A code orange for traffic-light-protocols as a communication mechanism in IGRT

Traffic-light protocols (TLPs) use color codes to standardize image registration and improve interdisciplinary communication in IGRT. Generally, green indicates no relevant anatomical changes, orange signals changes requiring follow-up but does not compromise the current fraction, and red flags unacceptable changes. This study examines the communication aspect, specifically the reporting accuracy for locally advanced non-small-cell lung cancer (LA-NSCLC), and identifies barriers to reporting.

We conducted a retrospective study on 1997 CBCTs from 74 LA-NSCLC patients. Each scan was in retrospect assessed blinded using the tailored TLP by an IGRT-RTT and subsequently by a second RTT for a subset of fractions. The assessment included both CBCTs from current clinical practice (TLP₂₀₂₃) and from the TLP implementation period (TLP₂₀₁₉). Accuracy of image registration was not evaluated. Reporting barriers were identified through focus group discussions with RTTs.

During TLP₂₀₂₃, 22 of the 63 (35%) patients received at least one code orange during therapy, with 2 of them having a systematic code orange, totaling 43 (2%) fractions with at least one code orange. The IGRT-RTT assigned code orange or red in 59 (94%) patients, 38 (60%) of which had systematic codes orange. In total, the IGRT-RTT reported 684 (40%) fractions with code orange and 13 with code red. During TLP₂₀₁₉, similar numbers are observed. In the subset reviewed by two IGRT-RTTs, reports matched in 77% of cases. Various factors contribute to a low reporting rate, originating both during the decision-making process such as lack of online reporting tools and within offline processes such as divergent feedback expectations.

While our TLP has successfully promoted the widespread adoption of CBCT-based RTT-led IGRT, it has not succeeded in establishing interdisciplinary communication. Our study reveals significant underreporting of flagged LA-NSCLC fractions in clinical practice using a TLP. This underreporting stems from multifactorial origins.