Radiation Therapy At Scale: How Mary Bird Perkins Cancer Center Cut Administrative Burden by 40%
Radiation therapy departments are under pressure that shows no sign of easing. Patient volumes continue to grow. Treatment protocols are more individualized. And yet the way most RT departments manage their daily operations, manually, through a combination of experienced staff, spreadsheets, and workarounds layered over an OIS, has not fundamentally changed. The result is a system that seems to work, until it doesn't. A senior therapist goes on leave. A machine goes down. A patient calls back with new availability constraints. Any one of these events can cascade through an already fragile schedule, consuming hours of coordination time that clinical staff cannot afford to spend on administrative work. Mary Bird Perkins Cancer Center, one of the leading cancer care organizations in the Gulf South, recognized this pattern at its Baton Rouge site, the center's highest-volume location, treating approximately 1,600 radiation therapy patients per year across five linear accelerators.
THE PROBLEM WASN’T EFFORT. IT WAS INFRASTRUCTURE.
The scheduling process at MBPCC required senior radiation therapists to manually assemble a full proposed schedule before calling the patient. If the patient's availability didn't match, the therapist had to end the call, rebuild the schedule, and call the patient back. Machine load across the five linacs was calculated manually, with no shared visibility and no automated distribution logic. And because the scheduling rules existed only in the minds of the people who had been doing it for years, there was no way to delegate the process to less experienced staff. The cumulative effect was predictable: the clinicians best positioned to support complex treatment decisions were spending a disproportionate share of their time on administrative coordination, not because of poor organization, but because no infrastructure existed to absorb that work.
This is a pattern seen across RT departments operating at scale. The operational knowledge that keeps the schedule running is person-dependent, fragile to turnover, and impossible to encode in a standard OIS. MOSAIQ, ARIA, and similar systems were built for clinical documentation and treatment delivery, not for the continuous rebalancing of patient demand against finite capacity that RT scheduling actually requires.
A DIFFERENT APPROACH TO RADIATION THERAPY SCHEDULING.
In October 2024, MBPCC deployed GrayOS at the Baton Rouge site. GrayOS integrates bidirectionally with MOSAIQ, which remained the system of record for all clinical activities throughout the deployment.
The operational change was structural. When a dosimetrist finalizes a treatment chart in MOSAIQ, GrayOS reads it automatically and translates the clinical plan into a complete set of appointments, accounting for machine eligibility, operational rules, patient priority, and available capacity. The scheduling RT reviews the proposed schedule, inputs patient preferences if needed, and GrayOS re-optimizes in real time. Once confirmed, everything writes back into MOSAIQ.
The process is now fast enough to run live, on the phone, with the patient. That was structurally impossible before.
Three operational shifts define what changed:
First, scheduling is now automated: when a dosimetrist finalizes a treatment chart, GrayOS reads it and automatically generates a complete appointment set in seconds, accounting for machine eligibility, operational rules, patient priority, and available capacity, work that previously required a therapist to navigate multiple calendars manually.
Second, machine load across all five linacs is visible in real time and factored into every schedule the system generates, replacing a manual and error-prone calculation.
Third, patient preferences are handled live: because the process runs in seconds, the scheduling RT can input a patient's availability constraints on the phone and re-run the algorithm instantly, getting a confirmed schedule before the call ends, rather than hanging up, rebuilding, and calling back.
MEASURED RESULTS.
The impact was evaluated through a formal analysis conducted with and approved by Mary Bird Perkins Cancer Center.
Scheduling time was reduced by 40%, representing approximately 15 minutes saved per patient. Administrative delay, defined as the time from task creation to the first treatment appointment being scheduled, was reduced by 22%.
The impact was evaluated through a formal analysis conducted with and formally approved by Mary Bird Perkins Cancer Center.
WHAT MAKES THIS APPROACH REPLICABLE.
Three factors shaped the outcome at MBPCC, and together they reflect something that applies beyond this deployment: a care orchestration platform only works if it adapts to the operational reality of each center, not the other way around.
The senior radiation therapists who carried the scheduling knowledge were not passive recipients of the deployment. They became the architects of the constraint logic in GrayOS, shaping how the system interprets their center's operational rules. That co-construction is what makes the system specific to a center's reality rather than generic.
Deployment was iterative by design. At MBPCC, the initial workflow triggered off physician prescriptions. When it became clear that prescriptions were often finalized after scheduling had already begun, the team reconfigured GrayOS to read from treatment charts instead. This kind of adjustment is not the exception, it is the rule. Operational constraints and tacit scheduling rules cannot all be captured in a kick-off workshop. They surface through real use, and the deployment methodology is built to accommodate that.
GrayOS was built to augment existing clinical infrastructure, not replace it. MOSAIQ remained unchanged as the system of record throughout the deployment. This matters beyond technical compatibility: cancer centers have invested years of work and significant resources digitizing their clinical processes into their OIS and EHR. GrayOS leverages those investments rather than displacing them, adding an operational layer above existing systems that translates clinical intent into a coordinated, constraint-aware schedule, without requiring centers to start over.
THE BROADER IMPLICATION.
MBPCC's deployment illustrates something that extends beyond one center's results. The operational fragility that most RT departments manage daily (person-dependent processes, manual coordination, schedules that break under routine disruptions) is not a failure of the people managing it. It is a structural consequence of the absence of infrastructure built specifically for continuous operational optimization in radiation therapy.
That layer now exists.
