The COLLISION trial: Interview with Prof. Martijn Meijerink

January 21, 2025

COLLISION is a randomized, international, multicentre, phase III non inferiority trial investigating surgery versus thermal ablation for small-size colorectal liver metastases. The final results of the COLLISION trial show that minimally invasive thermal ablation may indeed offer similar outcomes to surgical resection in terms of overall survival, representing a potential shift in clinical practice.

With the publication of the study in The Lancet Oncology on January 20, we spoke to primary investigator Prof. Martijn Meijerink to learn about the background of the trial.

CIRSE: From a personal point of view – can you tell us about your own history and experience beginning to use thermal ablation for CRLMs? At what point did you realize that this could be a viable first-line treatment comparable to surgical resection?

Meijerink: I began my radiology residency in early 2003, and I performed my first open radiofrequency ablations (RFAs) for liver tumours later that same year as a junior resident. This opportunity arose because the radiologist responsible for these procedures left the hospital, and there was no one else to assist in the operating room. From that moment, I began to form a strong professional relationship with the liver surgeons.

Together, we developed a comprehensive follow-up protocol, which allowed us to observe that most of the ablations were complete. By around 2008, we started to question why thermal ablation, a less invasive procedure, was only considered for unresectable liver tumours. With the advent of percutaneous ablation techniques, the procedure became even less invasive, enabling patients to go home either the same day or the following morning.

Prof. Meijerink at his hospital, Amsterdam University Medical Center.

CIRSE: Take us back to the beginning – what is the background of the COLLISION trial?

Meijerink: Despite the advantages, discussions during tumour board meetings — which, I know, are similar worldwide — often led to the frustrating realization that we would not be able to change global treatment guidelines without conducting a well-designed large-size randomized controlled trial. Before we would be able to design such a trial, we needed to do several things:

1. Claim equipoise and objectify the fact that retrospective comparative series previously labelled as ablation inferior to resection based on the overall survival data. With the publication of our systematic review and meta-analysis (Puijk et al. CVIR 2018), where the analyses were conducted by an independent epidemiology bureau, we could conclude that the risk of residual bias was very high because patients only received ablation when somehow surgery was not an option in all these retrospective series.
2. Create a consensus as a community of interventional radiologists and liver surgeons on what should be considered resectable and what should be considered ablatable. Using the Delphi consensus technique, we published clear definitions and a decision flowchart (van Nieuwenhuizen, Cancers 2020). Furthermore, with a lot of discussions on what the preferred endpoint should be in the field of ablation, we used a similar technique and created consensus on specific time-to-event endpoints in liver tumour ablation (Puijk et al. Radiology 2021)
3. Create a multidisciplinary review panel to assess potential cases where surgeons were merely asked if they would be able to resect and interventionists if they would be able to ablate the tumours, regardless of the specific approach used within that centre.

CIRSE: Can you give us a quick summary of the results of the trial? Was there anything that surprised you?

Meijerink: Between August 2017 and February 2024, 300 patients were randomly assigned to two groups: 148 to the experimental treatment and 148 to the control treatment. The trial was stopped at halftime because it met certain criteria for early benefit and it would be unethical to continue accrual. There already was a >90% chance that ablation would be non-inferior regarding overall survival when compared to surgical resection. As expected, thermal ablation resulted in fewer side effects (28 cases [18.9%] vs. 67 cases [45.9%]) and considerably shorter hospital stays. Serious side effects occurred in 11 patients (7.4%) in the ablation group and 29 patients (19.9%) in the surgery group. These included bleeding that needed intervention (1 patient [1.0%] vs. 8 patients [5.5%]) and infections requiring treatment (6 patients [4.0%] vs. 11 patients [7.5%]). Treatment-related deaths occurred in 0 patients with ablation and 3 patients with surgical resection (2 due to heart complications after surgery, 1 due to sepsis and liver failure). Surprisingly for all of us, local control was equal to superior following ablation.

Conclusion: The idea that thermal ablation should only be used for tumours that cannot be surgically removed needs to be reconsidered.

CIRSE: Were you able to define when thermal ablation might be more suitable than surgery?

Meijerink: Though I strongly recommend the preferred treatment to be individualized and based on clinical characteristics and available expertise, the recruiting centres, including the Amsterdam UMC, have changed their internal protocols and now favour thermal ablation over surgical resection if a patient is eligible for both. In short, in my opinion, based on the COLLISION trial results, surgical resection ought to be reserved for unablatable tumours in centres with sufficient expertise in both ablation and surgery.

CIRSE: What challenges have you faced in conducting the COLLISION trial? Conversely, what have been the high points for you personally along the way?

Meijerink: Conducting a direct comparison between the outcomes of different medical specialties often involves political debates and biases, which have been the primary reasons why previous attempts to recruit a significant number of patients have failed. However, I must acknowledge the invaluable multidisciplinary collaboration with liver surgeons and the support from several surgical and oncological professional societies, including the European-Asian HPB Association and the Dutch Colorectal Cancer Group, which have been crucial to the success of this effort. As the trial unfolded, the surgical panellists, realizing that most of the tumours ablated at their centres were fully treated, became more selective and started turning down cases with deeper-seated tumours. This change in focus created a real challenge in ramping up patient recruitment during the final two years of the trial.

CIRSE: What response do you expect from the oncological community on the COLLISION trial?

Meijerink: In my opinion, the single most important aspect of providing optimal care for our patients is fostering a close and professional relationship with your colleagues from surgery and medical oncology. This collaboration is essential for ensuring that we focus on what is truly in the best interest of the patient. By working together, sharing knowledge, and respecting each other’s expertise, we can make more informed decisions, tailor treatments more effectively, and ultimately improve patient outcomes. Strong interdisciplinary relationships are the foundation of a patient-centred approach, where the collective goal is always the well-being and care of the patients we treat. The COLLISION trial may have been a collision between two approaches; it also helped us to create a multidisciplinary community of physicians focused on improving outcomes.

CIRSE: The results from the COLLISION trial were published in The Lancet Oncology yesterday – what do you expect will be the impact of having your results published in such a highly visible journal?

Meijerink: A publication in Lancet Oncology will undoubtedly help reach the widest possible audience and raise awareness of the findings. While we believe that these results should influence the adoption of updated guidelines, we are fully aware that the outcomes largely depend on the dedication and local expertise of the treating centres. Therefore, a shift in the guidelines, with thermal ablation as the new standard of care, must be accompanied by the implementation of best practices from high volume dedicated centres to the less experienced centres. This will ensure that the high-quality care demonstrated in our study can be consistently delivered across all institutions and geographical regions.

CIRSE: COLLISION-XL and COLLISION RELAPSE are still ongoing. What additional information do you hope to gain from these two trials?

Meijerink: The COLLISION XL trial is a randomised controlled trial comparing thermal ablation to stereotactic body radiotherapy (SBRT) for larger-size and unresectable colorectal liver metastases 3-5cm. We hope this ongoing study will clarify if we should prefer SBRT or thermal ablation, and furthermore it will tell us the local control rates of these ablative modalities for larger-size tumours and shed some light on the question if we should adhere to the 3cm upper size limit for liver tumour ablations. The COLLISION RELAPSE trial is a trial comparing upfront repeat local treatment with thermal ablation and/or surgical resection versus neo-adjuvant chemotherapy followed by local treatment for patients with new colorectal liver metastases in the first year after the initial local treatment.

CIRSE: For you personally – is there anything you remain curious about or wish to investigate further? Where will your research interests take you next?

Meijerink: We appear to be entering a golden era for interventional oncology, with the majority of patients with liver (and many other) malignancies now being treated using minimally invasive interventional techniques. So, the simple answer to this question is a resounding yes—I remain deeply curious about many aspects of this field. There is still much work to be done, and we will need the collaboration of everyone in our community to make it happen.

The exciting paradigm shift lies in combining local ablations with local or systemic immunotherapies, using robotic devices and confirmation software to optimize outcomes, and expanding these techniques to other organs such as the prostate, kidney, lung, pancreas, musculoskeletal system, and, who knows, possibly even the brain in the future. These developments should be the focus of our future research.

Currently, we are designing two key trials: the COLLISION-2 implementation trial, a stepped-wedge cluster trial comparing current practices to best practices, and the COLLISION-3 trial, which will focus on conventional CT-guided ablation with or without the use of stereotactic robotic navigation devices to optimize results and reduce radiation exposure for both physicians and patients.

Prof. Meijerink will speak about the COLLISION trial at CIRSE 2025. Stay tuned!