Room: Exhibit Hall
Purpose: To identify the glands, which best correlate with the manifestation of the clinical endpoint of xerostomia. To determine the dosimetric metrics that correlate best with the outcome data. To estimate the parameters of two NTCP models that describe the dose-response relation of xerostomia based on two toxicity scoring systems.
Methods: Eighty-three patients were treated on a multi-institutional phase II study for patients with favorable risk, HPV-associated oropharyngeal squamous cell carcinoma. All the patients received 60Gy IMRT and they reported symptoms using two toxicity scoring systems (CTCAE-PRO and QLQ-C30). Xerostomia was defined as a ≥ 2 point increase from baseline (pre-RT value), independently assessed at 12 months post-RT. We derived the individual patient dosimetric data of the parotid and submandibular glands (SMG) as separate structures as well as combinations. The Lyman-Kutcher-Burman (LKB) and Relative Seriality (RS) NTCP models were used to fit the patient data.
Results: The AUC values were highest for the contralateral glands (parotid+SMG) (Dmean and V27: 0.69) and contralateral parotid (Dmean and V12: 0.69 and 0.67, respectively) for the CTCAE-PRO. For the QLQ-C30, the values for the contralateral glands were Dmean and V51: 0.64 and 0.65, respectively, whereas for the contralateral parotid they were 0.61 for both Dmean and V11. The calculated parameters of the LKB model were D50 = 35.8 Gy, m = 0.61, n = 0.68 for the combined contralateral glands, whereas for the RS model they were D50 = 33.9 Gy, γ = 0.51, s = 0.42.
Conclusion: A correlation was established between the mean dose and given dose volume metrics with xerostomia, which was stronger for the CTCAE-PRO system. The examined NTCP models could fit the clinical data well with similar accuracy. The contralateral glands and contralateral parotid appear to correlate equally well with the outcome data from the CTCAE-PRO and QLQ-C30 systems.