SRFA pioneer calls on interventional radiologists to start training now, “before it is too late”

Reto Bale

Exactly 20 years on from his first stereotactic radiofrequency ablation (SRFA) procedure to treat a cancerous liver tumour, Reto Bale (Medical University Innsbruck, Innsbruck, Austria) speaks to Interventional News about the future of this procedure, and warns interventionalists that this may be the next arena for a potential turf struggle: interventional radiologists should start learning now, he urges, as, in his view, “90% of current liver resections could be replaced by SRFA”. At Bale’s institution, proceduralists are able to use up to 36 coaxial needles to ablate tumours, achieving results that rival those of their surgical counterparts, leaving no residual mass behind.

When did you first start using this approach?

The story begins in 1994, when I was using navigation systems in ENT and neurosurgery. Back then, as a student, I developed a few things together with my student fellow Michael Vogele: We had a patent for a navigation device and a head fixation device, amongst other technologies. In 1995 I developed the first aiming device initially for neurosurgical work. I wanted to expand beyond neurosurgery, because I wanted to use navigation not just in the brain, but everywhere. That was why I moved to interventional radiology (IR), and in 1996, as an assistant doctor, I set up the SIP—the Stereotactic Interventional Planning lab in Innsbruck, Austria.

I think stereotaxy is the ideal tool for radiofrequency ablation (RFA). The problem is that there is often a discrepancy between the tumour volume and the ablation volume. This is the reason for recurrence. The simple solution to this is to use multiple coaxial needles. However, although the idea is simple, it is difficult to place multiple coaxial needles in order to ablate the entire tumour volume with conventional ultrasound (US) and computed tomography (CT) guidance. This is why, in my opinion, using stereotaxy is the way to go.

We did the first procedure in the liver in May 2001. I am sure that the way we do radiofrequency ablations has to change—I have always been sure, but when we started this everyone said ‘Oh that crazy Austrian, advancing 10 needles into the body!’ But now, 20 years later, people have changed their minds.

What value does stereotaxy add to radiofrequency ablation?

SRFA helps you not only to guide the needles and visualise placement, but to plan the procedure as well. In most cases, you need multiple overlapping zones in order to ablate the complete volume. Even with microwave ablation (MWA), the maximum short ablation diameter is approximately 3.6cm. So, if a lesion is larger than 2–3cm, it is very difficult to achieve an adequate ablation margin; you need overlapping ablation zones.

How do results from SRFA compare with surgical outcomes?

It [SRFA] is actually better than surgery!

From our results, it is at least as good as surgery in terms of local control rates, and it is dramatically less invasive, of course, and parenchyma-sparing. There is no doubt in my mind: if you can ablate the lesion completely, why would you do a resection? No patient would like you to open them up if there was a less invasive option with comparable outcomes.

In 2019, myself and several colleagues published a retrospective study in Hepatology that evaluated the efficacy of three-dimensional (3D)-navigated, multiprobe radiofrequency ablation (SRFA) with intraprocedural image fusion for the treatment of hepatocellular carcinoma (HCC) by histopathological examination.

From 2009 to 2018, 97 patients were transplanted after bridging therapy of 195 HCCs by SRFA. The rate of residual vital tissue could be assessed in 188 of 195 lesions in 96 of 97 patients by histological examination of the explanted livers using haematoxylin and eosin (H&E) and Tdt-mediated UTP nick-end labeling (TUNEL) stains. Histopathological results were compared with the findings of the last CT imaging before liver transplantation. Complete radiological response was achieved in 186 of 188 nodules (98.9%) and 94 of 96 patients (97.9%), and complete pathological response in the explanted liver specimen in 183 of 188 nodules (97.3%) and 91 of 96 patients (94.8%), respectively. In lesions ≥3cm, complete tumour cell death was achieved in 50 of 52 nodules (96.2%). Residual tumour did not correlate with tumour size (p=0.5).

We therefore concluded that multiprobe SRFA with intraprocedural image fusion represents an efficient, minimally invasive therapy for HCC, even with tumour sizes larger than 3cm, and without the need of a combination with additional treatments. For us, the results justify the additional efforts related to the stereotactic approach.

A very interesting thing about SRFA is that image fusion is part of the procedure. Image fusion is now—now [Bale’s stress]—a key issue. We have been doing this for 20 years.

What developments would you like to see in image fusion?

Firstly, we use image fusion to check the accuracy of needle placement. We do two CT scans: a planning CT, before we place any needles, and a control CT, which is taken with the needles in place. This native dataset is then superimposed onto the planning CT, and onto the virtual pathways we planned. Thereby, we can measure the accuracy of every needle.

The second step is the verification of the ablation zone—we want to know if there is a sufficient ablation margin covering the tumour. This is extremely important, and very helpful. It is better than the pathological report the surgeons receive, for example. One week following surgery, they get a pathology report telling them if they have achieved R0 or R1, whereas we see this immediately, while the patient is still under anaesthesia, and we see it in three dimensions, not only in a two- dimensional rendering. In most cases it is not possible for the pathologists to look at the whole margin under the microscope, it would take far too long! So, they just take a representative specimen. But when we use image fusion, it is a 3D tool, so we can assess the margin in all dimensions.

This allows us to be adaptive intraprocedurally. If we do not achieve the ablation zone, and we have some residual tumour, we can put another needle into the patient.

In our tumour board meetings, the first question used to be “Is it resectable?” Now the first question is – in most cases – “Is it ablate-able?” But this is only at Innsbruck! I hope it will change around the world—this is just the start for SRFA.

What are the barriers to this being adopted at other institutions?

Getting the word out! Our problem is that we do not get refunds for the additional effort. Our department gets money for the ablation probes, like you would need for an US- or CT-guided intervention, which is completely different from SRFA. We need support from insurance companies, and they will only finance this if there are enough data. Currently, we only have data [on SRFA’s safety, efficacy and long-term outcome] from our institution.

We have a prospective database. It includes more than 1000 SRFA patients with more than 4000 tumors treated by SRFA over the last 20 years. It is  linked to the hospital’s record of patient files, so it is hard data that is very transparent and auditable. All of our replications are based on this registry.

However, in order to support reimbursement for SRFA, insurance companies need to see prospective, randomised data—my next aim is therefore to conduct a multicentre, prospective, randomised controlled study to obtain these data and hopefully to get reimbursement for the procedure.

What would your advice be for interventional radiologists wishing to do SRFA at their own centres?

Get in touch! I would be happy to assist them to set up a programme. I have done it in Innsbruck, and it works very well, so I would be delighted to assist. It is hard work at the beginning, organising the system and teaching everyone the technical details, but I think it is necessary that more interventional radiologists have this training.

The good thing is that, though it was initially challenging, now we have so much experience. We have many standards of practice; it is much easier to train someone to do SRFA than it is to perform liver surgery, for example. I have a colleague who has completed just three months training, and then he did a procedure completely alone, with great results. My other colleague has been training with me for a year, and he performs extremely complicated procedures that no one else does in the world, outside of Innsbruck.

A very important point: you can do a lot of training without even touching a patient, because we have such extensive datasets, with information (such as CT images) available. So you can train if you want on more than thousand patients before you start with real patients! That is really cool, because the path planning is the most difficult part. It is quite easy to use the navigation system to place a needle; I will turn you into an expert in two weeks, with that, as long as you have some experience with ablation already. But it is difficult to make the path plan; this can be learned in the virtual dataset. My aim is therefore to teach interventionalists with these datasets; I think we could make something big happen here.

It is extremely important that interventional radiologists start now, before it is too late, because I know that otherwise the surgeons will start performing SRFA themselves. In my view, 90% of current liver resections could be replaced by SRFA. If it is possible to completely ablate the tumour and leave no residual mass behind, then there is no argument for a more invasive procedure.

What does the future look like for SRFA?

My hope is that it will become a lot more common practice at several institutions around the world. In addition, at Innsbruck, we aim to make the preprocedural path-planning process at least semi-automatic, through the use of machine learning. We have just recently secured funding for a PhD candidate to get started on this work, so it will be exciting to see what they can achieve.


Please enter your comment!
Please enter your name here