Breast magnetic resonance imaging (MRI) is a well-established and indispensable modality for assessing suspicious breast lesions, for evaluating response to neoadjuvant chemotherapy, and for breast cancer screening in people with an increased risk of developing breast cancer or with dense breasts.1–3

Currently, breast MRI screening is performed for people with high risk of developing breast cancer (eg, BRCA mutation carriers and their first-degree relatives or people with overall lifetime cancer risk of 20%–25%) according to several (inter)national guidelines.4–9 However, supplemental MRI screening for people at average risk of breast cancer was also proven beneficial,10 and various studies showed that, compared with conventional mammography, sensitivity for breast cancer detection in diverse subgroups increased when using breast MRI.10–12 Consequently, guidelines have been expanded based on scientific evidence,1,13–15 and now also advise MRI screening for people with extremely dense breast tissue or personal/familial history of breast cancer.5,6,16 Advantages of using MRI are that the image is not affected by breast density, the sensitivity for detecting breast cancer compared with other modalities is higher, hemodynamics is evaluated, and compression of the breast is not required.1,3

Although the screening interval that results in low risk of interval cancer in most people may last longer for MRI screening than what is observed in mammographic screening, and screening frequency may thus be reduced, it is still likely that the required capacity for MRI screening needs to increase.10,17 Moreover, according to cost-effectiveness analyses on the use of breast MRI in screening, the cost per MRI has the strongest influence on the total costs of a screening program.18 The major step to minimize costs of breast MRI is thus to reduce scanning times and turnaround times for the examination itself.17

One way of increasing the capacity of MRI scanners for breast screening and reducing the costs per examination is by using abbreviated breast MRI (abMRI) protocols, referring to shortened standard breast MRI protocols, enabling performance of a diagnostic MRI scan within minutes (<10 minutes).19–24 In addition, abMRI protocols could make MRI examinations less burdensome for the person being scanned by reducing the amount of time spent in prone position. It is important to note that motion artifacts are associated with discomfort, further highlighting the importance of comfort during breast MRI scans.19,20,25,26 However, the use of abMRI protocols alone is insufficient to truly increase the throughput of breast MRI. Turnaround times for preparing a person for the scan are in the order of 10 minutes (intravenous [IV] cannulation for contrast media injection, positioning on MRI table)27 and remain, even with the introduction of shorter breast MRI protocols, a bottleneck for cost reduction. By optimizing the workflow around a breast MRI examination, the number of people that can be screened with MRI could potentially be increased. Albeit not aimed at breast MRI, Recht et al28 described that reengineering of the MRI workflow in a new imaging center saved roughly 5 minutes per person.

The aim of this project is to use real-life data to create a model of an ideal scanning environment for breast cancer screening with MRI. By simulating the processes involved, we aim to show how adapting several factors, such as workflow, personnel, equipment, and scanning protocol, may impact the efficiency of breast MRI procedures.

METHODS Study Design

In this study, an innovative concept for a breast MRI screening facility was developed. The envisioned screening facility is person centered and aims for an efficient workflow-oriented design. A 3D simulation software package (FlexSim Healthcare Software Version 21.2.4; FlexSim Software Products, Inc, Orem, UT) was used to design and evaluate the new concept. The software provides a realistic virtual logistic pathway for MRI screening with a 3D model, determining actual throughput while assuming different scenarios. The scenarios can be tested and compared to evaluate performance, find bottlenecks, and optimize processes and resource allocation.

Collection of Input Data

The 3D simulation was based on prospective monitoring of the workflow of breast MRI screening. To obtain realistic variation in measurements, 1 academic hospital (site 1), 1 cancer institute (site 2), and 1 clinic specialized in breast care (site 3) were included as participating hospitals. All monitored participants in sites 1 and 3 provided signed informed consent; the institutional review board in site 2 waived the informed consent.

Sites 1 and 2 have a 20-minute time slot per person for breast screening with an abbreviated scanning protocol of approximately 8 minutes, whereas site 3 uses 45-minute timeslots per person with a scanning protocol of approximately 24 minutes. At all 3 considered sites, a team of 2 radiographers currently performs the procedures. For all sites, the breast MRI workflow was divided into the following 12 steps following the participant (also described in Supplemental Digital Content 1, https://links.lww.com/RLI/A891):

Participant check-in at front desk Participant entering changing room Participant exiting changing room Start of cannulization End of cannulization Positioning participant on MR table by radiographer Start time scan End time scan Participant exiting MR room Participant entering changing room Postscan procedures MR room (eg, cleaning) by radiographer Participant exiting changing room

For each step, the starting time was recorded in an hour:minute:second format. The organization and duration per step of the MRI workflow, which depends on the availability of staff, appointment time, and available rooms per appointment, varied per site. By subtracting the timestamps, the duration of every step within the workflow was calculated. The simulation of the various scenarios made use of statistical distribution methodology, depicting process times with minimum, maximum, and average duration.

Data were collected during 62 MRI examinations (site 1: 20 scans, site 2: 32 scans, site 3: 10 scans), over the course of several days (2–7 days). This provided sufficient data on average examination times including time needed for preparation, scanning, and actions after completion of the scan (Supplementary Digital Content 1, https://links.lww.com/RLI/A891).

Major differences in scheduled time per MRI were observed between site 3 and the other 2 sites. Sites 1 and 2 had similar workflows and scheduled times. To ensure comparability, data from site 3 was not included as input for the simulation.

Definition of Architectural Layout

In the FlexSim simulation software, an entry, reception, waiting rooms, changing rooms, and MR rooms could be freely positioned in an empty space with an area of 390 m2, enabling the creation of a dedicated high throughput screening unit. Blueprints for an optimized layout that put the main focus on throughput were drafted by medical planners/architects and were refined in several feedback loops to ensure that the layout is also practical. One of the elements used in the simulation was a dockable table concept. The advantage of using a dockable table is that a person can be prepared for the scan outside of the MRI room, thus optimizing the availability of the MRI room for scans. One of the drafted layouts was, therefore, optimized for the use of dockable tables (Fig. 1).

FIGURE 1:

Example layouts. A, Reference layout for an MRI setup within a hospital. B, First proposal for a new layout; however, it is not the best for use with dockable tables as the corners are too sharp to comfortably and quickly maneuver the dockable table. C, Image was changed by curving the walls of the changings/preparation rooms, and therefore more optimized layout for use with dockable tables. Ultimately panel C was used for simulating the different scenarios. The facility blueprint is available upon request from the authors.

Workflow Simulation

Various screening scenarios—including the measured baseline—were compared. Each scenario was characterized by different combinations of layouts, staff composition, scanning protocols, and equipment (Table 1). In addition, multiple variations in peoples' arrival patterns were modeled (Supplementary Digital Content 2, https://links.lww.com/RLI/A892).

TABLE 1 - Overview of Main Scenarios (Baseline, 1, and 2) Used as Input Parameter for the Simulations Baseline Scenario 1 Scenario 2 Equipment  MRI system MAGNETOM Prisma (Siemens Healthcare, Erlangen, Germany) MAGNETOM Prisma (Siemens Healthcare, Erlangen, Germany) MAGNETOM Prisma (Siemens Healthcare, Erlangen, Germany)  Coil 18-Channel breast 18-coil (Siemens Healthcare, Erlangen, Germany) or equivalent 18-Channel breast 18-coil (Siemens Healthcare, Erlangen, Germany) or equivalent 18-Channel breast 18-coil (Siemens Healthcare, Erlangen, Germany) or equivalent  Contrast agent injector Conventional system Advanced system with 24 h depot, automatic venting, etc Advanced system with 24 h depot, automatic venting, etc Changing room/dockable table 2 Changing rooms 2–4 Changing rooms 2–4 E-drive powered dockable tables (speed: 30 m/min) and changing/preparation rooms Scanning protocol Baseline protocol scan time: 08:07 min; average actual scan time: 12:49 min Optimized protocol scan time: 02:12 min; average actual scan time: 05:52 min Optimized protocol scan time: 02:12 min; average actual scan time: 05:52 min Workflow process Baseline workflow: 2 radiographers; no specified tasks per radiographer (radiographers rotate in person preparation tasks); execution of tasks in parallel Optimized workflow: 3–5 radiographers; dedicated tasks per radiographer (ie, 1 radiographer for scanning only, 2 radiographers for person preparation); execution of tasks in parallel Optimized workflow: 3–5 radiographers; dedicated tasks per radiographer (ie, 1 radiographer for scanning only, 2 radiographers for person preparation); execution of tasks in parallel

The goal of the simulation was to define and simulate an optimal setting for breast MRI screening with the best high-end equipment available to ensure the highest possible throughput. The simulation software aims to take all relevant parameters into account as best as possible, to accurately reflect a realistic scenario. Some boundaries were predefined. The simulated changes consisted of optimizing (1) facility layout, (2) scanning protocol, and (3) workflows and resources. Simulated waiting times were modeled around the same level as the baseline scenario but should never be longer than 30 minutes. Furthermore, waiting time for a person on the dockable table (ready to be scanned) was maximum 5 minutes. Therefore, the waiting times should be at an acceptable level. In terms of staff satisfaction, the utilization rate of staff needed to be in between the range of ±10% of the baseline scenario. Walking speed of staff (average speed: 75 m/min without dockable table, 30 m/min with dockable table), duration of process steps, and distances were simulated accordingly. The duration of the different process steps was based on the measurements performed at the different participating hospitals. Table 3 shows the optimized workflow durations used as input for the simulations.

Data Analysis

The simulation runtime was 10 h/d (from 08:00 to 18:00) without including any staff break. This was repeated for 90 days and 60 replications in total for every simulated scenario. Variations of the different process steps were based on the distributions of time in minutes for each process step according to the measurements. The simulation automatically generates outcomes of the breast MRI screening workflow. The following dashboards were created in FlexSim for the analysis as well as Excel export of the data from these dashboards:

Throughput: average number of people examined per day Utilization of staff: percentage of time staff is occupied by execution of tasks related to the breast cancer screening workflow Utilization of locations: percentage of time when MRI room is occupied Waiting time for locations: duration of waiting time for a preparation room and available MRI room in minutes Waiting time for staff: duration of waiting time for available staff to prepare and escort a person to and from the preparation room and MRI room in minutes

More information can be found in the Supplementary Digital Content 2, https://links.lww.com/RLI/A892.

RESULTS Prospective Workflow Monitoring

After the evaluation of the collected data, 2 measurements were excluded because their values were extreme. One seemed to be caused by measuring errors (ie, aftercare 12 minutes); in the other, it appeared impossible to complete the examination because the IV cannula could not be placed. Table 2 depicts the data that were obtained from sites 1 and 2.

TABLE 2 - Workflow Overview With the Collected Measurements From Clinical Practice Step Step Description Min (min:s) Max (min:s) Avg (min:s) 1 Waiting time in waiting area 00:00 29:00 06:00 2 Radiographer 1 person pick-up, instruction change of clothes 00:48 11:00 03:18 3 Radiographer 1 sets IV cannula 00:54 10:00 02:00 4 Radiographer 2 places person on table and prepares* 01:30 04:00 03:00 5 Radiographer 2 performs scan and processes images and documentation* 10:30 17:54 12:48 6 Radiographer 2 postprocessing on table* 01:42 10:00 02:18 7 Radiographer 2 prepares table for next person*,† 00:18 01:42 01:00 8 Person changes clothes*,† 00:30 04:30 02:30 See Table 1 for baseline scenario details; 2 radiographers, no dockable tables, and 2 changing rooms. Statistical distribution methodology is used to depict the process times for each workflow step: minimum, maximum, and average in MRI room occupation in baseline during steps 4–7.

*Workflow steps during which the MRI room is occupied.

†Steps occurring in parallel.

Simulation Optimization Process

Although sites 1 and 2 were already efficient, further improvements were possible. The results are divided in the following categories: (1) facility layout, (2) scanning protocol, and (3) workflows and resources.

Facility Layout Optimization

First, an optimized facility layout was developed from scratch after discussion on the reference layout used at site 1 (Fig. 1A). The first optimized facility layout was based on factors such as optimal communication, accessibility, and workflow direction (Fig. 1B). An important change is to perform IV cannulation in the changing rooms. Second, in the situation where dockable tables are used, the layout had to be adapted in order to facilitate the workability for the personnel. To optimize the layout for use with dockable tables, the walls of the changing rooms were subsequently curved and the door to the MRI room was slightly widened (Fig. 1C).

Abbreviated MRI Scanning Protocol

Second, the MRI scanning protocol was further abbreviated (Table 4). An already implemented abMRI scanning protocol was used as a reference protocol. This scanning protocol had already been reduced compared with the full diagnostic MRI protocol and had a total duration of 08:07 minutes of scanning time. Based upon previous research that did not show significant improvement in cancer detection or specificity by the use of additional sequences beyond perfusion imaging in a screening setting,24 a perfusion-only protocol with a duration of 02:12 minutes was created. This protocol provides dynamic inflow evaluation with a temporal resolution of below 5 seconds and a spatial resolution of less than 1 × 1 × 1.5 mm3, yields automatic subtractions, and generates temporal maximum intensity projections; however, it does not include either T2-weighted imaging, diffusion-weighted imaging, or late-phase dynamic evaluation. The time for intermediate instructions from the radiographers and injection of the contrast agent was included in the actual scanning duration of both the baseline abbreviated protocol and the optimized protocol.

Workflow and Resource Optimization

Finally, workflows and resources such as staff and equipment were optimized to assess the impact on throughput (Table 5). Varying arrival patterns showed that waiting times could be optimized. Also, by varying the number of radiographers and changing rooms/dockable tables, throughput could be influenced. Magnetic resonance imaging room occupation per person was reduced by implementing dockable tables, which allowed radiographers to prepare people in the changing rooms, so that the MRI room is only used for pure scanning purposes. This also facilitated the possibility to perform certain workflow steps in parallel. The baseline situation with 2 radiographers allowed a throughput of approximately 36 people a day, with 85.08% utilization of the MRI room and 73.34% of radiographers. Without dockable tables, a throughput of 41 people per day seems to be the maximum, and a further increase of number of staff from 2 to 5 radiographers had only minor impact in this scenario. With dockable tables, a throughput of 74 people per day is theoretically possible. A scenario with 3 radiographers seems to be most resource-efficient (defined as most efficient throughput using the available resources as specified in the scenario such as staff, dockable tables/rooms, and imaging equipment), with a throughput of more than 68 people per day, and utilization of MRI room of 86.09% and radiographers of 77.8%. The results of the rest of the simulations can be found in the Supplemental Digital Content 2, https://links.lww.com/RLI/A892.

Appointment Duration

In the baseline scenario, the total average appointment time was 25:54 minutes, of which the MRI room is occupied for 19:06 minutes on average (Fig. 2). Total table times while scanning was on average 12:48 minutes when using an 08:07-minute protocol (Table 4). The optimized workflow executes more tasks in parallel, makes use of dockable tables, and applies an even more abbreviated scanning protocol (Table 3). This reduces the total average appointment time to 19:36 minutes, of which 06:21 minutes are spent in the MRI room with a total table time of 05:36 minutes (02:12-minute protocol).

FIGURE 2:

A, Comparison of baseline and the optimized appointment duration. B, Appointment time (excluding waiting times) comparison. Tasks happening in parallel are combined in the timeline (baseline steps 7 and 8 and optimized scenario steps 9 and 10, see tables above); the longest time is taken as the duration. *The black box depicts the duration of MRI room occupation.

TABLE 3 - Workflow Overview With Optimized Duration of Each Step Used as Input for the Different Scenarios in the Simulation Step Step Description Min (min:s) Max (min:s) Avg (min:s) 1 Radiographer 1 picks up and instructs person 00:06 00:42 00:15 2 Person changes clothes 00:33 10:48 02:42 3 Radiographer 1 sets IV cannula 00:25 07:30 02:11 4 Radiographer 1 positions person on (while radiographer 2 undocks ) 01:00 05:40 02:52 5 Radiographer 1 docks * 00:12 00:36 00:24 6 Radiographer 3 performs scan and processes images and documentation 04:36 07:54 05:42 7 Radiographer 1 undocks * (while radiographer 2 prepares next person) 00:06 00:24 00:15 8 Radiographer 1 aftercare person 01:00 05:00 02:23 9 Radiographer 1 cleans † 00:06 03:00 01:20 10 Person changes clothes† 01:08 05:30 02:53 See Table 1 for the scenario 1 and 2 details; 3 radiographers, 2 dockable tables and 2 preparation/changing rooms. MRI room occupation in scenario without dockable table: during steps 4–8. MRI room occupation in scenario with dockable table: during steps 5–7.

*Workflow steps that are specific to a scenario with dockable tables.

†Steps occurring in parallel.

TABLE 4 - Comparison of Reference Abbreviated Breast MRI Protocol for Screening Purposes Currently Used in Clinical Practice and Optimized Breast MRI Protocol Used in the Simulated Scenarios Sequence Baseline Reference Protocol (min:s) Optimized Protocol (min:s) 1 Localization 00:09 00:09 2 Without prothesis 3 Dixon pre 01:31 4 Contrast 5 First blanco 6 Perfusion 02:03 02:03 7 Dixon post 01:31 8 Diffusion 02:53 9 Subtraction Total sequence duration 08:07 02:12 Actual scanning duration 12:49 05:42

Actual scanning time includes intermediate actions of radiographers such as interaction with the person, injecting contrast agent or repeating sequences if needed.

TABLE 5 - Overview of Best Scenarios When Using Different Combinations of Staff Numbers and Equipment Baseline Best Scenario Without Dockable Tables Best Throughput Scenario Most Cost-Efficient Scenario No. radiographers 2 3 5 3 No. rooms (dockable tables) 2 2 4 4 Throughput, people/d* 36.13 41.15 74.78 68.93 (66.7) Waiting for preparation room, h:min:s† 02:20:00 02:06:14 00:08:22 00:28:30 (00:01:58) Waiting for MR, h:min:s 00:01:08 00:03:56 00:03:35 00:02:16 (00:01:55) Utilization scanning room, % 85.08% 96.58% 92.59% 86.09% (83.61%) Utilization radiographer, % 73.34% 41.34% 41.47% 77.8% (75.42%)

Values in brackets are the results of an optimized arrival pattern to reduce waiting times: every full hour 2 people arrive, every 10 minutes 1 person arrives, last person at 17:40 hours, and maximum of 69 people per day simulated in total.

*Staff breaks were not considered in the simulation. Simulations make simplified assumptions about reality. Therefore, the results in reality may differ from the simulation.

†Arrival pattern was not optimized: 2 people per 15 minutes, last person at 17:15 hours, and maximum number of 76 people per day.

DISCUSSION Summary of Evidence

In this study, we simulated a proposed design for a new MRI screening facility and examined various modifications of the workflow to increase efficiency. Turnaround times could become a bottleneck for implementation of breast MRI as a screening modality. By changing the location of certain steps from the MRI room to the changing room, upgrading layout and equipment, and implementing abMRI scanning protocols in a simulated environment, we were able to reduce total appointment time, to reduce MRI room occupation per appointment, and to increase the total number of people that can be scanned per hour or day. The total average appointment time and average time in the MRI room per person could be considerably reduced by, respectively, 06:18 and 13:45 minutes. This is also likely to reduce the costs associated with MRI screening, as the costs per examination have previously been shown to be the most important determinant for cost-effectiveness. The most resource-efficient simulated scenario allowed a throughput of 68 people per day per MRI scanner, which is an increase of 91% compared with the baseline.

Changes in Resources and Staff

Workflow steps such as safety instructions, IV placement, positioning on the table, and postscan procedures combined take time, whereas the location where each step is performed also determines the MRI room occupation and potential turnover. Therefore, minor adaptions of parallelizing tasks were evaluated. By implementing dockable tables, preprocessing and postprocessing could be moved to the changing room, freeing up the MRI room for another scan. A study by Heacock et al27 has shown that inefficiencies were eliminated by implementation of dockable tables, dedicated preparation rooms, 2 doors in each MRI room, ideal position for the scanner with a direct path for the table, and all necessary equipment. These adaptations resulted in mean time savings of 05:28 minutes per person. In line with our study, it enhanced the efficiency of MRI by boosting the accessibility and productivity of both the MRI scanner and technical teams, while reducing the duration people spend in the radiology department and MRI room.27 Our study also shows that when enough changing rooms with dockable tables are available, the MRI can be used more continuously compared with the baseline situation, and ultimately increases throughput and MRI efficiency.

In this study, we did not assess the actual experience or satisfaction levels of radiographers and the general population. Nonetheless, we incorporated outcomes such as radiographer utilization and waiting times to replicate a workflow that matches or surpasses the current one. By emulating the existing workflow, our aim was to approximate similar satisfaction rates.

Changes in Scanning Protocol: Abbreviated MRI Sequences

In the simulation, the implementation of an abMRI scanning protocol also turned out to be a big-time saver, reducing total appointment time, MRI room occupation, and time one needs to lie in prone position. Various studies have been investigating ideal abMRI and ultrafast protocols, in order to obtain optimal image quality while shortening the acquisition time.23,29–32 Currently, abMRI protocols are clinically already in use. However, most abbreviated protocols still are a compromise between a multiparametric approach and time.19,20,27,33 Our study shows that further shortening of the protocol allows for greater efficiency and cost-effectiveness, because throughput can be increased while keeping staff and MRI utilization the same. Other benefits are improving comfort (sternal discomfort, motion) and producing fewer images, which allow for less storage space and faster image retrieval.19,20 Also, interpretation time is reduced with abMRI, which implies that the actual benefits extend beyond the acquisition phase. However, it must be acknowledged that breast abMRI sequences are designed to identify disease in a screening setting and not fully classify eventual lesions.21 It could be envisioned that in some cases a full diagnostic MRI is still performed after an abMRI to better classify or describe the disease. Nonetheless, for breast cancer screening, the use of abMRI protocols seems to be the most logical approach.

Our study shows that implementing abMRI scanning protocols alone does not increase throughput sufficiently. Although it reduces MRI room occupation (more scans per hour), other factors such as availability of changing rooms/dockable tables and inefficient workflow steps can become bottlenecks (unavailability of rooms/tables). Despite the fact that some workflow steps had actually slightly longer duration with the dockable tables, for example, due to slower walking speed, other type of hygiene protocols, and so on, a combination of optimized workflow steps makes the most impact on process efficiency and maximizes scanning capacity compared with other scenarios.

Cost-Effectiveness

The perceived high-costs and limited availability of MRI units are 2 factors that prevent widespread use of MRI as a screening modality for breast cancer.3,34 However, multiple studies have shown that the use of breast MRI as a screening tool is cost-effective, even in people who are not at very high risk.35–39 One way of maintaining cost-effectiveness is to elongate the screening interval when screening with MRI in lower risk people. This saves substantial costs and reduces the burden of screening, while maintaining a very low rate of interval cancers.18 The optimization steps presented in this study may help to further reduce the costs and improve availability of breast MRI without the need for more scanners. Based upon the modeling results, performing the screening with 3 radiographers seems optimal. Further increasing the personnel leads to only a moderate improvement in throughput and reduces the utilization of radiographers strongly. Because, at this rate of throughput, the costs are most dominated by the personnel wages, maximizing the throughput by bringing in more staff seems inadvisable. Although a formal cost-effectiveness evaluation still needs to be performed, we estimate that the high throughput that can be achieved with 3 radiographers would already reduce the costs to below €200 per examination based upon the current advocated price of €343.68 for a regular breast MRI in the Netherlands.

A further method to reduce the costs may be the use of less expensive equipment, although it should be realized that the equipment costs only play a marginal role compared with staffing costs. Nonetheless, use of a 1.5 T scanner may further reduce costs of this setup by lowering purchase and operational costs. Although it may currently be difficult to achieve the performance of an ultrafast protocol at 1.5 T at sufficiently high spatial resolution, an alternative is to implement a fast protocol consisting of only precontrast and postcontrast acquisitions and subtractions, which can be obtained in the same timeframe. This may have a slight impact on the sensitivity in cases of moderate and marked background parenchymal enhancement,29 but would still allow offering of a much more sensitive screening technique to many eligible women who are currently only offered mammography.

Limitations of This Study

The findings in this research should be interpreted with consideration of the following limitations. This study analyzed how to optimize breast MRI screening by maximizing the throughput in a single MRI scanner, in the context of a new dedicated MRI screening facility. However, as is also clear from the assessment of the logistics at site 3, the logistics at many existing sites are not yet optimized for screening. A first improvement might, therefore, be achieved by simply optimizing the workflow processes at existing scanners for breast screening. One recommendation here is to schedule people that will take presumably longer (eg, silicone breast implants, disabilities, etc) at the end of the working day. This will likely already generate substantial additional capacity that can be used to implement MRI screening for many more people without substantial demands on changing the facility layout or workflow.

In addition, our study involved a limited number of institutions, which may not fully represent the broader variation of workflows. The simulation might also not capture all potential interactions between different components of the workflow or potential unexpected events such as equipment breakdowns or staff shortages. The further optimized results presented in this study are the product of a simulation, a simplified version of reality, based on multiple assumptions and estimates. Thus, the real-life situation could turn out to be different. Initial focus primarily lies on logistical and practical aspects of the workflow; therefore, perspectives of staff and visiting people have only been considered partly.

However, by using a simulation, the (dis)advantages of different scenarios can be explored and compared without having to build the actual setup and test with human subjects. The simulation was based on real-life workflows in order to attempt a realistic simulation outcome. This way, potential bottlenecks can be identified ahead of time. Furthermore, these results can be used as a base of discussion for stakeholders in order to map their perspectives, in order to be able to implement those in the final layout and workflow proposals. A potential strategy for research in practice can therefore be specified early on.

Future Research

In terms of logistical and practical aspects, further evaluation of workflow processes, assessment of the actual effectiveness of the most abbreviated scanning breast MRI protocol, and cost-effectiveness evaluation are indicated. A business case study is necessary in order to judge whether building the optimal screening facility turns out to be financially viable. If needed, another round of simulations could be performed using this newly generated knowledge. Furthermore, regarding guidelines for personalized screening, person subgroups that benefit most from MRI screening need to be defined, as well as taking age to begin and end screening with MRI and the optimal screening interval with MRI into account. Furthermore, the experiences of radiographers (and other relevant staff) and participants of the screening program are relevant to be studied and included in future research.

CONCLUSIONS

In conclusion, the results of this simulation study suggest that, by implementing changes in workflow, facility layout, and resources in a breast MRI screening situation, throughput could theoretically be increased by 91%. Reducing MRI scanning time by implementing an abMRI protocol and reducing MRI room occupation by implementing dockable tables resulted in shorter appointment duration and increased availability of the MRI scanner.

ACKNOWLEDGMENTS

This article and the research behind it would not have been possible without the participation of all radiographers and participants in the 3 participating hospitals. We greatly appreciate their contribution and participation to this study.

REFERENCES 1. Bakker MF, de Lange SV, Pijnappel RM, et al. Supplemental MRI screening for women with extremely dense breast tissue. N Engl J Med. 2019;381:2091–2102. 2. Bougias H, Stogiannos N. Breast MRI: where are we currently standing? J Med Imaging Radiat Sci. 2022;53:203–211. 3. Mann RM, Cho N, Moy L. Breast MRI: state of the art. Radiology. 2019;292:520–536. 4. Mann RM, Balleyguier C, Baltzer PA, et al. Breast MRI: EUSOBI recommendations for women's information. Eur Radiol. 2015;25:3669