Respiratory Adaptive Computed Tomography: Feasibility Study on Real-Time Gated 4DCT for Lung Cancer Radiotherapy

Four-dimensional computed tomography (4DCT) is considered the standard of care for modern high-precision stereotactic ablative body radiotherapy (SABR) techniques. Clinical 4DCT works by acquiring CT slices and breathing data synchronously as the patient moves through the rotating X-ray imager. The CT slices are sorted into 5-10 'phase images' (mid-exhale, peak-exhale, mid-inhale, peak-inhale, etc.) and stitched together to form a 3D+breathing induced tumour/organ motion or '4D' representation of the breathing anatomy. Current best practice in radiation therapy is for clinicians to perform manual delineation or 'contouring' of lung tumours on one or more 4DCT phase images to ensure proper coverage by the treatment beam despite breathing motion. In the presence of irregular breathing, mismatches arise between CT slices that are in the same breathing phase but imaged at different couch positions. As a result, the images produced suffer from various kinds of discontinuities including truncation, duplication or overlap. Yamamoto found that irregular breathing caused at least one error >4mm in 90% (45 out of 50) abdominal and thoracic patient scans. The main consequence of 4DCT image errors is that they can introduce tumour volume and position uncertainties as large as 30% between different 4D phase bins or different observers.

This can potentially reduce the likelihood of tumour control by up to one-third and lead to unnecessary irradiation of healthy lung tissue, contributing to major dose-limiting side effects like radiation pneumonitis, which is symptomatic in up to 30% of patients and fatal in 2%. The proposed solution is Real-Time Gated 4DCT using prospective regularity gating, which will be implemented for the first time on lung cancer patients.

The Real-Time Gated 4DCT method detects and then pauses the CT beam during irregular breathing events. We will analyse the patient's breathing pattern to prospectively gate acquisition in real-time. Real-time gated 4DCT ensures that the X-ray imager is switched on only when the breathing is regular; the scan gets paused otherwise. The benefits for the patient of gating a 4DCT scan have been demonstrated in previous studies:

• a reduction of breathing motion errors by 50%
• a reduction of imaging dose. Delineation of tumours and organs in CT scans is the first step in a lung cancer patient's radiation therapy journey. Accurate delineation is pivotal to patient outcomes because errors propagate throughout treatment and can lead to radiation dose missing parts of the tumour or unnecessarily irradiating healthy tissue or organs. Unfortunately, the most important imaging modality used for the delineation of lung cancers, 4DCT, is error prone. Radiation Oncologists (ROs) are faced with an enormous challenge and are required to delineate tumours when 90% of clinical scans suffer severe imaging errors. These errors obfuscate the true tumour position/shape and reduce the chance of tumour control by up to one third.

The main culprit for imaging errors is irregular breathing: i.e. natural breathing variations, which are heightened for lung cancer patients but not accounted for by clinical scanners. A solution to this problem was proposed, which is to gate the CT scanner (i.e. pausing the beam) whenever a breathing condition occurs that is likely to produce an imaging artefact.

Research Question The primary endpoint is that the number of image artifacts using the Real-time gated scan will produce fewer artefacts than the conventional scan.

Secondary endpoints involve examining image quality and the impact on treatment planning with the benefits to the patient being:

1. Consistent/predictable image quality: Consistent image quality will assist radiation oncologists in identifying tumour and organ interfaces and therefore reduce delineation errors.
2. Improvements in treatment plans: More accurate delineation will lead to improved treatment plans and therefore improved patient outcomes.

This clinical trial is a first-in-humans pilot feasibility study and the aim is therefore to prove feasibility of Real-Time Gated 4DCT, with the aim of leading to an efficacy trial. As such, patient scans will be acquired across a broad range of patient breathing conditions to optimise the Real-Time Gated technique and to plan a larger, hypothesis-driven phase II clinical trial to follow. Automatic cut-offs for these erratic breathing traces will be built into the Real-Time Gated 4DCT software.

Rationale for Current Study In-silico studies suggest that the Real-Time Gated 4DCT approach, can reduce imaging errors by up to 50% whilst also reducing imaging dose by >20%. It has been estimated that the current rate of artifacts using conventional 4DCT is at least 60% and possibly up to 95%. A reduction to at most 35% (25% of images with fewer artifacts) would be clinically worthwhile.