In order to improve and individualize cancer treatment personalized treatments developed much further. Colon cancer is treated with surgery and thereafter adjuvant oncological treatment. The selection of chemotherapy is today mainly done according to best guess. Today only a small fraction of oncological treatment may be known to be effective in a person before treatment start, most often it is trial and error. A fast reliable system for looking at response to different treatments in each unique patient is much needed and would, if successful, completely change the way we give oncological treatment today. This system would also be possible to use to evaluate new treatments and if successful, implement in the clinical setting. In this project we will implant a part of the patient's tumour tissue into a zebrafish embryo and evaluate tumour growth and frequency of metastatic disease as well as response to given oncological treatment.

2.2 Objective: The objective of this project is to explore the usefulness of zebrafish (Danio Rerio) embryo models to determine tumor biology and treatment response in colon cancer. An overarching goal would be, before start of any oncological treatment in a patient, to have evaluated the response of oncological treatment in the zebrafish avatar and only treat with a combination of drugs known to have effect against the patient's own tumour.

2.3 Study design: This protocol describes a series of prospective studies in different cohorts of patients with colon cancer to investigate the applicability of zebrafish embryo models. The common denominator of the sub-studies is prospective collection of tumor tissue implanted in zebrafish embryos in order to evaluate if the model is robust enough for growing colon cancer tissue and evaluate growth pattern and response to chemotherapy. This study protocol is designed according to and in adherence with the SPIRIT guidelines.

2.4 Intervention: In all sub-studies the intervention is inoculation of tumor cells in zebrafish embryos followed by observation of tumor behavior and testing of treatments.

The possibilities to predict oncological tumor response for individual patients with GI malignancies in order to tailor a personalized oncological treatment, are currently limited. This means that most of the information used to decide upon oncological treatment pertains to a group of patients rather than individuals. Similarly, most of the chemotherapy regimens used is not individually tailored although it has been shown to be beneficial for certain cohorts of patients. In colorectal cancer, KRAS and BRAF mutation status predict the response to drugs targeting various steps in the EGFR/KRAS/BRAF signaling pathway, whereas MSI/MMR status may predict the outcome of immune checkpoint inhibitors [1]. As chemotherapy, as well as other cancer treatments, may have severe side effects the treatment may not only be ineffective but may also cause severe adverse events even shortening the patient's life and in neoadjuvant treatment inhibit the chance of curative tumour resection. For this reason, pre-treatment response evaluation in an avatar model of zebrafish, would be most valuable.

From a clinical point of view the easiest situation is in the adjuvant setting. At this point the whole tumor is available for analysis and larger pieces of tissues may be used in the avatar model. 5-fluorouracil or capecitabine (a 5-FU prodrug) either alone or in combination with oxaliplatin is the current standard adjuvant treatment in stage III colon cancer and stage II colon cancers with 2+ risk factors. Other types of combination regimens including e.g. irinothecan have so far failed to show benefit in the adjuvant situation.

It is still only in exceptional cases that neoadjuvant chemotherapy is given, then in a try to shrink a non-respectable tumor before surgery.

(Nils ska ev lägga till refs enl Pers dokument) 10 Version 1.0, 2021-04-30 Different models have been explored to evaluate tumour specific response to treatment, but most are slow or have shown mixed results[8]. Today, mouse patient-derived xenograft (PDX) model is the most used and validated to predict response to therapy, but evaluation of oncological therapy takes months [9]. Therefore, the mouse PDX model cannot be used for clinical decision-making. Organoid cultures using patient-derived cancers is a well used in vitro screening tool, with promising results for different tumors[10]. Organoids maintain the genetic characteristics of the original tumor tissue. Still, these models are slow when it comes to evaluating patients´ tissue for treatment decisions and lacks the ability to evaluate metastatic or angiogenic potentials.

Recently zebrafish embryos have been used as avatars for human cancer, first in haematogenous cancers but lately also in solid cancers like PDAC, breast cancer and colorectal cancer [11-14].

The model has a number of advantages, the strongest is response evaluation only 3-5 days after the tumour tissue is implanted in the embryos. The embryos are transparent so tumour growth and spread can be visualized in detail and also quantified in a semi-automatic manner [15]. As the zebra fish embryos own immune system does not respond until day 8 [14], the immune response against the tumour comes from the patient's own immune cells, this is also the reason why xenograft engraftment works in this model. On the other hand. So far, biopsy tissue is not enough, about a cubic centimetre of tumour tissue is needed to generate these avatars. For this reason, needle biopsies are not providing enough material but resected colon tumour tissues are needed in the current protocols.

In zebrafish embryos, colorectal cancer cells have been shown to grow and response evaluation of different chemotherapy combinations have been successful with very good correlation to the response in humans [14].

4.2 Available evidence The value of implanting tumor cells in zebrafish embryos to observe tumor behavior and to test the effects of different anti tumoral agents is mainly unknown in clinical settings.

The earlier studies came from blood malignancies, but much later from solid tumors [11, 12].

The method has been shown to induce a very high success rate of implantation in pancreatic adenocarcinoma with response evaluation of oncological treatment [13]. This has also been shown in small cohorts for CRC cell lines and single cell studies [14]. Still, so far very few studies have been performed in the clinical setting and even fewer where the oncological response to chemotherapy has been compared between zebra fish and the patient.

4.3 Need for prospective trial To get real life evidence for the usefulness of zebrafish models to determine prognosis and to individualize treatment prospective human trials are needed.

11

Inclusion Criteria:

• Colon cancer planned for radical colon resection signed informed consent

Exclusion Criteria:

• Not willing to participate