Functional and Patient-reported Changes in Swallowing and Voice After Combined Chemotherapy and Radiotherapy for Limited Stage Small Cell Lung Cancer

Purpose: To describe the nature and impact of dysphagia (diculty swallowing) and dysphonia (impaired voice) in patients with limited-stage small cell lung cancer (SCLC) before and after chemoradiation. Methods: A prospective cohort study was conducted on patients receiving chemoradiotherapy for limited-stage SCLC. Patients received either 40Gy in 15 fractions, 45Gy in 30 fractions (delivered BD) or 50Gy in 25 fractions, commencing the second cycle of carboplatin/etoposide or cisplatin/etoposide chemotherapy. Outcomes included: videouoroscopy swallowing studies (VFSS) to investigate aspiration, swallowing function, and oesophageal motility disorders; limitations to oral intake; patient-reported swallowing problems; and patient-reported voice problems. Data were collected before treatment and at one, three and six months post-treatment. Results: Twelve patients were enrolled. The oropharyngeal swallow was safe and functional at all time points. Three patients exhibited oesophageal motility disorders before treatment, while a further three exhibited disorders post-treatment. Oral intake was most compromised one month post-treatment with ve patients being either tube dependent or eating only very limited diets. At all other times patients were eating a normal or near-normal diet. Despite normal oropharyngeal swallowing on VFSS, three patients reported moderate or severe swallowing diculties one month post-treatment. Three additional patients reported moderate or severe diculties three and six months post-treatment. Patients who reported swallowing diculties one month post-treatment had all received a mean and maximum radiation dose to the oesophagus of ≥ 15.7Gy and ≥ 42Gy respectively. Dose-response relationships were no longer apparent three and six-months post treatment. Voice problems were variable, with the worst scores reported one month post-treatment. Conclusions: Although patient numbers are small, this study identied discordance between observed swallowing function and patient-reported problems, which has clinical implications for the management of patients with SCLC, as well as identifying areas for future research. Ongoing efforts to reduce mucosal toxicity in lung cancer patients are essential. Across all types, best between NCI CTCAE ratings and patient-reported outcomes 38 . In lung cancer patients specically, a discrepancy between clinician-rated dysphagia (based on the NCI CTC) and patient-reported QOL and pain measures has also been reported 39 . This is the rst time that a discrepancy has been identied between physiological swallow function (using VFSS) and patient-reported dysphagia in lung cancer. The value of instrumental assessment – specically VFSS – cannot be understated. These tools can identify functional problems but they can also differentiate between a physiological and issues, such as residual sensory These ndings highlight the importance of measuring different aspects of swallowing in research that attempts to investigate the nature or extent of dysphagia in a given population, including lung cancer. There are also implications for the clinical management of patients following chemoradiotherapy, as identifying the nature and extent of dysphagia will be highly dependent on the tools used. The subsequent management of dysphagia relies on this accurate classication of the problem, and assists in patient education and support approaches such as speech pathologists and dietitians working together to optimise oral intake and nutritional status.


Introduction
Lung cancer remains the most common cancer in the world, with over 2 million estimated new cases and 1.7 million deaths worldwide in 2018 1 . Mortality and morbidity -both from the disease as well as its treatment(s) -remain signi cant challenges for clinicians, and it has been identi ed that patients with lung cancer experience more symptom distress than those with other types of cancer 2 . Dysphagia (di culty swallowing) and dysphonia (impaired voice) have been identi ed in patients with lung cancer as being signi cant problems, which may be caused by direct tumour invasion (from mediastinal disease or cervical lymphadenopathy), nerve compression (causing unilateral vocal-cord palsy), age, dyspnoea or deconditioning, and these problems may be further compounded by the treatment itself, which often comprises intensive chemoradiotherapy 3,4 . Dysphagia can result in malnutrition, aspiration pneumonia, associated hospital admissions, anxiety and depression, and a signi cant reduction in Quality of Life (QOL) [5][6][7][8] .
Acute dysphagia has been recognised as a dose-limiting toxicity in patients with lung cancer receiving concurrent chemoradiotherapy, with reports that between 30-51% of patients develop moderate (Grade 2) dysphagia, while 11-13% develop severe (Grade 3) dysphagia during the acute phase of treatment 9,10 .
However, the data is limited by the somewhat crude measurement of 'dysphagia', using the NCI Common Terminology Criteria for Adverse Events (CTCAE) scoring system which is based on subjective symptoms and does not identify the actual cause or nature of the swallowing problem. 'Dysphagia' may refer to swallowing di culties as a result of oesophagitis-related pain, aspiration of uids or a physical inability to clear solid food through the pharynx or oesophagus. These varying presentations have vastly different management strategies and impacts on patients, so an accurate appraisal of the nature of the problem using a variety of tools that measure all aspects of dysphagia is essential.
The need for well-designed, prospective research on the nature of dysphagia and dysphonia in lung cancer patients has been previously identi ed 4,11 . Understanding exactly what types of swallowing and voice problems patients develop and when, ensures accurate information can be provided to patients prior to treatment and allows clinicians to provide more timely preventative or therapeutic intervention, potentially avoiding dysphagia-related malnutrition or aspiration pneumonia, or the daily challenges resulting from dysphonia.
Most studies investigating symptoms in patients with lung cancer only include those with non-small cell lung cancer (NSCLC) 2 . Although small cell lung cancer (SCLC) makes up only about 13-15% of all lung cancers 12,13 , management of limited stage disease often involves a hyperfractionated radiotherapy schedule delivered twice daily; a schedule typically associated with increased rates of oesophagitis 14 .
Those with SCLC also have a poor prognosis with a 5-year relative survival rate of 31-34% 12,15 . This higher morbidity and mortality pro le highlights the importance of maximising QOL at all stages of treatment planning and delivery, so accurate data on the outcomes and risks of this vulnerable patient group is urgently required.
There are currently no studies that have reported on swallowing or voice outcomes in patients with SCLC, and the use of instrumental measures of swallowing function for any lung cancer patients is also lacking. The purpose of this prospective study therefore was to investigate the nature and impact of dysphagia and dysphonia in patients with limited-stage SCLC before and after treatment, using a range of measures to examine potential associations between post-treatment outcomes and treatment-related characteristics.

Study Design and Participants
This was a prospective longitudinal cohort study, using repeated measures on a convenience sample of newly diagnosed patients with SCLC from [removed for blind review]. The study was approved by the [removed for blind review] ethics committee prior to commencement (HREC approval: [removed for blind review]) and all participants provided written informed consent prior to participation.
Patients with newly diagnosed limited-stage SCLC were recruited prior to receiving chemoradiotherapy at one of [removed for blind review]'s three radiotherapy sites across [city, country; removed for blind review].
Patients were eligible for inclusion if they were: over 18 years of age; had limited stage disease (de ned as no known distant metastatic disease and appropriate for radical chemoradiotherapy); able to read English; no previous radiotherapy to the head or neck; and no previous conditions or treatment that could cause swallowing or voice impairment. At each time point data were collected on: swallowing function, limitations to oral intake, patient-reported swallowing and voice problems, health-related quality of life (QOL) and nutritional outcomes. Study data were entered into and managed using the REDCap electronic data capture tool hosted at [removed for blind review] 16,17 .
Video uoroscopic swallowing studies (VFSS; also referred to as the modi ed barium swallow) were used to analyse swallowing function. Swallowing images were recorded using the Philips Allura FD20 X-ray system and converted to .avi format for later analysis. During the VFSS participants completed the following with the patient seated in the lateral position: three swallows of 5 ml and one swallow of 10 ml liquid, three swallows of 5 ml semi-solid (pureed fruit), and a piece of solid food (½ cracker). All boluses were mixed or coated with barium to allow clear visualisation of the bolus through the oral cavity, pharynx and upper oesophagus. An anterior-posterior view was also taken while the participant swallowed 5 ml bolus of semi-solid, to allow observation of the presence of oesophageal motility disorder(s). Measures taken from the VFSS images were: (i) aspiration of liquids, using the Penetration-Aspiration Scale 18 , an 8-point scale ranging from 1 (material does not enter the airway) to 8 (material enters the airway, passes below the vocal folds, and no effort is made to eject), (ii) overall swallow function of liquids, semi-solids and solids, using the Swallowing Performance Status Scale (SPSS) 19 , a 7point rating scale which takes a number of different swallowing features into account, including how safely and effectively the bolus moves through the oral cavity and pharynx, and (iii) oesophageal motility disorder(s), rated as present or absent.
Oral intake was measured using the Functional Oral Intake Scale (FOIS) which rates the degree to which a person needs to modify the consistency/texture of food they eat and/or requires tube feeding. The scale ranges from 7 (total oral intake with no restrictions) to 1 (nothing by mouth) and has demonstrated validity and reliability 20 .
Patient-reported swallowing and voice problems were measured using the Dysphagia Handicap Index (DHI) 21 and the Voice Handicap Index (VHI) 22 respectively. These are questionnaires which, for the DHI, comprise 25 items with three response options (never; sometimes; always) and one global item, and for the VHI, comprises 30 items rated on a 5-point Likert-type scale.
The University of Washington QOL questionnaire (UW-QOL) 23 was used to measure health-related QOL. It includes 16 questions regarding pain, appearance, activity, recreation, swallowing, chewing, speech, shoulder, taste, saliva, mood and anxiety.
Demographic and treatment-related data were collected from the medical records. Data of interest included: gender, age, nutritional status (using the Patient-Generated Subjective Global Assessment [PG-SGA] 24,25 ), radiotherapy regimen, chemotherapy agent, mean radiation dose to the oesophagus, maximum radiation dose to the oesophagus, oesophagitis (worst CTCAE score recorded and duration of worst CTCAE score recorded) and required admission for neutropenia.

Statistical Analysis
Descriptive statistics were used to summarise the demographic and clinical characteristics of the participants. These included counts and percentages for nominal valued variables; and means and standard deviations or medians, interquartile ranges and ranges for continuous valued variables.
For the purposes of analysis, the global item scores of the DHI were rated as normal (score of 1), mild (score of 2 or 3), moderate (score of 4 or 5) or severe (score of 6 or 7) 26 . Figure 1 shows the patient ow and data availability from recruitment to nal data collection at 6 months post treatment. From a potential cohort of 34 patients diagnosed with limited-stage SCLC, complete -or near complete -data were collected on 12. Figure 1 outlines the reasons for non-participation or dropout and highlights the challenges of recruiting to studies with cancer types, as well as recruiting to supportive care research, particularly in the period immediately following diagnosis and prior to commencing treatment. Although we had intended to recruit 30 participants in total, the study was closed early as time and nancial limitations did not allow for ongoing recruitment. Table 1 outlines the demographic and treatment characteristics for the 12 study participants. Three patients exhibited oesophageal motility disorders before treatment. These problems persisted posttreatment for those three patients, and there were an additional three patients who also exhibited these disorders at various times post-treatment -one patient at 3 months post-treatment only, one patient at 1 and 3 months post-treatment and one patient at all three post-treatment time points (see Appendix A).

Results
Oral Intake Table 2 outlines the oral diet that patients were managing at each time point. All patients were managing a normal or near normal diet consistency pre-treatment. The most marked limitations occurred at 1 month post-treatment when one patient remained feeding tube dependent (score 3), two managed an oral diet of a single consistency only (score 4), and two managed an oral diet with multiple consistencies, but requiring special preparation (for example, only eating food that was pureed or very soft with extra sauces; score 5). Oral intake returned to a normal or near normal consistency at 3 and 6 months posttreatment with the exception of one patient who remained limited to a special preparation diet (score 5).
Treatment-related toxicity data revealed that all ve patients who required a signi cant modi cation of their diet at 1 month post-treatment (scores 3-5) had experienced Grade 3 oesophagitis and required hospital admission for neutropenia during treatment. Nevertheless, pain had completely resolved for all patients at 1 month post-treatment, other than the one who remained feeding tube dependent due to persistent pain. Of the seven patients who were eating a normal or near-normal diet consistency at 1 month post-treatment (scores 6-7) all had experienced a maximum of Grade 1-2 oesophagitis during treatment and were not diagnosed with neutropenia at any time. There were no relationships readily apparent between FOIS scores and treatment data at 3 and 6 months post-treatment.

Patient-Reported Swallowing Problems
For all domains -physical, functional and emotional -scores were worst at 1 month post-treatment and had improved by 6 months but not back to baseline (pre-treatment) levels (see Table 3). Scores from the global item of the DHI (a rating that "best describes the severity of your swallowing problem") identi ed that two patients reported moderate or severe swallowing di culties at baseline, three patients reported moderate or severe swallowing di culties 1 month post-treatment, but these di culties were reported as no more than mild at follow-up assessments for those patients, and three different patients reported a new onset of moderate or severe swallowing di culties at 3 and 6 months post-treatment (see Fig. 2). Interquartile range 0 to 4.5 0 to 7.3 0 to 2 0 to 2.7 Range 0 to 6 0 to 10 0 to 5 0 to 7

Lower scores represent better outcomes
Exploration of treatment-related toxicity data revealed that patients who reported any swallowing di culties at 1 month post-treatment (mild, moderate or severe) had all received a mean radiation dose to the oesophagus of 15.7 Gy or more, and a maximum dose of 42 Gy or more. In contrast, those who reported no swallowing di culties at 1 month all received mean and maximum radiation doses to oesophagus less than these doses. The relationship between DHI scores and treatment data was no longer apparent at 3 and 6 months post-treatment and there were no obvious relationships between oesophagitis or neutropenia and DHI scores.

Patient-Reported Voice Problems
Patient-reported voice problems were minimal, with median scores of 0 or 1 ('never' or 'almost never' experience the problem) at all time points for the functional, physical and emotional domains (see Appendix B). The most frequent or highest scores were provided in response to questions rating the statements: "My voice sounds creaky and dry", "I run out of air when I talk" and "The sound of my voice varies throughout the day". The greatest spread of scores occurred at 1 month post-treatment although patients continued to report problems (though to a lesser degree) across voice-related physical, functional and emotional domains at 3 and 6 months post-treatment.

Health-Related Quality of Life (QOL)
Scores for all QOL domains varied at all time points with no clear patterns evident. See Appendix C for the scores for each symptom scale presented as per UW-QOL v4 guidelines.
Nutritional Outcomes PG-SGA scores were worst at 1 month post-treatment with a median point score of 8, indicating the need for nutrition intervention, bordering on a 'critical need for improved symptom management and/or nutrition intervention' (see Table 4). Scores at 1 month post-treatment also identi ed that 50% of patients were moderately malnourished (PG-SGA category B), with the remaining 50% being well nourished (PG-SGA category A). No patient was identi ed as being severely malnourished (PG-SGA category C) at any point. Exploration of treatment-related toxicity data revealed that patients who were moderately malnourished and in need of nutrition intervention at 1 month post-treatment had all received a mean radiation dose to oesophagus of 15.7 Gy or more, and a maximum dose of 42 Gy or more.

Discussion
This study is the rst time that detailed functional and patient-reported data on swallowing and voice outcomes in patients with SCLC has been reported. Although patient numbers were small due to challenges recruiting from this uncommon patient cohort, the prospectively collected data is comprehensive and has implications for the multidisciplinary management of these complex patients.
Detailed assessment of swallowing function using VFSS identi ed that in our cohort swallowing was safe and effective at all times. There is limited data in the literature regarding the prevalence of dysphagia in patients with lung cancer. In a study of 72 patients with advanced lung cancer (8% of whom had SCLC) receiving palliative chemotherapy, 18% identi ed as having dysphagia on the self-rated EAT-10 tool 27 . This could represent an over-estimation, as some patients had neurological co-morbidities or a history of radiotherapy to the head and neck, or it could be an under-estimation, as only outpatients undergoing palliative chemotherapy were recruited. A small number of subsequent studies have attempted to identify the prevalence of dysphagia in heterogenous cohorts of cancer patients (including those with lung cancer). Kenny et al. 5 found that of the 59 lung cancer patients in their cohort of 385 cancer patients, 25% had dysphagia as con rmed through clinical evaluation. In a recent study from our centre, 78% of lung cancer patients reported dysphagia for solids and 33% reported dysphagia for liquids 28 . While these symptoms were usually rated as mild, responses to the statement "I have trouble eating certain solid foods" were often severe. The low prevalence of oropharyngeal dysphagia in the current study may be an artefact of the small sample and selection bias; in particular, our careful exclusion of patients with dysphagia-associated comorbidities.
Oral intake limitations, patient-reported swallowing function and nutritional status were most compromised at 1 month post-treatment, even though oropharyngeal swallow function was safe and normal (or near normal) at this time. For the patient with the most signi cant swallowing issues at this time their UW-QOL pain score was 'moderate', but for all other participants there was no clear association between swallowing complaints and reported levels of pain, indicating an alternative cause to their issues. It is possible that residual sensory changes following the resolution of treatment toxicities contributed to these patients' limitations to their diet, reports of dysphagia symptoms and compromised nutritional status. The relationship between radiation dose to the oesophagus and acute oesophagitis is well-established 29 , as is the relationship between oesophageal dose and nutritional status 30 . More recently, neutropenia has also been identi ed as a risk factor for higher grades of acute oesophagitis 10,31 .
In the current study, limitations in oral intake, patient-reported dysphagia and compromised nutritional status at 1 month post-treatment were only observed in patients who had received a higher oesophageal dose, as well as those who had experienced Grade 3 oesophagitis and neutropenia. It therefore seems plausible that more severe toxicities and mucosal damage contribute to residual sensory changes, even after those toxicities have resolved, and this hypothesis warrants further investigation. Reducing this risk requires ongoing efforts to optimise mucosal protection during radiotherapy, and approaches such as IMRT that keep the oesophageal dose as low as reasonably achievable are essential. The discordance between observed and reported dysphagia has been well-documented in the head and neck cancer literature [35][36][37] . Across all cancer types, a modest agreement at best has been shown between NCI CTCAE ratings and patient-reported outcomes 38 . In lung cancer patients speci cally, a discrepancy between clinician-rated dysphagia (based on the NCI CTC) and patient-reported QOL and pain measures has also been reported 39 . This is the rst time that a discrepancy has been identi ed between physiological swallow function (using VFSS) and patient-reported dysphagia in lung cancer. The value of instrumental assessment -speci cally VFSS -cannot be understated. These tools can identify functional problems but they can also differentiate between a physiological problem and other issues, such as residual sensory changes. These ndings highlight the importance of measuring different aspects of swallowing in research that attempts to investigate the nature or extent of dysphagia in a given population, including lung cancer. There are also implications for the clinical management of patients following chemoradiotherapy, as identifying the nature and extent of dysphagia will be highly dependent on the tools used. The subsequent management of dysphagia relies on this accurate classi cation of the problem, and assists in patient education and support approaches such as speech pathologists and dietitians working together to optimise oral intake and nutritional status.
Voice problems in our cohort were minimal overall. Previous research investigating lung cancer patients reported that 90% of them were perceptually dysphonic, however only 27.5% were concerned about their voice 40 . This again highlights the discrepancy between observed and reported problems; indicating an application to dysphonia as well as dysphagia. The minimal problems found in our study may re ect this discrepancy as we only collected data on patient-reported voice problems without any instrumental or clinician-rated validation.
The small numbers in this study presented a number of limitations. We were unable to characterise the prevalence of dysphagia in SCLC or to robustly explore the potential associations between dysphagia and radiation treatment factors. Nevertheless we were able to demonstrate the feasibility of objective dysphagia assessment using VFSS and the ability to collect data across a variety of functional and patient-reported domains. Further research is now required to better understand the apparent discrepancy between observed and reported dysphagia in this population, and to investigate the potential link between mucosal damage and its impact on swallowing function or sensory changes.

Conclusions
This study is the rst time that patient-reported dysphagia following chemoradiation for SCLC has been reported, and it has been identi ed despite an absence of impaired swallow function. Patient-reported dysphagia is multi-factorial, and our ndings suggest that treatment factors such as radiation dose to the oesophagus and oesophagitis may all have an impact. Despite our best efforts, side effects from intensive treatment are often unavoidable, so clinicians must acknowledge and identify the potential for patient-reported dysphagia in lung cancer patients, so that appropriate and timely support and intervention can be provided.

Declarations
Consent for publication:

Not applicable
Availability of data materials: The datasets during and/or analysed during the current study available from the corresponding author on reasonable request.

Competing interests:
The authors declare that they have no competing interests Funding: This work was supported by the Victorian Cancer Agency (VCA) through which the rst author was a recipient of an Early Career Seed Grant Author contributions: JF was a major contributor to study design, data collection, data analysis and interpretation, and manuscript writing.
KG assisted with study design, statistical analysis and was a major contributor in writing the manuscript.
RH assisted with data collection, data interpretation and manuscript writing.
AD assisted with study design and data analysis.
MD assisted with data collection and interpretation.
NK assisted with study design, data interpretation and manuscript writing.
JP-N assisted with study design and manuscript writing.
SS assisted with study design, data collection and manuscript writing.
BS assisted with study concept and design.
DB assisted with study concept and design, data collection, data interpretation and manuscript writing.