2268 - Qualitative and Quantitative Evaluation of Adaptation Strategies for Maximal Dosimetric Benefit in Patients Receiving CBCT-Guided Online Adaptive Radiotherapy for Lung Cancer

J. Duan, J. Harms, H. Boggs, M. Soike, A. J. Kole, N. Viscariello, R. A. Popple, C. Cardenas, D. N. Stanley, and J. A. Pogue; Department of Radiation Oncology, University of Alabama at Birmingham, Birmingham, AL

Purpose/Objective(s): Cone-beam computed tomography (CBCT)-guided online adaptive radiotherapy (oART) has shown benefits for lung cancer patients, yet the optimal adaptation frequency for oART remains unclear. We hypothesized that the magnitude of benefit from oART correlates with the frequency of adaptation and is patient-dependent for conventionally-fractionated lung cancer treatment. Materials/

Methods: Eight lung cancer patients treated with daily CBCT-guided oART in 30/33 fractions (n=7/1) were included in this study. For a total of 243 fractions, daily contours on CBCT and adaptive/non-adaptive plans on synthetic CT (sCT) were used to simulate four different adaptation frequencies: (1) no adaptation; (2) single adaptation (replanning at fraction 21); (3) weekly adaptation (adapted plan used remainder of week); (4) daily adaptation. Dose volume histograms (DVH) from each treatment fraction were compared between adaptation frequencies, and each plan was compared. Patient-specific analysis of different adaptation frequencies was conducted using an open-source PlanScoreCard tool, which quantifies individual plan quality by equally weighting target and OAR dosimetry (50% each). To evaluate the accuracy with which the PlanScoreCard reflected plan quality, three attending radiation oncologists performed manual review of a subset of patients’ plans using a five-point Likert scale.
Results: Across 972 plans, daily-adaptation significant improved (p=0.01, Wilcoxon paired test) clinical target volume (CTV) V100%, planning target volume (PTV) D98%, PTV D0.03cc, PTV homogeneity index, lungs-gross tumor volume (GTV) V20%, heart Dmean, and esophagus Dmean, as compared to all other adaptation strategies. Compared to no adaptation, daily adaptation lead to mean [range] plan score increase of 12.5% [1.7-29.3%], substantially benefiting 6/8 patients (improvement =5.0%). Weekly adaption achieved an increase of 3.3% [-0.9-12.5%], with 3/8 patients receiving nominal benefits. Single-adaptation only showed a mean score increase of 0.5% [-3.6 – 3.3%]. Plan scores showed strong monotonic correlation with Likert ratings (Spearman correlation (R) = [0.67-0.82]).
Conclusion: Physician reviews indicated that an in-house scorecard was adept at measuring increasing plan quality with adaption frequency. Only daily adaptation consistently achieved initial plan quality and showed improvements compared to no adaptation, while the effectiveness of other adaptation frequencies varied and was patient-dependent. Abstract 2268 – Table 1: Comparison of plan quality across different adaptation frequencies and R values using PlanScoreCard tool. A negative score indicates improvement with adaption

	Initial plan score - adaptation strategy score				Spearman R
Patient #	Daily	Weekly	Single	None
1	1.5	4.1	2.9	3.2	0.77
2	-3.7	9.5	16.2	17.5	0.67
3	0.6	17.4	26.5	29.9	0.82
4	2.3	20.8	21	21.6	-
5	-0.2	8	9.1	7.6	-
6	-0.6	1.1	3.4	4	-
7	-0.9	8.5	12	8.4	-
8	-4.9	-1.4	-0.5	2	-
Avg	-0.7	8.5	11.3	11.8	-