biomarkers have been long sought to increase HCC detection rates, for guiding individualized patient treatments [19, 20]. Some examples of biomarkers examined include G-test, AFP, and AAR. In this study, we also investigated these 3 biomarkers, both alone and in combination, to determine their predictive capabilities for diagnosing HCC. We found that G-test, AFP, and AAR levels were closely related to the occurrence of HCC, and that out of the 3 parameters, G-test had the greatest diagnostic value for predicting HCC, compared to AFP and AAR, as determined by AUC values. In particular, G-test had the highest specificity and sensitivity, compared to the other 2 markers, for LC and healthy individuals. However, for CH patients, AAR had the greatest sensitivity. Overall, though, the combination of all 3 biomarkers provided the most optimal sensitivity, specificity, and predictive capabilities for diagnosing HCC.
AFP was first proposed as a tumor marker for HCC in the 1960s and has since been widely used for clinical HCC detection. However, its utility for screening and diagnosis of HCC has been criticized, due to its low sensitivity and specificity [21,22,23]. It has been noted in multiple studies, though, that AFP sensitivity and specificity for HCC varied with the AFP threshold value; some studies showed that an AFP threshold of 20 ng/mL yielded sensitivity and specificity values for HCC detection of, respectively, 41-65% and 80-94% [24]. Increasing the AFP threshold from 20 to 50 ng/mL, though, yielded a significantly increased specificity value of 96%, along with a positive predictive value of 75%. On the other hand, sensitivity was only 47% [25]. Therefore, a lower AFP threshold value would yield increased sensitivity, but lower specificity, possibly leading to a greater risk for HCC false positives. These findings were in line with what was observed from 77 HCC patients in our study, in which at AFP thresholds of 7, 100, and 400 ng/mL, sensitivity for HCC was, respectively, 58.4%, 41.6% and 27.3%, while the rate of missed diagnosis increased from 41.6 to 72.7%. All these observations thus indicate the necessity of combining AFP with more effective biomarkers to obtain a better detection strategy for HCC. AAR has been used as an indicator to evaluate liver fibrosis in chronic liver disease [9, 26], and more recent studies have suggested that it can also be used to distinguish cirrhosis from HCC [10], with a sensitivity of 75.9% and specificity 55.7%. Additionally, AAR has been independently associated with early recurrence of HCC [11]. In this study, we evaluated the diagnostic ability of AAR, in terms of AUC, for HCC, and found that it had the highest value among CH patients, compared to G-test and AFP. This higher measurement is in line with AAR also having the greatest sensitivity among those patients. However, AAR, compared to G-test and AFP, had the lowest specificity for detecting HCC in CH.
Abnormal structural changes in liver glycosyltransferase have been noted as a key feature in the development of HCC, which could be reflected in the resulting serum N-glycan branching [27]. A study from Liu et al. [13] evaluated changes of the serum N-glycan spectrum among HCC patients using DNA sequencer-aided fluorophore-assisted carbohydrate electrophoresis, and the ratio of log (peak value 9/peak value 7 in the N-glycan spectrum) was named the GlycoHCCTest, or G-test for short. G-test has been found to be an effective and non-invasive means for detecting HCC among cirrhosis patients [28]. Our G-test results demonstrated that the test values was much higher among HCC, compared to CH and LC patients, which was consistent with the HCC developmental process progressing from hepatitis to cirrhosis to liver cancer experienced by most HCC patients. G-test was found by Wan et al. [16] to be superior to AFP in screening for liver cancer among patients with chronic hepatitis B and cirrhosis. Additionally, the combination of both parameters further improved the diagnosis rate for hepatitis B virus-related liver cancer. These findings were in line with our study, which evaluated a larger sample size and added AAR as a biomarker of interest, along with G-test, and AFP. We found that G-test was significantly better than AFP in distinguishing between those who developed HCC from those who did not, including healthy, CH, and LC individuals, while AAR was the most optimal only for differentiating between HCC and CH individuals. Nevertheless, the combination of G-test, AFP and AAR demonstrated the highest diagnostic capability, suggesting that this was likely the optimal approach for detecting HCC onset.
Cirrhosis and inflammation during HCC development complicate the early diagnosis of HCC. Due to the high rate of false negatives from AFP for HCC, biomarkers for AFP-NHCC have recently become a significant topic of interest. A study from Li et al. found that the gamma-glutamyl transpeptidase (GGT) to alkaline phosphatase ratio, combined with GGT to AST ratio and AAR, were effective diagnostic markers for AFP-NHCC [12]. In our study, we further analyzed the diagnostic ability of G-test and AAR for AFP-NHCC, and found that G-test, as well as the combination of G-test and AAR, was able to effectively detect AFP-NHCC. By contrast, AAR was significantly less effective for diagnosing AFP-NHCC. Furthermore, although both approaches used AAR levels as part of diagnosing HCC, our method was able to diagnose HCC onset using G-test and AFP, on top of AAR. The Li et al. method focused on detecting AFP-NHCC, while our method focused more on diagnosing HCC in general, being able to predict the occurrence of HCC [12], whether AFP-positive or negative (AFP-NHCC), among healthy, cirrhotic, and hepatitis (non-cancerous) patients. In addition, the Li et al. method was most effective during the early stages of AFP-NHCC, when the tumor size is small [12]. Nevertheless, additional studies are needed to unravel the true association between AAR and AFP-NHCC.
Changes in IgG antibody-linked oligosaccharides, in terms of both types and levels, have also been used as diagnostic markers for the onset and progression of various types of cancer. For instance, Kanoah et al. found that NSCLC progression was associated with significant decreases in mono- (Fr1) and digalactosyl (Fr2) IgG oligosaccharide levels, coupled with increases in agalactosyl IgG oligosaccharide (Fr4) [29]. Similar changes, entailing decreased Fr1 and F2, coupled with increased Fr4, were previously observed among that research group for prostate cancer [30]. More recently, changes in glycosylation patterns were observed in epithelial ovarian cancer patients, compared to healthy ones, with respect to IgG1, 2 and 3. In particular, IgG1 had significantly lower sialylation, and higher fucosylation, among cancer patients, while those patients also had increased agalactyosylation, along with decreased digalactosylation and sialylation, for IgG3. These alterations, especially for agalactosylation, were also found to be positively correlated with the widely utilized diagnostic marker CA125 [31]. However, the methods used to identify the oligosaccharide chains, such as fluorophore-associated carbohydrate electrophoresis and matrix-assisted laser desorption/ionization coupled with time-of-flight mass spectrometry are approaches that may be difficult to apply for widespread clinical use. Therefore, less cumbersome methods may need to be developed before IgG oligosaccharide chains could be utilized as a clinical diagnostic marker.
Moreover, it is worth noting that other methods for diagnosing HCC through serum diagnostic markers, such as exosomal DNA containing the TP53 gene mutation [32], phenylalanyl-tryptophan [33], miRNAs, such as miR-10b [34], prothrombin induced by vitamin K deficiency or antagonist-II [35], lncRNA-D16366[36], des-gamma-carboxyprothrombin [37], and dickkopf-1 [38], etc., have been documented. However, the widespread adoption for a number of these biomarkers as a diagnostic tool has been limited, owing to low sensitivity, which is exacerbated by HCC often being found alongside chronic liver disease and inflammation [39]. Furthermore, determining the appropriate cut-off values for detecting HCC onset with high specificity and sensitivity, as well as developing cost-effective approaches for measuring serum miRNA, lncRNA, and exosomal DNA levels, is a continued work in progress [39].
Our results provide a new predictor for diagnosing HCC, particularly AFP-NHCC. However, there are still a number of limitations in our study, one of which is its retrospective nature, which may reduce the predictive value of the results. Additionally, the sample size was small, possibly resulting in sampling biases. Lastly, there is a lack of sufficient HCC staging and follow-up data, limiting our ability to evaluate the association between the screening value of indicators with different HCC stages and follow-up findings. Therefore, future prospective studies with large sample sizes, multiple centers and adequate follow-up data collection are needed to validate the results.