Radiation Accident - IMRT in NY state

2005, a New York (state of) Hospital (unpublished)

A word of warning: this narrative is presented as a lesson in what can occur. One would hope that the vendor has met their responsibility in altering their software to prevent a recurrence of this accident. I have documentary evidence of the incident, it is not made up. The name has been withheld as this is presented as a learning exercise. If you know the case and I have presented something factually incorrect, please let me know.

A site using linacs, R&V and planning software all from the same manufacturer treated a H&N patient with 14Gy to large open fields on three consecutive days, i.e., 42Gy in 3Fx to large fields (for this manufacturer's IMRT process, the primary and secondary jaws are open at the start)

How did this happen?


A H&N patient was planned for a course of sliding window (or dynamic) IMRT. After the first week, the oncologist asked for the patient to be replanned that day.


The dosimetrist made the required modifications, gained approval of the dose distribution and saved the plan. The save process 'hung' in the software and to not lose time the dosimetrist hit the 'off' button on the computer. The plan was then re-opened on another workstation, the remaining work completed and transferred to the R&V. The displayed dose distribution looked the same.


On the treatment machine, the radiation therapist noticed that there was a change in the MLC configuration appearing on the screen (i.e., now there was no leaf motion where there had been before) but the screen did say that MLCs were being used.

After 4 days of this, it was detected that the sliding window was in fact an open window (i.e., the lack of movement of the MLCs was very accurate).


Reconstructive dosimetry undertaken by the site's physicist showed that the patient had received ~14Gy with each fraction on 3 subsequent days (42Gy/4Fx).

The fate of the patient at last information was unknown but this dose to large H&N fields would not be expected to be compatible with life.

What were the problems demonstrated?

  • Radiation Oncologist
    • starting a patient on an inadequate plan in the first place (why the rush? is this the result of commercial interests that occur in privately run institutions?)
    • asking for a rapid re-plan of a complicated technique at short notice (I would think this is obvious.)
  • Dosimetrist
    • agreeing to provide a rapid re-plan on a complicated technique (satisfying professional procedures is more necessary than keeping the oncologist happy. Sometimes the safest thing is to say "no".)
    • not checking all parameters (But this is very unfair, because all parameters were checked, just not on the second computer.)
  • Radiation Therapist
    • not investigating an obvious paradox in the events of delivery (assuming the computer must be right?)
  • Vendor
    • supplying software capable of saving a partial plan (an application specialist from the company nodded knowingly when I related the accident, the specific advice offered was that the cause arose because the MLC settings were the last parameters to be saved by the software. So the turning-off of the computer left the dose pictures intact but did not include the MLC settings, which is why the fields were open)

As you can see, all the participants played a part in the production of this accident.

here is a follow up on this story published in the New York Times.


"Mr. Jerome-Parks approached his illness as any careful consumer would, evaluating the varied treatment options in a medical mecca like New York. Yet in the end, what led him to St. Vincent’s, the primary treatment center for Sept. 11 victims, was a recommendation from an acquaintance at his church, which had become an increasingly important part of his life.

Its cancer unit, managed by Aptium Oncology, a unit of one of the world’s leading pharmaceutical companies, AstraZeneca, was marketing a new linear accelerator as though it had Mr. Jerome-Parks specifically in mind. Its big selling point was so-called smart-beam technology.

“When the C.F.O. of a New York company was diagnosed with a cancerous tumor at the base of his tongue,” promotional material for the new accelerator stated, “he also learned that conventional radiation therapy could potentially cure him, but might also cause serious side effects.”

The solution, the advertisement said, was a linear accelerator with 120 computer-controlled metal leaves, called a multileaf collimator, which could more precisely shape and modulate the radiation beam. (View an interactive graphic demonstrating how multileaf collimators work, and how problems at St. Vincent's caused a fatal overdose.) This treatment is called Intensity Modulated Radiation Therapy, or I.M.R.T. The unit St. Vincent’s had was made by Varian Medical Systems, a leading supplier of radiation equipment.

“The technique is so precise, we can treat areas that would have been considered much too risky before I.M.R.T., too close to important critical structures,” Dr. Anthony M. Berson, St. Vincent’s chief radiation oncologist, said in a 2001 news release."

Unless otherwise stated, the content of this page is licensed under Creative Commons Attribution-NonCommercial-NoDerivs 3.0 License