A Nomogram to Predict Severe Late Toxicity after Definitive Reirradiation for Squamous Carcinoma of the Head and Neck
Ward MC, Lee N, Caudell JJ, Chagin K, Awan MJ, Koyfman S, Dunlap NE, Zakem SJ, Hassanzadeh CJ, Marcrom S, Boggs H, Isrow D, Vargo JA, Heron DE, Siddiqui F, Dault J, Bonner JA, Beitler JJ, Yao M, and Riaz N. A nomogram to predict severe late toxicity after definitive reirradiation for squamous carcinoma of the head and neck. Int J Radiat Oncol Biol Phys 2017; 99(2):S235-S236.
Int J Radiat Oncol Biol Phys
Purpose/Objective(s): Severe late toxicity is common after re-irradiation for recurrent or second primary (RSP) squamous carcinoma of the head and neck. Progressive disease after re-IMRT is also common, and the relationship between late effects and competing risks is critical to both patient and physician decision-making. From an updated multi-institution dataset, we developed a nomogram for patient-level prediction of radiation-related late toxicity while accounting for the competing risks of progression or death. Purpose/Objective(s): Patients with RSP squamous carcinoma originating in a field previously irradiated to at least 40 Gy and treated non-operatively with IMRT-based re-irradiation to at least 40 Gy were collected. Late toxicity developing beyond 3 months after the end of re-IMRT were collected using CTCAE v4.0 and specifically included osteoradionecrosis, aspiration pneumonia, esophageal strictures, carotid blowout syndrome, fistula and tissue necrosis. Feeding tube dependence beyond one year in the absence of progressive disease was also considered late toxicity. A multivariable competing-risk model was then fit to the actuarial risk of late toxicity with subsequent progression or death treated as competing risks. Variable selection for the model was performed based on reverse selection using the bootstrap optimized concordance statistic and predictive ability was plotted to generate a calibration curve. The final model was created into a nomogram for visual use. Results: From 9 institutions, 505 patients were included. The overall cumulative incidence of late toxicity was 16.1% at 2-years and the cumulative incidence of progression or death was 65.5%. The final model included the following factors, with the listed groups experiencing a higher rate of late toxicity: radiotherapy to the neck, BID fractionation, systemic therapy with re-irradiation, a shorter time between radiation courses, pre-treatment organ dysfunction, a lower rN-stage, better KPS, second primary tumor (rather than recurrence), pharynx/larynx tumors, higher first dose of radiation, non-IMRT RT during the first course and increased pack-years smoked. The model was well-calibrated between 0% and 40% risk of late toxicity and the concordance statistic of the final model was 0.764. Conclusion: Only 1 in 5 patients treated with definitive re-IMRT is alive at 2 years without progression, death, or late toxicity. The risk of severe late toxicity from re-IMRT is dependent on both treatment factors and factors that modulate the competing risk of progression or death. Modifiable treatment factors such as twice-daily fractionation and neck irradiation may affect this risk. Patients likely to survive without progression are also more likely to experience severe late toxicity. This model has significant implications for treatment selection and consent discussion for re-irradiation in the IMRT-era.