PV QA 1 - Poster Viewing Q&A 1
Purpose/Objective(s): Model-based approach using normal tissue complication probability (NTCP) model has been proposed to select patient for new treatment techniques, such as proton beam therapy (PBT), based on a certain NTCP difference (∆NTCP) threshold. The goal of this study is to make a reliable dose-response curve to predict the probability of radiation-induced liver toxicity (RILT) as well as the ∆NTCP with confidence interval (CI) between treatment modalities for Child-Pugh A (CP-A) primary liver cancer patients.
Materials/Methods: A total of 180 patients (105 hepatocellular carcinoma and 75 intrahepatic cholangiocarcinoma) were treated with 3D conformal radiotherapy (N=87), intensity modulated radiotherapy (IMRT, N=67), or stereotactic radiotherapy (N=26) between June 2007 and February 2017. Hepatitis B or C virus infection was found in 64 patients. Common toxicity criteria for adverse events (CTCAE) grade≥2 RILT was scored. The mean fraction size equivalent dose of normal liver (mean FED), corrected using reference fraction size=2 Gy/fraction, an α/β ratio=2 Gy, was used as a proxy of generalized uniform equivalent dose for the Lyman-Kutcher-Burman (LKB) NTCP model with fixed volume effect parameter (n) to be 1.0. The estimates of LKB NTCP parameters were estimated by the maximum likelihood method. The variance of ∆NTCP of certain combinations of FEDs was evaluated by Delta method for estimating a CI of ∆NTCP.
Results: CTCAE grade≥2 RILT occurred in 83 patients. The median of mean FED was 19.7 Gy in RILT vs 15.9 Gy in non-RILT groups (p=0.006). The identified NTCP parameters were TD50(1)= 20.6 Gy and m=1.35. There were RILT in 33/64 patients with hepatitis infection compared with 50/116 patients without infection. The TD50(1) and m specifically estimated in hepatitis infection subgroup were 16.2 Gy and 0.68, respectively, and 27.9 Gy and 1.97 in non-infection subgroup. Table showed the estimated NTCP and ∆NTCP with 68%CI of an example patient who received mean FED 33 Gy from IMRT plan and 20 Gy from PBT plan assuming different hepatitis infection status.
Conclusion: Viral hepatitis infection status had substantial effect on the prediction curve in CP-A primary liver cancer patients, suggesting the use of specific NTCP curve in subgroups. The uncertainties of NTCP and ∆NTCP curve improve the reliability of NTCP model-based approach which could be useful to guide patient selection for new treatment techniques such as PBT.
|NTCPIMRT (68%CI)||NTCPPBT (68%CI)||∆NTCP (68%CI)|
|Child-Pugh A with hepatitis virus infection (derived from 64 patients’ data)||93.5% (87.5%-99.5%)||63.3% (56.2%-70.5%)||30.2% (25.5%-34.9%)|
|Child-Pugh A without hepatitis virus infection (derived from 116 patients’ data)||53.7% (43.5%-63.9%)||44.3% (39.5%-49.0%)||9.4% (1.3%-17.5%)|
The asset you are trying to access is locked. Please enter your access key to unlock.