Physico-biological treatment plan optimization for individualized dose escalation in advanced non-small cell lung cancer radiotherapy: Maximally achievable tumor dose versus risk of radiation pneumonitis

Purpose: Radiation dose escalation was shown to reduce local relapse and to improve overall survival in advanced stage nonsmall cell lung cancer (NSCLC) patients. Dose-enhancement of an existing treatment plan can be achieved by re-normalization until a pre-defined normal tissue dose constraint is reached, by re-optimization using a trial-and-error approach or by physicobiological optimization (PBO) where the tumor control probability (TCP) and normal tissue complication probability [e.g., for Grade ≥2 radiation induced pneumonitis (NTCPRIP)] drive the inverse treatment planning process. Methods and Materials: This in silico study included ten randomly selected patients with stage IIIA/B NSCLC. The initial treatment plans at standard fractionation were re-normalized or re-optimized using PBO for an iso-toxic (IT) or maximally tolerable (MT) approach to determine the maximally achievable TCP as function of the NTCPRIP, with and without target dose homogeneity constraint, while satisfying dose constraints for the esophagus, spinal cord and heart. Results: In the initial treatment plan, the TCP was 34% and NTCPRIP 6.4 16.4%. For the homogeneous approaches, IT re-optimization yielded the least benefit, while MT re-optimization resulted in a median [range] TCP of 53.5% [37.8 73.8%]. MT heterogeneous re-optimization produced a median TCP of 72.8% [58.8 91.4%]. Fraction sizes for the MT homogeneous and heterogeneous reoptimized study arms were 2.2 [2.0 2.3] Gy and 2.3 [2.1 2.7] Gy, respectively. Conclusions: PBO provides an individual trade-off boundary of TCP versus NTCPRIP as a means to generate standard optimized treatment plans that are independent from the local institute's dose-prescription protocol, and allows for 'customized' dose prescription.