234 - Analysis of Protocol Deviations at a Large Multi-Center Academic Radiation Oncology Department to Improve Patient Safety

M. Shah¹, A. Kapur², L. Potters², and L. Tchelebi²; ¹Northwell Health, Lake Success, NY, ²Northwell, Lake Success, NY

Purpose/Objective(s): While peer review is considered an essential component of radiation therapy quality assurance (RT QA) for cancer patients, peer review of contours prior to treatment planning is rarely done in the United States. This study reviews the results of contour peer review identifying the leading causes of deviations at our large academic medical center. Materials/

Methods: All cases within our department are treated per a uniform set of system-wide directives, akin to abbreviated clinical trial protocols. Once contoured, cases undergo prospective peer review with each case assigned a grade (A, B, or C). “A” cases are considered per protocol (i.e. directive) requiring no changes, while “B” and “C” cases are felt to have protocol deviations requiring minor or major modification, respectively, prior to treatment planning. We reviewed all radiation cases with curative intent from 2017-2022 and identified reasons for protocol deviations. We analyzed protocol deviations by primary tumor site, physician tenure (> or < 10 years at our department), and generalist vs specialist. The two-proportion Z-test was performed to determine differences in the proportion of B scores.
Results: Of the 13,091 curative cases treated from 2017-2022, 41.3% were scored as B and 0.1% were scored as C. Of the cases receiving a B score, 52.2% were due to contour modifications (of target volume or organ at risk, OAR), 37.4% were due to directive variance (discrepancy between the prescription and treatment directive), and 10.4% were due to other errors (Table). The 5 anatomical sites with the highest proportion of B scores (out of all cases receiving a B) were head/neck (20.2%), prostate (13.2%), breast (9.1%), lung (7.9%), and endometrial (5.7%). There was no significant difference in the percentage of B scores between tenure (n=9) and non-tenure (n=25) physicians (40.5% vs. 43.2%, p = 0.05). Of the top 10 anatomical sites, generalists (n=22) had a significantly higher percentage of B scores than specialists (n=12) (46.8% vs. 32.8%, p < 0.002).
Conclusion: This study outlines deviations from intended treatment identified during contour peer review for cancer patients treated at our large academic medical center. We found that inaccuracies in target volume and OAR contouring were responsible for most protocol deviations, regardless of disease site, physician tenure at our center, and particularly among generalists. The overall discovery of deviations in 41% of curative cases shows the value of contour peer review rounds and the ability to catch errors upstream of treatment planning. Table. Reasons for B Grade Assignment

Peer Review Category	Reason	Number of Reports	Fraction, %
Contour Modifications:		2821	52.2%
	Modify Contour - PTV	1771	32.8%
	Modify Contour - OAR	928	17.2%
	Modify Contour - PTV & OAR	122	2.3%
Directive Variance:		2021	37.4%
	Rx & Directive Mismatch	1848	34.2%
	Dose Constraint	96	1.8%
	Laterality	66	1.2%
	Rx/Directive Modification	11	0.2%
Other:		562	10.4%
	Miscellaneous	488	9.0%
	Contour Delinquent	74	1.4%
Total:		5404	100%