Design, Customization, and Implementation of an Oncology Information System-Based Paperless Proton Therapy Workflow

Abstract

Purpose

Proton therapy centers present distinctive workflow challenges that demand seamless coordination, documentation, and collaboration across multidisciplinary teams. Given the limited published experience with paperless workflows in proton therapy, this study describes the design, customization, and 2-year clinical implementation of a comprehensive paperless workflow at a newly commissioned multi-room proton therapy center using a commercial oncology information system (OIS).

Methods

A multidisciplinary team of radiation oncologists, medical physicists, and radiotherapy technologists led the development and implementation of a department-wide paperless workflow with the Elekta MOSAIQ OIS. The process followed a structured 5-phase strategy and leveraged native OIS functionalities, including Quality Control Lists (QCLs) for task management, virtual locations, IQ Scripts for automation, customized assessment forms, and eScribe templates. User acceptance was evaluated via a multidisciplinary satisfaction survey conducted 26 months post-implementation.

Results

The paperless workflow covered the entire patient pathway, from pre-planning audit to treatment completion and discharge summary generation, and incorporated 22 QCLs, 13 virtual locations, 44 IQ scripts, 10 customized assessment forms, and 2 eScribe templates. Implementation reduced the reliance on paper, improved workflow traceability, standardized multidisciplinary approvals, reduced missed tasks, and enhanced interdisciplinary coordination. A 3-tier physics plan review process ensured safety and quality assurance in treatment planning and evaluation while reducing variability among planners. The satisfaction survey (n = 36) yielded an overall mean score of 3.79/5.0, with safety and quality rated highest (3.87/5.0), while system usability identified opportunities for future optimization (3.53/5.0).

Conclusions

A fully paperless, OIS-driven proton therapy workflow was designed and successfully implemented using the native capabilities of the MOSAIQ OIS system. The structured, multidisciplinary approach and proton-specific checklist framework presented here provide a practical and reproducible model for other centers establishing particle therapy programs.

Introduction

The global expansion of charged-particle radiotherapy centers has introduced unique operational challenges, requiring workflow solutions that go beyond those developed for conventional photon radiotherapy.1, 2, 3 These centers are characterized by complex, multidisciplinary workflows that place exceptional demands on coordination, documentation, and quality assurance. In this context, traditional paper-based patient charts are inadequate—they slow interdisciplinary communication, increase documentation errors, and introduce avoidable risk into complex clinical processes.⁴ Dedicated Oncology Information Systems (OIS) address these limitations by centralizing patient data, automating task assignment, and providing structured electronic documentation, enabling safer and more efficient radiotherapy workflows.5, 6

Drawing on the American Association of Physicists in Medicine (AAPM) Task Group 262 (TG-262)⁷ and TG 275⁸ we developed and implemented a dedicated paperless workflow for our newly established proton therapy center using the MOSAIQ OIS (Elekta Inc, California, USA). By customizing various functions and tools within MOSAIQ, we established workflows spanning the entire patient journey, from pre-planning audit (PPA) to treatment completion, with dedicated checklists for physicists and radiotherapy technologists. Despite the clear need for OIS-based paperless workflows in particle therapy, published implementation experiences from dedicated proton therapy facilities remain limited. This article presents our 2-year implementation experience, evaluated through workflow performance metrics and a multidisciplinary user satisfaction survey.

Materials and methods

General infrastructure and staffing

The MOSAIQ OIS (Version 2.83) was utilized to establish the paperless workflow, with licenses installed across various locations within the department. The specific modules and functions utilized are described in Table 1. The proton therapy facility is equipped with an IBA Proteus Plus system (Ion Beam Applications, Belgium) comprising 3 360° rotating gantry treatment rooms dedicated to pencil beam scanning (PBS) delivery, each fitted with a robotic patient positioning system (PPS; Leoni Cia Cable Systems S.A.S., France). Patient position verification is performed using 2D orthogonal radiographs and 3D cone-beam CT (CBCT) acquired with 2 orthogonal X-ray systems (Varian Medical Systems, Palo Alto, CA, USA). Treatment planning is performed using RayStation (RaySearch Laboratories, Stockholm, Sweden; Version 12A).

Table 1.

OIS licenses and functions utilized for workflow customization and implementation.

MOSAIQ Tool/Function	Role or Function	Examples	Count
IQ Scripts	Workflow automation	Automatic task generation, skip-and-redirect logic, QCL triggering	44
Quality Control Lists (QCL)	Task assignment and workflow communication	Image export/import, simulation, treatment planning, pre-treatment verification	35
Observation Views and Assessments	Electronic forms and checklists	Simulation chart, prescription form, plan technical checklist, PSQA results form	20
Virtual Locations	Staff- and process-specific locations	Treatment machines, CT simulator, Physics Whiteboard, RO Whiteboard	13
Schedules	Patient appointment and scheduling	Proton therapy treatment, CT simulation	—
eScribes and eScan	Document preparation and document scanning	Patient consent form, treatment discharge summary	—
Plan Evaluation module	Treatment plan evaluation	—	—
Audit and reporting tools	Audit trails and data reporting	—	—
Diagnosis and Problems	Patient diagnosis and toxicity recording	—	—

Count, number of instances utilized; QCL, Quality Check List; PSQA, Patient-Specific Quality Assurance; RO, Radiation Oncology.

The proton therapy center operates as a specialized sub-center within an established parent radiotherapy department, treating approximately 4000 patients annually. Medical physicists and radiotherapy technologists are assigned exclusively to the proton facility, whilst radiation oncologists share the workload with the parent center. The center is supported by 21 radiation oncologists, 2 radiation oncology residents, 11 medical physicists, 21 radiotherapy technologists, 1 clinical coordinator, and 3 receptionists. A dedicated OIS Core Team, comprising 3 radiation oncologists (including the clinical lead), 2 medical physicists, and 3 radiotherapy technologists, was formed to lead the design, configuration, and implementation of the paperless workflow.

Consensus process and timeline

The OIS server, database, and client modules were installed alongside the RayStation TPS during the proton therapy facility installation phase. The workflow design and implementation process proceeded in 5 phases, as illustrated in Figure 1. Phase 1 encompassed system acceptance and installation, formation of the OIS Core Team, delineation of individual responsibilities, and iterative brainstorming and review sessions to finalize the departmental workflow design, spanning approximately 4 months. Phase 2 comprised staff training, including 1 week with an Elekta applications specialist, completed in 3 weeks. Phases 3 and 4 required intensive effort from the OIS Core Team to customize the workflow using MOSAIQ tools and functions, as detailed in Table 1; this spanned approximately 2 months. Workflow customization was performed first in the Practice database, followed by testing, validation, and replication into the Clinical database. Before going live, the finalized workflow was presented to all staff during category-specific training sessions. In total, the period from the formation of the OIS Core Team to the first patient treatment spanned approximately 11 months.

Figure 1.

Five-phase implementation framework for the OIS-based paperless workflow at the proton therapy center.

Design of checklists

Six specialized proton-specific checklists were developed using the MOSAIQ assessment module and embedded into the workflow via IQ script-driven automation. These comprises of (a) Plan Technical Checklist: Independent physics review prior to physician assessment; (b) Planners Checklist: Self-audit by treatment planners prior to clinician approval; (c) Physics Plan Review: physics review following plan approval; (d) OIS scheduling and PSQA results review; (e) Pre-treatment Checklist: technologist verification prior to the first fraction; and (f) Weekly treatment chart check. The electronic workflow ensures that downstream processes (QA, scheduling, treatment delivery) cannot proceed until all mandatory physics review items are completed and electronically signed within MOSAIQ. The framework was validated during Phase 4 and iteratively refined based on end-user feedback prior to clinical implementation.

Results

Workflow design and incorporation in OIS

The departmental workflow was stratified by professional category: radiotherapy technologists, physicians, physicists, and reception staff, and decomposed into individual work profiles and locations, as illustrated in Figure 2. The workflow implementation used a range of MOSAIQ tools and functions, including IQ scripting for automation, Quality Control Lists (QCLs) for task assignment, Observation Views and Assessments for electronic forms and checklists, and virtual locations for staff-specific workflow visibility, as detailed in Table 1. A detailed list of the IQ script summary is given in Supplementary Tables S1-S3. The workflow spans pre‐planning audit (PPA) to treatment completion; initial consultation and patient selection for proton therapy, which may occur across different institutional campuses, fall outside the scope of this OIS-based workflow. Once a patient is accepted for proton therapy, the treating physician schedules a PPA—a multidisciplinary case review conducted within a virtual MOSAIQ location accessible to all staff. This review documents the clinical rationale for proton therapy selection, simulation dates, immobilization requirements, and treatment-specific technical specifications, with dedicated fields for patient-specific considerations such as the presence of cardiac implantable electronic devices, high-density metallic implants, and motion management requirements. These details are recorded in an auto-generated assessment form (Supplementary Figure A), completed by the senior resident and countersigned by the treating consultant, ensuring dual-level physician documentation. Patient-specific allergies and relevant clinical alerts are additionally flagged through the MOSAIQ Alerts and Allergy module, ensuring visibility across all staff workspaces throughout the workflow. Where such considerations have direct implications for planning or delivery, these are cross-documented in the prescription form and simulation chart to ensure the planning physicist and treating radiotherapy technologist are explicitly informed. Completion of the Simulation QCL generates the simulation chart (Supplementary Figure B), documenting the complete proton-specific immobilization setup, including device specifications and indexed couch-top positioning details. Following independent sign-off by a second radiotherapy technologist, an image export task is automatically triggered, subsequently generating an image import task at the TPS Physics Whiteboard. Upon completion, fusion and contouring QCLs are automatically assigned to the TPS RO Whiteboard.

Figure 2.

Departmental workflow implemented in the OIS using various system functionalities.

When a QCL cannot be completed, users may skip it, automatically prompting an assessment form requiring documentation of the reason. All such forms are tabulated under the “Re-works” tab, supporting event tracking and workflow transparency. This skip-and-redirect functionality operates across multiple workflow stages, including planning, contouring, and image acquisition, and maintains a complete audit trail of all deviations.

Completion of the fusion and contouring QCL generates the dose prescription form (Supplementary Figure C), which requires sequential review and electronic sign-off by the senior resident, followed by the treating radiation oncologist. Upon approval, the treatment planning QCL is automatically assigned to the TPS Physics Whiteboard. Each plan undergoes multi-level physics verification through dedicated checklists, as described in the following section. Completion of the planning QCL notifies physicians for plan evaluation, with 2 independent radiation oncologists completing sequential Clinical Approval-1 and Clinical Approval-2 QCLs. Any plan modification is documented through a “Reasons for Re-planning” assessment form, which automatically generates a re-planning QCL and maintains a comprehensive audit trail.

Upon dual-physician plan approval, the Physics Review QCL is automatically triggered; on completion, PSQA generation and scheduling tasks are assigned to the Physics Whiteboard. To facilitate handover between physics team members, 2 dedicated QCLs were developed: the “QA Generation and Scheduling” QCL, which on completion prompts the PSQA Execution Instructions form (Supplementary Figure D), capturing phantom specifications, beam count, and measurement depths; and the “QA Execution and Analysis” QCL, which on completion prompts the PSQA Results Record form (Supplementary Figure E), documenting measurement and analysis results. In our institutional protocol, beam-specific 2D fluence measurements are performed at a minimum of 2 depths within the spread-out Bragg peak (SOBP). QA results outside acceptable tolerances trigger either a “Repeat-QA” or “Re-planning” QCL based on the nature and magnitude of the deviation. Successful QA validation automatically triggers the Clinical Plan Scheduling QCL.

Completion of the Clinical Plan Scheduling QCL generates the “Review QA and Plan Scheduling” QCL, prompting an independent second physicist to verify plan scheduling parameters and PSQA results via the PSQA and Plan Scheduling Review form (Supplementary Figure F) prior to treatment initiation. This simultaneously generates 2 concurrent QCLs: the “Pre-treatment Checklist” assigned to the RTT Whiteboard and the “Call Patient for Treatment” QCL assigned to the Reception virtual location. Prior to the first fraction, radiotherapy technologists complete the Pre-treatment Checklist (Supplementary Table S4), which covers patient identity verification, setup confirmation, plan and prescription verification, consent validity, accessory checks, and a collision-detection dry run. During treatment, technologists may generate an “Offline CBCT Review” QCL to prompt physician evaluation of acquired CBCT images. Upon completion of the final fraction, a “Treatment Completion” QCL notifies the treating physician. Dedicated eScribe templates for consent forms and discharge summaries automatically fetch relevant treatment details from the MOSAIQ database, ensuring standardized documentation.

The adaptive replanning workflow was not part of the original clinical go-live configuration but was developed and incorporated into the OIS following accumulated clinical experience, reflecting the iterative nature of the implementation process. The adaptive replanning process begins with the “Offline CBCT Review” QCL, generated by the radiotherapy technologist and assigned to the treating physician's whiteboard. The physician reviews the CBCT images and, if adaptive replanning is indicated, documents this decision in the patient record notes. A “Check Scan” QCL is then assigned to the CT simulator virtual location. Following check scan acquisition, the physician fuses the new scan with the original dataset and propagates target and organ-at-risk contours. A “Check Scan Analysis” QCL is subsequently assigned to the Physics Whiteboard, where the physicist recalculates the original plan on the new scan and determines whether full re-optimization and a new treatment plan are required. If no significant dosimetric changes are observed, the original plan is continued; otherwise, the workflow proceeds through the standard planning pathway—including prescription update, multi-tier physics plan review, PSQA, and independent plan scheduling verification—without requiring a separate parallel adaptive process. This design was intentional: rather than establishing a dedicated adaptive workflow requiring separate maintenance and staff training, the existing automated QCL cascade was leveraged to integrate adaptive replanning seamlessly into the established framework, reducing complexity whilst preserving comprehensive documentation and audit trails.

Plan and treatment parameter review checklists process

A 3-tier comprehensive treatment plan review process was designed and implemented using MOSAIQ assessment forms, QCLs, and IQ scripting, as summarized in Table 2. Level 3 is intentionally streamlined relative to Levels 1 and 2 to reduce review time at the post-approval stage, having been progressively refined from an initially comprehensive review based on accumulated clinical experience. The framework encompasses 8 categories of verification: general demographics and plan identification; patient data management; contouring and structure review; geometric parameters; plan design and calculation; beam parameters; optimization and robustness; dosimetric evaluation; and administrative and compliance checks. Notably, the 3-tier framework represents an evolution from the original 2-tier design; a dedicated Plan Technical Checklist (Level 2) was introduced following clinical experience that demonstrated the need for independent physicist review prior to physician evaluation, given the complexity of proton treatment planning.

Table 2.

Three-tier physics plan review checklist for proton therapy

Category	Specific Checkpoints	Description	Level 1: Planner Self-Check	Level 2: Peer Review (Independent Physicist)	Level 3: Post-Approval Check
General Demographics	Patient & Plan ID	Verify patient demographics (name, case number, gender, DOB) match between TPS and OIS.	✓	✓
	Verify plan name corresponds to the generated QCL.			✓	✓
Patient Data Management	CT Simulation Parameters	Confirm the correct CT image set is selected and named per institutional nomenclature; verify for adaptive/phase-wise planning.	✓	✓
		Confirm that the correct CT scanner specifications are applied.	✓	✓
		Confirm the appropriate HU-to-density conversion curve for the scanner and protocol.	✓	✓
		Confirm scan date corresponds to the correct simulation or mandated imaging dataset.	✓	✓
Contouring & Structures	Artifacts, Targets, OARs & Support Structures	Review CT for artifacts and foreign implants; assign density overrides to implants and deformation artifacts.	✓	✓
		Confirm target contours are smoothed, geometrically consistent, and named per departmental SOP.	✓	✓
		Confirm OARs are delineated with correct laterality per institutional protocols.	✓	✓
		Confirm external body contour includes immobilization devices and ancillary equipment.	✓	✓
		For ARO: confirm CT contour propagation, additional structure delineation, and density overrides.	✓	✓
		Confirm fiducial markers are contoured with appropriate density overrides.	✓	✓
Geometric Parameters	Setup & Collision Checks	Verify patient orientation (e.g., HFS, FFS) in the treatment plan.	✓	✓	✓
		Confirm the couch rail position (Rail-In/Out) does not intersect the planned beam path.	✓	✓
		Confirm the correct couch top model is inserted and defined within the calculation volume.	✓	✓
		Confirm CBCT collision check structure assesses gantry/CBCT proximity.	✓	✓
		Confirm that the CBCT FoV check structure ensures target coverage within the imaging volume.	✓	✓
		Confirm localization fiducial point (Origin POI) and isocenter (ISO POI) location, type, and XYZ coordinates.	✓	✓	✓
		Verify patient setup shifts relative to the plan isocenter and origin.	✓	✓	✓
Plan Design & Calculation	Prescription, Algorithm & Settings	Verify that the prescription (dose/fraction, fractions, total dose) matches the physician-approved prescription form.	✓	✓	✓
		Verify the beam model corresponds to the allocated treatment machine.	✓	✓	✓
		Confirm that the calculation grid size is appropriate for the site and beam characteristics, and that the calculation volume encompasses all relevant anatomy, including immobilization devices.	✓	✓
		Confirm an appropriate validated algorithm is selected (Monte Carlo preferred).	✓	✓
		Verify Monte Carlo computational settings, including statistical uncertainty thresholds.	✓	✓
		Confirm final dose calculation is complete and verified.	✓	✓	✓
Beam Parameters	Beam Nomenclature, Configuration, Energy & Modulation	Verify beam nomenclature adheres to departmental SOP.	✓	✓	✓
		Confirm the number of beams is clinically appropriate and in accordance with departmental planning policy.	✓	✓
		Confirm that the energy layer and spot spacing are appropriate for the target dimensions.	✓	✓	✓
		Confirm range shifter selection is appropriate for the treatment geometry.	✓	✓	✓
		Verify that the minimum MU per spot and per energy layer meet the machine-defined delivery thresholds.	✓	✓	✓
		Confirm ions per spot are within optimizer-defined thresholds.	✓	✓
		Verify per-beam settings: block structure, radiological depth, energy margins (lateral/proximal/distal), and repainting cycles.	✓	✓
		Confirm Bragg peak positions are clinically appropriate relative to the target and critical structures.	✓	✓
		Verify beam parameters, air gap distance, collision checks, and accessory positions (snout, range shifter) for each field.	✓	✓	✓
		Confirm air gap measurements and collision checks are performed and documented for each beam.	✓	✓	✓
Optimization & Robustness	Optimization & Robustness Settings	Confirm that optimization objectives for all targets and OARs are clinically appropriate.	✓	✓
		Confirm optimization settings respect the allocated machine's delivery limitations.	✓	✓	✓
		Confirm robustness settings (setup and range uncertainty margins) are appropriate for the site and consistent with the prescription.	✓	✓	✓
		Confirm robustness objectives are applied to all relevant targets and OARs.	✓	✓
		For ARO: confirm the correct secondary CT, cavity-filling delineation, and density assignments.	✓	✓
Dosimetric Evaluation	Target Coverage, Nominal and Robust Dose, Radiobiological Evaluation (LET, Variable RBE)	Confirm nominal dose, robust scenarios, LET distribution, and McNamara variable RBE dose are calculated prior to review.	✓	✓	✓
		Confirm target coverage meets departmental acceptance criteria under nominal and robust scenarios.	✓	✓	✓
		Confirm target hotspots are within acceptable limits.	✓	✓	✓
		Confirm OAR dose goals are met or documented with justification; constraints consistent with applicable protocol or institutional guidelines.	✓	✓	✓
		Confirm that the LET evaluation is performed for critical structures in applicable cases.	✓	✓	✓
		Confirm variable RBE-weighted dose (McNamara model) evaluation is performed.	✓	✓	✓
		Confirm robust dose evaluation (including voxel-wise worst-case scenarios) per prescription and departmental policy.	✓	✓	✓
		Confirm the maximum dose per beam is within the acceptance criteria.	✓	✓	✓
Administrative & Compliance	Documentation & Approval	Confirm planning comments are documented in designated fields.	✓	✓	✓
		Confirm all prior review-level comments have been addressed.			✓
		Confirm that the beam path structure (20% beam-specific isodose) is available for CBCT-based image-guidance matching.			✓
		Confirm the plan is ready for Clinical Approval-1, or document the reasons for re-planning.	✓		✓

QCL, Quality Control List; ARO, Anatomical Robust Optimization; OAR, Organ at Risk; TPS, Treatment Planning System; LET, Linear Energy Transfer; RBE, Relative Biological Effectiveness; POI, Point of Interest; MU, Monitor Units; CBCT, Cone Beam CT; FoV, Field of View; SOP, Standard Operating Procedure; HFS, Head First Supine; FFS, Feet First Supine.

The combined PSQA result review and plan scheduling checklist (Supplementary Figure F) was subsequently integrated into the automated QCL-driven workflow following recognition that OIS-based plan scheduling is performed independently of the TPS and is therefore susceptible to transcription errors not captured by physics plan review alone. The Pre-Treatment Checklist (Supplementary Table S4), completed by the radiotherapy technologist prior to the first fraction, provides structured verification of patient identity, simulation setup, treatment plan parameters, consent validity, accessory availability, and a collision detection dry run.

Workflow deviations and error tracking

An audit of errors and deviations identified during the 2-year implementation period is summarized in Table 3. We identified 21 error categories spanning registration, simulation, contouring, prescription, treatment planning, and plan scheduling. The most frequent deviation was non-adherence to institutional beam nomenclature (n = 34), followed by incomplete contours (n = 24) and dose-prescription inconsistencies (n = 12). Most errors were intercepted by the automated workflow; however, 2 incidents were detected later. In 1 case, the wrong number of fractions was scheduled and was only identified by the technologist after the first fraction had been delivered. In the other case, a motion-management technique was not attached to the beam but was corrected before beam delivery. Both incidents prompted targeted workflow refinements and updates to the checklist.

Table 3.

Summary of errors and deviations identified during the 2-year workflow implementation period.

No.	Error / deviation	Freq.	Stage at which error occurred	Stage at which error was detected (n)	Corrective action
Registration / Simulation Errors
1	Patient nomenclature error	6	Registration	PPA (3); CT simulation (3)	Corrections made
2	Metal indexer present within the treatment region, resulting in image artifacts	1	Simulation	Contouring (1)	Re-simulation
3	Wrong image set selected for contouring	4	Contouring	Contouring (1); Treatment planning (3)	Image set selection verified and corrected
4	Simulation without dental evaluation (dental artifact)	3	Simulation	Contouring (1); Treatment planning (2)	Re-simulation with metal artifact reduction algorithm; artifact evaluation added as a checklist element
5	Simulation without appropriate immobilization device	4	Simulation	Contouring (3); Treatment planning (1)	Re-simulation
6	Insufficient scan length acquired	2	Simulation	Contouring (1); Treatment planning (1)	Re-simulation
Contouring / Prescription Errors
7	Dose prescription inconsistent with target structures drawn	12	Contouring	Treatment planning (8); Plan Technical Check - Tier 2 (4)	Prescription revision or contour renaming
8	Incomplete contours	24	Contouring	Treatment planning (15); Plan Technical Check Tier 2 (7); Physics Review - Tier 3 (2)	Recontouring; recontouring and re-planning
9	Incorrect dose prescription entered in the prescription form	2	Prescription generation	Treatment planning (2)	Prescription correction
Treatment Planning Errors
10	Incorrect dose prescription selected while planning	4	Treatment planning	Plan Technical Check - Tier 2 (2); Physics Review - Tier 3 (2)	Re-planning
11	Incorrect beam model selected	3	Treatment planning	Plan Technical Check - Tier 2 (3)	Re-planning
12	Beam path traversing metallic implant or artifact not addressed during planning	8	Treatment planning	Plan Technical Check - Tier 2 (5); Physics Review - Tier 3 (3)	Re-planning
13	Suboptimal air-gap selection	8	Treatment planning	Plan Technical Check - Tier 2 (5); Physics Review - Tier 3 (3)	Re-planning
14	Plan non-deliverable due to inconsistent beam parameters (energy layers/spots)	2	Treatment planning	Plan Technical Check - Tier 2 (2)	Re-planning
15	Plan non-deliverable due to insufficient gantry clearance	2	Treatment planning	Plan Technical Check - Tier 2 (1); Physics Review - Tier 3 (1)	Re-planning
16	Wrong choice of methodology for anatomical robust optimization	2	Treatment planning	Plan Technical Check - Tier 2 (2)	Plan revision and physicist discussion
17	Non-adherence to institutional beam nomenclature conventions	34	Treatment planning	Physics Review - Tier 3 (20); PSQA and Plan Scheduling Review (10); Pre-treatment Checklist (4)	Nomenclature corrections
Plan Scheduling Errors
18	Beam path structures not sent for CBCT-based soft tissue matching	11	Plan scheduling	Physics Review - Tier 3 (9); Pre-treatment Checklist (2)	Rescheduling with updated structure set
19	Incorrect number of fractions scheduled in OIS	6	Plan scheduling	After the first fraction (1). PSQA and Plan Scheduling Review (4); Pre-treatment Checklist (1);	Plan rescheduling; workflow refined by adding mandatory PSQA and Plan Scheduling review step
20	Treatment field omitted during plan scheduling (similar field nomenclature)	1	Plan scheduling	Pre-treatment Checklist (1)	Plan rescheduling
21	Motion management parameters were not updated in the treatment beam during scheduling	5	Plan scheduling	During delivery at the machine console (1); PSQA and Plan Scheduling Review (3); Pre-treatment Checklist (1);	Plan rescheduling; PSQA and Plan Scheduling Checklist updated to include motion management parameters

User adoption and satisfaction survey

A staff satisfaction survey was conducted 26 months post-implementation, with 36 respondents representing various categories of departmental staff, and the results are shown in Figure 3. The full questionnaire and detailed results are provided in Supplementary Table S5. The overall mean satisfaction score was 3.79/5.0, with radiotherapy technologists (n = 16), medical physicists (n = 9), and radiation oncologists (n = 11) reporting mean scores of 3.89, 3.78, and 3.60, respectively, indicating broadly positive acceptance across all groups. Safety and Quality received the highest domain score (3.87/5.0), reflecting confidence in the system's error-trapping and data reliability. Workflow Efficiency (3.74/5.0) and System Integration (3.72/5.0) also scored favorably, whilst System Usability (3.53/5.0) was the lowest-rated domain across all groups.

Figure 3.

Mean user satisfaction scores by functional category and professional group, assessed 26 months post-implementation (n = 28). All items were rated on a 5-point Likert scale (1 = Strongly disagree, 5 = Strongly agree). The dotted line indicates the combined overall mean score (3.79/5.0).

Qualitatively, QCL-driven automation, centralized data access, and physics plan checklists were consistently highlighted as strengths. Technologists commended the pre-treatment checklist; physicists valued the PSQA forms and transparent task assignment; oncologists appreciated the prescription form and improved interdisciplinary communication. The principal concern across all groups was system performance, with intermittent crashes, sluggish responses during peak hours, and excessive window-switching identified as barriers to efficiency. Staff suggestions included automated IQ scripting for adaptive workflows, integration of QA analytics, and streamlined diagnosis entry.

Discussion

This study presents the design, customization, and 2-year clinical implementation of a comprehensive paperless workflow at a newly commissioned proton therapy center using the MOSAIQ OIS. Beyond operational benefits, the elimination of paper-based documentation across the entire proton therapy workflow represents a meaningful step towards environmentally sustainable clinical practice. To our knowledge, it is one of the first detailed reports of a fully integrated, OIS-driven paperless workflow developed specifically to meet the technical and dosimetric demands of proton therapy, including PBS, robust optimization, and multi-tier physics verification. The workflow was designed from the outset around these demands, as no member of the OIS Core Team had prior clinical proton therapy experience at commissioning. Rather than adapting existing photon radiotherapy templates, proton-specific forms and checklists were developed at every stage of the patient pathway. The simulation chart captures indexed couch-top accessory positioning details critical for reproducible PBS setup, whilst the prescription form incorporates dedicated fields for robustness margins and worst-case scenario dose parameters. PSQA was implemented as a mandatory, non-bypassable step, with proton-specific forms requiring beam-specific 2D fluence measurements at a minimum of 2 depths within the SOBP. The 3-tier physics plan review incorporates checkpoints with no photon equivalent, distinguishing this framework from standard OIS implementations in conventional radiotherapy. The structured approach and lessons learned provide a replicable model for other proton therapy centers undertaking similar OIS implementations.

The workflow was constructed entirely from MOSAIQ's native tools—QCLs for task assignment, virtual locations, customized assessment forms, and IQ scripting for automation. These features successfully eliminated paper-based processes across the entire patient journey. The overall user satisfaction score of 3.79/5.0 (assessed 26 months post-implementation) indicates strong clinical acceptance by multidisciplinary staff, with particularly favorable ratings in the Safety and Quality domain (3.87/5.0).

An important finding was that deploying an OIS-based workflow is iterative, not a one-time design. During the 2-year rollout we refined multiple components in response to clinical experience: we restructured the 3-tier physics review, added an independent plan-scheduling verification, integrated spirometry-based motion management, optimized QCL timelines, and developed dedicated adaptive-replanning and incident-reporting pathways. These changes closed gaps identified in routine practice while accommodating new technologies and operational needs. These refinements, implemented without disrupting workflow, demonstrate the framework’s flexibility and the need for ongoing monitoring and multidisciplinary review.

The workflow design and implementation share several key principles with published OIS implementation experiences in conventional radiotherapy.6, 9, 10, 11, 12, 13 Our consensus-driven OIS Core team approach and phased implementation strategy, incorporating dedicated practice database testing prior to clinical go-live, are consistent with the recommendations of Ban et al.⁹ for MOSAIQ Care Plan development. The use of OIS-driven workflow automation to standardize clinical processes has been associated with measurable improvements in efficiency and error reduction across different platforms10, 11, 12; Kovalchuk et al¹⁰ reported a reduction in physics plan check error rates from 19.1% to 5.2% and improved physician compliance with critical workflow tasks from 45.9% to 85.8% following ARIA Visual Care Path implementation. A direct pre-post comparison was not performed in the present study, as the proton therapy center commenced operations in a fully paperless environment without a paper-based baseline. However, satisfaction scores and strong safety and quality domain ratings suggest that the workflow contributes meaningfully to error prevention and treatment data integrity.

The error tracking results indicate that the multi-tier automated workflow functions as intended, with most deviations identified during the plan technical check (Tier-2) and independent physicist review before treatment initiation. The persistence of beam nomenclature, contouring, and prescription errors suggests these are recurring human- and process-related failure modes that can be minimized through standardized naming conventions, targeted training, and stronger electronic validation. The 2 errors not intercepted by the existing workflow—an incorrectly scheduled fraction detected after the first treatment fraction and a motion management technique not linked to the scheduled beam—highlight a specific vulnerability: OIS-based plan scheduling, when managed independently of the TPS, remains a susceptible step that is not fully addressed by physics plan review alone. This finding directly led to the inclusion of PSQA and Plan Scheduling Review as a mandatory, non-bypassable pre-treatment checkpoint. Overall, the error tracking data support the effectiveness of the multi-tier physics review process while emphasizing the need for continued refinements in scheduling integration, nomenclature enforcement, and periodic staff retraining to reduce recurring deviations.

The complexity of proton therapy planning and delivery necessitates robust plan review mechanisms. The 3-tier physics plan review checklist implemented in this study has proven effective in standardizing the planning process and reducing errors. As none of our medical physicists had prior clinical proton therapy experience at commissioning, the checkpoints were developed iteratively, drawing on onsite training visits to established centers and undergoing continuous refinement over the 2-year implementation period. The incorporation of variable RBE models and the evaluation of the high-LET region exemplify this experience-driven, proton-specific adaptation. Literature specifically guiding the design of proton therapy plan review checkpoints remains limited. Park et al¹⁴ recently introduced Auto-IniCheck, an automated script for initial physics plan review in proton therapy based on AAPM TG-275 recommendations. They classified checkpoints using Failure Mode and Effects Analysis as described in AAPM TG-100¹⁵ and quantified risk using Risk Priority Numbers. We intend to refine our framework further using these principles to eliminate redundancy and prioritize high-risk checkpoints. In practice, the stringent Tier 1 (planner self-check) and Tier 2 (independent physicist peer review) checkpoints have substantially reduced the need for replanning after physician approval. Furthermore, the independent Tier 2 and Tier 3 reviews have enhanced overall plan quality and fostered valuable knowledge sharing amongst physicists—a significant benefit given the diversity of clinical cases treated at our center.

A significant limitation of the OIS customization is the lack of direct integration with the hospital's existing electronic medical record system, which is shared across the photon therapy department and other centers in the same city. Consequently, patient follow-up information and examination details must be entered separately into both the existing electronic medical record and the new OIS, introducing redundancy and increasing administrative burden. Additionally, billing management falls outside the scope of the current OIS workflow, as it is handled through a separate institutional system. Integration of these systems with MOSAIQ faces considerable logistical and technological challenges, which we anticipate addressing in the coming months. Implementing the automated OIS workflow presented multiple challenges, particularly regarding staff training and system reliability. The highly automated nature of the workflow, in which a single user action triggers a cascade of dependent processes, creates vulnerability to disruption if errors occur at any step. This poses particular challenges for rotational staff who work across both proton and photon facilities, as well as newly recruited personnel. Comprehensive Standard Operating Procedures developed for all new staff have significantly reduced training-related errors and improved workflow adherence. Intermittent system crashes, corroborated by the satisfaction survey, represent a further operational challenge, disrupting workflow continuity and requiring vendor or IT support for resolution. Addressing system stability remains a priority for optimizing platform reliability.

Conclusions

This study demonstrates the successful implementation of a comprehensive paperless workflow using MOSAIQ OIS at a newly commissioned proton therapy facility. Successful OIS customization required a dedicated multidisciplinary team comprising medical physicists, technologists, and physicians. Critical to this process was the systematic planning phase, in which existing departmental workflows were carefully analyzed and deconstructed into discrete components before incorporation into the OIS. The checklists introduced during treatment planning and patient treatment effectively standardized processes and reduced errors, whilst the staff satisfaction survey revealed predominantly positive perceptions of the system. By eliminating paper, proton therapy workflows gain improved data accessibility, enhanced communication, and greater operational efficiency and environmental sustainability.

Ethics

This study describes the design and implementation of a paperless clinical workflow and does not involve patient experimentation, clinical intervention, or collection of identifiable patient data. The staff satisfaction survey was conducted anonymously with voluntary participation. This work constitutes a quality improvement and service evaluation initiative; formal ethics committee approval was therefore not required.

Funding

The open access publication fee for this article was funded by the Advanced Center for Treatment, Research, and Education in Cancer (ACTREC), Tata Memorial Center, Navi Mumbai, India. This research received no other specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

Author contribution

Vysakh Raveendran and Kantaram Darekar contributed equally to this work and share first authorship. Both authors were responsible for conceptualization, study methodology, workflow design, customization and implementation, formal analysis, and manuscript preparation. Shwetabh Sinha and Jifmi Jose Manjali: Clinical oversight and critical manuscript revision. Rachana Anil Kulkarni, Abhijeet Ashok Powar, Nanasaheb Shriram Khatik, and A Chithambara Prabu: Workflow implementation and operational data collection. Lalit Chaudhari: Workflow implementation and staff training coordination. Siddhartha Laskar: Project supervision and senior clinical guidance.

Data Availability Statement for this Work

The workflow design, checklist frameworks, and implementation details are described in full within the manuscript and its supplementary materials. Anonymized satisfaction survey data are available from the corresponding author upon reasonable request.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix A. Supplementary material

Supplementary material