IMRT, AI, and fewer cases: Is radiation oncology facing oversupply?
Industry Buzz
AI has not caused a significant change in my clinic but I anticipate it will, particularly for our dosimetrist. The auto-contouring does a reasonably good job outlining most normal organs, requiring minimal adjustments... I do not consider it reliable for outlining treatment volumes.
—Kelly Orwat, MD
Radiation oncology sits in an unusual position. Clinical demand still exists, especially outside urban centers.
At the same time, fewer fractions per patient, automation, and reimbursement pressure are changing how much work each physician does. The result is a job market that feels tighter than workforce models alone would predict.
Residency expansion and pipeline pressure
Between 2003 and 2018, radiation oncology residency programs grew from 54 to 91, and annual positions rose from 106 to 193, an 82% increase.[] Among graduates surveyed through 2021, market saturation was described as an emerging concern, with more graduating residents than jobs posted on ASTRO’s Career Center over a 5-year period.[]
The current Match data looks less dire than the low point. In the 2025 NRMP Main Match,[] radiation oncology offered 179 advanced positions and filled 174, a 97.2% fill rate. The physician reserved track offered 9 positions and filled 6. Those numbers mark a stronger year than the unmatched shock seen earlier in the decade, but they do not erase the structural concern underneath.
ASTRO’s workforce analysis landed in a more measured place than the harshest forecasts. In its 2023 report,[][] the Society projected relative balance between supply and demand through 2030 under its most likely scenario. The caveat mattered.
The report also stated that oversupply becomes a concern if physician productivity reaches the highest modeled range. The post-2030 period was flagged as more vulnerable if supply keeps growing while Medicare beneficiary growth slows.
Related: Inside the shift: How oncologists are using AI for trial matchingHypofractionation is reducing clinical volume per patient
Why does productivity matter so much now? Start with fractionation.
An ABR workforce review stated that the average number of treatment fractions has been roughly halved compared with a decade ago.[] Fewer fractions mean fewer on-treatment visits and less machine time per patient.
For patients, this is good medicine. For workforce planning, this cuts against older demand assumptions built on longer treatment courses.
Reimbursement pressure continues to reshape practice economics
IMRT economics add another layer. A Medicare reimbursement analysis found major declines from 2010 to 2019, with IMRT treatment delivery down 39% and IMRT planning down 31%.[]
The authors concluded that falling IMRT reimbursement was the main driver of the overall decline. In a capital-intensive field, shrinking payment per course changes hiring, ownership, and the viability of smaller practice groups.
Kelly Orwat, MD, a board-cCertified radiation oncologist at Mercy, in Baltimore, told MDLinx, “CMS [Centers for Medicare & Medicaid Services] introduced IMRT reimbursement changes which were particularly drastic for 2026. These reimbursement cuts, combined with increasing use of hypofractionation, strain practice operating budgets at a time when costs to provide high-quality care have continued to increase.”
The ramifications are significant, as she explained: “Radiation oncology clinics require not only a costly treatment machine but also a service contract, a CT for treatment planning, computer software licenses, multiple specialized pieces of equipment for the treatment of individual sites, in addition to staffing."
"In 2026, Medicare introduced multiple cuts in reimbursement, including a new “efficiency adjustment,” which eliminated many radiation oncology billing codes, including those for daily image guidance utilized by radiation therapists and [which are] reviewed daily by the physician to ensure accuracy of radiation delivery," she said. "The cuts also eliminated the IMRT codes, consolidating them into complexity-based codes more. A related but separate challenge both doctors and patients struggle with is the large number of insurance denials and prior authorizations which delay patient care and require the use of further, already strained resources.”
AI is increasing throughput per physician
Auto-contouring and planning tools are no longer pilot projects. British researchers reported in 2025 that AI auto-contouring reduced time from CT to plan completion by up to nine days and improved clinician productivity. In addition, it led to less contouring time and less after-hours work.[]
Dr. Orwat gave her perspective on AI’s place.
“Currently,” she said, “AI has not caused a significant change in my clinic but I anticipate it will, particularly for our dosimetrist. The auto-contouring does a reasonably good job outlining most normal organs, requiring minimal adjustments. Based on my experience, I do not consider it reliable for outlining treatment volumes. With continued improvement, it may decrease the work load for our dosimetrist, potentially enabling a clinic to employ a lower number to care for their volume of patients.”
Related: Biomarker testing: The gap between guidelines and what actually happens in the clinicConsolidation is limiting job flexibility
Practice structure is shifting, too. According to one analysis, the number of practicing radiation oncologists rose 16% from 2015 to 2023.[] Large practices grew 51%, and solo practices fell 27%, resulting in an overall 13% decrease in the number of practices employing radiation oncologists.
Sifan Grace Lu, MD, a lead author on the study, said rates of retirement, new entry, and job change were all “pretty stable” and all below 5% year-to-year. Stable churn does not mean easy hiring. It means fewer doors, especially outside consolidated systems.
Dr. Orwat notes another consideration, saying, “Another change which could impact both workforce and the clinic environment is the allowance for physicians to supervise the office by being virtually available rather than physically present while patients are treated. This could potentially decrease the need to hire locums or fill-in doctors for vacation coverage at the expense of having a doctor available in the clinic.”
Commenting on practice consolidation, she said, “Practice consolidation does continue to occur, with private radiation oncology practices being less common. A private radiation oncology practice that was in the Baltimore region until last year was employed by a hospital system in the fall of 2025. The reasons that private practices are becoming less common include the more challenging financial environment, where it is difficult to absorb the higher operating costs along with the decreasing rates of reimbursement. In addition, many large hospital systems prefer to employ the physicians directly rather than contract with private practitioners.”
So, is the market about to get brutal? Short answer: Not yet.
The better answer is that radiation oncology has a mismatch problem, not a simple surplus problem. Rural access gaps and staffing strain still exist. At the same time, fewer fractions, more automation, reimbursement pressure, and consolidation all squeeze demand for new hires in many urban markets.
The field still needs radiation oncologists. It also needs a tighter grip on how many it trains, where they train, and what a modern workload now looks like.