Quaglia, Alberto; (2002) Mechanisms of progression in hepatocellular carcinoma. Doctoral thesis (Ph.D), UCL (University College London).

Text Mechanisms_of_progression_in_h.pdf Download (47MB)

Abstract

In order to investigate the genesis of hepatocellular carcinoma (HCC) and the nature of its putative precursor nodular lesions in cirrhotic liver, a systematic light microscopical analysis was combined with a molecular approach including investigation of clonality and genomic heterogeneity. Histological scoring system: The current histological classification of HCC and its putative precursor nodular lesions is unsatisfactory. Most of the histological criteria for the evaluation of early HCC and its putative precursors have not been validated properly and the diagnosis of these lesions still depends on subjective interpretation. If the hypothesis of the progression from cirrhotic liver, to putative precursors and eventually to HCC is accepted, then the division of this evolutionary biological continuum into mutually exclusive categories becomes problematic. With this in mind, a histological scoring system was designed with the intention of making systematic the histological assessment of HCC and its putative precursor nodular lesions in cirrhotic livers. This histological scoring system was based on a methodical and detailed analysis of those individual histological features considered in the literature to be helpful for this purpose. To each category (e.g. capillarisation, arterialisation, reticulin loss, cellular atypia etc) a grade was assigned and at the end of the assessment the individual grades were added up to generate a final score. The scoring system was tested on 106 nodular lesions from 35 patients, which were also designated according to standard diagnostic categories. The histological scores were compared to the diagnostic classification. There was good correlation between the two assessments. In particular, there was good correspondence between the score ranges and the diagnostic assignments (i.e. most macroregenerative nodules (MRN) scored between 0 and 3, most dysplastic nodules (DN) scored between 4 and 8, and all HCC scored more than 9 points). In addition a category of nodular lesions with features reminiscent of focal nodular hyperplasia, in cirrhotic livers, became evident using this scoring system. This final histological score should give an indication of the position of a lesion in the benign to malignant evolutionary spectrum. Multiple correspondence analysis showed that hepatocellular nodules in cirrhosis assessed using the scoring system, (as well as the individual histological variables used for the calculation of the score) can be represented as a continuum on a "malignancy scale". In order to validate this scoring system, further work was performed to compare scores with molecular and clinical follow-up data. The correlation between each individual component of this histological score and molecular and clinical data was also examined. Combined molecular and histological analysis: This histological scoring system was used to assess a group of nodular lesions from cirrhotic patients, on which molecular analysis was also carried out. This molecular analysis consisted of the assessment of clonality and genomic heterogeneity. Clonality using the HUMARA technique was carried out on 35 nodules obtained from 15 female cirrhotic patients. Genomic heterogeneity was assessed on 26 nodular lesions from 11 cirrhotic patients. The combined analysis of clonality and genomic heterogeneity was carried out on 14 nodular lesions from 8 cirrhotic patients. Monoclonality is said to identify neoplastic cell populations but does not indicate malignancy. Malignancy is associated with tumour progression, with increasing tumour grade and tumour heterogeneity as malignant neoplasms acquire the capability of invasion and metastasis. The identification of genomic heterogeneity within a neoplastic population (defined for present purposes by monoclonality) implies that spontaneous genomic mutations have occurred which may eventually lead to frank malignancy. Clonality: My analysis of clonality showed that the majority of HCC (90%) are monoclonal, approximately half of the MRN are monoclonal (53%), and the majority (70%) of samples of regenerative nodules are polyclonal. The methods available to define clonality are limited by the recent appreciation of the concept of "patch size", i.e. tissues are composed of groups of cells of variable size that have the same pattern of X-chromosome inactivation depending on the embryonic stage at which this "Lyonization" process occurs. For many human tissues, including liver, this patch size is not known. An attempt was made to explore this question with glucose-6- phosphate dehydrogenase (G6PD) enzyme histochemistry on liver samples obtained from Sardinian female patients heterozygous for the Mediterranean variant of G6PD deficiency. Genomic heterogeneity: Genomic heterogeneity was studied using a genomic fingerprinting technique (arbitrarily-primed polymerase chain reaction) on 26 nodular lesions (11 HCC, 2 DN and 13 MRN) from 11 cirrhotic patients. The fingerprints of two halves of each nodular lesion were compared and each lesion was defined as homogenous when the two sets of fingerprints were identical and heterogeneous when they were different. All HCC and DN were heterogenous. Two MRN were heterogeneous. 8 MRN were homogeneous. Three MRN for which only one set of genomic fingerprints was available, showed a different pattern when compared to normal reference tissue. The degree of genomic alteration was also quantified calculating the genomic damage fraction (GDF), i.e. the degree of difference between lesional and normal reference tissue. This showed that the genomic damage of HCC (average GDF= 0.28, std =0.11) is greater than DN and MRN (average GDF= 0.08, std =0.1) (p value < 0.0001). When all nodular lesions (MRN, DN and HCC) were considered together, there was a positive correlation (correlation coefficient 0.53, p value < 0.001) between genomic damage and lesional size. This positive correlation was also seen when HCC alone were considered (correlation coefficient 0.33, p value <0.02), but not for MRN or DN. Combined analysis of clonality and genomic heterogeneity: Combined analysis of clonality and genomic heterogeneity was carried out on 14 nodular lesions (5 HCC and 9 MRN) from 8 cirrhotic patients. Three MRN were polyclonal and showed intralesional genomic homogeneity and were identical to normal reference tissue. The remaining 6 MRN were monoclonal and showed intralesional genomic homogeneity. Five of these showed some degree of genomic difference when compared to background regenerative nodules and normal reference tissue. Four out of five HCC were monoclonal, and all showed intralesional heterogeneity. These results suggest that the evolution from putative precursor nodular lesions to HCC involves an initial polyclonal and genomically stable phase, a presumably subsequent monoclonal phase with evidence of genomic mutation, and a third monoclonal phase with increasing intranodular genomic heterogeneity, corresponding to the morphological category of overt HCC. Correlation between molecular analysis, histological scoring system and clinical data: When the genomic damage was correlated with the histological scoring system, there was a positive correlation (correlation coefficient = 0.76) between increase of genomic damage and increase in histological score (p value < 0.001). This correlation was maintained when the size component of the histological scoring system was ignored (correlation coefficient = 0.76), indicating that there is a correlation between genomic damage and histological changes assessed by the scoring system, independently of lesional size (p value < 0.001). Moreover, when the nodules were reassigned as "benign" or "malignant" using the scoring system (nodules with score 9, respectively), the correlation between histological score and genomic damage improved (MRN correlation coefficient from 0.34 to 0.66, HCC correlation coefficient from 0.68 to 0.73, p value < 0.001). No significant correlation was found between histological score and genomic damage when DN were considered (however only two examples of DN were available for molecular analysis). No correlation was evident between clonality and histological score. There was no significant correlation between histological score, genomic damage and survival and tumour recurrence data. At follow-up, 9 patients (including 5 patients with HCC) were all alive, without any evidence of recurrent tumour (average follow-up 610 days, range 124-1322 days). One patient with a single 20 mm diameter HCC in the liver removed at transplantation (histological score =15, GDF=0.31), died 12 days after transplantation due to graft failure. One patient with a single 70 mm HCC in the explant liver (histological score 39, GDF=0.39) had intrahepatic tumour recurrence diagnosed histologically at 958 days post-0LT and died at day 1140 post-0 LT of carcinomatosis. Further clinical follow-up is needed to assess the outcome of the other patients. Conclusion: A combined approach including systematic histological assessment and molecular analysis is helpful in the investigation of the genesis and evolution of HCC and its precursors and in the accurate identification of the significant morphological features that are expressed during human hepatocarcinogenesis. More work is needed to identify those individual histological features predictive of clinical progression and the features with prognostic significance

Download activity - last month

Export as CSVJSONXML

Download activity - last 12 months

Export as XMLJSONCSV

Downloads by country - last 12 months

Export as CSVXMLJSON

Archive Staff Only

View Item