Liver Resection for Colorectal Cancer Metastases. Prognostic Factors and Interventions Affecting Outcome

Sammanfattning: Background: With a yearly incidence of 56 per 100,000 inhabitants, colorectal cancer (CRC) is one of the most common malignancies in the Western world, representing the third leading cause of cancer-related mortality. Within three years after the diagnosis of CRC, 29% of patients will have developed liver metastases. Liver resection is a possibly curative treatment, and several preoperative strategies have been developed to make initially irresectable tumours resectable. Preoperative chemotherapy may decrease the tumour burden, but also damages the liver. Portal vein embolisation (PVE) is used to increase the future liver remnant prior to advanced liver resection, but has been suggested to promote tumour growth. The influence of risk factors and interventions on outcome after liver resection needs to be further investigated. This includes studying prognostic factors concerned with histopathological features of the primary, as well as the expression of tumour markers. Identifying prognostic factors may help to predict outcome and to tailor treatment to the individual patient. Aim: The aim of this thesis was to analyse prognostic factors and interventions affecting the outcome, both short-term and long-term, after resection of CRC liver metastases.Patients and methods: I: The perioperative course in 97 patients treated with preoperative chemotherapy prior to resection of CRC liver metastases was compared to that in 136 patients without preoperative chemotherapy. II: 17 patients with preoperative chemotherapy and PVE were compared with 17 matched controls, as well as with 75 unmatched controls, with preoperative chemotherapy only. III: Tumour progression after PVE was studied in 34 patients with CRC liver metastases, comparing tumours in the embolised and those in the non-embolised liver lobes. IV: An artificial neural network (ANN) model was constructed on 241 patients, to predict survival after liver resection. V: Predictive factors for overall survival (OS) and disease-free survival (DFS) after liver resection were analysed in 100 patients, with focus on histopathological features of the primary colon cancer, such as lymph node ratio (LNR). VI: Expression of cancer stem cell markers CD44, CD133 and CD24 was analysed in colonic liver metastases in 67 patients, and its predictive value for OS and DFS was determined.Results/Conclusions: I: No significant differences were found in perioperative course, when comparing patients with and without preoperative chemotherapy. II: Perioperative course was similar between patients with preoperative chemotherapy and PVE as compared to those with preoperative chemotherapy only, except for operation time. III: Tumour progression was seen in the embolised lobe in 3/34 patients and in the non-embolised lobe in 3/23 patients, with a median decrease in tumour volume of 16% and 11%, respectively. IV: The ANN model selected six risk variables and had a C-index of 0.72, showing a better performance than the Cox regression model. V: LNR and perineural invasion of the primary colon cancer can be used as predictors for DFS after liver resection, whilst vascular and perineural invasion of the primary colon cancer are predictive for OS. VI: CD133 expression in colonic liver metastases was selected as a predictor for DFS.