Association Between Objective Measures of Oral Health and Salivary Gland Irradiation with Patient Reported Outcomes Following Head Neck Radiation Therapy

Abstract

Purpose:

To determine whether objective measures of oral health and salivary gland irradiation explain subjective measures of eating, drinking and salivation in patients following head neck radiation therapy (HNRT).

Methods:

This cross-sectional study included 112 patients following HNRT with a completed patient reported outcome (PRO) scale. Objective measures at post HNRT visit included decayed-missing-filled teeth (DMFT) scores, periodontal disease condition, oral hygiene status and dental prosthesis use and prescribed radiation dose to salivary glands. Data was collected and statistical analysis was performed.

Results:

There was no significant association seen between PRO scales and dental prosthesis use, periodontal disease, and oral hygiene. Although some significant findings were seen with DMFT and prescribed radiation dose to salivary glands, this explained only very small amounts of the variation in eating, drinking and salivation measures in these patients.

Conclusion:

PRO measures should be integrated in the routine care of head and neck cancer patients.

INTRODUCTION

Head neck radiation therapy (HNRT) is known to cause significant acute and chronic oral adverse effects that can have profound impact on patients’ general and oral health related quality of life (QOL).^1,2 The most common is xerostomia, decreased salivary output that can affect patients’ swallowing and masticatory functions.^3,4 Previous study has demonstrated that the rate of HNRT-induced xerostomia is highest within the first half-year after treatment and decreases with time.⁵ Trismus - interincisal distance less than 35mm - is another common complication of HNRT⁶ and can lead to improper mastication function, difficulty in maintaining adequate oral hygiene and associated with poorer QOL.^7,8 The long-term side effects of HNRT that impact QOL is typically assessed using PRO tools which are designed to assess how a patient evaluates his or her functioning (subjective assessment) rather than actual performance (objective assessment).⁹ Various PRO tools have been used for accurate measurement of symptom burden that is essential in order to provide individualized supportive care in head and neck cancer patients.^10,11Several studies have been published in the literature evaluating long term QOL in patients with head and neck cancer such as swallowing and mastication,¹² salivary flow rates,¹³ oral intake,¹⁴ impact of post radiation caries,¹⁵ trismus,¹⁶ xerostomia,¹⁷ and physical health and psychosocial wellbeing¹⁸

While it is evident that head and neck cancer patients are highly prone to radiation associated toxicities within the oral cavity, there exist only fewer studies correlating objective measures of oral health following HNRT.^19,20 A longitudinal study by Hammerlid et al. demonstrated that oral health related measures such as dentition status, dry mouth and trismus persisted beyond 3 years from diagnosis in head and neck cancer patients and continued poor QOL.¹⁹ An observational study by Duke et al. demonstrated that dentition status such as occlusion, decayed-missing-filled teeth, trismus, and denture use has a persistent impact on subjective QOL in long term head and neck survivors.²⁰ Another study reported that oral health condition including caries activity, presence of periodontal disease and incidence of missing teeth was higher in patients following HNRT with a direct impact on their QOL.^21,22 In addition, there is paucity of literature on association between salivary glands irradiation with PROs that can directly or indirectly affect the objective measures of oral health. Arbab et al. retrospectively evaluated PROs and radiation dose delivered to the organ at risk (OAR) in 169 patients with head and neck cancers and identified only weak association between PROs and OARs including major and minor salivary glands.²³ A prospective study by Hayakawa et al. demonstrated that radiation dose to the parotid gland were associated with severe deterioration in health related QOL attributable to difficulty with social eating in patients following HNRT.²⁴

As a part of a multidisciplinary team effort to improve functional outcomes in patients following HNRT at our institution, a cross-sectional study was conducted on 112 head and neck cancer patients treated with IMRT comparing the objective measures of trismus (maximal interincisal opening) and salivation (unstimulated and stimulated salivary flow rate) with the PRO scales from the validated FACE Q Head and Neck Cancer Module.²⁵ It was found that the objective measurements of trismus and salivation explained only a small fraction of the variation in the eating-drinking-salivation measures. The present study aims to further examine how much variation is explained with oral health objective measures in the same cohort of head and neck cancer patients with their existing trismus and xerostomia measures. Additionally, we investigated how salivary glands irradiation affected the PRO scales in these patients.

PATIENTS AND METHODS

This cross-sectional study was approved by Memorial Sloan Kettering Cancer Center (MSKCC) Institutional Review Board.

Patient population:

Between January 2016 (when PRO scales were first implemented in association with HNRT objective measurements) and March 2021 (when the study database was closed for analysis), we identified 115 head and neck cancer patients who underwent intensity modulated radiation therapy (IMRT) at MSKCC with a pre-HNRT objective measurement (unstimulated salivary flow rate, stimulated salivary flow rate and interincisal mouth opening), and completed a PRO scale within 6 months of post-HNRT objective measurement during follow-up at our institution. Patients with head and neck cancer recurrence and prior history of HNRT were excluded from the study. Patients were not excluded based on tumor histology or head neck subsites. Three patients missing all three subjective measures on the PRO questionnaire or who had PROs captured later than 6 months of post-HNRT evaluation were excluded, leaving 112 patients in the analysis.

Objective measurements:

Standard post-HNRT objective measurement included unstimulated and stimulated salivary flow measurement. The patient was instructed to spit into the pre-weighed cup once every minute for five minutes. This cup was then weighed. For stimulated salivary flow measurement, the salivary flow was first stimulated by applying 2% citric solution using two cotton tip applicators on the lower sides of the patient’s tongue every 30 seconds for 2 minutes followed by the same method of measurement. Maximal interincisal opening was measured with a Boley gauge using an established reference point (usually maxillary and mandibular central incisors). If there were no established reference points or if the patient had an incomplete set of teeth, the immediate next tooth was taken into consideration.

Objective measures of oral health:

As standard practice at our institution, patients are referred to our Dental Service for a pre-radiation dental evaluation and necessary dental extractions prior to HNRT.²⁶ These patients also undergo long-term dental follow-up after HNRT completion. In this study, objective measures of oral health were performed by the dental oncologist using clinical and radiographic assessment. The oral health objective measures included decayed-missing-filled teeth (DMFT) scores, periodontal disease condition, oral hygiene status and presence of any dental prosthesis at the post-HNRT visit. The dental charts were retrospectively reviewed for DMFT scores. This is a standard dental evaluation tool that adds the number of teeth that are decayed, missing, and filled. The scores range from 0 to 32 with 32 meaning all the teeth are decayed, missing, or filled. Periodontal disease condition was recorded as “present” if there was localized or generalized mobility of the teeth and/or localized or generalized bone loss seen on a panoramic radiograph. Oral hygiene condition was graded as “good” (minimal or no interproximal plaque accumulation and overall healthy gingiva in the absence of gingival inflammation) or “suboptimal” (moderate or heavy interproximal plaque accumulation with either localized or generalized gingival inflammation). Presence of any form of dental prosthesis (such as partial removable denture, fixed partial denture or implant retained prosthesis) was recorded at the time of post-HNRT visit.

Radiation dose to the salivary glands

CT based delineation of OARs for HNRT including the major and minor salivary glands is a standard practice for the radiation oncologists at our institution.²⁷ As standard of care, parotid sparing technique was employed during radiation treatment planning of head neck cancer patients.²⁸ For each patient, prescribed mean radiation dose to ipsilateral parotid gland (iPG), contralateral parotid gland (cPG), ipsilateral submandibular gland (iSMG), contralateral submandibular gland (cSMG), and minor salivary glands in the oral cavity (OC) were retrieved from the IMRT plan using the Eclipse treatment planning system (Varian Medical Systems) (Figure 1).

Figure 1.

CT-based delineation of major and minor salivary glands in a patient treated with IMRT with a total dose of 7000cGy for squamous cell carcinoma of the left oropharynx.

PRO Questionnaire

Head and neck cancer patients routinely complete questionnaires as standard practice in Head and Neck Service at MSKCC. The FACE-Q Head and Neck Cancer is a new generation validated PRO measure developed at our institution.^29,30 This PRO module includes 14 independently functioning scales measuring 4 domains: health related QOL, function, appearance, and experience of care. The following 3 scales were used in this study: (i) Eating and Drinking (function domain), an 8-item scale that measures the ability to eat and drink, with items that ask about chewing, drinking, tasting and the effort required to perform these activities); (ii) Eating Distress (health-related QOL domain), a 7-item scale measures how eating problems affect the patients by asking them how often they do not enjoy eating meals, feel self-conscious when eating or avoid eating in public; (iii) Salivation (function domain), an 8-item scale that measures problems with saliva, with items that ask about having a dry mouth, need to drink liquids and trouble eating. For this PRO, independent scale item responses were added to provide a total score which was then converted into a Rasch score from 0 to 100, with higher scores indicating a better outcome and lack of dysfunction or discomfort.

Statistical Analysis:

We were interested in evaluating the association between the 3 subjective measures and 9 variables of interest, corresponding to 27 pairwise associations. We visualized the relationships using raincloud plots depicting the density distribution and dot plot for categorical variables, and scatterplot for the continuous variables. The association between objective measures of oral health and PROs were evaluated based on linear regression. For each of the linear regressions with a continuous predictor, we tested for nonlinearity using restricted cubic splines with knots at the tertiles. As we found no evidence of a nonlinear association, the continuous variables were included in the linear regression model without nonlinear terms in all cases. We then reported the amount of variation in PRO scores which could be explained by the dentition measure. Confidence intervals (CI) for the coefficients of determination (R²) were calculated based on a noncentral F-distribution. Descriptions of the data included median (interquartile range, “IQR”, which presents the 25^th percentile and 75^th percentile values) and frequency (percentage) for continuous and categorical variables, respectively. All statistical analyses were performed using R, version 4.1.0 software.

RESULTS

Our cohort matches the typical patient population of 112 head and neck cancer patients with a histology of squamous cell carcinoma and treated with a median radiation dose of 68Gy (IQR, 60–70), predominantly male, median age of 61. The most common primary tumor site was oropharynx (66/112; 59%) followed by oral cavity (17/112; 15%). There was HPV positivity for 62/64 (97%). Patient characteristics are summarized in Table 1.

Table 1.

Patient characteristics. Results are presented as median (interquartile range) and frequency (%).

	N	N=112
Age	112	61 (53, 68)
Sex	112
Male		86 (77%)
Female		26 (23%)
Race	112
White		89 (79%)
Black		6 (5.4%)
Asian		12 (11%)
Other		4 (3.6%)
Unknown		1 (0.9%)
Ethnicity	112
Hispanic		9 (8.0%)
Non-Hispanic		101 (90%)
Unknown		2 (1.8%)
Primary Tumor Site	112
Oropharynx		66 (59%)
Oral cavity		17 (15%)
Larynx		11 (9.8%)
Nasopharynx / nasal cavity		8 (7.1%)
Hypopharynx		3 (2.7%)
Parotid glands		4 (3.6.%)
Others (maxilla, eyelid, thyroid)		3 (2.6%)
Smoking Status	112
Former		61 (54%)
Never		51 (46%)
HPV positive status in oropharynx*	64	62 (97%)
Prior Surgery	112	40 (36%)
Mandibulectomy	112
No		108 (96%)
Marginal		3 (2.7%)
Segmental		1 (0.9%)
Neck Dissection	112
Unilateral		29 (26%)
Bilateral		3 (2.7%)
None		80 (71%)
Level 1(B) node dissection	112	22 (20%)
Concurrent chemotherapy	112
Yes		80 (71%)
No		32 (29%)
Radiation dose to the primary tumor site (Gy)
Oropharynx	66	70 (60, 70)
Oral cavity	17	60 (60, 66)
Remaining head neck subsites	29	64 (60, 70)

^*HPV Status only presented among patients with primary site of disease of oropharynx.

Abbreviations: N: number; HPV: Human Papilloma Virus; Gy: Gray.

The oral health objective measures at post HNRT visit, PRO scores and mean radiation dose to salivary glands are summarized in Table 2. As demonstrated in our former study, objective measures of salivary flow rate and maximal interincisal opening were performed at a median of 9 months (IQR, 4–26 months) following HNRT.²⁵ PRO measures were completed at a median of 11 months (IQR, 4–27 months) following HNRT and at a median of 42 days (IQR, 8–93 days) from the post HNRT objective measurements.

Table 2.

Summary of objective measures of oral health, salivary gland irradiation and PRO scores following HNRT. Results are presented as median (quartiles) and frequency (%).

	N	N = 112
Oral health objective measures
Decayed-missing-filled teeth (DMFT)	112	18 (14, 23)
Periodontal disease	112	20 (18%)
Any dental prosthesis	112	50 (45%)
Good oral hygiene	112	73 (65%)
Post HNRT objective measurements of trismus and xerostomia
Unstimulated salivary flow rate (grams/5mins)	112	1.7 (1, 2.5)
Stimulated salivary flow rate (grams/5mins)	112	1 (0.6, 1.5)
Maximum interincisal opening (in millimeters)	112	43 (37, 48)
Prescribed HNRT dose to the salivary glands (Gy)
Ipsilateral parotid gland (iPG)	110	21.4 (16.3, 25.7)
Contralateral parotid gland (cPG)	109	11.7 (7.9, 16.9)
Ipsilateral submandibular gland (iSMG)	86	54.2 (42.4, 65.4)
Contralateral submandibular gland (cSMG)	102	36.7 (24.6, 52.1)
Minor salivary glands in the oral cavity (OC)	106	37.3 (30.7, 43.7)
PRO scores
Eating & Drinking Score	112	100 (62, 100)
Eating Distress Score	112	100 (62, 100)
Salivation Score	112	57 (46, 100)

Abbreviations: PRO: Patient Reported Outcome; HNRT: Head and Neck Radiation Therapy; DMFT: Decayed/missing/filled teeth; Gy: Gray.

We found a similar distribution in eating-drinking-salivation related PRO scores based on presence or absence of dental prosthesis, periodontal disease and good oral hygiene in patients following HNRT (Figure 2), and formal testing likewise indicated no evidence of a difference in scores based on groups. This is further depicted by the flat regression lines showing the association between PRO scores and DMFT (Figure 3). The R² was particularly low for the Salivation score (0.1%; 95% CI 0, 4.7). Table 3 presents the coefficient of determination (R²) and confidence interval, where we do not have good evidence that the explained variation in patient-reported outcomes by the respective oral health measure is not zero.

Figure 2.

Rain-cloud plots showing the distribution of PRO measures based on objective measures of oral health such as periodontal disease, oral hygiene and prosthesis use.

Figure 3.

Scatter plots (with linear regression line and corresponding 95% CI) showing the association between PRO measures and DMFT.

Table 3.

Coefficient of determination (R²) and 95% confidence intervals from linear regression with patient-reported outcome score as the outcome and oral health objective measures and salivary gland irradiation as the predictor.

	Eating and Drinking	Eating Distress	Salivation
Dental Prosthesis	0.1 (0–3.9)	0 (0–2.9)	0 (0–3.0)
Periodontal Disease	0.1 (0–3.6)	0.3 (0–5.4)	1.1 (0–7.8)
Good Oral Hygiene	2.1 (0–9.8)	1.6 (0–8.8)	2.1 (0–9.9)
Decayed-missing-filled teeth	0 (0–2.5)	0.8 (0–6.9)	0.1 (0–4.1)
HNRT dose to ipsilateral parotid gland (iPG)	3.7 (0–12.7)	2.0 (0–9.8)	1 (0–7.6)
HNRT dose to contralateral parotid gland (cPG)	4.7 (0–14.3)	5.8 (0.3–16)	3.7 (0–12.8)
RT dose to ipsilateral submandibular gland (iSMG)	0.1 (0–5.4)	0.5 (0–7.2)	0 (0–4.1)
RT dose to contralateral submandibular gland (cSMG)	3.6 (0–13.1)	6.3 (0.3–17)	1.3 (0–8.7)
RT dose to minor salivary glands in the oral cavity (OC)	0.3 (0–5.6)	0.3 (0–5.7)	0.3 (0–5.5)

Eating and Drinking scores were inversely related to iPG and cPG. Eating Distress scores was inversely associated with cPG and cSMG. Decreased Salivation scores were associated with increased cPG only (Figure 4). We did not see any association between 3 PRO scores and OC, which is further supported by the low R² (all 0.3, with the highest upper bound of the CI at 5.7%; Table 3).

Figure 4.

Scatter plots (with linear regression line and corresponding 95% CI) showing the association between PRO measures and prescribed radiation dose to the salivary glands.

DISCUSSION

This cross-sectional study was initiated as a part of a prior study from our own institution to further determine the correlation of PRO with oral health measures and salivary gland irradiation. We found that neither objective measures of oral health measures nor estimates of radiation dose to salivary gland explain variations in patient-reported eating and drinking related function following HNRT.

Association between oral health objective measures and PRO

There was no significant association found between the DMFT and PRO scores. In our study population, good oral hygiene was seen in the two-third patients and only less than one-fifth patients had evidence of either localized or generalized periodontal disease or radiographic bone loss. None of the PRO scales were affected with the presence or absence of good oral hygiene in patients following HNRT (highest R² 2.1). Nearly 50% patients who were partially dentate and without any prosthesis had similar PRO scores as those with some form of prosthesis for their missing teeth (highest R² 0.1). Our findings were in contrast with an earlier study which showed that dentition status had a persistent impact on QOL in the long-term head and neck cancer survivors.¹⁹ This could be explained by the fact that pre-HNRT dental evaluation is a standard of care at our institution. In this visit, patients are informed about the importance of maintaining good oral hygiene and the potential risks associated with HNRT that could arise due to poor oral hygiene. Fluoride therapy (neutral sodium fluoride 1.1% with 5,000 parts per million in the form of a dentifrice toothpaste) is also recommended to patients for preservation of remaining dentition and to prevent demineralization of teeth following HNRT.²⁶ Studies have shown that the involvement of dental professionals in the multidisciplinary team of head and neck cancer treatment have a positive impact on patients’ oral health status.³¹ Implementation of a comprehensive dental care program for patients entailing pre HNRT stabilization of oral disease, daily fluoride exposure, regular dietary-counselling, regular oral hygiene advice, and close post HNRT follow-up, has shown significantly decreased levels of caries 3 years post HNRT.²⁶

Association between salivary gland irradiation and PRO

In recent years, efforts have been made to reduce the severity of xerostomia in patients treated with HNRT using parotid sparing technique for radiation treatment planning.^32,33 A study from our own institution demonstrated that limiting iPG and cPG to <25Gy will likely decrease the rate of xerostomia.⁴ In addition, limiting cSMG below 39Gy has also been found beneficial in decreasing patient scored xerostomia with longer follow up.³⁴ Studies have demonstrated that Dmean <40Gy to the OC is associated with favorable patient-reported and observer-rated xerostomia.³⁵ In our study, mean prescribed RT dose to the iPG, cPG, iSMG, cSMG and OC was found to be 21Gy, 12Gy, 54Gy, 37Gy and 37Gy, respectively (Table 2). We identified weak association between PRO and salivary gland irradiation. Higher Eating and Drinking scores were associated with lower iPG and cPG. Higher Eating Distress scores was associated with lower cPG and cSMG. High Salivation scores were associated with increased cPG only (Figure 4). We did not see any association between 3 PRO scores and OC. Our findings were consistent with a recent study demonstrating weak association between PROs and OARs including salivary glands in 169 head and neck cancer patients who received definitive or post operative radiation therapy.²³

Our finding of a weak association between objective measures of oral health and PRO scores is also consistent with study by Long et al.²⁵, which showed that subjective scores improve over time. Patient adaptation might explain why objective measures of oral health do not predict PROs in head and neck cancer patients. Patients modify behavior and psychologic expectations to facilitate eating and to maintain social interaction. They also modify or avoid foods that could be challenging yet report enjoyment with eating. Vermaire et al. similarly found a weak association between objective test outcomes of mastication, swallowing and salivary flow in comparison with PRO scores.¹¹ Studies have identified psychological, functional, and social dimensions of food that impact eating experience in long term survivors of head and neck cancers.³⁶ This observation has led to the necessity for development and evaluations of interventions which address the psychological or social aspects of eating and drinking beyond improving swallowing and jaw mobility in patients following HNRT.³⁷

There are limitations to our study. Firstly, our study was limited by its cross-sectional study design. This means that we are unable to examine relative changes over time or to investigate the hypothesis of whether patient adaptation explains the poor predictiveness of objective function. Secondly, our cohort consisted of a heterogenous group of head and neck cancer patients and radiation dose to the oral structures resulting in varying degree of severity of oral toxicities. Lastly, objective measures of oral hygiene and periodontal disease were recorded dichotomously without using standard tools to measure and quantify the severity of periodontitis or oral hygiene index. Future studies should include prospective study with larger sample sizes using standard periodontal or oral hygiene indices to allow evaluation and better correlation in different subgroups of head and neck cancer patients.

CONCLUSION

We found that subjective experience of eating, drinking and salivation related problems in patients following HNRT are not well defined by oral health status and salivary gland irradiation. PRO should be integrated along with objective measures in assessing symptom burden in patients following HNRT.

Availability of data and material (data transparency)

The authors affirm that this manuscript is an honest, accurate and transparent account of the study being reported; that no important aspects of the study have been omitted; and that any discrepancies from the study as planned have been explained.