Automated clinical decision support method and system for treatment re-planning

Abstract

An automated clinical decision support method for treatment re-planning that enhances a CBCT scan using information from an initial CT scan of the patient to create a synthetic CT scan; adapts contours of organ(s) and tumour(s) from the initial CT scan to align with the synthetic CT scan; and calculates a delivered radiation dose from the synthetic CT scan; compares these against an original treatment plan to detect deviations exceeding preset thresholds, and triggers alerts with assigned priority levels for a re-plan decision.

Description

iAUTOMATED CLINICAL DECISION SUPPORT METHOD AND SYSTEM FOR TREATMENT RE-PLANNING

[0001] FIELD OF THE INVENTION

[0002] The present invention relates generally to clinical decision support for the radiotherapy treatment re-planning process, in particular, automating analysis, consolidating clinical information, providing actionable decision-support, and allowing configurability for broad regional, departmental, and patient-specific applicability.

[0003] BACKGROUND OF THE INVENTION

[0004] A patient’s radiotherapy treatment plan is optimised on a CT scan taken days (or even weeks) before treatment starts. Adhering to this treatment plan across each fraction of treatment is important for maximising the therapeutic radiation dose to cancerous tissue while minimising the exposure of surrounding healthy tissue to this dose.

[0005] Anatomic variations caused by factors such as weight loss, tumour regression, and other changes to the shape of the tumour or surrounding organ(s) can occur during treatment that follows a treatment plan. This treatment may span from 3 to 7 weeks, or in some cases, even longer. These anatomic variations can significantly diminish the effectiveness of the treatment plan. When such variations are substantial, the treatment plan may become unsuitable, necessitating the formulation of a new treatment plan. This new treatment plan aims to re-optimise the dose for the targeted area and minimise exposure to adjacent organs at risk (OAR).

[0006] The necessity for a treatment re-plan often arises in response to significant anatomical changes (for example, weight loss, tumour progression or regression, etc) in the patient beyond those discussed previously. For example, for head & neck cancer patients, if the spinal cord receives more than 50 Gy (max dose), this is considered a major deviation. These changes increase the risk of either insufficient treatment of thecancerous tissue or excessive damage to nearby healthy tissue or organs. The challenge lies in the lack of a systematic approach to assess whether the original treatment plan has become ineffective. Additionally, there is a need for actionable clinical decision support to define and identify what constitutes a significant anatomical change warranting a treatment re-plan. Fig. 1 depicts an example of how structures at planning are no longer representative mid-way through their course of treatment.

[0007] In the case of Head & Neck (H&N) cancer patients, one or two mid-treatment replanning events may result in significant clinical benefits. Several studies support this. A two-year study found that local control rates were higher in H&N patients who underwent re-planning (88%) compared to those who did not (79%), with a statistically significant difference (p = 0.01). Over a five-year period, local recurrence-free survival rates were greater for H&N patients who underwent re-planning (96.7%) than for those who did not (88.1%), a notable difference with a p-value of 0.022. Re-planning resulted in a notable reduction of the average radiation dose to the parotid glands by 5.1 Gy, corresponding to an 11% decrease in the predicted risk of xerostomia, a common side effect. Additionally, re-planning effectively reduced excessive radiation ('hot spots') in the brainstem and spinal cord, bringing them back within established protocol tolerance. Overall, treatment re-planning during the course of radiotherapy has been proven to be clinically beneficial. For H&N cancer patients, specifically, re-planning not only improves outcomes but also minimises the dose to critical OARs and reduces the risk of treatment-related toxicity.

[0008] Although treatment re-planning is beneficial for adapting to anatomical changes during radiotherapy, it poses significant challenges for most treatment centres. The replanning process is often hindered by the substantial demand for human and material resources and existing technical limitations. These factors contribute to making treatment adaptation to anatomic variations resource-intensive and also time-consuming, which impacts the overall efficiency of patient care in radiotherapy.

[0009] Beyond the specific issue of re-planning, there are other major challenges in radiotherapy. In medical departments and clinics, the need for re-planning imposes additional demands on already limited resources and time. The increasing technicali complexity of radiotherapy advancements further compounds it. Deploying, training, and maintaining these increasingly complex technologies incur higher effort and costs, complicating the treatment process. Furthermore, the workflows for deciding when to replan are often cumbersome, characterised by subjective, unclear processes or the need for extensive manual coordination across various applications and teams, making integrating quantitative data into re-planning decisions laborious. These challenges are interconnected, with inefficient workflows exacerbating resource limitations and the advancement of complex treatment procedures further complicating these workflows.

[0010] These challenges during the re-plan process can occur under three different scenarios: (a) the standard clinical workflow re-planning is not implemented for treatment;(b) re-planning midway through the treatment as the standard of care for all patients; and(c) through ad hoc re-planning, triggered by observations made during pre-treatment imaging.

[0011] When a re-plan is implemented, current standards of care can result in additional time requirements. For example, an additional 11 days when a re-plan is necessary, and an additional 4 days, even in cases where a re-plan is ultimately not required. These extended timeframes place a significant burden on the radiotherapy team. They necessitate intricate and manual analyses using various analytical tools, consuming considerable resources, whether or not a re-plan is eventually carried out.

[0012] Beyond the clear need for an effective and resource-efficient decision-support system of informing re-plan decisions, radiotherapy centres increasingly recognise the need for an effective decision-support system that enhances the efficiency of re-plan decisions. A key part of this need is the systematic storage of results for all patients and treatment sessions, particularly for cases where re-planning was not required. A significant gap exists in maintaining comprehensive records, an important factor for verifying the accuracy and efficacy of treatments. Addressing this gap may improve the quality of patient care in radiotherapy and enhance resource allocation, making the treatment process more efficient and effective.

[0013] The current process for deciding on treatment re-planning is cumbersome and relies heavily on manual efforts, involving the use of multiple distinct software tools that each only provide fragments of the necessary overall information. This makes the clinical decision-making process performed piecemeal, resource-intensive, poorly defined, and lacking adequate clinical decision support. This complexity is exacerbated with additional challenges. Firstly, the daily imaging technique used, namely, a CBCT scan, provides images of low quality with a limited field of view and lacks the essential electron density information necessary for accurate dose calculation. Secondly, the estimation of organ and target contours on a CBCT scan is a highly skilled, manual, and time-intensive task, complicating the dose-volume reporting process. Thirdly, calculating the radiation dose itself requires intensive computing resources. Lastly, direct measurement of the dose within internal organs is typically impractical or too onerous to perform effectively. Currently, no available solutions comprehensively address re-plan decision support. Specifically, there's a gap in available solutions that provide integrated analysis of dose and volume metrics, comparisons and insights into planned versus delivered dose differences, trend analysis of treatment fractions, and the capability to automate these functions to reduce the resource demands of the re-plan decision-making process.

[0014] Reference to any prior art in the specification is not an acknowledgment or suggestion that this prior art forms part of the common general knowledge in any jurisdiction or that this prior art could reasonably be expected to be understood, regarded as relevant, and / or combined with other pieces of prior art by a skilled person in the art.

[0015] It is, accordingly, an object of at least one embodiment of the present invention to provide an alternative approach to radiotherapy treatment re-planning that is holistic and more comprehensive, reduces the need for resource-intensive manual investigations, is better defined and provides clinical decision support.

[0016] SUMMARY OF THE INVENTION

[0017] In one aspect, there is provided an automated clinical decision support method for treatment re-planning of a patient. The method comprises enhancing a Cone BeamiComputed Tomography (CBCT) scan of the patient using information from an initial CT scan of the patient to create a synthetic CT scan of the patient. The method also comprises adapting contours of organ(s) and tumour(s) from the initial CT scan to align with the synthetic CT scan. The method also comprises calculating a delivered radiation dose from the synthetic CT scan. The calculated delivered radiation dose and adapted contours are compared to an original treatment plan to detect whether a deviation in at least one dose-based metric from the original treatment plan has occurred, if a detected deviation exceeds a predetermined threshold value for the respective dose-based metric, a notification with an assigned priority level is generated to alert a clinician for a re-plan decision, the assigned priority level corresponding to the extent of the deviation from the predetermined threshold value.

[0018] The predetermined time interval may be 24 hours.

[0019] The method may further comprise configuring at least one protocol template on a patient basis, and if a new patient is treating using the at least one protocol template, a protocol of the at least one protocol template is applied to triage the delivered radiation dose into triage categories, wherein the protocol is used to identify which structures a clinician is most concerned about to populate a dose-volume score card and histogram, and to identify a decision threshold per structure.

[0020] The at least one protocol may be determined according to any one from the group consisting of: a type of radiotherapy centre, a type of cancer and a type of treatment.

[0021] The method may further comprise an initial step of predicting the synthetic Computed Tomography (CT) scan using one or more from the group consisting of: planning CT scan, structure, dose file, treatment plan data, Cone Beam Computed Tomography (CBCT) scan, Magnetic Resonance Imaging (MRI) scan, Positron Emission Tomography (PET) scan.i

[0022] The method may further comprise performing Deformable Image Registration (DIR) to obtain the transform between the initial CT scan or subsequent CT scans, and the CBCT scan.

[0023] The method may further comprise generating trend data of treatment fractions.

[0024] The notification of a clinical change for a re-plan decision may include a visual indicator with a plurality of sections comprising a first section having a first colour indicating major thresholds exceeded; a second section having a second colour indicating minor thresholds are exceeded; and a third section having a third colour indicating thresholds are within thresholds.

[0025] At least one dose-based metric may comprise any one from the group consisting of: dosimetric performance, dose distribution, tumour control probability (TCP) and normal tissue complication probability (NTCP), dosiomics, CTV D98, PTV D95, Cord Max Dose, other organs at risk (OARs) mean dose and conformity indices (Cl).

[0026] The enhancement of the CBCT scan may further comprise performing field of view masking to compensate for limited CBCT field of view and enable accurate anatomical representation for dose calculation.

[0027] The enhancement of the CBCT scan may further comprise performing slabmatching, wherein intensity matching is performed in discrete image slabs to improve alignment between the synthetic CT scan and the CBCT scan.

[0028] The enhancement of the CBCT scan may further comprise processing different anatomical components, including soft tissue, bones, and air pockets, separately to refine the synthetic CT scan.i

[0029] The enhancement of the CBCT scan may further comprise incorporating updated patient setup information, including couch and bolus structures, to ensure that dose calculations accurately reflect actual treatment conditions.

[0030] The contours may be adapted by applying body-site-specific transformations based on anatomical deformation patterns.

[0031] The selection of the body-site-specific transformation methods may be determined by predefined anatomical site-specific protocols, wherein different deformation models are applied to different regions.

[0032] Calculating the delivered radiation dose may include performing a singleresolution or multi-resolution dose calculation, wherein in the case of multi-resolution dose calculation, a low-resolution dose calculation is performed over the synthetic CT scan and a high-resolution dose calculation is performed over a smaller region of interest within the synthetic CT scan.

[0033] A final dose calculation result may be obtained by interpolating between the low- resolution and high-resolution dose calculations.

[0034] The delivered radiation dose may be calculated using a computer-based dose calculation algorithm from any one of the group consisting of: Collapsed Cone Convolution (CCC), Pencil Beam, Monte Carlo, and the Linear Boltzmann Transport Equation, and wherein the computation is performed using a central processing unit (CPU) and / or a graphics processing unit (GPU).

[0035] Detecting whether a deviation exceeds the predetermined threshold value may further comprise analysing dose trends across multiple treatment fractions.

[0036] The analysing dose trends may include predicting future dose deviations based on analysing prior dose variations.i

[0037] The assigned priority level in the generated notification may be determined based on the severity and recurrence of dose deviations.

[0038] The assigned priority level in the generated notification may be determined based on a scoring system that assigns higher priority to patients with more significant or repeated deviations.

[0039] The assigned priority level in the generated notification may be determined based on custom weighting of different types of deviations, wherein a user-defined protocol enables assigning different priority levels to minor and major threshold violations.

[0040] The method may further comprise determining dose discrepancies by comparing calculated dose distributions to reference values from the treatment planning system and flagging inconsistencies for review by the clinician.

[0041] In a second aspect, there is provided an automated clinical decision support system for treatment re-planning of a patient. The system comprises a CBCT enhancement component to enhance a Cone Beam Computed Tomography (CBCT) scan of the patient using information from an initial CT scan (or subsequent CT scans) of the patient to create a synthetic CT scan of the patient. The system also comprises a contour adaptation component to adapt contours of organ(s) and tumour(s) from the initial CT scan to align with the synthetic CT scan. The system also comprises a dose calculation component to calculate a delivered radiation dose from the synthetic CT scan. The calculated delivered radiation dose and adapted contours are compared to an original treatment plan to detect whether a deviation in at least one dose-based metric from the original treatment plan has occurred. If a detected deviation exceeds a predetermined threshold value for the respective dose-based metric, a decision support module issues a notification with an assigned priority level is generated to alert a clinician for a re-plan decision, the assigned priority level corresponding to the extent of the deviation from the predetermined threshold value.

[0042] In a third aspect, there is provided a computer software product comprising a sequence of instructions storable on one or more computer-readable storage media, said instructions when executed by one or more processors, cause the processor to: enhance a Cone Beam Computed Tomography (CBCT) scan of the patient using information from an initial CT scan of the patient to create a synthetic CT scan of the patient; adapt contours of organ(s) and tumour(s) from the initial CT scan to align with the synthetic CT scan; and calculate a delivered radiation dose from the synthetic CT scan. The calculated delivered radiation dose and adapted contours are compared to an original treatment plan to detect whether a deviation in at least one dose-based metric from the original treatment plan has occurred. If a detected deviation exceeds a predetermined threshold value for the respective dose-based metric, a notification with an assigned priority level is generated to alert a clinician for a re-plan decision, the assigned priority level corresponding to the extent of the deviation from the predetermined threshold value.

[0043] The method and system of the present invention are an improvement over the prior art because of its comprehensive analyses, automation of tasks, and its seamless clinical workflow integration and configurability. Existing available solutions only partially address the requirements of the re-plan process. For example, Medical Image Merge (MIM Software Inc’s software) does not calculate dose, but instead “deforms” the dose and generally contribute to major problems in the re-plan process, such as being inaccurate, cumbersome, demanding extensive resources, and lacking robust decision support.

[0044] The traditional treatment re-planning process in radiotherapy relies heavily on manual, resource-intensive investigations. These investigations typically use various separate products that individually offer only a limited scope of the necessary information for determining if a treatment re-plan is needed. Furthermore, these existing products do not provide comprehensive support for decision-making. In contrast, the system of the present invention transforms this process in a unique transformative way. It introduces aspects of automation, consolidation, and decision support into the radiotherapy re-plan workflow. This significantly eases the burden on medical departmental resources and optimises the entire treatment re-planning process. As a result, the system of the present■ invention ensures that all patients consistently receive care that is specifically adjusted to their needs throughout their radiotherapy treatment.

[0045] The present invention has industrial applicability in the field of radiation oncology, where precision and adaptability in retreatment planning are useful for effective cancer therapy.

[0046] As used herein, except where the context requires otherwise, the term "comprise" and variations of the term, such as "comprising", "comprises" and "comprised", are not intended to exclude further additives, components, integers or steps.

[0047] Further aspects, advantages, and features of embodiments of the invention will be apparent to persons skilled in the relevant arts from the following description of various embodiments. It will be appreciated, however, that the invention is not limited to the embodiments described, which are provided in order to illustrate the principles of the invention as defined in the foregoing statements and in the appended claims, and to assist skilled persons in putting these principles into practical effect.

[0048] BRIEF DESCRIPTION OF THE DRAWINGS

[0049] Embodiments of the invention will now be described with reference to the accompanying drawings, in which like reference numerals indicate like features, and wherein:

[0050] Figure 1 depicts a pair of CT images showing anatomic variations observed for a head & neck cancer patient during their course of radiotherapy (right image) compared to the planning CT image (left image).

[0051] Figure 2 is user interface screen of a dose review system in accordance with an embodiment of the present invention showing how to assess whether a re-plan is warranted.

[0052] Figure 3 is a labelled user interface screen in accordance with an embodiment of the present invention showing how relevant clinical information is consolidated and provides actionable information for informing re-plans.

[0053] Figure 4 is a flow chart showing clinical workflows (and additional time required for re-plans) for a system-assisted re-plan decision in accordance with an embodiment of the present invention.

[0054] Figure 5 is a flow chart showing clinical workflows of the dose review system relating to consolidating, automating, and providing decision support in the re-plan process, in accordance with an embodiment of the present invention.

[0055] Figure 6 is a flow chart showing an expanded view of the dose review system’s automated analysis and data flow.

[0056] Figure 7 depicts the dose review system’s automated analysis and data flow mapped to the dose review system’s communication of results via a user interface screen in accordance with an embodiment of the present invention.

[0057] Figure 8 is a type of alert generated by the dose review system.

[0058] Figure 9 is a triage list dashboard is displayed in the dose review system.

[0059] Figure 10 is a triage list dashboard with a filter applied to display only “major deviations”.

[0060] Figure 11 is a screen for configuring patients in the dose review system.

[0061] Figure 12 is a screen for configuring thresholds in the dose review system.ii

[0062] Figure 13 is a block diagram showing the workflow between an Oncology Information System (OIS) and the dose review system.

[0063] Figure 14 is a set of examples of implementing a silent listener module of the dose review system with two different OIS.

[0064] Figure 15 is a block diagram of the dose review system in accordance with an embodiment of the present invention.

[0065] Figure 16 is a flow chart of the slab intensity matching process used for aligning input images with a reference image, in accordance with an embodiment of the present invention.

[0066] Figure 17 is a data flow diagram depicting a processing pipeline for treatment data within the dose review system in accordance with an embodiment of the present invention.

[0067] Figure 18 is a chart of dose calculations for different slab materials, specifically air (-1000 HU) and lung (-700 HU), across varying depths and off-axis distances where the comparison is performed using multiple dose calculation algorithms, including Collapsed Cone Convolution (CCC), an another implementation of CCC (TruckCCC), Analytical Anisotropic Algorithm (AAA), and AcurosXB(AXB).

[0068] Figure 19 is a chart of slab dose gamma results, comparing the fail rates of different dose calculation models (AAA, AXB, and TPS-CCC) within a treatment planning system (TPS) across various material slabs characterised by Hounsfield Unit (HU) values, and illustrates the gamma error rate of the dose review system’s CCC-based dose calculation method against the listed dose calculation models, including CCC vs. AAA, CCC vs. AXB, and CCC vs. TPS-CCC, providing a direct evaluation of the accuracy and consistency of the dose review system's approach.11

[0069] DETAILED DESCRIPTION OF EMBODIMENTS

[0070] Figs. 2 to 15 depict a method and dose review system 100 for automated dose replanning that may implement an embodiment of the invention described herein.

[0071] In one embodiment, a dose review system 100 is provided that streamlines the treatment re-planning process via automated processing by calculating and comparing the delivered radiation dose with the original treatment plan. This functionality generates and presents daily patient dose information. This functionality enables the treatment team or clinician to detect and determine when a re-plan is needed, make timely, well-informed clinical decisions based on actual evidence and medical data, and streamlines the overall treatment management process. It addresses challenges associated with the offline review of treatment plans by providing an automated workflow that supports clinical decision-making, augmenting the accuracy of identifying instances that necessitate replanning while reducing unnecessary effort in scenarios where re-planning is not required. Designed to fulfill clinical demands, system 100 incorporates decision support, re-plan process automation, and protocol configurability to tackle prevalent challenges in the re-planning process. The system 100 is a comprehensive solution that advances the efficiency and efficacy of treatment re-planning, making it superior to current standards and available solutions.

[0072] The system 100 includes a CBCT enhancement component to enhance a Cone Beam Computed Tomography (CBCT) scan of the patient using information from an initial CT scan of the patient to create a synthetic CT scan of the patient. The system 100 also includes a contour adaptation component 122 to adapt contours of organ(s) and tumour(s) from the initial CT scan to align with the synthetic CT scan. The system 100 also includes a dose calculation component 121 to calculate a delivered radiation dose from a synthetic Computed Tomography (CT) scan. The calculated delivered radiation dose and adapted contours are compared to an original treatment plan to detect whether a deviation in at least one dose-based metric from the original treatment plan has occurred. If a detected deviation exceeds a predetermined threshold value for the respective dose-based metric, a decision support module 120 issues a notification with■ an assigned priority level is generated to alert a clinician for a re-plan decision, the assigned priority level corresponding to the extent of the deviation from the predetermined threshold value.

[0073] Contours are adapted from an initial CT scan. Oncologists and radiation therapists draw contours on the initial CT scan because it has high image quality. Frequently, this is done by integrating data from MRI or PET imaging to ensure accuracy and detail. On each day of treatment, a Cone Beam Computed Tomography (CBCT) scan is taken and used, however, a CBCT scan has limited field of view, lower dose, and comparatively poorer image quality compared to a CT scan. To address these limitations, the CBCT scan is processed to fill in missing information, enhancing it to match the scale and coordinates of the CT scan as closely as possible. This enhancement involves deformable image registration, which is performed using traditional methods, such as bspline or demon algorithms, or more advanced Al techniques, including Transformers or unsupervised ll-nets, to establish an accurate transformation between the CT and CBCT scans. The same transform (with additional interpolation) is applied for the contours to obtain adapted contours, which are the contours on the CBCT scan.

[0074] In another example, another type of initial planning scan may be used, for example, Magnetic Resonance Imaging (MRI) scan or Positron Emission Tomography (PET) scan. DIR is performed to obtain the transform between the initial planning scan and the daily CBCT scan.

[0075] The system 100 also comprises a user interface (III) 110, a protocol customisation module 130, and a listener module 140.

[0076] User interface

[0077] The Ul 110 displays the calculated dose and contour results in addition to metrics, and compares them to the original treatment plan. An example Ul for a patient’s fraction 10 of treatment is shown in Figs. 2 and 3. The Ul 110 functions as a central user interface for clinicians, consolidating and categorising clinically-important treatmentli information using a visual "traffic light system" to provide clear, actionable insights. The “traffic light system” functions as a triage mechanism, where red indicates results outside the thresholds set via the protocol customisation module 130, amber indicates results trending towards red, and green indicates results are within acceptable thresholds.

[0078] The III 110 integrates data processed by the system 100, which has calculated the delivered radiation dose and contours the relevant organs and tumors for each day of a patient's treatment. Additionally, the III 110 displays comparisons between these contours and the delivered dose against the contours and prescribed dose from the original treatment plan. Therefore the III 110 has an important role in the overall streamlining of the decision-making process by providing actionable indicators that signal when a re-plan of the treatment is warranted, thereby enhancing the efficiency and effectiveness of patient care.

[0079] The III 110 is in the form of a graphical user interface (GUI) designed for reviewing daily treatment results against the original treatment plan for each fraction, including trend results where possible. It features a dashboard view for offline patient reviews, presenting both patient-specific and aggregated data, and indicates whether a fraction or patient review is pending. This GUI setup facilitates efficient monitoring and analysis of treatment progress and outstanding reviews.

[0080] Heatmap Visualisation

[0081] The dose heatmap visualisation is described. The dose 'heatmap' is a graphical representation designed to visually depict the distribution and intensity of a dose across a specified area. This visualisation method uses a colour-coded scheme to convey information about the dose levels, facilitating an intuitive understanding of spatial dose variations. The dose 'heatmap' includes a colour gradient, visual interpretation cues, and colour gradient precision.

[0082] The heatmap uses a colour gradient that transitions from blue to red. Each colour within the spectrum corresponds to a specific dose level, with blue representing the lowerBi end of the dose spectrum and red indicating the higher dose levels. A heatmap visualisation component may be configured to provide interactive functions including zooming-in, zooming-out, panning, adjusting the dose level thresholds for the clinician to alter the colour gradient dynamically. The colour gradient may be calibrated against dose measurements according to a linear relationship from minimum dose to maximum dose.

[0083] For visual interpretation cues, low and high dose representations are communicated. Areas receiving a low dose are depicted in shades of blue. This colour choice is deliberate, as blue is commonly associated with cooler temperatures, symbolising lower intensity or concentration. Conversely, areas subjected to a high dose are shown in shades of red. The colour red is chosen for its association with warmth or heat, effectively indicating increased intensity or concentration.

[0084] The colour gradient precision provides precise shades of blue (low dose) and red (high dose), along with the colours in between, are calibrated to represent varying degrees of dose intensity. This allows for a nuanced visual understanding of dose distribution, highlighting areas of both high and low concentration with significant accuracy.

[0085] Alerting mechanism

[0086] The alerting mechanism of the system 100 is described. The system 100 categorises patient treatment alerts into three distinct categories based on the urgency and severity of action required. In one embodiment, the III 110 communicates this to the clinician via a colour scale: green, amber and red. In another embodiment, it may be a numerical scale. The green category Indicates that no immediate action is necessary. Alerts in this category signify that the patient's treatment is proceeding as planned, without any deviations warranting attention. The amber category indicates minor deviations from the planned treatment parameters that are trending towards exceeding the set thresholds. These alerts suggest that users should monitor the specified conditions closely, although immediate action is not required. The red category indicates that clinical thresholds for significant targets or organs have been reached or exceeded,necessitating immediate action by the users. The red category is critical and indicates situations where patient safety could be compromised without prompt intervention.

[0087] Alerts categorised as green and amber are displayed on the dashboard of the system 100, providing users with an overview of current statuses at a glance. In contrast, red alerts are communicated directly to users through more immediate channels such as pop-up notifications, emails, or entries on their worklists, ensuring that these urgent notifications receive prompt attention by clinicians.

[0088] Users or clinicians have the ability to customise the sensitivity of the system 100 in generating alerts for specific targets or organs. This customisation includes setting stricter thresholds for high-priority targets or organs, aligning the alert system with the departmental clinical protocols and ensuring that alerts are generated to reflect the clinical importance of different treatment parameters.

[0089] The system 100 alerts users to deviations from the original treatment plan and informs users when the treatment for a daily fraction significantly deviates from the plan, prompting them to review the comprehensive analysis provided by the system 100. This review process is required for determining whether a treatment re-plan is necessary, which remains a decision that the clinician is responsible for by using the detailed information provided by the system 100.

[0090] Fig. 8 depicts an example of an alert message 80 - alerting the user of the detected deviation for what target / organ and how big a difference is detected. Whenever the system 100 detects a significant deviation from the planned treatment, it notifies the relevant clinicians. This alert 80 prompts the clinician to review the case in the Ul 110, as outlined in the decision support module 120. The alert 80 may contain a link 81 to enable the clinician to click through to see the specific dose review. The system 100 is designed to provide clear, actionable alerts 80 when the treatment deviates from the planned metrics. This feature ensures that any significant changes are promptly identified, allowing for timely interventions.If

[0091] Referring to Fig. 9, a triage list dashboard is depicted. Referring to Fig. 10, a triage list dashboard filtered to “major deviations” only (i.e. urgent cases for review) is depicted. The dashboard view can distinguish between data specific to individual patients and data that is aggregated across multiple patients by showing the traffic lights for the last three treatments of each patient.

[0092] Decision Support Module

[0093] The decision support module 120 is designed to calculate the radiation dose delivered to a patient daily. It also delineates relevant organs and the tumour (s), comparing these daily results with the original treatment plan. The goal is to identify and alert the clinician when the data indicates that a re-planning decision may be necessary.

[0094] The system 100 uses a variety of data sources to enable accurate and effective treatment planning tracking and monitoring, including: planning data, daily fraction data, commissioning data, additional image modalities and additional treatment modalities. Planning data includes the planning CT scans, structural data, and dose files. This type of data is required for comparing the daily fractions of the treatment.

[0095] The system 100 streamlines the dose investigation phase of the treatment replanning process, significantly aiding clinicians in their decision-making. It provides robust decision support by aggregating relevant metrics and analytics in a unified location, presenting clear, actionable indicators for deviations from the original treatment plan, and providing temporal analytics to visualise trends across treatment fractions. The detailed temporal analytics presents trend data for the results of the set metrics over the entire course of the treatment. This trend analysis across all treatment fractions gives a clearer picture of changes over time, providing deeper insights and aiding in more informed decisions about whether a treatment plan needs to be revised. This enhances the clarity and effectiveness of clinical decisions in the treatment re-planning process.

[0096] The system 100 reconstructs 503 daily doses and contours daily structures, accurately mapping treatment by comparing these daily results with the original treatmentIB plan based on configured thresholds. When deviations exceed these thresholds, it alerts clinicians through notifications 80, enabling timely and effective adjustments to the treatment plan.

[0097] The system 100 provides an efficient, data-driven approach to treatment planning and adjustments. The types of actionable indicators and metrics the system provides include a red alert for any dose volume parameter that surpasses the clinical threshold set by the user during installation via the protocol customisation module 130. Additionally, the system 100 displays trends indicating whether a parameter is approaching a "red alert" level. The trends towards a "red alert" level may be determined by calculating the linear extrapolation based on the last three treatment fractions. It is at the user’s discretion to make a clinical decision on whether to proceed with a re-planning decision.

[0098] Protocol Customisation Module

[0099] The system 100 has a protocol customisation module 130 to support multiple protocols of radiotherapy centres and the individual requirements of patients, as depicted in Figs. 9 and 10. This module 130 facilitates the adjustment of system settings to match both department-specific protocols and patient-specific requirements. The system 100 integrates automated functions that are necessary for effective cancer treatment management, in turn providing personalised and efficient processes. Its metric and threshold configurability allows for precise customisation to cater to various departmental requirements, treatment objectives, various clinical indications, patient cases, patientspecific needs. The key functionalities of the protocol customisation module 130 include configurable metric thresholds that are manually set at the commissioning period or modified as needed, significantly enhancing its versatility, flexibility, adaptability and applicability in a clinical environment.

[0100] The importance of this customisation provided by the protocol customisation module 130 is underscored in the "Customer Acceptance Testing" workflow 501, illustrated in Figs. 5 and 6, which requires a configuration phase. Thisensures that the system 100 provides tailored results and notifications, enabling re-plan decisions that conform to the specific standards of care for each department and patient.

[0101] The protocol customisation module 130 facilitates the manual setting of customisable thresholds at installation and subsequent adjustments as needed.

[0102] Users can configure and adjust the thresholds in the system 100 via two main steps: during instal lation / initial setup and for each patient. During the initial setup, users import their department's planning protocols for each specific disease site, such as oropharyngeal cancer. These protocols include both significant and minor variations that are clinically acceptable. Users can manually adjust these templates if they prefer slightly different alert thresholds than those provided by the primary protocol variation. For configuring individual patients, users have the ability to manually set a distinct threshold or assign a higher or lower priority to each anatomical structure. An example Ul 100 is provided, showing how to configure which patients to monitor without generating active alerts ("silently listen as provided for by the silent listener module 140") and how to set specific thresholds and priorities for various targets / organs in terms of results and alerts.

[0103] Upon configuration, the process is designed to be intuitive and flexible within the system 100. Initially, at the point of installation or at any subsequent time, a clinic is empowered to create a protocol template within the system 100. This functionality is required when a new patient is admitted and identified for treatment, as it leverages a pre-established protocol template. Specifically, when a patient's treatment corresponds to an existing protocol template, the system 100 seamlessly applies this protocol to categorise the patient's daily radiation dose delivery into three distinct triage categories: green, indicating that all parameters are within expected norms; yellow, signalling minor deviations in certain regions of interest; and red, denoting significant violations in these critical areas.

[0104] These protocols also identify which anatomical structures are of paramount importance to clinicians, thereby facilitating the generation of a targeted dosevolume scorecard and histogram. Additionally, these protocols assist in determining the11 decision-making thresholds for each structure, tailoring patient care to the specific clinical context.

[0105] The system 100 offers a personalised approach to patient treatment. Since each patient presents a unique set of medical circumstances, the system 100 allows clinicians to modify the configurations of the protocol templates. These adjustments are tailored to align with the individual patient's pathology and physiology, ensuring that the treatment is as effective and appropriate as possible.

[0106] Silent Listener Module

[0107] The system 100 is designed to simplify and enhance the dose investigation phase of the treatment re-planning workflow. It functions as a clinical decision support tool, aiding healthcare professionals in the re-plan process. It comprises a silent listener module 140 that operates in the background, automatically analysing data for all indicated patients. It minimises the need for manual intervention, efficiently handling the bulk of the analysis without disrupting the existing workflow.

[0108] Re-plan process automation is implemented at least in part by a silent listener module 140 in the system 100 that continuously collects and analyses data from patients under treatment. Using the silent listener module 140, the system 100 analyses data from all relevant patients while keeping manual interactions to a minimum. It compares each fraction of the delivered dose against the original plan.

[0109] Existing solutions require users to manually export and import relevant DICOM series (such as CT, CBCT, Plan, Dose, etc.) into the appropriate systems. This process requires the user to interact with multiple systems in order to review. This is time-consuming, error prone and requires expertise at each step to accurately navigate and manage the import / export of the correct parameters to achieve the desired outcome. In contrast, the system 100 includes the silent listener module 140 which minimises manual interactions by the clinician. These steps are performed and completed autonomously, without requiring any user intervention. The system 100 uses the silentlistener module 140 to listen for new CBCT scans arriving at the clinical database. This occurs daily when the patient receives a new radiation treatment. If the patient is monitored by the system 100, the silent listener module 140 will query the new CBCT scan to be transferred into the system 100 for analysis. Other rules for queries can be applied in regards to patients with specific treatments, with a specific disease and so forth.

[0110] Existing solutions also require multidisciplinary effort to produce individual analyses for review, making the process resource-intensive. No available solution performs automatic analysis for all indicated patients within the clinical workflow with the functionality of the silent listener module 140. While solutions like MIM can be set up for automatic analysis of certain patients, they still necessitate manual verification of each result. Furthermore, none consolidate all the necessary information to inform a re-plan decision, with MIM providing dose results but lacking insights on anatomical variations or clear indicators for when a re-plan is necessary. Additionally, no existing solution automatically flags instances where a patient’s treatment significantly deviates from the original plan, necessitating a review for potential re-planning. The system of the present invention has superior advantages from its capability to configure patient-specific levels for these automatic comparative analyses and notifications, a functionality not found in existing available solutions.

[0111] Referring to Figs. 13 and 14, are examples of the silent listener module 140 implemented for different vendor OIS setups 200. The silent listener module 140) detects when new data is pushed to the oncology centre’s database. Also, the silent listener module 140 it detects which data is indicated for the system 100. The silent listener module 140 facilitates the transfer of relevant data to be processed by the system 100.

[0112] Automated Workflow and Analysis

[0113] Referring to Fig. 6, the automated portion of the workflow of the system 100 is described in further detail with reference to the data and algorithm flowchart. Theli automated analysis function generates a synthetic CT image and daily anatomical outlines (contours) by using the daily Cone Beam Computed Tomography (CBCT) scan 602 in conjunction with the initial planning data 601 from the original treatment plan. The CBCT scan is taken by a medical linear accelerator (Linac) 201, for example, a Linac integrated cone beam CT (CBCT) scanner. This synthetic CT image is then used to determine the actual dose of radiation delivered based on the planning data. Using this approach, the system 100 can compute daily radiation dose metrics for each of the generated contours. Having access to this data, the system 100 compares the analysed results for each contour with those specified in the original treatment plan.

[0114] During the commissioning period, synthetic CT data (data that is estimated or predicted) is used. This is an example of commissioning data which is used performing DIR and calculating delivered doses. Synthetic CT scans allow for comparing the system’s dose and contour performance against the ground truth provided by historical planning data. This process ensures the system's accuracy and reliability.

[0115] Referring to Fig. 7, the sequence of steps in the automated analysis workflow of system 100 is illustrated, along with the correlation of these steps to the outcomes displayed in the product. In Fig. 7(a), the system 100 calculates 605 the daily delivered dose, visually represented as a dose "heatmap" superimposed over daily images, demonstrating the distribution and intensity of the delivered dose. Fig. 7(b) displays daily contours generated 606 by the system 100, depicted as overlays on these images, providing a visual representation of treatment areas. In Fig. 7(c), the analysis for each contour is translated 607 into quantifiable data (i.e. dose analytics per structure), and presented in a results table for easy interpretation. Lastly, Fig. 7(d) compares 608 the delivered results with the original treatment plan and shows the comparison in three formats: (i) a dose difference "heatmap" in the image panel illustrating the variance from the plan, (ii) quantified differences in the results table for precise numerical data, and (iii) the colour-coded "traffic light system” (using colours like red, amber, green) to visually indicate the magnitude of the difference, facilitating quick interpretation and clinical decision-making by indicating the severity of any deviations.11

[0116] The system 100 performs an automated analysis by creating an estimated synthetic CT scan. This is achieved using daily Cone Beam CT (CBCT) scans and the initial planning data from the original treatment plan. The synthetic CT then allows for the calculation of the actual dose delivered based on the planning data. Furthermore, the system 100 calculates daily dose metrics for each contoured area. This process involves comparing the daily dose measurements of each contoured area with the doses outlined in the original treatment plan.

[0117] In contrast to the prior art workflows representing the current standard of care, the system 100 automatically analyses each patient’s treatment data in parallel to the clinical workflow without interruption. When the system 100 detects that a metric from a delivered dose deviates from the original plan, it automatically notifies the relevant clinician to review this patient in the system 100 due to the detected deviation. The clinical workflow of the system 100 is shown in Fig. 4. Some detected dose-based metrics include dosimetric performance, dose distribution, tumour control probability (TCP) and normal tissue complication probability (NTCP), dosiomics, CTV D98, PTV D95, Cord Max Dose, other organs at risk (OARs) mean dose or conformity indices (Cl).

[0118] Clinical Workflow Integration

[0119] Fig. 5 shows how, in the clinical workflow, the system 100 automates, consolidates 502, and provides decision support in a radiotherapy centre’s re-plan process. The system 100 aligns with the clinician’s needs with the following benefits. The system 100 flags patients for review only upon significant deviations in delivered dose or contour metrics compared to the original treatment plan, minimising resources when replanning is not required. It consolidates metrics and visual data on a single screen of the III 110, eliminating manual report writing and analysis, thereby accelerating decisionmaking.

[0120] The system 100 also has a database 135 to provide comprehensive record-keeping, documenting every treatment fraction to capture detailed records of the treatment process, whether adjustments are needed, or treatment is proceeding asl planned. By systematically storing the results for all patients and treatment fractions, including those where a re-plan was not necessary, the system 100 builds a comprehensive evidence base. This extensive record-keeping is invaluable for verifying the accuracy and appropriateness of the course of treatment administered. It ensures that healthcare providers have a detailed and accurate treatment history at their fingertips, which can be useful for continuous quality improvement, patient care monitoring, and internal audits. This level of detail in record-keeping not only enhances the transparency of the treatment process but also reinforces the trust and confidence of patients and stakeholders in the healthcare system.

[0121] The system 100 outputs comprehensive reports, results, and alerts 80 to the OIS 200. The system 100 logs treatment fractions that have been opened or reviewed, or triaged cases that need reviewing. This logging enhances tracking and accountability.

[0122] The system 100 closely monitors patient treatment by automatically calculating the daily treatment dose and comparing it with the original treatment plan. This occurs seamlessly alongside the standard treatment workflow via the silent listener module 140 operating in the background. This automation significantly reduces the time and effort traditionally required for manual analysis and report generation.

[0123] The system 100 is a comprehensive tool for managing and monitoring patient treatment undergoing radiation therapy, ensuring adherence to treatment plans, and facilitating timely interventions when deviations occur. The system 100 includes a range of features to support and enhance the accuracy and efficiency of treatment planning and monitoring: configurable thresholds, automated reconstruction of daily dose, automated daily contouring of structures, automated comparison with original treatment plan, automated notification generation, automated generation of reports and alerts to the Oncology Information System (OIS), logging of review activity, triage of cases for review using the traffic light system, and the III 110 for performing reviews.

[0124] The system 100 enhances the treatment re-planning process by automatically computing dose-based metrics and the daily treatment dose for each patient, comparing these to the original treatment plan using the silent listener module 140 to reduce manual effort significantly. It can shorten the re-planning process by up to 11 days when necessary and by up to 4 days otherwise, improving overall efficiency.

[0125] In cases where an original treatment plan needs to be revised and adjusted, the system 100 can reduce the time needed for this re-planning process by up to 11 days. Even when a re-plan isn't necessary, the time savings can be up to 4 days. The system 100 intelligently identifies patients who may require further review only if it detects significant deviations in dose or contour metrics compared to the planning data. This targeted approach minimises unnecessary resource usage in situations where replanning isn't needed.

[0126] If the system 100 issues a re-plan notification, it specifies that the currently applied treatment plan (original treatment plan or preceding re-planned treatment plan) is inadequate, which may be caused by changes in the patient's anatomy affecting the targeted or surrounding structures. The system 100 indicates the necessity for a re-plan and identifies which structures are no longer receiving the appropriate dose. It is then up to the treating department to devise a new plan based on their established protocols. Typically, the process of re-planning is akin to the original treatment planning phase, with most departments capable of performing this task once the need is identified.

[0127] The way the system 100 detects significant deviations between planned and delivered dose 605 and between contour metrics 606 is described. The system 100 performs image registration 603 between two key images: the "image of the day," obtained via Cone Beam Computed Tomography (CBCT) scan 602, and the "image at Plan," derived from the synthetic Computed Tomography (CT) scan 604. This step aligns the daily treatment image (CBCT scan) with the baseline planning image (CT scan), ensuring that changes in patient anatomy can be accurately accounted for. It is important to determine what may be considered a significant anatomical for the system 100, namely, establishing the criteria or threshold that indicates the need for a treatment planrevision. Clinics are encouraged to set and configure their own thresholds for this purpose. Dose analytics are determined 607 per structure. Additionally, the evaluation process involves comparing 608 the delivered dose 605 against the Quantitative Analyses of Normal Tissue Effects in the Clinic (QUANTEC) database limits for each organ to determine the absolute limit of the dose received by each organ. This comparison may be a benchmark for deciding when the original treatment plan needs re- evaluation.

[0128] Following image registration, the system 100 recalculates the radiation dose distributed within the tissues. This recalculation uses the parameters of the Linac as defined in the original treatment plan but applies them to the anatomical structure captured in the "image of the day" (CBCT scan). A key aspect of this process is the transfer of the physician-defined contours from the planning CT to the daily CBCT. This ensures that the areas of interest, such as organs or target volumes, are consistently defined across both sets of images.

[0129] For each daily fraction of the treatment, Cone Beam CT (CBCT) scans are used. These are an example of daily fraction data. CBCT scans are used because it is the imaging technology available in the treatment room and typically taken during the treatment session. The CBCT scans are used for Deformable Image Registration (DIR) using bsline or demon algorithms, or using Al (e.g. a Transformer or unsupervised ll-net) and for calculating the dose on a daily basis. For daily fraction data, which is important for performing Deformable Image Registration (DIR) and dose calculation, CBCT scans are used. Image registration is performed to propagate the contour. This is because CBCT scans face several limitations in comparison to CT scans, including a restricted field of view, streaking artifacts, photon starvation artifacts (areas appearing darker than expected), and reduced resolution. Additionally, CBCT scans do not provide accurate electron density information, making dose calculation based on CBCT scan a highly complex task. A CBCT scan has such low image quality that performing contouring on a CBCT scan can be very erroneous and accurate, even with the use of machine learning models to assist with the process.i

[0130] With the recalculated dose information, the system 100 can provide detailed reports on the actual radiation dose delivered to and received by each organ or target volume. These reports compare the delivered dose against the planned dose for each specific area. If the system 100 detects a deviation between the delivered dose and the planned dose that exceeds a predetermined threshold, this discrepancy is highlighted and reported to the clinician as clinical decision support. This ensures that the intended treatment adheres closely to the planned protocol, allowing for adjustments as necessary to optimise patient care.

[0131] The system 100 may generate a comprehensive report in PDF format. This report includes the dose-volume scorecard, a dose map superimposed on relevant images, and the dose volume histogram. This report provides a robust basis for justifying any potential adjustments in the treatment plan. Also, the report facilitates the escalation process, enabling clinicians to present a well-substantiated case for review, whether for reconsideration of the current treatment plan or for seeking advice from a senior medical officer regarding the necessity to replan or maintain the existing treatment strategy.

[0132] The system 100 has a dual-faceted approach designed to integrate and display data through the unified III 110 to clinicians. The first facet involves the execution of image registration between a Cone Beam Computed Tomography (CBCT) scan and a synthetic Computed Tomography (CT) scan. This process enables accurate dose calculations to be performed on the more frequently taken CBCT scans by aligning them with corresponding synthetic CT scans, thus facilitating precise dose evaluation for each day's treatment (more frequent interval). The second facet focuses on the direct querying and retrieval of data from the clinic’s Digital Imaging and Communications in Medicine (DICOM) database, as a background process. This includes but is not limited to, data related to CT and CBCT scans, treatment plan(s), and dose information. By leveraging this capability, the system 100 ensures that all relevant parameters are seamlessly integrated and accessible within the III 110, thereby enhancing the efficiency and effectiveness of treatment planning and evaluation.IB

[0133] The system 100 provides a comprehensive solution that gathers all necessary metrics and analytics required for making decisions about treatment plan revisions. It consolidates this information in one central location, streamlining the process and reducing dependence on more cumbersome and less efficient methods.

[0134] The contours that the system 100 analyses daily include: variation on volume and source-to-surface distance (SSD). These evaluations are considered supplementary functions of the system 100. The principal functionality of the system 100 is assessing variations in radiation dose. This involves analysing dose distribution within the contours to ensure optimal delivery.

[0135] It is envisaged that the system 100 is expandable and can be enhanced by supporting additional imaging modalities such as Magnetic Resonance Imaging (MRI) and Positron Emission Tomography (PET). The system 100 may accommodate various treatment modalities, including other radiation therapy options like proton therapy, further broadening its applicability and effectiveness in cancer treatment. The system 100 is applicable to additional patient populations by implementing additional updates, beyond patients with head & neck cancer.

[0136] Referring to Figs. 18 and 19, in another embodiment of the present invention, the system 100 uses a multi-resolution dose calculation approach using the Collapsed Cone Convolution (CCC) algorithm to optimise computational efficiency while preserving dose calculation accuracy. Fig. 18 illustrates dose calculations for different slab materials, specifically air (-1000 HU) and lung (-700 HU), across varying depths and off-axis distances. The comparison is conducted using multiple dose calculation algorithms, including Collapsed Cone Convolution (CCC), TruckCCC, AAA, and AXB. The top row 180, 181 , 182, 183 represents dose calculations for air slabs, displaying depth dose profiles and dose distributions at different depths (23.8 mm, 73.8 mm, and 148.8 mm). This highlights variations in dose attenuation and scatter behaviour among the different algorithms when radiation passes through low-density, air-equivalent materials. The bottom row 184, 185, 186, 187 presents similar dose distributions for lungequivalent slabs, demonstrating how each algorithm models dose penetration throughmoderately low-density tissue. Differences in calculated dose profiles may result from variations in how the algorithms handle electron transport and scattering effects in lung tissue. Fig. 18 illustrated the performance of different dose calculation algorithms in handling heterogeneous materials, which ensures accurate radiation therapy treatment planning in complex anatomical regions. Fig. 19 illustrates the slab dose gamma analysis results 190, comparing the fail rates of different dose calculation algorithms across various material slabs characterised by Hounsfield Unit (HU) values. The x-axis represents different slab types, including air, HU-980, HU-700, HU-350, and water, while the y-axis indicates the fail rate percentage. Fig. 19 compares three different dose calculation models, AAA, AXB, and TPS-CCC, within a treatment planning system (TPS). Also, it illustrates the gamma error rate of the CCC-based dose calculation method of the system 100 against these models, including CCC vs. AAA, CCC vs. AXB, and CCC vs. TPS-CCC. The results indicate variations in fail rates depending on material density, with air and low-density slabs exhibiting higher failure rates compared to higher-density slabs such as water. This analysis provides a direct evaluation of the accuracy and consistency of the CCC-based approach of the system 100 and highlights the performance of different dose calculation models in handling heterogeneous tissue environments.

[0137] The system 100 first performs a coarse dose calculation over a large volume corresponding to the planning CT at a resolution of 10 mm3, generating an initial estimate of radiation distribution across the full anatomical region. A subsequent high- resolution dose calculation refines the dose distribution within a smaller volume of interest, corresponding to the CBCT region, at a resolution of 2.5 mm3to 3 mm3. The system 100 interpolates between the low- and high-resolution calculations using a weighted interpolation method (for example, purely heuristic) that accounts for local variations in regions where primary photon interactions extend beyond the smaller field of view (FOV). For regions completely outside the FOV that do not receive radiation traces, the system 100 applies a fill-in approach to reconstruct missing dose data. Preferably, the same interpolation technique is applied across all anatomical sites (head & neck, pelvis, etc), as no site requires a distinct interpolation method. This technique ensures that CCC’s accuracy in modeling radiation transport and interactions is preserved while significantly reducing computational requirements compared to full-resolution dose calculations over the entire anatomical volume.

[0138] The system 100 calculates the delivered dose on a synthetic CT scan rather than directly on the CBCT scan to account for tissue heterogeneity and anatomical deformations. The synthetic CT scan is generated by applying deformable image registration (DIR) to the planning CT, ensuring that anatomical structures are consistently aligned across imaging modalities. The system 100 uses a region-specific DIR strategy optimised for different anatomical sites to enhance registration accuracy. In prostate treatments, where bladder volume fluctuations contribute significantly to organ motion, the system 100 precomputes one hundred bladder volume variations and dynamically selects the best-matching configuration for each fraction. The registration process follows a prioritisation hierarchy, where anatomical components are registered sequentially based on their influence on deformation. The most significant driver of deformation is processed first, while less influential components are registered later to refine the alignment. For example, in prostate treatments, bladder registration is prioritised before overall transformation, whereas in breast treatments, soft tissue alignment is performed before finalising bony structure registration. This sequential approach ensures that dominant sources of anatomical variation are accurately incorporated into the final synthetic CT reconstruction. The bladder model incorporates shape-based deformation by first identifying the most likely deformed bladder from the precomputed library based on the current bladder shape and size. Once selected, the deformed bladder replaces the original bladder within the planning CT, which is then used as the reference for deformable image registration (DIR) against the CBCT scan, ensuring an anatomically accurate alignment and eliminating the need for continuous time-series modeling of bladder motion. In breast treatments, the system 100 first aligns bony structures before refining soft tissue registration to minimise localisation errors. The system 100 processes different tissue components, including soft tissue, bone, and air pockets, separately before merging them into a unified transformation field, ensuring that localised distortions are accurately represented in the synthetic CT reconstruction.

[0139] The system 100 determines default threshold values for dose deviation detection based on published treatment planning guidelines and large-scale clinical trial data. These threshold values define acceptable dose variation limits and provide a reference for identifying clinically significant deviations and to issue a trend-based alert. The system 100 allows clinicians to manually adjust deviation thresholds within patient-specific protocols, enabling personalised treatment management. The system 100 does not automatically adjust threshold values based on historical trend data but presents historical dose deviation metrics for clinician review. Future implementations will support both predefined rule-based predictive models and Al-driven machine learning regression models trained on historical patient data. The integration of Al-based models is contingent on the availability of a sufficiently large dataset to ensure robustness and reliability in trend-based predictions.

[0140] The system 100 tracks dose deviation trends over multiple treatment fractions, enabling trend-based analysis without accumulating dose metrics across fractions. Instead of aggregating dose deviations over an entire treatment course, the system 100 maintains a rolling window of past fractions, allowing clinicians to visualise longitudinal trends in dose distribution. The system 100 applies linear extrapolation to estimate potential deviations based on prior fractions, providing an early indication of cases at risk of exceeding predefined thresholds. These trends are displayed through numerical tables and graphical visualisations, ensuring that clinicians can interpret deviation patterns efficiently. Preferably, a user interface screen displays a structured visualisation of dose deviations across multiple treatment fractions, assisting clinicians in monitoring delivered radiation and identifying potential concerns. At the top of the interface, the “Structure and Metric” section displays the selected anatomical structure, such as the parotid gland, alongside the corresponding dose metric being analysed. Directly below, the “Analysed Metric and Unit” field specifies the exact measurement being tracked, such as mean dose (Dmean) in Gray (Gy), ensuring clarity in dose interpretation. The “Date of Analysis” indicates when the data was last reviewed or updated. Each fraction's delivered dose is represented by a black dot on the trend graph, with the most recent result highlighted under “This Fraction’s Result”, allowing clinicians to track variations in dose delivery over time.

[0141] The interface uses a color-coded threshold system to visually distinguish between different levels of deviation. The “Major Deviation” region, displayed in red, signifies doses exceeding acceptable clinical thresholds, potentially requiring treatment adjustments. The “Minor Deviation” region, shown in yellow, represents doses approaching the threshold but remaining within a tolerable range. The “In Tolerance”region, marked in green, indicates doses that fall within the expected safe range, ensuring treatment remains within planned parameters. To help clinicians interpret the plotted data, a “Plot Data Legend” differentiates between the planned dose per fraction and the delivered dose, facilitating comparisons. Additionally, a “Dropdown Menu for Structure and Metric Selection” allows users to select different anatomical structures and dose metrics for trend visualisation, enabling customised dose analysis based on specific clinical needs. This user interface enhances decision-making by offering a clear, intuitive representation of dose trends, enabling early detection of deviations and improving overall treatment accuracy.

[0142] The system 100 uses a priority-based triage mechanism to assist clinicians in making timely treatment re-planning decisions. Patients are dynamically ranked based on a predefined sorting hierarchy, wherein major deviations are prioritised above minor deviations, and cases without an assigned protocol or analysis errors are ranked highest for immediate resolution. The system 100 applies the following sorting logic: (1) Patients with unavailable results due to missing protocols or analysis errors are ranked highest, followed by (2) patients with major deviation results, (3) patients with minor deviation results, and (4) patients within tolerance or without analysis. Sorting is applied to the patient’s most recent three fractions, ensuring that cases with recurrent deviations remain visible in the priority queue. Unlike a weighting-based system where clinicians define custom ratios for different violations, the system 100 maintains a fixed hierarchy in its triaging logic.

[0143] The system 100 prioritises patients using a triage-based sorting mechanism that ranks patients based on the presence and severity of detected dose deviations. The system 100 sorts patients using a fixed hierarchy, wherein patients with missing protocols or analysis errors are assigned the highest priority, followed by those with major deviations, minor deviations, and cases within tolerance. The system 100 applies this sorting logic over the three most recent treatment fractions for each patient, ensuring that recurrent deviations remain visible for clinical review. The system 100 does not currently apply a weighting mechanism that escalates priority based on the cumulative occurrence of minor deviations over time. In one embodiment, the system 100 may include a dynamic priority adjustment mechanism that increases priority when apatient exhibits a progressive pattern of minor deviations. The system 100 may also ensure that all detected deviations are visible in the patient list for clinician review.

[0144] The system 100 retrieves patient imaging data from a DICOM database through a periodic query-based retrieval mechanism, ensuring timely access to the latest CT and CBCT scans for dose calculation and treatment analysis. If a queried image is missing or incomplete, the system 100 automatically attempts up to three retries to retrieve a valid dataset. The system 100 logs the failure in a retrievable audit log for post processing review. If the image remains unavailable after all retries, the system 100 flags the missing data in the patient list and logs a retrieval failure notification for clinician review. Referring to Fig. 17, the system 100 also provides a manual upload mode, allowing clinicians 170 to directly input missing imaging data if automatic retrieval is unsuccessful. The user 170 manually uploads data through the III web upload interface 171 , initiating the processing. Alternatively, the system 100 can retrieve imaging and treatment data from the clinical system database 179 via a gateway 178, which handles data retrieval and export operations. Both manually uploaded and retrieved data are then directed to a watch folder 172, which serves as a staging area for new data awaiting processing. The system 100 then performs data pre-processing 173 to refine and prepare the incoming data for further analysis. Once pre-processed, the data is temporarily stored in a staging area 174 before proceeding to results calculation 175, where the system 100 computes the dose distribution and performs analytical evaluations. The final dose results are then stored in the dose review database 176 for further clinical assessment, while additional processed results are archived in a separate results store 177 for reference and future analysis.

[0145] In cases where image registration fails due to missing reference scans, the system 100 automatically flags the failure in the III, providing an error message indicating the nature of the issue. The system 100 performs automatic matching of a reference scan based on the date of acquisition, selecting the closest available planning CT scan for alignment with the daily CBCT scan. If an appropriate reference scan cannot be identified, the system 100 allows clinicians to manually select a substitute reference scan if one is available.11

[0146] When multiple scans are acquired within the same treatment session, the system 100 determines the latest scan using metadata timestamps, prioritising the scan with the most recent acquisition time. The system 100 verifies imaging completeness by ensuring that the scan contains all expected slices and associated metadata before proceeding with analysis. If a scan is flagged as incomplete, the system 100 selects the most recent fully completed scan within the session. If no complete scans are available, the system 100 issues a warning and defers processing until a valid dataset is available.

[0147] Clinicians can manually review patients in any order through a search and filtering interface. While minor deviations cannot be manually escalated in the scorecard system, the system 100 allows clinicians to add case-specific notes, such as “Cord dose reviewed and replan required,” ensuring that significant observations are documented for clinical decision-making.

[0148] The system 100 retrieves imaging data from a DICOM database using a periodic query-based retrieval mechanism, polling the database at scheduled intervals to detect newly acquired imaging studies. The system 100 queries the database using metadata parameters, including patient ID, study ID, imaging modality, and acquisition time, ensuring that only CT and CBCT scans are processed. The system 100 filters out irrelevant imaging modalities such as MRI and PET to maintain workflow efficiency. If multiple scans are acquired within a single treatment session, the system 100 processes only the most recent scan to prevent redundant analysis. If multiple scans are acquired across separate treatment sessions on the same day, the system 100 processes each session independently to ensure that all relevant data is analysed.

[0149] The system 100 preprocesses CBCT scans to correct imaging artifacts and improve dose calculation accuracy before generating the synthetic CT. The system 100 applies field-of-view masking to compensate for CBCT’s limited imaging range, reconstructing missing anatomical regions using planning CT data. Referring to Fig. 16, the system 100 uses a slab-based histogram matching technique, normalising intensity values within localised image regions instead of applying global normalisation, thereby improving registration consistency between CBCT and planning CT images. The slabintensity matching process begins with two images: an input image 160 and a reference image 164, which need to be aligned to ensure consistency for further processing. To achieve this, the system 100 first performs slab extraction 161 , where image slices are divided along the matched x / y / z directions, breaking the images into localised regions rather than applying a global adjustment. Each extracted slab then undergoes histogram matching 162, a technique that aligns intensity distributions between the input and reference images. This step 162 compensates for variations in imaging modalities or acquisition conditions, ensuring a more accurate alignment. Once all slabs have been processed and their intensity values adjusted, the system generates a final matched image 163, which exhibits improved consistency with the reference image, facilitating more accurate deformable image registration and dose calculations. The slab parameters, including slab size, stride, and orientation, are preconfigured based on anatomical site and refined through empirical testing to achieve optimal registration performance. The system 100 separately registers different tissue components, including soft tissue, bone, and air pockets, before merging them into the final synthetic CT. The system 100 incorporates treatment setup components such as the couch and bolus by extracting their dimensions and material properties from the DICOM plan and structure set, ensuring the reconstructed synthetic CT accurately reflects the patient’s positioning during treatment.

[0150] The system 100 determines the appropriate beam model for dose calculation by extracting treatment parameters from the DI COM dataset. The system 100 selects the correct beam model based on machine ID, beam energy, multi-leaf collimator (MLC) configuration, and the presence or absence of a flattening filter. The system 100 validates the selected beam model against the treatment planning system dataset, flagging any discrepancies for clinician review. The validation process involves comparing the system’s dose calculation on the planning CT with the corresponding dose calculation performed by the treatment planning system over the same volume. Any discrepancies between the two calculations are reported and traceable, allowing the treatment team to take appropriate action. If a discrepancy is detected, the system 100 does not provide automatic recommendations but instead flags the issue for clinician review. The treatment team may then choose to (1) re-optimise for a new beam model, or (2) accept and document the difference, ensuring that the identified discrepancy is notedand excluded from future analyses. CCC dose calculations are precomputed for efficiency but are not cached for future retrieval, as each fraction's dose calculation is independently performed without carrying forward prior results beyond the planning CT dose distribution. The system 100 automatically compares CCC dose calculation results with those from the treatment planning system (TPS) and alerts clinicians to major deviations. Such a comparison is depicted in Fig. 18. These comparisons are displayed on a dedicated "Quality Assurance" page, where users can view side-by-side results of the system 100’s dose calculations against TPS-calculated doses. The system 100 highlights any discrepancies between the two calculations, allowing clinicians to assess the differences. Users also have the option to ignore dose metrics that exhibit significant deviations, ensuring that such discrepancies do not affect future analyses. The system 100 presents dose visualisation using heatmaps, where dose distribution is mapped based on isodose levels, using the prescribed dose as the reference. Clinicians can assess radiation exposure through isodose contours at predefined thresholds, such as 80%, 90%, 100%, and 110% of the prescribed dose, providing an intuitive representation of dose coverage and potential hotspots.

[0151] The system 100 verifies DIR accuracy using multiple validation techniques. The system 100 reports Jacobian metrics to quantify local volume changes induced by deformation, providing clinicians with a measure of registration accuracy without requiring external reference data. To validate the accuracy of the DIR algorithm itself, the system 100 performs manual landmark-based validation using Target Registration Error (TRE) measurements, ensuring that registration performance aligns with clinical expectations. TRE validation is primarily conducted during the system’s clinical evaluation as part of the manufacturer's responsibility. For users, TRE validation is only recommended during system commissioning to verify initial registration accuracy. The system 100 does not perform real-time TRE validation for each fraction due to practical constraints. Unlike segmentation-based validation methods such as the Dice similarity coefficient, which require ground-truth segmentations, Jacobian-based validation allows the system 100 to provide registration quality metrics for every processed case without additional manual input. The system 100 includes confidencebased warnings for Jacobian metrics, issuing an alert if more than 10% of voxels exhibitdeformation values outside the acceptable [0,1] range, ensuring clinicians are aware of potential inaccuracies in image registration.

[0152] The system 100 provides a comprehensive clinical decision support framework for automated dose review and treatment re-planning in radiation therapy. By integrating multi-resolution dose calculation, automated imaging analysis, deformable image registration, and trend-based deviation detection, the system 100 enhances treatment accuracy, workflow efficiency, and patient safety. Through automated workflow components and clinician-driven decision-making, the system 100 ensures that treatment deviations are detected early, enabling timely intervention and improved patient outcomes.

[0153] In this specification, terms such as ‘processor’, ‘computer’, and so forth, unless otherwise required by the context, should be understood as referring to a range of possible implementations of devices, apparatus and systems comprising a combination of hardware and software. This includes single processor and multi-processor devices and apparatus, including portable devices, desktop computers, and various types of server systems, including cooperating hardware and software platforms that may be colocated or distributed. Hardware may include conventional personal computer architectures, or other general-purpose hardware platforms. Software may include commercially available operating system software in combination with various application and service programs. Alternatively, computing or processing platforms may comprise custom hardware and / or software architectures. For enhanced scalability, computing and processing systems may comprise cloud computing platforms, enabling physicali hardware resources to be allocated dynamically in response to service demands. While all of these variations fall within the scope of the present invention, for ease of explanation and understanding, the exemplary embodiments described herein are based upon single-processor general-purpose computing platforms, commonly available operating system platforms, and / or widely available consumer products, such as desktop PCs, notebook or laptop PCs, smartphones, tablet computers, and so forth.

[0154] In particular, the term ‘processing unit’ is used in this specification (including the claims) to refer to any suitable combination of hardware and software configured to perform a particular defined task, such as generating and transmitting authentication data, receiving and processing authentication data, or receiving and validating authentication data. Such a processing unit may comprise an executable code module executing at a single location on a single processing device, or may comprise cooperating executable code modules executing in multiple locations and / or on multiple processing devices. For example, in some embodiments of the invention authentication processing may be performed entirely by code executing on a server, while in other embodiments corresponding processing may be performed cooperatively by code modules executing on the secure system and server. For example, embodiments of the invention may use application programming interface (API) code modules, installed at the secure system, or at another third-party system, configured to operate cooperatively with code modules executing on the server in order to provide the secure system with authentication services.

[0155] Software components embodying features of the invention may be developed using any suitable programming language, development environment, or combinations of languages and development environments, as will be familiar to persons skilled in the art of software engineering. For example, suitable software may be developed using the C programming language, the Java programming language, the C++ programming language, the Go programming language, and / or a range of languages suitable for implementation of network or web-based services, such as JavaScript, HTML, PHP, ASP, JSP, Ruby, Python, and so forth. These examples are not intended to be limiting, and it will be appreciated that convenient languages or development systems may be used, in accordance with system requirements.

[0156] In the exemplary system, the devices each comprise a processor. The processor is interfaced to, or otherwise operably associated with, a communications interface, one or more user input / output (I / O) interfaces, and local storage, which may comprise a combination of volatile and non-volatile storage. Non-volatile storage may include solid-state non-volatile memory, such as read only memory (ROM) flash memory, or the like. Volatile storage may include random access memory (RAM). The storage contains program instructions and transient data relating to the operation of the device. In some embodiments, the device may include additional peripheral interfaces, such as an interface to high-capacity non-volatile storage, such as a hard disk drive, optical drive, and so forth.

[0157] The processor of a computer is interfaced to, or otherwise operably associated with a non-volatile memory / storage device, which may be a hard disk drive, and / or may include a solid-state non-volatile memory, such as ROM, flash memory, or the like. The processor is also interfaced to volatile storage, such as RAM, which contains program instructions and transient data relating to the operation of the server.

[0158] In a conventional configuration, the storage device maintains known program and data content relevant to the normal operation of the server. For example, the storage device may contain operating system programs and data, as well as other executable application software necessary for the intended functions of the server. The storage device also contains program instructions which, when executed by the processor, instruct the server to perform operations relating to an embodiment of the present invention, such as are described in greater detail. In operation, instructions and data held on the storage device are transferred to volatile memory for execution on demand.

[0159] The processor is also operably associated with a communications interface in a conventional manner. The communications interface facilitates access to a data communications network.■

[0160] In use, the volatile storage contains a corresponding body of program instructions transferred from the storage device and configured to perform processing and other operations embodying features of the present invention.

[0161] It should be appreciated that while particular embodiments and variations of the invention have been described herein, further modifications and alternatives will be apparent to persons skilled in the relevant arts. In particular, the examples are offered by way of illustrating the principles of the invention, and to provide a number of specific methods and arrangements for putting those principles into effect.

[0162] Accordingly, the described embodiments should be understood as being provided by way of example, for the purpose of teaching the general features and principles of the invention, but should not be understood as limiting the scope of the invention, which is as defined in the appended claims.

Claims

CLAIMS:

1. An automated clinical decision support method for treatment re-planning of a patient, comprising: enhancing a Cone Beam Computed Tomography (CBCT) scan of the patient using information from an initial CT scan of the patient to create a synthetic CT scan of the patient; adapting contours of organ(s) and tumour(s) from the initial CT scan to align with the synthetic CT scan; and calculating a delivered radiation dose from the synthetic CT scan; wherein the calculated delivered radiation dose and adapted contours are compared to an original treatment plan to detect whether a deviation in at least one dosebased metric from the original treatment plan has occurred, and if a detected deviation exceeds a predetermined threshold value for the respective dose-based metric, a notification with an assigned priority level is generated to alert a clinician for a re-plan decision, the assigned priority level corresponding to the extent of the deviation from the predetermined threshold value.

2. The method according to claim 1, wherein the predetermined time interval is 24 hours.

3. The method according to claim 1, further comprising configuring at least one protocol template on a patient basis, and if a new patient is treating using the at least one protocol template, a protocol of the at least one protocol template is applied to triage the delivered radiation dose into triage categories, wherein the protocol is used to identifyi which structures a clinician is most concerned about to populate a dose-volume score card and histogram, and to identify a decision threshold per structure.

4. The method according to claim 3, wherein the at least one protocol is determined according to any one from the group consisting of: a type of radiotherapy centre, a type of cancer and a type of treatment.

5. The method according to claim 1, further comprising an initial step of: predicting the synthetic Computed Tomography (CT) scan using one or more from the group consisting of: planning CT scan, structure, dose file, treatment plan data, Cone Beam Computed Tomography (CBCT) scan, Magnetic Resonance Imaging (MRI) scan, Positron Emission Tomography (PET) scan6. The method according to claim 1, further comprising: performing Deformable Image Registration (DIR) to obtain the transform between the initial CT scan or subsequent CT scans, and the CBCT scan.

7. The method according to claim 1, further comprising: generating trend data of treatment fractions.

8. The method according to claim 1, wherein the notification of a clinical change for a re-plan decision includes a visual indicator with a plurality of sections comprising: a first section having a first colour indicating major thresholds exceeded; a second section having a second colour indicating minor thresholds are exceeded; anda third section having a third colour indicating thresholds are within thresholds.

9. The method according to claim 1, wherein at least one dose-based metric comprises any one from the group consisting of: dosimetric performance, dose distribution, tumour control probability (TCP) and normal tissue complication probability (NTCP), dosiomics, CTV D98, PTV D95, Cord Max Dose, other organs at risk (OARs) mean dose and conformity indices (Cl).

10. The method according to claim 1, wherein the enhancement of the CBCT scan further comprises performing field of view masking to compensate for limited CBCT field of view and enable accurate anatomical representation for dose calculation.

11. The method according to claim 1 , wherein the enhancement of the CBCT scan further comprises performing slab-matching, wherein intensity matching is performed in discrete image slabs to improve alignment between the synthetic CT scan and the CBCT scan.

12. The method according to claim 1, wherein the enhancement of the CBCT scan further comprises processing different anatomical components, including soft tissue, bones, and air pockets, separately to refine the synthetic CT scan.

13. The method according to claim 1, wherein the enhancement of the CBCT scan further comprises incorporating updated patient setup information, including couch and bolus structures, to ensure that dose calculations accurately reflect actual treatment conditions.if14. The method according to claim 1, wherein the contours are adapted by applying body-site-specific transformations based on anatomical deformation patterns.

15. The method according to claim 14, wherein the selection of the body-site-specific transformation methods is determined by predefined anatomical site-specific protocols, wherein different deformation models are applied to different regions.

16. The method according to claim 1, wherein calculating the delivered radiation dose includes performing a single-resolution or multi-resolution dose calculation, wherein in the case of multi-resolution dose calculation, a low-resolution dose calculation is performed over the synthetic CT scan and a high-resolution dose calculation is performed over a smaller region of interest within the synthetic CT scan.

17. The method according to claim 16, wherein a final dose calculation result is obtained by interpolating between the low-resolution and high-resolution dose calculations.

18. The method according to claim 17, wherein the delivered radiation dose is calculated using a computer-based dose calculation algorithm from any one of the group consisting of: Collapsed Cone Convolution (CCC), Pencil Beam, Monte Carlo, and the Linear Boltzmann Transport Equation, and wherein the computation is performed using a central processing unit (CPU) and / or a graphics processing unit (GPU).

19. The method according to claim 1, wherein detecting whether a deviation exceeds the predetermined threshold value further comprises analysing dose trends across multiple treatment fractions.I20. The method according to claim 19, wherein the analysing dose trends includes predicting future dose deviations based on analysing prior dose variations.

20. The method according to claim 1, wherein the assigned priority level in the generated notification is determined based on the severity and recurrence of dose deviations.

21. The method according to claim 20, wherein the assigned priority level in the generated notification is determined based on a scoring system that assigns higher priority to patients with more significant or repeated deviations.

22. The method according to claim 20, wherein the assigned priority level in the generated notification is determined based on custom weighting of different types of deviations, wherein a user-defined protocol enables assigning different priority levels to minor and major threshold violations.

23. The method according to claim 1, further comprising determining dose discrepancies by comparing calculated dose distributions to reference values from the treatment planning system and flagging inconsistencies for review by the clinician.

24. An automated dose re-planning system for treatment re-planning of a patient comprising: a CBCT enhancement component to enhance a Cone Beam Computed Tomography (CBCT) scan of the patient using information from an initial CT scan of the patient to create a synthetic CT scan of the patient;il a contour adaptation component to adapt contours of organ(s) and tumour(s) from the initial CT scan to align with the synthetic CT scan; and a dose calculation component to calculate a delivered radiation dose from the synthetic CT scan; wherein the calculated delivered radiation dose and adapted contours are compared to an original treatment plan to detect whether a deviation in at least one dosebased metric from the original treatment plan has occurred, and if a detected deviation exceeds a predetermined threshold value for the respective dose-based metric, a decision support module issues a notification with an assigned priority level is generated to alert a clinician for a re-plan decision, the assigned priority level corresponding to the extent of the deviation from the predetermined threshold value.

25. A computer software product comprising a sequence of instructions storable on one or more computer-readable storage media, said instructions when executed by one or more processors, cause the processor to: enhancing a Cone Beam Computed Tomography (CBCT) scan of the patient using information from an initial CT scan of the patient to create a synthetic CT scan of the patient; adapting contours of organ(s) and tumour(s) from the initial CT scan to align with the synthetic CT scan; and calculating a delivered radiation dose from the synthetic CT scan; wherein the calculated delivered radiation dose and adapted contours are compared to an original treatment plan to detect whether a deviation in at least one dosebased metric from the original treatment plan has occurred, andif a detected deviation exceeds a predetermined threshold value for the respective dose-based metric, a notification with an assigned priority level is generated to alert a clinician for a re-plan decision, the assigned priority level corresponding to the extent of the deviation from the predetermined threshold value.

Images