Dose to Larynx Predicts for Swallowing Complications After Intensity-Modulated Radiotherapy
Caglar. IJROBP 2008;72:1110


To evaluate early swallowing after intensity-modulated radiotherapy for head and neck squamous cell carcinoma and determine factors correlating with aspiration and/or stricture.

Methods and Materials

Consecutive patients treated with intensity-modulated radiotherapy with or without chemotherapy between September 2004 and August 2006 at the Dana Farber Cancer Institute/Brigham and Women's Hospital were evaluated with institutional review board approval. Patients underwent swallowing evaluation after completion of therapy; including video swallow studies. The clinical- and treatment-related variables were examined for correlation with aspiration or strictures, as well as doses to the larynx, pharyngeal constrictor muscles, and cervical esophagus. The correlation was assessed with logistic regression analysis.

Results

A total of 96 patients were evaluated. Their median age was 55 years, and 79 (82%) were men. The primary site of cancer was the oropharynx in 43, hypopharynx/larynx in 17, oral cavity in 13, nasopharynx in 11, maxillary sinus in 2, and unknown primary in 10. Of the 96 patients, 85% underwent definitive RT and 15% postoperative RT. Also, 28 patients underwent induction chemotherapy followed by concurrent chemotherapy, 59 received concurrent chemotherapy, and 9 patients underwent RT alone. The median follow-up was 10 months. Of the 96 patients, 31 (32%) had clinically significant aspiration and 36 (37%) developed a stricture. The radiation dose–volume metrics, including the volume of the larynx receiving ≥50 Gy  and volume of the inferior constrictor receiving ≥50 Gy  were significantly associated with both aspiration and stricture. The mean larynx dose correlated with aspiration Smoking history was the only clinical factor to correlate with stricture (p = 0.05) but not aspiration.

Conclusion

Aspiration and stricture are common side effects after intensity-modulated radiotherapy for head-and-neck squamous cell carcinoma. The dose given to the larynx and inferior constrictors correlated with these side effects.


The addition of concurrent chemotherapy to radiotherapy (RT) has become a standard of care for the treatment of locally advanced head and neck squamous cell carcinoma (HNSCC). Randomized studies have demonstrated high local control, overall survival, and organ preservation with chemoradiotherapy (CRT) . However, with improving outcomes, concerns have been raised about the increased acute and long-term side effects of CRT. An important side effect of CRT is acute and long-term swallowing difficulty. In randomized trials comparing RT alone with CRT, swallowing difficulty and feeding tube dependence was significantly increased for those undergoing the combined modality treatment . In addition, the swallowing dysfunction observed after treatment can result in weight loss and aspiration pneumonia, and adversely effect patients' quality of life

Although dysphagia and swallowing complications are often reported in clinical studies, few studies have focused on swallowing dysfunction and its causes. The incidence of dysphagia was reported in only 12% of HNSCC studies and feeding-tube rates in only 3% of trials addressing the efficacy of treatment in head and neck cancer

Intensity-modulated radiotherapy (IMRT) is a new standard in the radiotherapeutic management of HNSCC. It can deliver high radiation doses to the tumor while minimizing the dose to the surrounding normal tissue. If the anatomic structures responsible for causing swallowing dysfunction (aspiration and stricture) could be identified, the dose to these structures could be modified to reduce toxicity.

This study evaluated early (1–2 months after therapy) swallowing dysfunction after IMRT with or without chemotherapy and attempted to determine the clinical and/or dosimetric (radiation dose–volume) factors correlating with swallowing toxicity.

 

Multiple toxicity concerns exist for patients undergoing combined modality therapy, including xerostomia, mucositis, taste changes, and so forth. However, among these toxicities, the most serious and distressing to patients are related to swallowing function. Multiple studies have directly related patient-assessed quality of life after head-and-neck treatment to their swallowing function

With the advent of IMRT and the flexibility to achieve more complex dose distributions, radiation oncologists have the opportunity to attempt to more precisely sculpt the dose distributions toward the targets and away from critical structures. To use IMRT to prevent swallowing complications, one needs to determine which clinical factors or anatomic structures, if changed or spared, would lead to a reduction in these toxicities. Therefore, the primary aim of this study was to find a relationship between aspiration and/or stricture formation and the radiation dose to structures involved in swallowing function.

The only clinical predictor of swallowing toxicity in our study was a history of smoking. This finding was not surprising, because previous work has associated smoking with more toxicity during CRT. What was interesting was the lack of significance of the primary site and type of chemotherapy, either in regimen or sequence. It is possible that we did not have sufficient power in this study to identify these relationships. However, we believe that the dose effects of RT are so powerful that they minimize any small effects of these other clinical factors.

Published reports directed at correlating the radiation dose to the specific anatomic structures involved in swallowing have been limited. Eisbruch hypothesized that the possible candidate structures for predicting complications related to swallowing included the pharyngeal constrictors and glottic and supraglottic larynx. When tested in a recently reported prospective study, they found that the mean dose to the pharyngeal constrictors and the partial organ dose for both constrictors and larynx >50 Gy both correlated significantly with the occurrence of aspiration. They found that the dose to the superior constrictor was of greatest significance

When we examined our data, our results mirrored those of the University of Michigan group. We found that the mean dose to the larynx and inferior constrictors and the volume of the larynx receiving ≥50 Gy were the most significant predictors for aspiration. However, we did not find as great a correlation with the superior constrictor dose. Because their study (n = 36) was limited to oropharynx and nasopharynx patients, we performed a subset analysis of our data limiting it to the same patient group (n = 66). This did not change the significance of the doses to the larynx and inferior constrictor; however, we did find a stronger association with the dose to the superior constrictor (p = 0.01). The implications of these two data sets are that avoidance of either the pharyngeal constrictors or the laryngeal structures can prevent aspiration. Clinicians should choose the structure that best fits the clinical scenario and the limits of their optimization system. However, it is our contention that the larynx is a more realistic avoidance structure than the pharyngeal constrictors. In addition, two recent reports have correlated swallowing complications with the dose to larynx, one correlating the dose to the laryngeal edema and the other relating laryngeal dose to speech-related quality of life

The result of our study (i.e., that the dose to the larynx correlated with aspiration) should not be surprising. In two-dimensional head-and-neck treatment planning, a larynx block is traditionally inserted in the low anterior neck to shield the larynx from RT. How to translate this practice to the IMRT setting is a subject of some controversy. Some have advocated split-field IMRT preserving the low neck anterior approach with the larynx block. Others have argued a full-neck IMRT program because of concerns for low-neck failure with split-field techniques. We share the concern for the use of split-field IMRT, especially when contouring the posterior neck nodes in Level V. We have found that only full-neck IMRT adequately addresses these nodes because they often lie deeper than the depth of a single anteroposterior photon field. In addition, the present study results have demonstrated no incidences of aspiration in nearly 100 patients when the dose to the larynx was limited to 48.2 Gy. The implication is that if our constraints are followed, it is not necessary to use the split-field technique. However, we recognize that this will continue to be an area of controversy and good reasons exist to justify both approaches.

In the present study, we found a 37% radiographic stricture rate in our patients. All these strictures were at the laryngeal–cricoid level. None of the patients who had received a mean inferior pharyngeal constrictor muscle dose of <53.9 Gy developed a stricture. It was difficult to correlate these data with the published data, because robust data on stricture vs. dose are limited. In a study by Lee  of 199 patients after concurrent CRT for HNSCC, 21% of the patients developed stricture. When they analyzed the prognostic factors, twice-daily radiation fractionation, female gender, and hypopharyngeal primary site were significant. Our study is the first, to our knowledge, to find a dosimetric relationship between the dose to specific structures and stricture formation. We did not find that patients with laryngeal/hypopharyngeal primary cancer or patients who had undergone surgery before RT were more likely to develop a stricture. Importantly, we chose to be aggressive with the management of these strictures as detected on video swallow study and 31 of the 36 patients underwent immediate dilation, with only two long-term strictures. Our overall stricture rate might appear high; however, this could have been a result of a detection bias. The rate of long-term strictures in this study was consistent with both our own experience and that of other institutes

Finally, it is important to emphasize that in addition to the dosimetric avoidance of the swallowing structures, a critical component to improving swallowing in these patients is the use of swallowing exercises and interventions by speech and swallowing therapists. Studies have shown the importance of early intervention by speech and swallowing therapists and the improvements that swallowing exercises can have on restoration of function

An important limitation of our study was the retrospective nature of the analysis. As such, we did not have good data on pretreatment swallowing function. However, even if we had had this information, it might not have contributed significantly to the results because the swallowing dysfunction of patients before treatment is more likely to be related to tumor burden and might actually improve during treatment rather than contribute to acute on chronic toxicity and could have confused the present results. In addition, we used video fluoroscopy only for patients who had clinical signs of swallowing dysfunction at their initial swallowing evaluation. This might have missed some “silent aspirators.” However, the patients we were primarily concerned were with those with clinically relevant swallowing toxicity; thus, patients with no swallowing deficit at all after treatment were excluded. In, addition in the study by Feng  from the University of Michigan, >90% of their patients were free of any pathologic findings on the pretreatment evaluation, confirming our recollection of our own patients that these deficits were quite limited. Nevertheless, more prospective studies in which all patients undergo detailed swallowing analyses are needed to validate our results.