PhD Defence Erik Tanis

colorectal liver metastases - spectroscopy, immune response and radiofrequency ablation

Erik Tanis is a PhD student in the Nanobiophysics group. His supervisor is prof.dr. T.J.M. Ruers from the Faculty of Science and Technology.

Colorectal cancer is the most common form of cancer and the 2nd most frequent cancer related death in Europe. Up to half of patients with colorectal cancer develop liver metastases during their lifetime. This thesis concerns quality improvements in care of patients with colorectal liver metastases.

The first chapter is a general introduction and overview of the care for patients with colorectal liver metastases (CLM), the implementation of spectral tissue sensing in treatment of CLM and the importance of quality assurance in surgical trials. The treatment approach of patients with resectable liver metastases differs from those with unresectable liver metastases. Unresectable metastases can be treated by chemotherapy in combination with local ablative procedures, where radiofrequency ablation (RFA) is the most frequently applied technique. Currently, RFA is only used for patients with unresectable CLM, due to an increased local recurrence rate compared to resection. Although quality improvements (more experience, better imaging and next generation probes) have resulted in lower local recurrence rates, generally RFA remains for most doctors an unacceptable alternative for patients with resectable CLM based on those historical studies. In chapter 3 we describe the local recurrence rates of RFA and resection of patients with CLM enrolled in two randomized controlled trails performed by the European Organisation for Research and Treatment of cancer (EORTC). In these well-documented patient groups we saw that the local recurrence rate per patient was higher for RFA compared to resection due to the more advanced disease status of the RFA treated patients. These unresectable patients had a higher median number of metastases compared to the resectable patients (4 vs. 1). Local recurrence rate per lesion did not differ between resection and RFA, emphasizing the need for future randomized trials on resection and RFA in resectable patients.

In chapter 4 and 5 we describe two translational studies where we use spectral tissue sensing (STS) integrated at the tip of a needle during routine RFA procedures for patients with CLM. STS is a relatively new technique that uses the individual tissue properties of reflectance and absorption of light in order to characterize a tissue (‘optical fingerprint’). Integrated at the tip of a needle, STS can be used to probe and investigate tissue at locations where visual or manual feedback is not possible, e.g. centrally in the liver, and possibly could improve the local control of a RFA procedure. In chapter 4 we have investigated whether STS could detect the heat-induced changes of tissue during and after RFA procedures in order to find a spectral marker for irreversible tissue damage. First we have investigated these changes in human blood samples and continued our study during ex vivo and finally in vivo measurements during routine RFA procedures for patients with CLM. We have shown STS could be used safely in vivo, and it could reliably determine tissue coagulation. In chapter 5 we have investigated the STS technique into clinical practice and used the STS needle during eight routine surgical RFA procedures for CLM. STS measurements were taken at the edges and in the tumor tissue and correlated to histological biopsies of these sites. Radiological markers were also placed for follow-up CT scans in order to evaluate whether the STS measurements were actually taken in the ablation zone. We have shown that STS could accurately differentiate (>96% accuracy) ablated tissue from non-ablated tissue (regardless whether it was tumor tissue or liver tissue). In a subset of patients continuous STS measurements were taken. It showed that there was no interference of the STS measurements due to the RFA procedure and also that the change in spectrum over time correlated well with the histological degree of thermal tissue damage using nicotinamide adenine dinucleotide diaphoresis (NADH staining). In chapter 6 we have investigated whether STS could be useful during routine hepatic resections for CLM. Similar to our previous described study, the STS measurements were taken inside and around the tumor and were correlated to histology. It showed STS could accurately identify and differentiate normal liver tissue from tumor tissue (sensitivity 95%, specificity 92%) and that bile was the most discriminative parameter. Moreover, pretreatment with systemic therapy did not hamper the diagnostic accuracy. In chapters 4 to 6 we have shown that STS, integrated at the tip of a needle, can accurately identify and discriminate tumor tissue and ablated tissue from normal liver tissue and could therefore be a useful technique for further quality improvement for both resection and RFA of patients with CLM.

It has been known for decades that our own immune system plays an important role in cancer recurrence and survival. Several studies in colorectal cancer and CLM show immune cells (e.g. lymphocytes) in the tumor or at the tumor border are associated with prolonged progression free survival (PFS) and overall survival (OS). In recent years there have been multiple studies on perioperative chemotherapy for patients with CLM, however there are no studies on the effect of chemotherapy on this (peri)tumoral immune response. In chapter 7 we conducted a study on the peritumoral immune response of 82 patients included in the EORTC EPOC trial (resectable CLM treated with or without perioperative FOLFOX chemotherapy). In our two patient groups (38 chemo vs. 44 chemo-naïve patients) we have investigated 6 types of immune cells (CD3+, CD4+ and CD8+ lymphocytes, B-lymphocytes, macrophages and mast cells) in three zones around the tumor border (normal liver tissue, at the tumor border, and tumor tissue). We have investigated the absolute number of immune cells and correlated it to pathological response to chemotherapy and PFS. Our study showed that lymphocytes (CD3+ and CD4+) and macrophages were the most common immune cells overall. Higher densities of CD3+ lymphocytes at the tumor border and mast cells inside the tumor are associated with increased PFS. Chemotherapy treated patients had significantly higher densities of CD3+ lymphocytes and mast cells inside the tumor and less CD4+ lymphocytes and macrophages in the surrounding normal liver tissue as well as less macrophages at the tumor border. Similarly, higher densities of CD3+ lymphocytes inside the tumor and mast cells overall was correlated with pathologic response to chemotherapy according to the ‘tumor regression grade’ (TRG) scoring system. Concluding (1) chemotherapy seems to positively influence a peritumoral immune response and (2) a specific peritumoral immune response is associated with improved PFS. This increase in PFS could partly be assigned to the chemotherapy-induced immune response.

In this thesis we investigate several aspects on quality improvement in care for patients with CLM. In general, quality assurance (QA) in scientific trials is an essential aspect in improving care as change in medical practice can only be achieved by high quality scientific trials. Chapter 8 concerns a study on the assessment of quality assurance in surgical trials, where we investigate surgical parameters reported in trials performed by the EORTC and develop a QA framework for future surgical trials. Fifty-one EORTC trials that included any kind of surgery, conducted between 1980 and 2013 were included in this review. Protocols, case record forms (CRF) and final publications were reviewed for surgical parameters and compared between eras (80-ies, 90-ies and 2000-2013). Surgical trials varied from a true surgical comparison (comparison of 2 surgical techniques or comparison of surgery vs. no surgery) and non-surgical comparisons where surgery was part of the protocol, but not subject of investigation or surgery was optional. Overall in more than 90% of protocols the surgical technique was described, but over the years a trend of more detailed description was seen (e.g. surgical technique, resectability and margins). However important surgical parameters, like reasons for unresectability, reoperation or other surgical complications were not consistently described, collected or published. The lack of a standardized surgical chapter and not enough surgeons participating in protocol development could account for these inconsistencies. We have developed a standard surgical chapter template for future surgical trials in which important (surgical) perioperative parameters and aspects for quality assurance are described. With this surgical chapter we hope to further improve the quality of surgical trials en subsequently future patient outcome.