Imaging and histopathological data.
Of the 175 patients who were included,
122 (69,7%) patients had evidence of viable tumor on imaging examination,
79 of them having undergone treatment before LT.
The remaining 53 patients had no evidence of viable tumor and 40 of them had undergone treatment before LT.
On histopathology,
125 patients (71,4%) had evidence of viable tumor,
while there were no viable HCCs present on the explanted liver in 50 patients.
The other collected imaging and histopathological features are summarized in table 2.
Correlation between imaging and histopathology.
A strong association was found between imaging and histopathology in the detection of the presence of viable tumor (p<0,001).
This correlation was demonstrated to be highly significant also in patients who had undergone treatment before LT (p<0,001) and among them in those with previous TACE (p<0,001) and radiofrequency ablation,
RITA (p=0,006) .
It remained in patients who had CT as dynamic-imaging evaluation performed closest to LT (p<0,001) as well as in those who had MR (p=0,007).
Strong associations between imaging and histopathology were also found in the assessment of the number of viable nodules (p<0,001),
diameter of the largest nodule (p<0,001),
overall viable tumor size (p<0,001) and lobar distribution of HCCs (p=0,03).
Accuracy of imaging techniques.
Of the 125 cases with pathologically proven viable tumor,
102 were found on both imaging and histopathology while 23 were missed at imaging.
Twenty cases were overdiagnosed (positive on imaging and negative on pathology) while in 30 patients there was no viable tumor on both imaging and pathology.
Therefore,
in the detection of viable tumor,
the sensitivity of imaging techniques was 81,6%,
specificity was 60% and accuracy was 75%.
The total population was then divided into patients who had MDCT as dynamic-imaging evaluation (n=138) and patients who had MR (n=37).
Results about the performance of imaging,
MDCT and MRI in the detection of viable tumor are summarized in table 3.
The McNemar test showed no statistically significant difference between imaging and histopathology in the detection of viable tumor (p=0,75).
Representative cases of concordance and discordance of MDCT and MRI with histopathology are shown in figure 1 and 2.
Recurrence rate and predictive factors.
The median follow-up period was 39,1 months.
Of the 175 patients included,
24 (13,7%) had recurrence of HCC after the LT.
Time to development of the recurrence ranged from 1,55 to 41,85 months after LT,
with a median value of 12,36 months.
The 1-,
3-,
5-years cumulative disease-free survival rates according to the Kaplan-Meier method were 92,96%,
83,9% and 82,84%,
respectively.
The mean duration of recurrence-free survival was 40,15 months.
Univariate analysis revealed various factors significantly associated with recurrence (Table 4).