MR-Guided Adaptive Radiotherapy in Localized Prostate Cancer

Abstract

MR-guided radiotherapy (MRgRT) is novel treatment modality in Radiation Oncology that could allow a higher precision and tolerability of Radiation Treatments. This modality is possible due to dedicated systems consisting of a MR scanner mounted on a conventional linac and software that permit daily online treatment plan adaptation. Prostate cancer (PC) is one of the most common malignancies in RO clinical practice and currently under investigation with this new technology. The focus of this review is to describe the current state of the art and clinical results of MRgRT in the treatment of PC. The available technology are briefly described, as well as the published literature and possible future applications

Introduction

Technology in radiotherapy (RT) constantly evolved in the last years. In less than 40 years, we moved from bidimensional radiotherapy (2D RT) to volumetric and rotational modulated arc therapy and image-guided radiotherapy (IGRT) that improved treatment precision and tolerability by reducing healthy tissue exposure. ¹ The introduction of integrated on-board imaging such as Cone Beam Computed Tomography (CBCT) or Megavoltage Computed Tomography (MVCT) has improved the positioning precision by enhancing tissue contrast compared to 2D images. However, the quality of those images is still not suitable for an accurate definition of anatomical boundaries, in many cases. Moreover no online plan adaptation has been possible on conventional linacs.

A step towards higher treatment personalization is represented by adaptive RT (ART), a novel RT technique aimed to further optimize the benefits of IGRT. This concept was introduced in 1997^2,3 and consists of dedicated systems and software able to perform a daily replan of RT treatment by accounting for the anatomical changes in target and organs at risk risk position and volume.

The MR-linac is one of those novel treatment systems. This technology consists of a conventional linear accelerator (linac) mounted on a magnetic resonance and enables to exploit the high soft tissue contrast provided by MR imaging (MRI) for an accurate daily replanning by means of dedicated software. This new treatment modality called magnetic resonance-guided radiation therapy (MRgRT) promises to be a game changer in radiation oncology. MR-linacs have several peculiar characteristics compared to conventional linacs like an improved image quality over CBCT, the possibility of a real-time treatment monitoring by means of cinematic-MR that enables an accurate tumor tracking/gating, and the possibility of using MR functional imaging to guide focal treatment or to predict tumor response (ie: radiomics). All these innovations have the potential not only to deliver more accurate and safer treatment, but also to progressively reduce treatment margins or escalate treatment dose with the aim to improve both tolerability and tumor control.^4,5,6,7

The peculiar characteristics of this new technology also carried changes in the typical treatment workflow of Radiation Oncology departments in terms of patient positioning, imaging, contouring, planning, gating and/or tracking, requesting the presence of dedicated adaptive teams of Radiation Oncologists, Physicists, and Radiotherapists. ⁸ There are currently four MR-linac systems: two of them (Unity by Elekta and MRIdian by ViewRay) are already available for clinical usage while the other two are under development.^9,10

Several clinical trials already demonstrated the potential of MR-linac, in anatomical sites like prostate, pancreas and liver, but also in oligometastases, ⁸ and several ongoing trials are demonstrating the applicability of MRgRT in other malignancies (ie: rectum, lung…). ¹¹

In particular, there are various advantages in the treatment of prostate cancer (PC). In fact the prostate might be challenging to define with precision on CT images or CBCT, especially the apex, the prostate/seminal vesicles interface, and the urethra. In this scenario, MRgRT might improve the sparing of critical structures (ie: urethra, neurovascular bundles), administering a simultaneous integrated boost (SIB) to the dominant intraprostatic lesions (DIL) or control intrafraction motion

This mini-review reports the current state of the art of MR-linac, underlying its main points of strength, limitations and future perspectives, especially in relation to the treatment of localized PC. Due to the complex and multifaceted nature of the topic some technical aspects, such as the different MRI sequences and the motion management techniques, will not be discussed in depth. Moreover other aspects such as the radiobiological considerations will not be addressed, as there's already examples in literature of more in depth discussions on the topic. ¹²

MRI-Linacs Technology Developments

Elekta Unity (Elekta AB, Stockholm, Sweden)

Elekta's Unity received FDA approval for clinical use in 2019. ¹³ The linac implemented a 1.5 T single magnet (Philips, Ingenia) with a small gap in the gradient coil and the magnet winding. ¹⁴

Regarding image acquisition, Unity provides a wide range of pulse sequences for planning and treatment, with anatomy-specific imaging sequences, encompassing T1, T2, Balanced, FLAIR, SPAIR and 3D-Vane being utilized to optimize target and healthy tissue visualization.

As for treatment delivery, this system utilizes a 7 MV flattening-filter free (FFF) beam for static step-and-shoot intensity-modulated radiotherapy (IMRT). ⁶ Dose rate is 500 MU/min and the gantry rotation speed is 6.0 rpm with no collimator rotation. ⁷

Contouring and planning are performed through Monaco TPS (Elekta, AB, Stockholm, Sweden). ¹⁵ Adaptive radiotherapy can be performed with two different models: Adapt to Shape (ATS) and Adapt to Position (ATP). ¹⁵ ATS provides a fully adaptive workflow, with online recontouring and replanning, while the patient is on the treatment table. ⁶ ATP on the other hand, is a model allowing for a virtual isocenter shift with no recontouring, before going into the optimization phase, through four possible models. This option offers a shorter treatment time and is the option of choice in favorable and reproducible anatomic sites or when longer fractions could induce larger intra-fraction uncertainties. ¹⁵ Since 2021 a hybrid workflow method called ATS-Lite has been implemented. ¹⁶

In February 2023 Unity received FDA approval for Comprehensive Motion Management tool (CMM). This tool allows one to choose from four distinct motion management strategies. The integration of CMM routines enables continuous real-time tracking and the correction of intrafraction tumor drifts by means of the BaseLine Shift (BLS). ¹⁷

ViewRay MRIdian (Viewray Inc., Oakwood, OH)

The MRIdian Linac System was the first MRI-Linac system to receive FDA approval on February 2017.

MRIdian mounts a 0.35 T split magnet which provides an adequate imaging quality while minimizing the interactions with the magnetic field compared to other systems. ¹⁸ This system uses a steady state free precession (SSFP) sequence for treatment planning and delivery with a very high resolution. It is also possible to acquire T1, T2 and DWI sequences to be used in conjunction with the SSFP image. The first model mounted a robotic 3-headed 60Co system for RT delivery. ¹⁸

After that, the system was updated with the integration of Linac, which can generate a 6 MV flattening filter-free photon beam at the nominal dose rate of 600 MU per minutes at 90 cm isocenter ¹⁹ with a gantry rotation speed of 0.5 rpm. ⁷

Since 2021, the MRIdian system allowed for intrafraction tracking via rapid acquisition of images several cutplanes (sagittal, coronal and axial) or in a single cutplane at 8 frames per second, while radiation beams can be gated by setting a prescribed physical boundary around the target volume. The radiation beam stops automatically when the volume exceeds the defined margins.

Clinical Applications of MRI-Linac in Prostate Cancer

MRI is already playing a central role in the diagnosis and staging of PC, as it provides the best morphological details, allowing for better definition of the structures surrounding the gland and the prostatic capsule and better guidance for biopsy. ²⁰ Also, in the RT planning phases it is advisable to use diagnostic MRI images as support for contouring. ²¹ Therefore, it is natural to assume that MRgRT could possibly improve the precision, safety and therapeutic ratio of prostate radiotherapy, since it was shown that an online adaptive strategy can improve planning target volume (PTV) coverage while reducing dose to OARs. ²²

On this topic, a recent study on 254 PC patients showed a mean volume increase of the gland of 6.4% during ultra-hypofractionated treatment. ²³ More in detail: prostatic swelling was reported in 50% of patients, with a mean volume increase of 15.4% of the starting volume. Stable volume, defined as ≤5% volume change, was observed in 39% of patients, and prostate shrinkage was observed in 11% of patients. These results should make radiation oncologists question in regard to current PTV expansions in a non-adaptive setting.

The MIRAGE trial is the only randomized phase III trial on PC SBRT comparing conventional linac and 0.35 T MR-linac. ²⁴ More in detail, 156 patients were randomized 1:1 to either groups. 37 patients (24%) also received pelvic irradiation and 89 (44%) were treated with a rectal spacer. Moreover, 26% received a SIB to the dominant lesion and 68% were treated with androgen deprivation therapy (ADT). The MRgRT treatments were planned with a reduced PTV (2 mm isotropic expansion from CTV), while the CT-guided treatments had a 4 mm isotropic expansion from CTV to PTV. Both groups received 40 Gy in 5 fractions, the SIB was 42 Gy in 5 fractions. Dose to pelvic lymph nodes was 25 Gy in 5 fractions. Acute GU ≥ 2 toxicity (primary end-point) was 24.4% in the MRgRT group compared to 43.4% in the CT-guided group (p = 0.006). As for acute gastrointestinal (GI) toxicity, in the MRgRT group there was a 0% incidence versus 10.5% in the control group (p = 0.01). Long-term results are awaited to see if these advantages are going to be confirmed for late toxicities too as well as for biochemical control.

As for safety, another recent study compared PC SBRT on MRI-linac (72) to conventional linac (63). ²⁵ In particular, patients were treated with prostate SBRT in 5 fractions, for a total dose of 35 Gy or 36.25 Gy for low or intermediate risk PC, according to NCCN risk classification. The PTV consisted of an expansion of CTV + 5 mm margins in each direction, except 3 mm posteriorly in both treatment groups. Acute toxicity was the primary end-point. The study found no acute toxicity differences between the two groups.

On the 1.5 T MR-linac side, the results of the prospective registry study MOMENTUM have been published in 2023. ²⁶ Four-hundred twenty-five PC patients were treated with SBRT (total dose 36.25 Gy/5 fx) reporting acute toxicity results. Acute grade 2 GI toxicity and 3 and 6 months was 1.7% with no grade ≥ 3 events. Acute grade 2 GU toxicity at 3 and 6 months was 18.1% and 7.5%, with only 1 (0.6%) grade 3 case.

Anyway, the meta-analysis of Leeman et al compared acute toxicity between MRgRT and CT-based SBRT. ²⁷ This study pooled results from twenty-nine prospective studies, including a total of 2547 patients. The analysis demonstrated a significant acute grade ≥ 2 toxicity in patients treated with MR-linac. In particular, acute grade ≥ 2 GU toxicity was 16% versus 28% in the MRgRT and conventional linac group, while acute grade ≥ 2 GI toxicity was 4% versus 9%, respectively, with a reduced risk of urinary side effects of 44% and 60% of bowel side effects.

It might be interesting to note that acute toxicities are lower in MIRAGE compared to the other trials and the metanalysis, despite the dose escalation to 40 Gy. This might be related to the smaller PTV expansion compared to conventional linac and/or to the adoption of rectal spacers for selected patients ²⁸ and raises some interesting points about the possible balance between dose escalation and PTV reduction. In this context, a correct patient setup might benefit from rectal spacers, which are clearly hyperintense in T2 sequences and, with other patient setup modifications, ²⁹ might allow for an even safer sparing of OARS, but also a more accurate treatment delivery. ³⁰

While MRgRT seems to effectively reduce toxicity in PC SBRT, a different consideration could be done for hypofractionated RT. In another study of the MOMENTUM registry, 146 patients treated with moderate hypofractionated RT (60 Gy/20 fx) on 1.5 T MR-linac reported acute grade ≥ 2 GI toxicity in 3% cases and acute grade ≥ 2 GU toxicity in 7% cases. ³¹ These results seem not dramatically different from previous randomized phase III clinical trials with the same setting and treatment schedule on conventional linac, like the PACE-B trial. ³²

Another interesting topic in PC is the boost to the DIL, which could be visualized during MRgRT. In 2023 an interesting pilot study was published, ³³ proposing SBRT to the prostate for a dose of 40 Gy in 5 fractions, with SIB to 45 Gy on the GTV. This treatment was delivered to 15 patients, all of whom underwent rectal spacer placement before the treatment. The end-point of the study was exploring the target coverage with daily-adaptive treatment and therefore no data of efficacy and safety are currently available. This might be an interesting topic for further research, considering that a previous phase I study in 2019, ³⁴ showed better biochemical control with a total dose escalation up to 40 Gy.

Recently the interim analysis of the phase II trial HERMES was published. ³⁵ In this single-center RCT, patients with intermediate-high risk PC were randomized to receive 36.25 Gy in 5 fractions or 24 Gy in 2 fractions over 8 days with an integrated boost to the magnetic resonance imaging (MRI) visible tumor of 27 Gy in 2 fractions. Both groups were treated using MR-Linac. The prespecified interim analysis was published after 10 patients for each group completed the treatment course. Preliminary results are showing a low incidence of GU toxicity in the 2-fraction group, with only 2 patients reporting grade 2 toxicity and no reports of grade 3 toxicity.

Another relevant topic is the application of MRgRT in the postoperative setting. A very interesting phase II trial is SCIMITAR. ³⁶ In this trial 100 patients underwent ultrahypofractionated radiotherapy on the prostatic bed +/- pelvic lymph nodes +/- boost to macroscopic lesions. Total dose was 30 to 34 Gy in 5 fractions and it could be delivered by CT-based LINACs or by MR-Linac based on physician's choice. The acute toxicity profile was good with both techniques, with a significant reduction of GI toxicity in the MRgRT group, in particular no grade 2 or higher acute GI events have been reported in the MRgRT group.

Current Limitations

There are also several limitations to this technique. The first and more obvious limitation is related to treatment costs. According to a cost-effectiveness study, ³⁷ initial capital investment is estimated to be $7,800,000, with estimated annual maintenance costs of $550,000, and SBRT treatments can be up to 18% more expensive than CT-based treatments.

Moreover, MRgRT is actually time consuming (ie MR sequences acquisition, IMRT treatment delivery…) with mean times per fraction ranging from 30 to 90 min in some cases, ³⁸ allowing for fewer treatments per workday compared to conventional linacs. From this analysis, an MRI-linac treating patients over an 8 h workday would take 2.4 up to 4.8 years to cover initial acquisition and installation costs.

Also, longer treatment sessions require a higher patient compliance, and the setup should be assembled as to provide comfortable positions, possibly also different than those for conventional Linacs treatments ³⁹

Longer treatment time should also raise questions about dosimetric factors due to a bigger position-related uncertainty, even though some studies ⁴⁰ are reporting an acceptable dose delivery with good CTV coverage. This can be partially compensated by motion management strategies already available for clinical use.

Conclusions

MRgRT is a technology capable of changing the treatment paradigms and workflows of radiation therapy, especially when applied to PC. The possibilities offered by a daily adaptive approach and real-time tumor tracking are pushing our boundaries in regards to target coverage, PTV reduction and possibly dose escalation and further hypofractionation. Current literature has already proved the safety of these treatments, even though long-term results are still awaited. The next step of the technique will have to pass through high-quality trials with efficacy endpoints as the main objective.