Stereotactic Ablative Treatment of Liver Metastases

What Does "SBRT" Mean Anyway?

The term can be applied if a radiotherapy treatment is

• delivering a high dose
• to a small target
• in the body region
• in one or few fractions
• in a very precise way
• using image guidance

This is a rather vague definition, but it should give a rough idea.

The confusion in terminology regarding SBRT, SABR and SRS and the consequences are nicely summarized in the following article:

Mohamed H. Khattab et al., "Recommendations for radiosurgery and stereotactic ablative radiotherapy terminology to guide clinical and research practices for the medical oncologist, radiation oncologist, and surgeon", Journal of Radiation Oncology (2020) 9: 1–3

What Does "High Dose" Mean?

In terms of biologically equivalent dose (BED), "high" means at least 100 Gy of total dose. This can be acheived with various fractionation schemes.

What Does "Ablative" Mean?

A treatment with a BED of 100 Gy and beyond may also be called "ablative", which means that the irradiated tissue dies and becomes necrotic.

There are several other ablative procedures. Radiofrequency ablation (RFA) and laser-induced thermal therapy (LITT) use heat to destroy the tumor.

Selective intraarterial radionuclide therapy (SIRT) with Yttrium-90 microspheres and trans arterial chemoembolisation (TACE) are other locoregional ablative therapy options.

Auto Beam Hold (ABH)

The ABH feature of the TrueBeam controls the MV beam automatically, based on the result of the marker detection on the current kV image.

If the marker is detected and within tolerance, it is marked with a green cross on the TrueBeam console:

If the marker is detected but outside tolerance, it is red:

Sometimes the marker is clearly visible to the human eye, but the software fails to detect it. This typically happens when the marker is viewed from the tip and the resulting projection on the kV image is much shorter than the nominal length:

If ABH is active, TrueBeam will stop the beam immediately if the marker is either not detected or out of tolerance.

We have a theory regarding the origin of this "glitch" in the marker detection algorithm. We assume that the algorithm was originally designed for the detection of prostate (gold) markers.

When markers are implanted in to the prostate, the direction of needle movement will be more or less from caudal to cranial. This will therefore be the orientation of the marker itself, once it is implanted.

If a patient is in the supine position on the treatment couch and a kV source rotates around him, the projection length of the marker will be more or less constant. If the marker is perfectly aligned with the cranio-caudal axis, the imaged marker length will always be nominal.

The algorithm tolerates a certain perspective shortening however. See for instance the animation: as soon as the marker appears shorter than about 50% of nominal, the detection fails.

We tried to fool the algorithm by specifying a shorter nominal marker length e.g., 3.0 or 2.5 mm instead of 5.0 mm), but then detection problems arise on the other side, because the marker candidate is longer than the specified object length.

The only solution we found so far for the first patient was to deactivate ABH altogether, and to evaluate the images visually.

Phantom Studies

To prepare for this new technique and to train the staff, we used different phantoms.

For studies which focussed on the moving marker inside a homogeneous phantom, we chose the RUBY from PTW with the film dosimetry insert. Instead of film, we used a cardboard containing the marker:

To assess marker visibility and detectability behind bony structures, we used the CIRS John head, which was also loaded with the marker.


(The square cardbord cutout is compatible with the "John" head.)

Both phantoms were moved around on the treatment couch with the help of the QUASAR Motion Platform and the dedicated software, simulating DIBH and free breathing movements.

The motion platform has a 4 cm longitudinal motion range (-20 mm to +20 mm). The vertical movement of the pedestal for the IR-marker block is always 1/4 of the longitudinal amplitude. Phase shift is constant.

 

 

Stereotactic Body Radiotherapy (SBRT) is one of several ablative therapy options for the treatment of liver metastases. In order to ensure a sufficient local tumor control rate, the biologically equivalent dose (BED) should be at least 100 Gy (assuming an α/β ratio of 10). See for instance the DEGRO recommendations¹.

Recently, we reported the first clinical treatment of this kind in our institute. All images shown on this page are taken either from this first patient or from preparatory phantom studies.

This series of kV images was acquired during the CCW half arc of the second fraction. Every 10° of Gantry angle, a kV image is triggered without interrupting the MV beam (which always comes from the right, orthogonal to the kV beam):

(kV marker tracking during MV treatment as seen by Offline Review. Circle diameter is 1 cm.)

Anticipated Challenges

The technical challenges when treating liver metastases stereotactically are at least twofold:

• Image contrast during IGRT procedures
• Motion

Lung metastases can easily be identified on CT and CBCT images, due to the large HU difference between the tumor and the surrounding lung tissue. In the liver however, the density difference is much smaller. While registered MR images are used to support target delineation in the planning phase, the daily treatement on a typical linac has no MR support. Therefore, the tumor is often not visible on the CBCT image. This is where fiducial markers come into play.

Tumor motion, mainly due to breathing, is another challenge. Two paths can be followed, depending on the capabilities of the patient:

• If the patient is capable of holding his breath for a certain amount of time, Deep Inspiration Breath Hold (DIBH) is the option of choice: the treatment is time efficient because the beam is always on².
• If the patient is unable to hold his/her breath but breathing rate is stable, a 4D planning CT can be acquired and phase gating to a certain fraction of the breathing cycle (e.g., the exhale part) can be considered. An average CT generated from the chosen exhale phases is used for treatment planning. During treatment, a Gated CBCT of the same exhale phases is acquired and used for online matching.

The motion amplitude during the breathing cycle depends on the location of the tumor within the liver, and can be up to 4 cm and more.

It must be noted that external surrogate markers like the IR marker block do not always work as expected: unlike with the QUASAR Motion Platform (see sidebar) and its fixed 1:4 amplitude ratio, there is no simple mechanical translation between external motion and internal motion in the patient. One may observe day-to-day variations in tumor motion even if externally detected breathing motion is exactly the same.

This is why it is important to track deep-seated moving targets directly during the treatment.

Our First Patient: Marker Choice, Implantation, Imaging

For our first patient, we decided to use a single platinum marker of dimensions 0.5 x 5.0 mm. Platinum is preferred to gold because it produces less artefacts in MR.

(The Visicoil platinum marker used in phantom studies.)

A specialist of our radiology department implanted the marker (needle: 21G, 15 cm length) with ultrasound guidance into the periphery of the tumor lesion, exactly as desired. This was preceded by some discussion about the ideal marker location and the required number of markers. Motion phantom studies on the TrueBeam with more than one marker and ABH (see sidebar) suggested that the automatic marker detection algorithm did not always distinguish correctly between the markers, and that it would be safer to use a single marker. On the other hand, there may be a certain risk of spreading tumor cells if the needle goes right into the center of the lesion. Therefore our choice.

One week after implantation, several MR studies were acquired. Only a moderate dark spot on the location of the marker appears on the MR image (MR contrast agent: 9 ml Primovist):

(Registered DIBH planning CT and MR images of our first patient.)

The MR studies were registered to our DIBH planning CT.

Since we did not know at this stage whether the patient would be able to hold her breath long enough for treatment, we also acquired a 4D-CT.

In the following animation, one can see that the marker moves in all three axes during a breathing cycle:

(4DCT with implanted platinum fiducial marker)

While the cranio-caudal amplitude is the largest with 1.2 cm, lateral and dorso-ventral movements are also visible. Both have an amplitude of 0.55 cm. This results in an overall amplitude of 1.43 cm between peak inhale and peak exhale.

As it turned out that DIBH treatment would be feasible, target delineation and planning proceeded on the DIBH-CT.

The GTV (volume: 6.9 cm³) was expanded to the PTV with an isotropic 1 cm margin (see the two structures in the next image).

Dose Planning

For our first patient, we chose a prescription of 5 x 6 Gy prescribed to the 60% isodose of the PTV. This gives a fraction dose of 10 Gy in the center of the lesion and a BED of 100 Gy.

The treatment plan contained two 6FFF half arcs, optimized with the Eclipse Photon Optimizer 16.1 and calculated with Acuros XB 16.1.

The plan was then dosimetrically verified with Varian Portal Dosimetry, see the 2 % / 2 mm Gamma results for the hCW and hCCW arcs.

Treatment Preparation

The DIBH breathing curve which had been recorded on the RGSC System at the time of planning CT image acquisition was attached to the (gated) treatment plan. Lower and upper limits for the Varian Gating System were set to 1.65 - 2.15 cm, which means that the vertical excursion of the IR-marker block on the patient's belly must be within a 5 mm wide band in order to enable beam on.

On the TrueBeam, certain choices have to be made during the first session which cannot be prepared in ARIA: The CBCT imaging procedure for instance is set to Gated CBCT. A reconstruction slice thickness of 1 mm in not strictly necessary for a precise match result. However, with 1 mm the shape of the marker and its orientation in space can be resolved.

To the (VMAT) treatment arcs, triggered "During" kV images have to be added. We chose a trigger interval of 10° of Gantry angle. After switching off Blade Tracking, the kV blades can be closed to a small area around the marker:

With the specification of marker dimensions (width and length) the preparation is complete.

From this point, treatment proceeds similar to a conventional DIBH, but with added security: while the breathing curve is monitored continuously and the MV beam is held off as soon as the gate is left, every 10° a kV image appears on the screen with instant marker detection and either ABH (if selected) or at least a clear visual information whether treatment is still on track.

The possibility of marker tracking renders high dose ablative treatments a very safe procedure.

Notes

¹ Florian Sterzing et al.: Stereotactic body radiotherapy for liver tumors: principles and practical guidelines of the DEGRO Working Group on Stereotactic Radiotherapy, Strahlenther. Onkol. 2014 Oct;190(10):872-81
² If some treatment interruptions occur because the patient wishes to exhale, this is no problem. TrueBeam Motion Management will hold the beam immediately. It is more important that the patient can breath in repeatedly to the same depth, and then hold the plateau for a certain amount ot time. Training sessions on the CT (without imaging) are used to train the patient.