This study reports on the feasibility and acute toxicity results of the first cohort of patients enrolled in the THUNDER 2 study, treated with an OA dose escalation protocol during treatment according to an ERI-based predictive model on a hybrid MR-Linac device. The results obtained appear very promising considering that only 1 patient experienced G3 toxicity, notably belonging to the observational arm.

The most common toxicities of nCRT treatment for LARC have already been widely evaluated in several studies. They often seem to correlate with the synergistic effect of CHT and depend significantly on the volume of healthy tissue irradiated [23].

Common acute toxicities include fatigue, nausea, diarrhoea, proctitis, mucorrhoea and hematological toxicity. However, their severity and incidence can significantly vary depending on the specific patient population and selected treatment regimen. Furthermore, they are closely associated with the techniques used for radiation delivery.

A promising technology, able to introduce an innovative way of planning and reduce unnecessary irradiation of OARs is MRIgRT. It permits real-time imaging during treatment, leading to an easier and more reliable identification of the target volumes, better management of motion, and improved planning of adaptive treatments. The precise targeting facilitated by MRIgRT also allows for the delivery of higher radiation doses to the tumor, while efficaciously minimizing the exposure to healthy tissue. These dosimetric advantages contribute to improved clinical outcomes and a better compliance for patients undergoing nCRT in rectal cancer [14, 15].

For intermediate-risk LARC patients with a low potential for distant metastases, such as those enrolled in this trial, there is a need to explore strategies that can improve CR rates while minimising adverse effects and avoiding overtreatment. In this context, escalated dose RT appears to be a logical treatment option, as the potential side effects are primarily limited to nearby OARs [7].

These are the assumptions behind the design of THUNDER 2 trial, which aims to safely increase the CR rate and thus OS in LARC patients. This interim analysis shows a very favorable toxicity profile, with only one patient of the non interventional arm (3.2%) experiencing G3 toxicity, which quickly resolved with a short treatment discontinuation.

This appears to be a very promising result when compared with historical data.

Previously, in a case series of 22 patients treated with MRIgRT using the MRIdian® system (in its tri-60Co-60 version), 22.7% developed G3 toxicity. Although this result may be explained by the less conformed dose distribution of the Cobalt version of the MRIdian system®, it appears to be in accordance with those reported in the literature regarding G3 toxicity, which range from 10.3 to 40% [6, 24, 25].

However, there is one key aspect of the study that needs to be considered: 50% of patients were considered “non-responders” based on the application of the ERI-based predictive model and the dose was escalated up to 60.1 Gy on the GTV. Interestingly there were no significant differences in toxicities between the two arms (Table 2).

When looking at RT dose escalation studies in rectal cancer, toxicity rates are variable in the literature, with G3 toxicity rates varying from 10 to 42.6% [6]. A recent meta-analysis focused on dose escalation using volumetric modulated arc therapy (VMAT) and intensity-modulated radiotherapy (IMRT) technique in LARC reported a mean ≥ G3 toxicity of 11.06% (range 0–44%), with a mean G2 toxicity of 27.08% (range 6.8-49%). No correlation was found between dose regimen and toxicity [26].

Interestingly, also the use of endorectal brachytherapy for very selective tumour boosting is burdened by G3 toxicity rates up to 19.7% [7, 27].

When evaluating the existing evidence for boost delivery by MRIgRT, the data available in the literature is limited and heterogeneous.

Boeke et al. [28] analysed the results of 5 patients who underwent long-course nCRT with weekly fractions of response-adaptive boost at 3 Gy per fraction, with 3 cases of G3 toxicity during treatment that resolved at 6-month follow-up. Specifically, according to the PRO-CTCAE scale one patient reported acute G3 toxicity of the type diarrhoea, abdominal pain, dysuria, fatigue and nausea and the other two reported acute G3 urinary frequency. Liu et al. [29] enrolled 43 patients who were randomised to receive short-course RT with boost (25 Gy in 5 fractions plus 4 Gy to the GTV, followed by four cycles of CHT) or the long-course nCRT group (50 Gy in 25 fractions with concurrent CHT). The authors noted acute G3 or higher toxicities according to CTCAE scale 4.0 in 10/21 (47.6%) patients in the experimental group and 4/21 (19.0%) patients in the control group. The most common severe toxicities in the experimental group were proctitis, pain, dermatitis, leukopenia and diarrhoea in the control group.

There is every reason to continue the trial given the observed G3 toxicity rate of only 3.2%. It is probable that the highly advantageous toxicity profile observed in the present study can be attributed to the narrow safety margins (0.3 cm for the GTV and 0.5 cm for the larger volumes) made possible by the OA workflow and real-time image guidance provided by cine MRI for direct gating purposes.

The use of OA MRIgRT in nCRT for LARC has therefore showed superior dosimetric outcomes in comparison to conventional RT techniques, offering improved protection of nearby critical structures (i.e. small intestine, bladder, and femoral heads), resulting in a decrease in radiation-related toxicities [14].

Despite the confirmed feasibility and promising toxicity profiles, our study still presents some limitations. The number of patients is certainly limited to draw definitive conclusions, so we are waiting for the full sample of patients to be enrolled.

Clearly, acute toxicity is an assessment of feasibility and adherence, but it will be more interesting to know the late toxicities in terms of quality of life and long-term organ function, especially in patients who have reached cCR and undergone conservative approaches. In this context, it will also be of interest to know the rates of post-operative complications in the two arms of the trial for the total sample of patients.