2018-19 Section 7-Neoplastic and Inflammatory Diseases of the Head and Neck eBook

Oral Oncology 73 (2017) 160–165

D.N. Margalit et al.

management for pain, dehydration, malnutrition and/or excess mucus (n = 6), fever of unknown origin (n = 1), obstipation (n = 1) and soft tissue infection (n = 1). For patients diagnosed with aspiration pneu- monia, 7 of the 8 had subjective reports of dysphagia and two had modi fi ed barium swallow (both showed aspiration). Tracheotomy was required for 18% of patients due to vocal fold paresis or laryngeal swelling at a median time of 6.0 months after the end of reRT. The risk of life-threatening hemorrhage, frequently cited as the most devastating risk of reRT, occurred at a median of 8.3 months after reRT. There were 7 patients that experienced hemorrhage, including 4 fatal cases. There were too few cases to identify predictors of hemorrhage. Yet 4 of the 7 patients had a local recurrence that was considered to contribute to the hemorrhage, 1 patient had isolated distant recurrence, and two were disease-free at last assessment. The 2 disease-free patients included one patient who had received 3 courses of radiation to the head and neck and another patient who received adjuvant reRT after salvage laryngectomy with an interval of 16 months from the initial course of RT. For all patients, the likely source of bleed was thought to be in the region of re-irradiation. This included one patient who died at home with oropharyngeal hemorrhage who had received 2 courses of reRT. Details are also provided in Table 3 . Due to the retrospective nature of the study and inconsistent doc- umentation of the precise cause of death, the ability to assess late grade 5 toxicity is limited. There were 13 patients that died without local or distant recurrence. These included 7 of unknown cause, 3 of likely as- piration, 2 of hemorrhage (as described above) and 1 of second primary lung cancer. At last follow-up, 52 (69%) patients had a PEG tube. Patients with a local recurrence after reRT were more likely to remain PEG tube-de- pendent with a rate of 86% vs. 53% PEG tube-use among patients without a local recurrence (p = 0.001). Predictors of serious toxicity are listed in Table 4 . Only receipt of postoperative reRT after surgical salvage was statistically signi fi cantly associated with a decreased risk of serious toxicity (OR = 0.28, p = 0.009) compared with de fi nitive reRT on univariate and multi- variate analysis. Patients who underwent postoperative reRT had lower odds of serious toxicity compared with de fi nitive reRT (37% versus 48%, p = 0.01) even after adjusting for site of reRT, local recurrence, and other variables as listed in Table 3 ( Fig. 1 ). While there was no signi fi cant association between site of reRT and serious toxicity, there was a trend toward increased toxicity with reRT Predictors of toxicity

of the hypopharynx/larynx compared to other mucosal sites (OR 2.19, 95% CI 0.79 – 6.05; p = 0.13) due predominantly to the need for op- erative intervention for necrosis and urgent tracheotomies amongst the patients with a yet intact larynx (n = 10). The toxicity rates by subsite are shown in Supplementary Table 1 . All patients treated with reRT to the oropharynx were PEG tube-dependent at last follow-up with simi- larly high-rates of PEG tube-use among oral cavity sites. Patients with neck-only treatment or larynx/hypopharynx tumors were less likely to be PEG tube-dependent. There was no statistically signi fi cant di ff erence between the odds of PEG-dependence among patients that received reRT to an intact larynx (n = 10) versus post-laryngectomy (n = 13), yet the numbers were small and underpowered (OR 2.0, 0.35 – 11.36); this di ff erence may be more pronounced in a larger cohort. Local re- currence after reRT was correlated with an increased risk of serious toxicity, yet this did not reach statistical signi fi cance. The sequencing of chemotherapy, whether induction or concurrent, was not associated with serious toxicity (OR 0.85, p = 0.83.) Additional radiation therapy-speci fi c predictors of toxicity Radiation-related factors were explored to determine if there was an identi fi able association with serious toxicity. The median laryngeal dose was higher among patients that had a serious toxicity compared with those that did not but this was not statistically signi fi cant (median 43.9 Gy vs. 30.8 Gy, p = 0.22). Additionally, patients that required a tracheotomy after reRT had a higher median laryngeal dose compared to those that did not (median 48.0 Gy vs. 40.1 Gy, p = 0.18). Unexpectedly, the interval between radiation courses was not as- sociated with risk of serious toxicity, whether time was analyzed as a continuous variable (as shown in Table 3 ) or using a cut-o ff of ≥ 5 years or < 5 years (OR = 1.43, p = 0.57). Additional surgical factors We identi fi ed the 18 patients that received fl ap reconstruction as part of the salvage surgery, including radial forearm (n = 5) osteocu- taneous fi bula (n = 3), pectoralis major (n = 7), rectus abdominis (n = 1), anterior lateral thigh (n = 1), and both radial forearm and pectoralis major (n = 1) fl aps. On univariable analysis, there was no statistically signi fi cant association or trend toward a di ff erence in toxicity rates between patients that did or did not receive fl ap re- construction ( Table 5 ). However, we were not able to adjust for the type or extent of surgery that may have warranted extended resection with free fl ap reconstruction.

Table 4 Logistic regression analysis for association of serious toxicity with patient and clinical characteristics (all cases, N = 75). Serious toxicity de fi ned as ≥ 1 of the following: hemorrhage from the head and neck (either mucosal or cutaneous), hospitalization during reirradiation, tracheotomy after the start of reirradiation, soft tissue complication requiring surgical debridement or reconstruction, or other grade ≥ 4 toxicity.

Variables

Serious toxicity:

Single variable analysis

No (n = 36)

Yes (n = 39)

OR (95% CI)

P-value

Gender - male

26 (72.2%) 59 (27, 83)

28 (71.8%) 63 (40, 92)

0.98 (0.36, 2.69) 1.04 (1.00, 1.08)

0.97

Age - years

0.078

Anatomic site of reRT

0.80 N/A

Neck-only

10 (27.8%) 3 (8.3%) 6 (16.7%) 9 (25.0%) 8 (22.2%)

9 (23.1%) 0 (0.0%) 6 (15.4%) 9 (23.1%) 15 (38.5%)

1.0 (Reference)

Nasopharynx Oral cavity Oropharynx

NE

NE

1.11 (0.26, 4.72) 1.11 (0.31, 4.04) 2.08 (0.60, 7.23) 1.03 (0.95, 1.12) 0.28 (0.11, 0.73) 1.02 (0.99, 1.04) 0.98 (0.93, 1.03)

0.89 0.87 0.25 0.47

Hypopharynx/ larynx

Interval between fi rst RT course and reRT - years

2.4 (0.2, 26.4)

2.6 (0.4, 25.2)

Postoperative reRT (vs. de fi nitive reRT)

24 (66.7%)

14 (35.9%)

0.009

Larynx dose (Gy)

30.8 (0.5, 60.7)

43.9 (1.4, 67.9)

0.22 0.48

Total irradiated volume (cc) *

329.7 (73.1, 1265.7)

289.5 (85.2, 991.6)

The continuous variables are given with sample median (range). Abbreviations – reRT = reirradiation, NE = not estimable, NA = not available, OR = odds ratio, CI = con fi dence interval. * = The odds ratio was calculated per 25 cc increase of total irradiated volume.

231