| Objective To investigate CT characteristics and compare the detectable rate of refined multi-phase contrast-enhanced CT in hepatocellular carcinoma (HCC) by using 64 multi-slice spiral CT (MSCT), and then optimize the scanning program. Meanwhile, to initially evaluate the value of quantitative indicators such as time-density curve (T-D curve), maximum enhancing rate (Emax) and washout rate in diagnosising HCC.Materials and Methods From April 2009 to January 2010,40 patients who were first diagnosed as HCC and confirmed afterwards pathologically or clinically entered and informed the study. Nonenhanced CT and dynamic contrast-enhanced CT were performed in all above-mentioned lesions before treatment. The scanning started at 22-24s(early arterial phase, EAP),29-31s(mid-arterial phase, MAP),37-39s(late arterial phase, LAP),60-65s(early portal venous phase, EPVP),80-85s(late portal venous phase, LPVP)and 150s(liver equilibrium phase)after initial injection of contrast medium.The CT findings of tumor were observed in nonenhanced CT and dynamic contrast-enhanced CT. The density difference values of the tumor and normal liver in each phase were calculated and compared. The display of tumor were analyzed in each phase, and also the numbers of detected lesions were calculated. The detectable rate and positive predictive value (PPV) were compared and evaluated byχ2 test in terms of HCC lesions with the largest tumor diameter≤3cm and>3cm as well. EAP, MAP and LAP were further optimized, and then four programs were worked out such as EAP plus MAP, MAP plus LAP, EAP plus LAP. The combined detectable rates were compared.HCC and hepatic metastases (HMs) cases would be performed controlled study, in which 40 HM cases clinically proved as the control group also be performed the above-mentioned examination with the same parameters as the HCC group. The density values of the solid part of tumor were measured and statistically compared for HCC and HMs. According to the density measurements, each of the T-D curve was drawn. The trends were analyzed and statistically classified, the distribution of T-D curve type were compared usingχ2 test. Peak time, maximum enhancing rate and relative/absolute washout rate were calculated and statistically compared between groups, where count data were analyzed usingχ2 test and measurement data using two sample t-test or Mann-Whitney rank sum test. We used ROC curve to evaluate the HCC diagnostic performance of enhancing rate and relative/absolute washout rate. The results were presented as mean±standard deviation (x±s) that the test was significant level ofα=0.05, p<0.05 indicated significance for all tests.Results1.61 HCC lesions were detected in 40 patients. The density difference values were significantly different between the solid part of tumor and normal liver parenchyma in each phase (H=45.208, p<0.01) in which the largest difference was in LAP. With the distribution of peak enhancement and the degree of enhancement in the different phases, we found the peak enhancement mostly appeared in LAP.2. The significant difference of detectable rate (χ2=32.910, p=0.000) was found among in simple EAP, MAP, LAP, EPVP and LPVP, with the highest one in LAP (78.689%).The lesions were divided into two groups by the largest diameter of tumor:≤3cm group (small hepatocellular carcinoma, SHCC) and>3cm group, both of which have a highest detectable rate in LAP with 85.000%,75.610%, respectively. The lesions which appeared as high-density at arterial phase and/or low-density at portal vein phase were classified as positive while the lesions which were confirmed clinically or pathologically were considered as true positive. The positive predictive value (PPV) in each phase of>3cm HCC group could reached 100.00%, but the positive predictive value (PPV) reached 100.00% only in PVP of≤3cmHCC group.3. The EAP, MAP and LAP could be combined to four scanning programs, the detectable rate of EAP plus LAP and all three arterial phase were consistent and the highest(≤3cm HCC Group:90.000%;>3cm Group:78.049%).4. The density values of the solid part of tumor were all significantly different in each phase between HCC and HMs group(p<0.01). Type A, B, C, D and E T-D curves were obtained on the basis of CT enhanced values. The curve type distribution was significantly different between HCC and HMs (χ2=67.037, p=0.000). Type A,C belonged to typical curves of HCC, while type B,D and E belonged to HMs. The peak time (PT) distribution between the two groups also existed a significant difference (χ2 =25.609, p=0.000).5. Significant difference was present in maximum enhancing rate (Emax) between HCC and HMs (t=2.944, p=0.004); the relative/absolute washout rate (Washr and Washa) also had statistically significant in the LPVP and liver equilibrium phase. The maximum enhancing rate (Emax) of HCC was higher than that of HMs, and the washout rate of HCC was faster than that of HMs in the LPVP and liver equilibrium phase. We evaluated the diagnostic performance by using ROC curve:The sensitivity and specificity in diagnosing of HCC by maximum enhancing rate (Emax) were 68.9%,60.0%, respectively; the sensitivity and specificity in diagnosing of HCC by Washr and Washa in liver equilibrium phase were 59.0%, 90.0%and 67.2%,82.5%, respectively. The maximum enhancing rate (Emax) of tumor≥88.02%, Washr≥15.22% and Washa≥24.30% in liver equilibrium phase were both suggestive of HCC.Conclusion1. In the dynamic contrast-enhanced CT examination, the best detectable time of hypervascular HCC was LAP, particularly significant for small hepatocellular carcinoma(SHCC).Optimization protocol of double arterial phase(EAP+LAP) plus (late) portal venous phase plus liver equilibrium phase was valuable for the detection and diagnosis of HCC.2. Dynamic contrast-enhanced CT examination could make a significant difference in qualitative and quantitative diagnosis for HCC. More specifically, the T-D curve could directly and objectively reflect the enhanced features of HCC and make a preliminary diagnosis for some atypical HCC.
|
PDF Full Text Request