Virtual Library

Start Your Search

L. Hartsing



Author of

  • +

    P1.08 - Poster Session 1 - Radiotherapy (ID 195)

    • Event: WCLC 2013
    • Type: Poster Session
    • Track: Radiation Oncology + Radiotherapy
    • Presentations: 1
    • +

      P1.08-011 - Dosimetric Predictors of Esophageal Toxicity in Patients with Non-small Cell Lung Cancer Receiving Chemotherapy and Radiotherapy (ID 1507)

      09:30 - 09:30  |  Author(s): L. Hartsing

      • Abstract

      Background
      Esophageal toxicity can be a dose-limiting event in patients with non-small cell lung cancer receiving chemotherapy and radiotherapy necessitating treatment breaks with potential to cause adverse treatment related factors. The objective of this study was to investigate those factors, both clinical and dosimetric, which predict for esophageal toxicity.

      Methods
      Patients (pts) with non-small cell lung cancer prospectively enrolled into an IRB approved database were retrospectively reviewed. Pts with biopsy-proven non-small cell lung cancer treated with radiotherapy alone, sequential or concurrent chemoradiotherapy had maximal esophageal toxicity scored per CTCAE 4.0 criteria. V5, V10, V20, V30, V40, V50, V60, V70, esophageal hot spot, and dose per fraction were the dosimetric variables and age, sex, race, chemotherapy, and stage were the clinical variables investigated. Data were analyzed using SAS (SAS INC, Cary, NC) Version 9.2 software package. Ordinal maximum reported esophageal toxicity was evaluated using logistic regression. A multivariable regression model was fit using important univariate predictors along with a backwards elimination stepwise regression. Probability of esophageal toxicity as a function of absorbed dose in a partial volume was modeled by the method of Lyman, by converting the dose volume histograms into an equivalent fractional volume receiving the maximum dose in the DVH, using the effective volume method of Kutcher and Burman. The parameters in this model (D50, slope m and volume exponent n) were determined by maximum likelihood estimation.

      Results
      A total of 100 pts were enrolled between 7/10 and 12/12 into a prospective database and eligible for analysis. Pts were excluded without a complete dose volume histogram data or were stage I disease leaving 71 eligible for analysis, 43 females and 28 males with a median age of 61 (range: 39-85). 14 pts were treated with radiotherapy alone while 23 received sequential treatment and 34 concurrent treatment. The median delivered dose was 66.6 Gy (range: 27.5-66.6) in a median of 1.8 Gy (range: 1.8-3.0) per fraction. Maximal esophageal toxicity was rated as 0: 12pts, 1: 21 pts, 2: 33 pts, and 3: 5pts. Univariate predictors of > grade 2 esophageal toxicity included, V5-V60 and use of concurrent chemotherapy. The maximum likelihood fit of the Lyman model parameters to patients with ≥ 2 esophageal symptoms were n=0.26 m=0.32, TD50=39.1 Gy when the α/β ratio was assumed to be 10 Gy. The maximal likelihood fit of the Lyman model parameters to patients when the α/β ratio was not set were n=0.26, m=0.32 and TD50 39.3 with the α/β calculated at 7.6 Gy. Patients not having chemotherapy had a higher TD50, 46.4 Gy as compared to patients having chemotherapy, TD50=37.1 Gy, p=0.09.

      Conclusion
      We have shown the TD50 for > grade 2 esophageal toxicity is lower for patients receiving chemotherapy and radiotherapy compared to patients receiving radiotherapy alone. This is first report to show the α/β ratio for esophageal toxicity may be lower than 10. Confirmations of these data are needed in an independent data set.