Long-term patient reported outcomes following radiation therapy for oropharyngeal cancer: cross-sectional assessment of a prospective symptom survey in patients ≥65 years old

Radiation therapy (RT) (+/− concurrent systemic therapy) is a well-accepted treatment for oropharyngeal carcinoma (OPC). However, the optimal treatment for older patients is not well defined [1]. In a large-scale meta-analysis, the survival benefit of adding chemotherapy to local regional therapy for OPC was shown to diminish with increasing age [2]. The lack of observed benefit in older patients could be due to increased treatment-related toxicity, poorer treatment tolerance, or confounding comorbid conditions, all potentially offsetting any survival benefit. Late sequelae of therapy remain concerning for older populations due to the potential of irreversible functional decline due to limited physical reserves or poor compensatory mechanisms [3]. Additionally, symptoms may be experienced synergistically along with age-related functional decline, thus compounding morbidity from treatment [4].

There is now clinical research emphasis on treatment de-escalation and toxicity reduction strategies in OPC, particularly for those with more favorable/human papillomavirus(HPV)-related disease, as the impact of late toxicities associated with current standard therapies in survivors is now recognized [5]. While HPV-related OPC skews younger, most OPCs in older patients are also HPV-related and incidence in this subset of patients is increasing. With highly curable disease, the importance of long-term toxicity reduction in older patients with OPC should be emphasized [6, 7].

Patients with OPC are increasingly being considered for primary surgical management using transoral and robotic surgery or de-escalation strategies with the goal of maintaining cure rates and reducing toxicity compared to current RT-based approaches [8]. At present, few prospectively collected late toxicity assessment series of OPC survivors treated with modern RT techniques (e.g. intensity modulated radiation therapy(IMRT)) exist to serve as benchmarks for estimation of symptom differentials for patients treated with alternate modalities (e.g. surgery) or advanced RT techniques (e.g. intensity modulated proton therapy).

To help define optimal treatment for older patients in the setting of continually evolving and potentially competing treatment modalities, OPC site-, treatment-, and age-specific patient reported outcomes (PROs) are needed to inform patients, clinicians, and investigators. Thus, we analyzed results of a validated multi-symptom PRO instrument in older OPC survivors in order to characterize the post-therapy symptom experience in patients treated at a tertiary academic medical center.

Consequently, the specific aims of the current study were to:

This prospective survivorship assessment study of older patients treated with (chemo)RT for OPC using contemporary standard approaches demonstrates a symptom profile which is generally favorably with identifiable subsets of patients with distinct post-therapy symptom constellations. While no one item or items had moderate-severe intensity recorded by a majority of respondents, items for dry mouth, difficulty swallowing/chewing, and taste received notable pluralities, with 35%, 29% and 29% of survivors reporting moderate-severe levels, respectively (Additional file 1: Table S1). Overall, these highest intensity symptom items reported were consistent with “classic” RT-related normal tissue late toxicities, namely xerostomia and dysphagia.

The majority of respondents were free from moderate-severe xerostomia, attributable to the well-characterized, more conformal, treatment offered by IMRT, yet even so, a detectable minority of 16% of the cohort reported severe ratings for dry mouth [18]. Likewise, approximately 10% of survivors reported severe late problems with dysphagia-related symptoms. The correspondingly lower proportion of patients with PEG tube at time of MDASI-HN suggest that the use of objective clinical endpoints alone do not capture the nuanced toxicity burden of this cohort. Additionally, previous studies have shown that even mild late dysphagia, which was present in our cohort, is strongly correlated with overall health related quality of life (HRQOL) (24). Thus, further efforts to improve xerostomia and swallowing dysfunction are likely to yield impactful gain and should be pursued aggressively [19].

Notably, in this older patient cohort, T-category and chemotherapy (either induction and/or concurrent) did not predict strikingly different toxicity profiles. However, individual item-level differences were associated with concurrent chemotherapy and T-category. Specifically, problem tasting food and voice symptoms were more severe in the subgroup of patients with T3-T4 disease which is in line with locally advanced disease infiltrating a larger area of normal structures and resultant larger RT volumes. Cancer subsites (i.e. base of tongue versus tonsil) did not differentially impact the mean item ratings or their distributions, suggesting that this factor alone does not determine the late toxicities experienced by older OPC patients.

Receipt of CCRT was not found to have an association with PROs in univariate analysis. The survival benefit of CCRT versus RT alone in older patients is not generally supported but recent studies have shown that CCRT should not be denied solely based on age [20]. While further study is necessary, the survival benefit and equivalent late toxicity suggested by our findings points towards expansion of CCRT into more aged populations. In a study using SEER-Medicare data regarding OPC in the elderly, as age increased from 70 to 81, treatment shifted toward surgery alone or no treatment [21]. Given the low symptom burden observed within our cohort, it seems subsets of older patients should be considered for standard therapies, as it appears that comorbidity and age did not have a detectable effect upon reported late symptom frequency nor severity.

Hierarchical cluster analysis was also informative with consistent representation of xerostomia at some level in the far majority of patients. While 64% of the cohort presented with generally no to moderate symptom burden, a dismaying ~33% of the cohort still reported moderate-severe ratings for two subsets of symptoms: one centered around classically RT-related symptoms (e.g. dry mouth, problems with mucus, swallowing/chewing, and taste) at moderate-severe ratings, and another around general symptoms (e.g. fatigue, memory, drowsiness) at more moderate ratings. This visualization technique reinforces the continued presence of these two categories of late toxicities and affirms that they tend to co-occur in symptomatic patients, rather than homogenously among the entire cohort.

One of these possible determinants was age-adjusted comorbidity status pre-RT, which has already been shown to have a negative association with overall survival (OS) [22]. Our analysis showed that cohort comorbidity, at least as measured by the age-adjusted CCI in our specific cohort, did not have an association with PROs. This finding may be due to the healthy, homogenous group of older patients that made up our cohort, evidenced by the median age-adjusted CCI of 3, which corresponds to older patients with few comorbidities, and low comorbidity burden outside of age-adjustment [23]. Our study shows that for these older patients judged fit enough to receive curative therapy, PROs remained favorable to a median follow-up period of nearly 4 years (46 months), suggesting that these patients, broadly, maintain global functionality after completion of therapy, as revealed in the low mean symptom interference reports observed here.

Inherent limitations of this analysis include data collected from a large-scale, tertiary, academic cancer institution, acknowledging that, patient self-selection and the patient profile of our specific patient population may not mirror the general OPC population. The patients in our study had comparatively high performance status, low comorbidity burden, and most received standard IMRT. Additionally, survey timing varied widely, 6–117 months, and while previous work has shown that most patients’ toxicity has stabilized 6 months after therapy, it is possible that the varying follow-up periods may not precisely capture late toxicity in this cohort and misrepresent true late toxicity [24]. However, given the low symptom burden observed overall, it is reasonable to presume that toxicity could only have been better with a narrower and later survey acquisition time range. Treatment strategies and expected toxicities differ between AJCC stages and the inhomogeneity of the cohort in this respect is a limitation for external validity. Due to the unbalanced distribution of AJCC stages within our cohort, T1/2 versus T3/4 was use for comparisons as a proxy for locally confined versus locally advanced disease. Baseline symptom data was unavailable, so it is difficult to ascertain what proportion of the symptoms reported were present pre-therapy and potentially persistent versus those that were in fact an actionable secondary side-effect from RT. The standard caveats from any cross-sectional analysis apply in that longitudinal follow-up or the toxicity deltas over time are lacking. Therefore, we recommend the pursuit of studies examining the specific dosimetric effects on normal structures, incorporating tumor HPV status, as well as, analyzing symptom item scores longitudinally with pretreatment baselines, which are already underway at our institution.

Nonetheless our data represents, a large single-site prospective cross-sectional interrogation of late survivorship in older OPC patients who received definitive radiotherapy. It provides a characterized multi-symptom profile for these older patients treated with contemporary techniques, using a standardized approach [25, 26]. These data provide a reference/benchmark dataset against which approaches leveraging alternate modalities (e.g. advanced surgical techniques), or advanced radiation therapy techniques such as proton therapy, may be compared.

In conclusion, older OPC patients with a median of nearly 4 years from completion of therapy exhibited a broad freedom from global symptoms, with a majority showing no more than mild-to-moderate intensity for experienced symptoms. However, 38% reported at least one severe symptom item. Moderate-severe xerostomia, difficulty tasting food, and dysphagia related symptoms were experienced by a plurality of patients (approximately 1/3 each), but efforts should be made to increase the currently small fraction (9%) of patients who are symptom free survivors.