Non-coplanar beam intensity modulation allows large dose escalation in stage III lung cancer

Purpose: To evaluate the feasibility of dose escalation in stage III non-small cell lung cancer, we compared standard coplanar (2D) with non-coplanar beam arrangements, without (3D) and with beam intensity modulation (3D-BIM). Materials and methods: This study was a planning effort performed on a non-selected group of 10 patients. Starting from a serial CT scan, treatment planning was performed using Sherouse's GRATIS(RM) 3D planning system. Two target volumes were defined; gross tumor volume (GTV) defined a high-dose target volume that had to receive a dose of at least 80 Gy and GTV plus the lymph node regions with >10% probability of invasion defined an intermediate-dose target volume (GTV + N). It was our intention to irradiate GTV + N up to 56 Gy or more. If the prescribed doses on GTV and GTV + N could not be reached with either the 2D or 3D technique, a 3D-BIM plan was performed. The 3D-BIM plan was a class solution involving identical gantry angles, segment arrangements and relative segment weights for all patients. Dose volume histograms for GTV, GTV + N, lung and spinal cord were calculated. Criteria for tolerance were met if no points inside the spinal cord exceeded 50 Gy and if at least 50% of the lung volume received less than 20 Gy. Under these constraints, maximal achievable doses to GTV and GTV + N were calculated. Results: In all 2D plans, spinal cord was the limiting factor and the prescribed doses for GTV and GTV + N could not be reached in any patient. The non-coplanar 3D plan resulted in a satisfying solution in 4 out of 10 patients under the same constraints. In comparison with 2D, the minimum dose in GTV + N was increased. Six patients had to be planned with the 3D-BIM technique. The theoretical minimum dose to GTV + N ranged between 56 and 98 Gy. The delivery of 80 Gy or more to GTV was possible in all patients. For a minimal dose of 80 Gy to GTV, the maximal dose to any point of the spinal cord varied between 27 and 46 Gy. The lung volume receiving more than 20 Gy ranged from 26 to 46%. Conclusion: The potential of 3D-BIM for dose escalation is explained as follows: (i) compared to other planning techniques, a larger amount of lung tissue can be spared by using beam directions that are well-aligned with the mediastinal structures. Such beam directions have narrow angles with the sagittal plane; (ii) dividing all beams into segments with well-specified geometrical restrictions in relation to the spinal cord and well-defined relative weights results in a lower dose to the spinal cord. (C) 1997 Elsevier Science Ireland Ltd.