Definitive (chemo)radiotherapy (CRT), i.e., radical radiotherapy with or without concurrent systemic chemotherapy, has emerged as the contemporary standard of care in the nonsurgical management of head and neck squamous cell carcinoma (HNSCC) treated with curative intent due to significant improvements in outcome [1–3]. A small proportion of patients who fail following definitive CRT can be offered salvage surgery, leading to comparable locoregional control and survival. The swallowing apparatus can be severely affected by radiation therapy (RT), resulting in significant impairment of swallowing form and function. It is becoming increasingly evident that although concurrent CRT improves tumor control, it is associated with high rates of severe early and late mucosal and laryngopharyngeal toxicity resulting in an increased incidence and severity of swallowing-related toxicities [4, 5], including dysphagia, aspiration, weight loss, and feeding tube dependence. Recently, conformal techniques such as three-dimensional conformal radiotherapy and intensity-modulated radiation therapy (IMRT), with their potential to generate highly conformal dose distributions around target volumes while sparing surrounding normal structures, have resulted in comparable locoregional control and survival [6] with favorable impact on quality of life (QOL).

Dysphagia is a common, multifactorial, and potentially life-threatening condition with a potential for aspiration and death [5, 7]. CRT-induced deterioration of swallowing function is dependent on several patient- and treatment-related factors [8]. Treatment-induced changes in the swallowing apparatus have been implicated in swallowing dysfunction [9]. Apart from obstruction by tumor itself, impaired mobility of dysphagia–aspiration-related structures (DARS), i.e., pharyngeal constrictors, upper esophageal sphincter, epiglottis, and supraglottic larynx, is responsible for the “anatomical” component of dysphagia [9, 10], whereas throat pain and xerostomia are possibly responsible for the “functional” component. Swallowing function may deteriorate following CRT due to impaired DARS mobility and worsening xerostomia; it may improve due to tumor regression and reduction in anatomic obstruction and throat pain. Conformal techniques can reduce the doses to DARS thus reducing the incidence and severity of severe dysphagia [11]. In the present study, an objective assessment of swallowing function was done using modified barium swallow (MBS) [12, 13] at baseline (pretreatment) and was compared with post-treatment assessments done at prespecified intervals following CRT.

Material and Methods

Forty-seven patients with nonmetastatic, non-nasopharyngeal HNSCC planned for definitive CRT were prospectively accrued and treated in an Institutional Review Board-approved swallowing function study after obtaining informed consent. The inclusion criteria were histologically confirmed squamous cell carcinoma of the oropharynx, hypopharynx, or larynx; stage T1–4, N0–3; age >18 years; Karnofsky performance status (KPS) ≥70; nutritional and physical condition compatible with planned therapy; and no prior history of treatment (surgery, irradiation, or chemotherapy) for the index disease, therapeutic irradiation or chemotherapy. Patients with neurological disease affecting swallowing were excluded. Most of the patients were treated with bilateral parallel opposing portals on a telecobalt unit after conventional two-dimensional fluoroscopy-based simulation and planning with reducing-field technique (66–70Gy/33–35 fractions) along with concurrent platinum-based systemic chemotherapy. Patients were initially assessed 2 months after completion of CRT and at 3–4 monthly intervals thereafter during the first year of follow-up as per routine clinical practice.

Objective assessment of swallowing function was done using the MBS at baseline (pre-CRT) and subsequently at 2, 6, and 12 months post-treatment. Microbar® barium suspension (Microbar, Mumbai, India) was used for MBS (Fig. 1). It does not coat and stick to the mucosa, making interpretation of the pharyngeal residue more apparent. “Thin” barium consisted of barium suspension diluted with water in a 1:2 proportion, “medium” barium was undiluted barium suspension, and “thick” barium was made with a mixture of barium suspension and barium powder in 1:1 proportion. Two interventional radiologists and a swallowing function rehabilitation specialist who were not blinded to the diagnosis objectively evaluated and scored the MBS using the penetration–aspiration scale (PAS) as shown in Table 1 [14]. For evaluation, the recordings were run back and forth at full speed (real-time) and in slow motion. Slow-motion, frame-by-frame analysis was used to evaluate oral, pharyngeal, laryngeal, and cricopharyngeal function. Abnormal swallowing was defined in terms of the incidence and severity of penetration–aspiration, pharyngeal residue, regurgitation of swallow, and posture change during swallow. Penetration is defined as any portion of the bolus entering the laryngeal vestibule to the level of (but not passing below) the vocal folds. Aspiration occurs once the penetrated material passes the level of the vocal folds and enters the subglottic region. Pharyngeal residue is defined as any portion of the bolus remaining in the vallecula and/or pyriform sinuses after the swallow. Postural change is defined as reflex change in the position of a patient’s head or body during the process of swallowing barium in MBS to eliminate risk of aspiration. Barium swallow was repeated using thin, medium, and thick barium. Subjective evaluation of swallowing was also done using a validated questionnaire [15] and performance status scale for head and neck (PSSHN), details of which have been published previously [16].

Fig. 1
figure 1

Modified barium swallow showing a normal swallow, b residual barium, c penetration, and d aspiration

Full size image
Table 1 PAS on MBS
Full size table

Statistical Analysis

Demographic characteristics were summarized using means, percentages, medians, and range. Baseline (pre-CRT), early post-treatment (2 months), and late post-treatment (6 and 12 months post-CRT) parameters were analyzed and compared using nonparametric tests (Mann–Whitney U test). The impact of patient-related factors on the 6-month PAS score was analyzed using Pearson’s χ2 test. Subjective assessment of dysphagia was also correlated with objective assessment using Pearson’s correlation coefficient. All analyses were done using SPSS 15.0 (SPSS Inc., Chicago, IL).

Results

Patient Characteristics

Patient characteristics have been described in detail in a previous report on subjective assessment of swallowing function using a standardized validated questionnaire on the same cohort of 47 patients [16]. Briefly, the study cohort consisted of 40 males and 7 females with a median age of 51 years (range = 40–65 years). The majority (72%) of the patients had a previous history of smoking but were actively dissuaded from tobacco use since diagnosis. Most of them had good KPS (≥70) and preserved nutritional status. Oropharynx (53%) and hypopharynx (34%) were the common primary tumor sites. The majority (76%) of the patients had locoregionally advanced disease (stage III–IV). The vast majority (40/47) of the patients were treated with conventional RT with tissue compensators for improving homogeneity; only seven patients received conformal RT. Forty patients also received concurrent weekly cisplatin chemotherapy as per institutional policy. All 47 patients completed baseline (pre-CRT) swallowing assessment. Although all patients completed the planned course of definitive CRT, the number of patients who underwent post-treatment evaluation decreased over time, with 46, 38, and 17 patients completing the objective assessment at 2, 6, and 12 months, respectively. Patients who succumbed to disease progression (n = 6), including one patient who died within 6 weeks of completion of CRT; had locoregional relapse (n = 12), refused or were unable to perform the full MBS study (n = 8) due to progressive disease or poor performance status; or lost to follow-up (n = 4) were excluded from further analysis at later time points.

Penetration–Aspiration Scale

Pre-CRT evaluation of swallowing status with PAS score showed poor aspiration, high residual, and postural change. However, at post-CRT evaluation there was further deterioration in these domains. This suggests that the tumor itself impairs swallowing function and that treatment with CRT causes further deterioration of objective swallowing function. At pre-CRT 9 (19%) patients had aspiration, 11 (23%) residual barium, 10 (21%) postural changes, and 5 (10.5%) regurgitation (Table 2). At the 2-month post-CRT follow-up, aspiration, residual, postural change, and regurgitation were observed in 11 (23%), 22 (48%), 13 (28%), and 6 (13%) patients, respectively. Aspiration, residual, postural change, and regurgitation were seen in 11 (29%), 11 (29%), 12 (32%), and 10 (26%) and in 5 (29.5%), 10 (59%), 4 (23.5%), and 1 (6%) patients at 6 and 12 months post-CRT, respectively. PAS score was evaluated with different barium concentrations (thin, medium, and thick) under fluoroscopy. Aspiration score was higher with thin barium, while residual and postural changes were higher with thick barium. At pre-CRT evaluation, aspiration with thin and thick barium swallows was seen in 20 and 10% of patients, respectively. On the other hand, residue with thin and thick barium was seen in 13 and 30% of patients, respectively. Postural change was also higher with thick barium.

Table 2 Objective assessment of swallowing function at baseline and serial follow-up with modified barium swallow
Full size table

Patient-related parameters and their impact on baseline pre-CRT objective swallowing function were also analyzed (Table 3). Patients with higher pre-CRT subjective dysphagia (worse PSSHN score) had significantly higher aspiration (P = 0.001), residual function (P = 0.002), regurgitation (P = 0.004), postural change (P = 0.004), and worse PAS scores (P = 0.004) compared with patients with no or minimal subjective dysphagia (better PSSHN score) prior to therapy. Hypopharyngeal primary and tumor volume (>15 cc) also predicted poor PAS scores at baseline. Other factors such as age, gender, habits, and stage did not significantly impact baseline (pre-CRT) objective dysphagia scores.

Table 3 Influence of patient-related factors on baseline (pretreatment) objective swallowing function
Full size table

Among the 34 patients with no or minimal pre-CRT objective swallowing dysfunction (PAS score ≤2), deterioration in PAS scores (3–7) was observed in 13/34 (38.5%) patients at 2 months, 9/30 (30%) at 6 months, and 10/14 (71%) at 12 months post-CRT (Table 4). In this patient cohort, impairment in aspiration was observed in 16, 18, and 13%; pharyngeal residue in 44, 19, and 28%; postural changes in 16, 16, and 11%; and regurgitation in 7, 19, and 2%, respectively, at 2-, 6-, and 12-month follow-up evaluation (Table 5). Thus, among patients without significant objective swallowing function compromise before CRT, a large proportion (44%) had additional impairment not only in the PAS scores but also in residue and posture change domains. Subjective dysphagia (PSSHN scores) correlated well (Pearson’s correlation coefficient 0.97) with objective swallowing dysfunction. At the 6-month post-CRT evaluation, subjective dysphagia (as assessed by PSSHN) was the only patient-related factor that impacted significantly upon objective swallowing dysfunction (Table 6).

Table 4 PAS scores on follow-up in patients with normal baseline scores (PAS ≤ 2)
Full size table
Table 5 Additional impairment in aspiration, pharyngeal residual barium, postural changes, and regurgitation domains on follow-up
Full size table
Table 6 Influence of patient-related factors on PAS scores at 6-month follow-up
Full size table

Discussion

Dysphagia is a common, multifactorial, and potentially life-threatening result of (chemo)radiotherapy [5, 7]. In the largest analysis [17] of swallowing dysfunction, Francis et al. [17] reported 40% dysphagia, 7% stricture, and 10% pneumonia within 3 years of treatment in over 8000 patients treated from 1992 to 1999 who were identified in the Surveillance Epidemiology and End-Results registry database. The rate of dysphagia increased by 10% during this time period (P < 0.05). Treatment-induced deterioration of swallowing function is dependent on several patient- and treatment-related factors. Prospective pretreatment and post-treatment longitudinal assessment of swallowing function using standardized methodology is imperative to establish the impact of various factors on swallowing. In a cohort of 122 patients treated with definitive RT, Caudell et al. [18] identified primary site (P = 0.01), concurrent chemotherapy (P = 0.02), and increasing age (P = 0.04) as factors significantly associated with long-term composite dysphagia. In another study [19], worse swallowing was associated with living in rural areas, history of heavy alcohol consumption, hypopharyngeal primary, large tumor volume, and nonconformal radiotherapy. In a large prospective study involving 529 patients treated with curative RT, Langendijk et al. [8] proposed a predictive risk model using logistic regression analysis. Tumor stage, bilateral irradiation, pharyngeal tumors, accelerated RT, and concurrent chemotherapy were independent prognostic factors for swallowing status at 6 months that were combined to calculate the Total Dysphagia Risk Score (TDRS) that stratified them into three risk categories. TDRS risk classification was significantly associated with acute dysphagia (P < 0.001) as well as with late swallowing dysfunction at 12, 18, and 24 months (P < 0.001 at all time points).

In the present study, a substantial proportion of patients had aspiration, pharyngeal residue, postural change, and regurgitation even at baseline, which is likely to be related to the site and volume of the primary disease. Following treatment, the disease burden was reduced thus potentially improving swallowing function. However, acute mucositis can worsen preexisting dysphagia, as was seen at the 2-month (early) evaluation. Late toxicity of the pharyngolaryngeal mucosa may further impair swallowing on follow-up, even with controlled primary. Decline in swallowing function assessed by PAS scores was persistent at the 6- and 12-month follow-up. Residue and aspiration were affected the most, which points toward impaired mobility of the DARS as being responsible for impairment post-CRT. An increase in aspiration may be associated with early aspiration-related death. Interestingly, several patients with impairment in residual function in early post-CRT period did not complete the 6-month assessment, suggesting that these patients had a greater number of events. Thin barium swallow was associated with higher aspiration, whereas thick barium had more residual, supporting the usage of a semisolid diet until complete recovery from swallowing dysfunction.

Comprehensive irradiation of a head and neck primary site with neck nodes results in salivary gland hypofunction leading to xerostomia [20]. Xerostomia results in decreased bolus lubrication and increased bolus transit time. Radiation can also affect the neuromuscular mechanism of swallowing resulting in multiple swallowing-measure abnormalities such as increased oropharyngeal transit time, incoordination of bolus movement through the oropharynx, reduced tongue-base contact with the posterior pharyngeal wall, and restricted laryngeal and hyoid elevation. Conformal techniques such as IMRT have consistently reduced xerostomia in single-institution studies as well as in multi-institutional and cooperative group trials [6]. In addition, IMRT can reduce the dose to the swallowing apparatus and DARS [11, 21], with the potential to reduce dysphagia [12]. The contouring and localization of these structures, however, remains somewhat challenging [12].

Videofluoroscopy is generally considered the gold standard for objective assessment of swallowing function [22, 23], but its utility is limited due to lack of widespread availability, high cost, and the need for a well-trained, experienced staff. MBS may yield results similar to videofluoroscopy but is less expensive and can be implemented by relatively inexperienced health-care personnel. An additional benefit of MBS is the ability to observe real-time movements of anatomic structures.

Conclusion

This is one of the few studies that addresses objective swallowing function using MBS in patients treated with mostly two-dimensional RT. Most patients have preserved swallowing function prior to therapy. Subjective dysphagia, hypopharyngeal primary, and tumor volume are patient-related factors significantly affecting objective swallowing function even before CRT. There is significant impairment of swallowing function in all domains following treatment, with pharyngeal residue and aspiration domains the most significantly affected. Despite relatively short follow-up and attrition bias, the present study provides useful information regarding patterns of objective swallowing dysfunction following (chemo)radiotherapy. Objective swallowing function preservation should be considered an important end point for future studies involving high-precision radiotherapy.