2017 Section 7 Green Book

M.C. Ward et al. / Oral Oncology 57 (2016) 21–26

Table 4 Univariate Fine-Gray competing risk regression for factors associated with severe late dysphagia.

HR

95% CI

p

Age

1.03

0.984–1.09 0.233–2.84 0.985–1.00 0.526–3.46 0.417–7.93 0.598–3.03 0.745–3.92 0.929–1.05 0.525–3.30 0.805–41.5 0.084–1.47 0.743–1.18 1.10–5.72 0.623–6.45 0.606–6.21 0.407–2.62

0.18 0.75 0.32 0.53 0.43 0.47 0.21 0.76 0.56 0.08 0.15 0.58 0.24 0.26 0.94

Smoking history

Current/After RT vs Never/Former

0.814 0.995

Pack-years smoking

Disease subsite

Supraglottic vs Glottic or NOS

1.35 1.82 1.35 1.71 0.99 1.32 5.78

T stage N stage

T3–4 vs T2

N2a-3 vs N0-1

Grouped stage Year treated Neck dissection

Stage IV vs Stage III

Yes vs No

Chemo type

Multiagent vs Single agent

RT type RT dose

IMRT vs 3D

0.353 0.937

0.028

Altered fractionation Feeding tube type

BID vs QD NG vs None PEG vs None NG vs PEG

2.51 2.00 1.94 1.03

the two reports together emphasize the point that patients should be closely followed by dedicated head and neck caregivers for the duration of their lifetime to screen for new onset severe late dysphagia. Our study differs from RTOG 91-11 not only in its retrospective nature but also in the radiotherapy fractionation and the intensity of chemotherapy. Most patients in the current study received com- bination cisplatin and 5-FU concurrent and conventional radio- therapy and 37% received BID fractionation. Regardless, oncologic outcomes in the current study appear similar or improved to those described in RTOG 91-11 in terms of 5-year overall survival (58% 91–11 vs. 70% current study), locoregional control (54.8% 91–11 vs. 80% current study) and distant control (85.3% 91–11 vs. 84% current study). A low rate of competing events combined with a more intense chemotherapy regimen may have increased the inci- dence of severe late dysphagia in the current study compared to the current era of treatment with IMRT and single-agent chemotherapy although power to detect this on univariate analysis is limited. This study highlights the challenge of measurement of late tox- icity following radiotherapy in the management of head and neck cancer. The insights RTOG 91-11 provided into the incidence, tim- ing and nature of late toxic events were limited by the method of recording a maximum grade late toxicity leading to the speculation that the increased late mortality in the concurrent chemoradio- therapy group reflected an increase in unmeasured or unrecorded severe late dysphagia. Our study provides a more detailed analysis accounting for competing risks of recurrence or death and suggests that this explanation is incorrect. The analysis method we used is therefore a strength of the current report in comparison to the methods used by most studies investigating physician-reported severe toxicity. The univariate regression demonstrating an association between twice-daily (BID) fractionation and severe late dysphagia ( Table 4 ) is likely related to practice patterns at our institution rather than an independent contribution of fractionation. Twice- daily radiotherapy with dose-escalation to 74.4 Gy was a regimen used during the earliest years of this study for patients with locore- gionally advanced disease in attempt to improve oncologic out- comes [5] . Nearly all of these patients received combination cisplatin and 5-FU along with conventional radiotherapy. There- fore, the result on univariate analysis may be a surrogate for older treatment regimens rather than a true independent contribution of fractionation on severe late dysphagia. This is generally supported by the long-term results of RTOG 90-03 which did not demonstrate a clear difference in severe late toxicity within the hyperfraction- ated arm compared to standard fractionation [12] .

The limitations of this study include its retrospective nature, a modest sample-size, patients with an unknown cause of death and the heterogeneity in radiotherapy and chemotherapy regimens used. While the study-size is modest, in comparison the concurrent chemotherapy arm of RTOG 91-11 analyzed 174 patients and expe- rienced more attrition due to locoregional recurrence or death. The follow-up in the current study, although less than RTOG 91-11, appears sufficient to capture the majority of severe late dysphagia events according to the cumulative incidence curves presented ( Fig. 2 ) but may be insufficient to capture very late dysphagia orig- inating beyond 5 years and further study is required. Additionally, six patients in our report died but the cause of death could not be determined. This is a common challenge with elucidating a cause of death and a similar rate was observed on RTOG 91-11 despite the prospective nature of the trial. Although it is impossible to say for sure, there is no reason to suggest that these patients suc- cumbed to severe late dysphagia at a higher rate than the rest of the cohort. In addition, the heterogeneity in chemotherapy and radiotherapy techniques was inherent to changes in practice pat- terns over the years. Although including older techniques limits the application to modern patients, the low incidence of deaths related to severe late dysphagia remains relevant in the modern era and enforces that good candidates for larynx-preservation should continue to be offered chemoradiotherapy. With modern IMRT techniques and chemotherapy regimens, the incidence of SLD is likely to continue to decrease. In conclusion, after a detailed time-to-event analysis accounting for the competing risks of recurrence or death in patients other- wise eligible for larynx-preservation strategies, we did not identify a contribution of severe late dysphagia to late mortality. Larynx preservation should continue to be offered to patients who meet criteria for RTOG 91-11. Because severe late dysphagia can occur beyond five years, patients should be followed closely by a dedi- cated head and neck caregiver to monitor for recurrence, second primary or toxicity for the remainder of their lifetime. Future clin- ical trials should carefully track the incidence of late toxicity to allow for clarity in elucidating the timing and incidence of radiotherapy-induced dysphagia. Conclusion

Conflicts of interest/Financial disclosures

None declared.

69