Fitting of tissue tolerance data to analytic function: improving the therapeutic ratio

Response of human tissues to ionizing radiation is a complex process. It is influenced by many factors, such as use of chemotherapy drugs and underlying diseases such as diabetes and/or lung emphysema. A phenomenological model such as Lyman's is an attempt to predict the complication, for a variety of tissues, in the absence of these factors. The use of the model requires the knowledge of the parameters to predict the response for a specific endpoint. Clinical response data are needed to determine these parameters. Emami et al. [6] have provided some data, based on pre-CT and pre-3-D information, for some of the most serious complications. Based on this information the parameters were determined [4]. However, to validate and further improve the predictive power of the model, improved clinical response data are needed. With CT-based 3-D treatment planning systems the dose-volume information is routinely produced. Efforts by the radiation oncology community are needed to collect this information and correlate it with the clinical outcomes in a uniform and systematic way, not only for the most serious complications but also for less severe radiation-induced complications that are routinely considered in radiation therapy. Also, the information about the tissue response with underlying disease and drugs will be useful. The use of NTCP for plan comparison is useful. However, the incorporation of TCP and NTCP for designing the plan is remarkable. A plan can be optimized for the best outcome for the patient. It is hoped that as the models and parameters are refined and predictive power of the model increases, better plans will be produced, significantly improving the therapeutic ratio.