Session Item

Adaptive radiotherapy (ART) enables correction of dose deviation in organs at risk (OARs) caused by anatomic changes during head and neck RT. The intensive workload for ART requires early identification of head and neck cancer (HNC) patients that benefit most during radiotherapy, not only to save workload/resource, but also to improve therapeutic ratio. The purpose of this study was to present a simple strategy to select candidate of ART in the first week of radiotherapy.

This retrospective analysis was based on 123 HNC patients who underwent photon radiotherapy and had at least 5 weekly in-treatment re-scan CTs (rCTs) available. In total, 15 OARs were segmented on each rCT using a semi auto-segmentation method (Gan et al., 2021). Then, the original treatment plan (VMAT or IMRT) was re-calculated on each rCT with full scale of prescription dose to obtain the actual weekly mean dose (D_mean) of OARs. The actual given D_mean was defined as the average of all actual weekly D_mean. We then applied the comprehensive NTCP-profile for HNC patients (van den Bosch et al., 2021)to translate the planning D_mean and actual given D_mean into nominal and actual NTCP values, respectively.  The ΔNTCP were obtained by subtracting nominal NTCP with actual NTCP values. The ΔD_mean  between the first weekly actual D_mean and planning D_mean was calculated and designated as ΔD_mean__week1. For each ΔNTCP, the maximum  ΔD_mean__week1 of OARs which were involved in the NTCP model was designated as maxΔD_mean__week1.  The maxΔD_mean__week1 was then used to develop different logistic regression models to identify suitable candidates of ART who presented large ΔNTCP with different criterion ΔNTCP from 0 to 10 of percentage point. For different ΔNTCP criterion, the cutoff maxΔD_mean__week1 was calculated (R program package of pROC) to obtain the maximum summation of sensitivity and specificity, patients presenting ΔD_mean__week1 of any OAR ≥ the cutoff maxΔD_mean__week1  were designated as classified candidate of ART, patients presenting any ΔNTCP ≥ criterion of ΔNTCP were classified as true candidate of ART. Major ΔNTCP was defined as ≥5% and considered clinically relevant.

In total, 7832 complete cases of NTCP values for different kinds of toxicity in the 6^th, 12^th, 18^th and 24^th month after the end of radiotherapy together with the corresponding maxΔD_mean__week1 were obtained. The Spearman’s rank correlation coefficient between maxΔD_mean__week1 and  ΔNTCP  was 0.42 ( p-value <0.001)(Fig.1). For major ΔNTCP, the cutoff maxΔD_mean__week1  was 5.14Gy, the proportion of true candidate of ART was 4.5%, the proportion of classified candidate of ART was 19.1% which included all the true candidate of ART. (Fig.2).

Fig.1 

Fig.2

Using 5.14 Gy as a cutoff value for ΔD_mean of OARs in the first week enables classifying all true candidates of ART who present ΔNTCP ≥ 5%, meanwhile correctly spares 80.9% of patients from labor-intensive procedure of ART.