Potential for reduced toxicity and dose escalation in the treatment of inoperable non-small-cell lung cancer: a comparison of intensity-modulated radiation therapy (IMRT), 3D conformal radiation, and elective nodal irradiation.

PURPOSE To systematically evaluate four different techniques of radiation therapy (RT) used to treat non-small-cell lung cancer and to determine their efficacy in meeting multiple normal-tissue constraints while maximizing tumor coverage and achieving dose escalation. METHODS AND MATERIALS Treatment planning was performed for 18 patients with Stage I to IIIB inoperable non-small-cell lung cancer using four different RT techniques to treat the primary lung tumor +/- the hilar/mediastinal lymph nodes: (1) Intensity-modulated radiation therapy (IMRT), (2) Optimized three-dimensional conformal RT (3D-CRT) using multiple beam angles, (3) Limited 3D-CRT using only 2 to 3 beams, and (4) Traditional RT using elective nodal irradiation (ENI) to treat the mediastinum. All patients underwent virtual simulation, including a CT scan and (18)fluorodeoxyglucose positron emission tomography scan, fused to the CT to create a composite tumor volume. For IMRT and 3D-CRT, the target included the primary tumor and regional nodes either > or =1.0 cm in short-axis dimension on CT or with increased uptake on PET. For ENI, the target included the primary tumor plus the ipsilateral hilum and mediastinum from the inferior head of the clavicle to at least 5.0 cm below the carina. The goal was to deliver 70 Gy to > or =99% of the planning target volume (PTV) in 35 daily fractions (46 Gy to electively treated mediastinum) while meeting multiple normal-tissue dose constraints. Heterogeneity correction was applied to all dose calculations (maximum allowable heterogeneity within PTV 30%). Pulmonary and esophageal constraints were as follows: lung V(20) < or =25%, mean lung dose < or =15 Gy, esophagus V(50) < or =25%, mean esophageal dose < or =25 Gy. At the completion of all planning, the four techniques were contrasted for their ability to achieve the set dose constraints and deliver tumoricidal RT doses. RESULTS Requiring a minimum dose of 70 Gy within the PTV, we found that IMRT was associated with a greater degree of heterogeneity within the target and, correspondingly, higher mean doses and tumor control probabilities (TCPs), 7%-8% greater than 3D-CRT and 14%-16% greater than ENI. Comparing the treatment techniques in this manner, we found only minor differences between 3D-CRT and IMRT, but clearly greater risks of pulmonary and esophageal toxicity with ENI. The mean lung V(20) was 36% with ENI vs. 23%-25% with the three other techniques, whereas the average mean lung dose was approximately 21.5 Gy (ENI) vs. 15.5 Gy (others). Similarly, the mean esophagus V(50) was doubled with ENI, to 34% rather than 15%-18%. To account for differences in heterogeneity, we also compared the techniques giving each plan a tumor control probability equivalent to that of the optimized 3D-CRT plan delivering 70 Gy. Using this method, IMRT and 3D-CRT offered similar results in node-negative cases (mean lung and esophageal normal-tissue complication probability [NTCP] of approximately 10% and 2%-7%, respectively), but ENI was distinctly worse (mean NTCPs of 29% and 20%). In node-positive cases, however, IMRT reduced the lung V(20) and mean dose by approximately 15% and lung NTCP by 30%, compared to 3D-CRT. Compared to ENI, the reductions were 50% and >100%. Again, for node-positive cases, especially where the gross tumor volume was close to the esophagus, IMRT reduced the mean esophagus V(50) by 40% (vs. 3D-CRT) to 145% (vs. ENI). The esophageal NTCP was at least doubled converting from IMRT to 3D-CRT and tripled converting from IMRT to ENI. Finally, the total number of fractions for each plan was increased or decreased until all outlined normal-tissue constraints were reached/satisfied. While meeting all constraints, IMRT or 3D-CRT increased the deliverable dose in node-negative patients by >200% over ENI. In node-positive patients, IMRT increased the deliverable dose 25%-30% over 3D-CRT and 130%-140% over ENI. The use of 3D-CRT without IMRT increased the deliverable RT dose >80% over ENI. Using a limited number of 3D-CRT beams decreased the lung V(20), mean dose, and NTCP in node-positive patients. CONCLUSION The use of 3D-CRT, particul mean dose, and NTCP in node-positive patients. The use of 3D-CRT, particularly with only 3 to 4 beam angles, has the ability to reduce normal-tissue toxicity, but has limited potential for dose escalation beyond the current standard in node-positive patients. IMRT is of limited additional value (compared to 3D-CRT) in node-negative cases, but is beneficial in node-positive cases and in cases with target volumes close to the esophagus. When meeting all normal-tissue constraints in node-positive patients, IMRT can deliver RT doses 25%-30% greater than 3D-CRT and 130%-140% greater than ENI. Whereas the possibility of dose escalation is severely limited with ENI, the potential for pulmonary and esophageal toxicity is clearly increased.