^

Patient Care and Research

RO-ILS Case Study 02

Adaptive Planning

 

Background:

Adaptive planning (AP) has had a major impact in radiation oncology. How AP is conducted and communicated varies by institution. One common approach is patients have a quality assurance repeat CT (QACT) during the first 10 fractions and it is typically repeated at two to three times during the course of radiotherapy. The frequency and timing can vary depending upon the clinical situation and the results of the QACT. A physicist/dosimetrist performs the primary review of the QACT and this is typically done within 24 hours of obtaining the QACT. The physicist performs a rigid registration of the QACT to the original CT simulation, transfers the target and normal tissue contours to the CT simulation, and recomputes the original plan on the QACT. Dose volume histograms (DVHs) and statistics are compared between the original and re-calculated plans. An electronic document is created in the radiotherapy-specific electronic medical record and is sent to the physician to review. This QA document contains comparison DVHs, values for institutional dose constraints (target, normal tissue), and a summary recommendation from the physicist regarding the need for re-planning.

Case Example:

An event occurred for a patient receiving definitive radiotherapy for his intact prostate cancer. The first QACT was performed on fraction 6. The QACT document was created and sent to the physician. The physics resident reviewing the QACT did not notice that the small bowel had moved and was adjacent to the target. The max dose to the small bowel was now exceeding the institutional standard (max dose ~ 5600 cGy). The QACT document was created and sent to the physician. This document did show the comparison DVHs but the physics resident did not explicitly state the new max dose to the small bowel and did not recommend replanning. The physician also did not notice the high max dose to the small bowel. The second QACT was performed on fraction 22. The same loop of small bowel that had moved adjacent to the target was again in the same location and the max small bowel dose was again exceeding institutional standards (max ~5600 cGy) but was not noticed by the physics resident. The second QACT document also did not explicitly highlight the new max dose to the small bowel. The physician reviewed the second QA document and this time noticed the high small bowel dose and requested a repeat QACT to occur as soon as possible. This request was communicated via adding a comment to the second QACT document and sending it back to the physics resident. The physics resident did not notice this request from the physician. At the time for planning the boost phase of treatment, the dosimetrist reviewed the QACT and noted the high max dose to the small bowel. Subsequently, the boost plan was optimized using a tighter objective goal for the small bowel, accounting for the high dose the small bowel received during the initial course.

Contributing Factors/Root Causes:

  1. Workload: Adaptive planning is done routinely on all patients at some institutions which significantly increases the workload. Most QACTs require no action (i.e., re-planning). Suboptimal workload (in this case too high) may have contributed to the physics resident and physician inadvertently not noticing the high dose to small bowel [1, 2].
  2. Documentation: The QACT is an automated document and there was no field for reporting the max dose to small bowel. DVHs can be difficult to review, especially when there are many structure sets overlaid in one graph.
  3. Communication: There was no standard method for physicians to communicate the order for repeat QACT.
  4. Process: For the initial CT simulation the individual loops of small bowel were contoured, but for the subsequent QACTs a bowel space was contoured. If the patient’s small bowel loops were contoured on the QACT then it may have improved the acuity of the physicist to notice that a loop of the small bowel had moved to be adjacent to the target.
  5. Supervision: A physics resident was conducting the review of the QACTs. A second check by a supervising physicist may have prevented this event.

Lessons Learned:

Adaptive planning is quickly becoming a standard part of the radiation oncology process (e.g., head and neck IMRT, MRI LINAC, proton therapy). The obvious value is in improving the precision/accuracy of radiation treatments, thus potentially reducing toxicity/morbidity and improving cancer control. However, there is a significant cost in that adaptive planning increases workload. Standard process and procedures as well as the automation of treatment planning and QA checks must be developed to improve the effectiveness and efficiency of adaptive planning. This event illustrates the need for the design of robust processes for the effective and efficient implementation and communication of adaptive planning.

References:

  1. Mazur LM, Mosaly PR, Hoyle LM et al. Subjective and objective quantification of physician's workload and performance during radiation therapy planning tasks. Pract Radiat Oncol. 2013;3(4):e171-7.
  2. Mazur LM, Mosaly PR, Moore C et al. Toward a better understanding of task demands, workload, and performance during physician-computer interactions. J Am Med Inform Assoc. 2016;23(6):1113-20.