Greater proportions of patients who performed swallowing exercises and/or maintained PO throughout treatment ate a regular diet at the conclusion of radiotherapy or chemoradiotherapy (P = .01). NPO indicates no oral intake; PO, oral intake. Symbols represent means, and error bars represent standard errors. Significant differences between individual subgroups are shown by P values within the figure.
Among the 313 patients who received a gastrostomy tube, exercise adherence and maintenance of some PO at the end of treatment was associated with significantly shorter duration of gastrostomy dependence (P = .03). NPO indicates no oral intake; PO, oral intake.
Hutcheson KA, Bhayani MK, Beadle BM, Gold KA, Shinn EH, Lai SY, Lewin J. Eat and Exercise During Radiotherapy or Chemoradiotherapy for Pharyngeal CancersUse It or Lose It. JAMA Otolaryngol Head Neck Surg. 2013;139(11):1127-1134. doi:10.1001/jamaoto.2013.4715
Data support proactive swallowing therapy during radiotherapy (RT) or chemoradiotherapy (CRT) for pharyngeal cancers. The benefits of adherence to a regimen of swallowing exercises and maintaining oral intake throughout treatment are reported, but independent effects are unclear.
To evaluate the independent effects of maintaining oral intake throughout radiotherapy and adherence to preventive swallowing exercise.
Retrospective observational study.
The University of Texas MD Anderson Cancer Center, Houston.
The study included 497 patients treated with definitive RT or CRT for pharyngeal cancer (458 oropharynx, 39 hypopharynx) between 2002 and 2008.
Main Outcomes and Measures
Swallowing-related end points were final diet after RT or CRT and duration of gastrostomy dependence. Primary independent variables included oral intake status at the end of RT or CRT (no oral intake, partial oral intake, or full oral intake) and adherence to a swallowing exercise regimen. Multiple linear regression and ordered logistic regression models were analyzed.
At the conclusion of RT or CRT, 131 patients (26%) had no oral intake and 74% maintained oral intake (167 partial [34%], 199 full [40%]). Fifty-eight percent (286 of 497) reported adherence to swallowing exercises. Maintenance of oral intake during RT or CRT and swallowing exercise adherence were independently associated with better long-term diet after RT or CRT (P = .045 and P < .001, respectively) and shorter duration of gastrostomy dependence (P < .001 and P = .007, respectively) in models adjusted for tumor and treatment burden.
Conclusions and Relevance
The data indicate independent, positive associations of maintenance of oral intake throughout RT or CRT and swallowing exercise adherence with long-term swallowing outcomes. Patients who either eat or exercise fare better than those who do neither. Patients who both eat and exercise have the highest rate of return to a regular diet and shortest duration of gastrostomy dependence.
The incidence of pharyngeal cancer is rising, with 13 930 new cases projected in the United States in 2013.1 A majority of patients with locally advanced pharyngeal cancer are treated with curative radiotherapy (RT) or chemoradiotherapy (CRT), the primary goals of which are locoregional control and functional organ preservation. Dysphagia is a common effect of nonsurgical organ preservation, with an estimated prevalence of 39% to 64% after RT or CRT.2,3 Swallowing function can be adversely affected by radiation-associated edema, fibrosis, and neuropathy. These toxic effects on normal tissue ultimately impair the range of motion of critical laryngeal and pharyngeal musculature, collectively referred to as dysphagia-aspiration–related structures,4 disrupting pharyngeal bolus transit and airway protection. Dysphagia significantly increases the risk of health problems after treatment.5 In severe cases of radiation-associated dysphagia, dietary restrictions and malnutrition necessitate lifelong gastrostomy dependence and chronic aspiration poses risk for potentially life-threatening pneumonia. Even in the era of conformal RT delivery for pharyngeal cancer, rates of long-term gastrostomy dependence after RT and CRT range from 5% to 12%,6- 8 and 11% of patients develop pneumonia after therapy.9,10
Proactive swallowing therapy, supported by observational studies and randomized trials, has become the standard of care at many institutions.8,10- 16 Acute toxic effects of RT and CRT including mucositis, salivary dysfunction, and dysgeusia make eating difficult. As a consequence, many patients require gastrostomy tube placement and dietary restrictions, such as avoidance of solid foods, during the course of RT or CRT. Limitations in oral intake (PO) lessen the normal resistive load on the swallowing mechanism and prompt disuse atrophy. Disuse encourages adverse remodeling of aerodigestive tract muscles that likely exacerbates the effects of radiation-associated edema and fibrosis.17,18 Thus, the central premise of proactive swallowing therapy is “use it or lose it” to mitigate the muscular wasting and remodeling that occurs after even brief intervals of disuse. Proactive swallowing therapy encourages patients to continue use of the swallowing musculature during treatment by (1) avoiding intervals of no oral intake (NPO) and (2) adhering to regimens of swallowing exercise. The benefits of these distinct swallowing goals (eat and exercise) are reported in separate studies,8,10- 16 but independent effects are unclear. The purpose of this study was to examine the independent effects of maintenance of PO (eat) and swallowing exercise adherence (exercise) during RT or CRT for pharyngeal cancers.
We conducted a retrospective study of patients treated with definitive RT or CRT for pharyngeal cancer. Eligible cases were selected from 659 consecutive patients with oropharyngeal and hypopharyngeal primary tumors of any histologic characteristics treated with definitive RT (with or without chemotherapy) at the University of Texas MD Anderson Cancer Center between January 2002 and December 2008. One hundred sixty-two patients were excluded from analysis in accordance with the following criteria: receipt of palliative RT, postoperative RT, or dose less than 66 Gy, or incomplete tumor response. Thus, 497 patients were included in the analysis. The study was approved by the institutional review board, and a waiver of informed consent was obtained.
Proactive swallowing therapy advocates maximal use of the swallowing musculature throughout RT and CRT. At our institution, patients receiving definitive RT or CRT are referred to speech pathologists as standard of care prior to treatment for baseline evaluation and initiation of proactive swallowing therapy. Two swallowing goals are outlined for each patient, guided by the findings of swallowing examinations: eat and exercise. Speech pathologists prescribe a standard swallowing exercise regimen targeting hyolaryngeal excursion, airway protection, and tongue base retraction. Specific exercises prescribed during the period reviewed included a modified Shaker exercise, jaw stretch, supraglottic, Valsalva maneuver, falsetto, lingual protrusion and retraction, yawn, gargle, Masako maneuver, and effortful swallows.19 As a component of routine follow-up during and after RT or CRT, patients are asked to demonstrate competency with swallowing exercises to the speech pathologist and report their adherence to a daily exercise regimen. These details are recorded in the medical record by the speech pathologist. Speech pathologists also prescribe individualized dietary modifications or swallowing compensations when needed to facilitate safe PO (ie, minimize aspiration) throughout RT and CRT. Gastrostomy tube practices in this cohort and rates of long-term dependence have been thoroughly described and examined in previous publications.8,14 Prophylactic gastrostomy tubes are not placed routinely at our institution; rather, gastrostomy tubes are placed when clinically indicated on the basis of the patient’s nutritional status and ability to safely maintain an oral diet.
The primary independent variables were (1) PO status at the end of RT or CRT (eat) and (2) self-reported swallowing exercise adherence (exercise). The PO status was defined as NPO (nothing by mouth, fully gastrostomy tube dependent), partial PO (tube feeding supplemented by consistent daily PO), or full PO (100% PO, regardless of dietary level). The patient’s self-reported swallow exercise adherence was taken by abstraction of the medical record. Patients who reported any (partial [<4 times/day] or full [≥4 times/day], per institutional protocols) exercise adherence were coded “adherent,” and those who reported no swallowing exercise or did not keep their speech pathology appointment for exercise training (ie, those who never saw the speech pathologist) were coded as “nonadherent.” To explore interactions of eat and exercise, patients were further stratified into 6 subgroups based on their swallowing status during RT or CRT as given in Table 1. Specific between-group comparisons were explored as follows: effect of exercise (vs no exercise) with eat held constant (NPO, partial PO, full PO), effect of partial PO (vs NPO) with exercise held constant (no exercise, exercise), and effect of full PO (vs partial PO) with exercise held constant (no exercise, exercise). Comparisons were also explored between those patients who maintained partial PO (gastrostomy tube + PO) and those who had no gastrostomy tube but restricted PO diets (liquid or pureed food only).
Two swallowing-related end points were examined: last diet level after the conclusion of RT or CRT and length of gastrostomy dependence. Diet level was defined by means of abstraction of the medical record at 6 to 12 and 18 to 24 months follow-up as NPO, tube + PO, liquid or pureed, soft, or regular. A regular diet was defined by no restriction of PO and no special preparation of foods such as blending or chopping of solids. The latest diet rating was coded for analysis. Duration of gastrostomy dependence was calculated from date of insertion to date of removal.
Demographic and treatment data were extracted from the electronic medical record. Data points included demographic characteristics, tumor site, tumor staging according to TNM classification, and treatment history. The primary treatment modality was reviewed, including method of RT (conventional 3-dimensional conformal fields or intensity-modulated radiotherapy [IMRT]), RT fractionation schedule (standard or accelerated), total RT dose (in grays), number of fractions, timing of CRT (none, induction, or concurrent), and agent. Detailed descriptions of this cohort have been previously published, along with predictors of gastrostomy placement and duration of gastrostomy dependence.8,14
Descriptive statistics were calculated. Bivariate associations were analyzed using χ2 tests for categorical outcomes (diet level) and t tests for continuous outcomes (length of gastrostomy dependence). Multivariable ordered logistic regression analyses were conducted to evaluate independent effects of eat and exercise on diet levels after RT or CRT, as coded in the Variables subsection. Multiple linear regression models were analyzed to assess independent effects of eat and exercise on length of gastrostomy dependence. Multivariable models were adjusted for clinically significant confounders including T classification, N classification, tumor subsite, therapeutic combination, age, and baseline (before RT or CRT) diet using stepwise backward elimination. Final multivariable models retained confounders that were independently associated (P < .05) with eat and exercise. A P value of <.05 was considered statistically significant. Statistical analyses were performed using STATA data analysis software, version 10.0 (StataCorp).
Four hundred ninety-seven patients met the eligibility criteria for this analysis. The median (range) age was 56 (38-80) years, and 87% were male. Most patients had oropharyngeal primary tumors; most had node-positive disease (81% ≥ N2). T classification was fairly evenly distributed. Most patients (452 of 497 [91%]) were treated with IMRT, and 121 (24%) were treated on accelerated (concomitant boost) fractionation RT schedules. Seventy-seven percent of patients (381 of 497) received systemic therapy, most often delivered concurrently (234 of 497 [47%]) with RT. Patient, disease, and treatment characteristics are provided in Table 2.
Seventy-four percent (366 of 497) of patients maintained PO (full or partial) throughout RT or CRT. At the conclusion of RT or CRT, 131 patients (26%) were NPO status, 167 (34%) were partial PO status, and 199 (40%) were full PO status. Of the 199 patients who maintained full PO status at the end of treatment, 87 ate only pureed or liquid diets and 112 ate masticated foods. Three-hundred eighty patients (76%) were seen by speech pathologists before or during RT and CRT. Fifty-eight percent (286 of 497) of all patients reported exercise adherence, with 128 patients reporting exercise more than 4 times per day and 158 patients 4 or fewer times per day. Patients were further stratified into 6 swallowing subgroups according to the interaction of eat and exercise as given in Table 1. Of the 2 swallowing goals (eat and exercise), only 13% of patients maintained neither, 64% met some swallowing goals with at least partial PO and/or exercise adherence, and 24% met both goals with full PO throughout treatment and swallowing exercise adherence.
Overall, 402 patients (81%) returned to a regular diet after the conclusion of RT or CRT. Greater proportions of patients who maintained PO throughout RT or CRT and/or performed swallowing exercises maintained a regular diet in long-term survivorship (P = .01; median [IQR] follow-up, 22.2 [17.8-24.2] months). Dose-response effects were suggested on subgroup analyses, as presented in Table 3 and Figure 1. According to adherence to swallowing goals during treatment, the proportion of patients returning to a regular diet after RT or CRT was as follows: 65% of patients who did neither (neither eat nor exercise), 77% to 84% of patients who maintained some swallowing goals (eat and/or exercise), and 92% of patients who met both swallowing goals (eat and exercise). Thus, only 65% of patients who were NPO status during RT or CRT and nonadherent to the exercise regimen returned to a regular diet after treatment, compared with 92% of patients who maintained maximal swallowing activity during treatment (full PO and swallowing exercise adherence).
Specific between-group comparisons revealed statistically significant differences in long-term diet according to the following stratifications: (1) full PO vs partial PO (P = .02) among patients who adhered to exercise and (2) exercise vs no exercise among patients who maintained full PO (P = .04). Among patients who did not exercise, PO status at the end of RT was not significantly associated with long-term diet (P = .10 for partial PO vs NPO and P = .24 for partial PO vs full PO). In addition, patients who maintained a restricted PO diet of pureed food or liquids during RT or CRT without tube feeding were more likely to return to a regular diet (76 of 87 [87%]) than those who received tube feedings supplemented with partial PO (132 of 167 [79%]; P = .19).
In adjusted models, both swallowing goals (eat and exercise) were independently associated with long-term diet levels (P < .05). Patients who maintained full PO during RT or CRT were 2.0 times more likely to eat a regular diet in long-term follow-up (adjusted odds ratio, 2.0 [95% CI, 1.0-4.6]) compared with those who were NPO status during treatment. Swallowing exercise adherence was associated with 4.0 times the odds of returning to a regular diet in adjusted models (adjusted odds ratio, 4.0 [95% CI, 1.9-6.4]). Results of multivariable regression analyses are described in Table 4.
Three hundred thirteen patients (63%) received a gastrostomy tube, and the median (IQR) duration of gastrostomy dependence was 5.0 (2.8-8.8) months. Swallowing during RT or CRT (eat and exercise) was associated with significantly shorter duration of gastrostomy dependence (P = .03), and dose-response effects were suggested in subgroup analyses (Table 3 and Figure 2). Among the 313 patients who required gastrostomy placement, the median duration of gastrostomy dependence was 222 days in patients who did neither (neither eat nor exercise), 151 to 157 days in patients who maintained some swallowing activity (eat and/or exercise), and 111 days in patients who did both (eat and exercise). Thirty-nine percent of patients who were NPO status and nonadherent to swallowing exercise were gastrostomy dependent for at least 1 year, compared with only 6% of those who maintained some PO throughout treatment and performed swallowing exercises. Kaplan-Meier estimates are illustrated in Figure 2.
In adjusted models, both swallowing goals (eat and exercise) were independently associated with duration of gastrostomy dependence (P < .05). The mean duration of gastrostomy dependence was 95 days shorter among patients who maintained PO during treatment and 109 days shorter among those who reported adherence to exercise, on the basis of regression coefficients in adjusted multivariable models. Results of multivariable regression analysis are provided in Table 4. Univariable analyses for adjustment variables associated with long-term diet and duration of gastrostomy dependence are given in Table 5.
Swallowing is the top functional priority rated by patients with head and neck cancers before and after treatment20 and is a driver of quality of life in survivors.21 Proactive swallowing therapy is prescribed to provide maximal use of the swallowing mechanism during treatment. Two goals can be given to patients under a “use it or lose it” paradigm: eat and exercise. The benefits of maintaining PO during RT and CRT (eat) and swallowing exercise adherence (exercise) have been demonstrated in separate studies,8,10- 15 but to our knowledge, the independent effects of these efforts have not been reported to date. In this study, we found that both swallowing goals—eat and exercise—were independently associated with significantly better swallowing-related end points. In addition, subgroup analyses suggested dose-dependent benefits. That is, these data imply that patients who either eat or exercise fare better than those who do neither, and swallowing end points are best among those who both eat and exercise.
Swallowing exercise during RT or CRT is associated with favorable functional outcomes. Four institutions have published outcomes of proactive swallowing exercises to date, finding superior swallowing-related quality of life,10,13 better tongue base and epiglottic movement,11 lower rates of percutaneous endoscopic gastrostomy tube placement,8,14 shorter duration of gastrostomy dependence,8,14 and higher diet levels after therapy15 among patients who receive early swallowing exercise regimens. Among these studies, 2 randomized controlled trials reported favorable muscle composition on acute post-CRT magnetic resonance imaging (MRI) and significantly better clinician-rated swallowing performance (Functional Oral Intake Scale and Performance Status Head and Neck Scale) in patients randomized to exercise therapy during treatment.12,15 In a sham-controlled trial, patients randomized to proactive swallowing exercise were 6 times more likely to have favorable swallowing outcome on the basis of a composite end point of weight, diet, and standardized swallow rating.12 Patients randomized to exercise also had a 36% absolute risk reduction for loss of functional swallowing ability during CRT.12 Although these results are encouraging, authors predominantly report outcomes in early recovery within 6 months of RT or CRT, and few studies have examined outcomes more than 1 year after therapy.13 In addition, to our knowledge, prior studies have not jointly examined both domains of proactive swallowing therapy (ie, eat and exercise) as independent constructs of swallowing behavior during treatment. In the present study, self-reported exercise adherence was associated on average with 3.5 months shorter duration of gastrostomy dependence and 4.0 times increased odds of eating a regular diet in long-term survivorship, after adjustment for PO status during RT or CRT. In addition, we found a 58% rate of self-reported exercise adherence, similar to that reported by Carnaby-Mann et al,12 who found that 68% of patients adhered to home exercise regimens. These data suggest that exercise adherence is an achievable goal for many patients despite the competing priorities and toxic effects encountered during treatment.
Maintenance of PO, or avoidance of NPO intervals, is recommended by many practitioners largely on the basis of empirical consensus. Effects of PO during RT or CRT have been studied less extensively than swallowing exercise. To date, 2 small retrospective observational studies have reported better swallowing outcomes in patients who maintain PO throughout treatment. Gillespie et al22 first reported that NPO intervals longer than 2 weeks during CRT significantly predicted long-term posttreatment swallowing-related quality of life scores (≥12 months) based on responses from the MD Anderson Dysphagia Inventory. Others have reported significantly better mean diet scores at 3-month, 6-month, and 12-month intervals among patients who maintain partial or full PO throughout treatment compared with patients who are NPO status.23 In the present study, we found that 74% of patients maintained partial or full PO status at the conclusion of RT or CRT, similar to other investigators who reported maintenance of PO by 76% of patients with head and neck cancers treated with definitive or postoperative RT or CRT.23 These data suggest that it is feasible for most patients to maintain at least some PO during treatment despite the acute toxic effects of RT and CRT. Multidisciplinary supportive care is critical to help patients avoid NPO intervals during treatment.
Proactive swallowing therapy seeks to counteract the loss of the normal resistive load that occurs when the acute effects of RT or CRT cause patients to stop eating solid foods. Skeletal muscles can begin to show evidence of disuse atrophy just hours after immobilization. Myoarchitecture changes rapidly with disuse, showing a decrease in muscle mass, infiltration of adipose tissue, and redistribution of fibers within the muscle. Over time, disuse atrophy manifests as a reduction in muscle strength, increased fatigability, and aberrant motor control. The extent of injury is dependent on the severity of restriction. Thus, greater use of the swallowing mechanism during RT should result in more normal muscle composition and function after treatment.17,18 This premise is supported by prior work in which patients randomized to proactive swallowing exercise regimens had significantly less deterioration in muscle mass and composition per T2-weighted MRI analysis of the genioglossus, hyoglossus, and mylohyoid after CRT.12 In addition, results of our subgroup analyses suggested dose-dependent associations; subgroups of patients with maximal use of the swallowing mechanism during RT or CRT had the best swallowing-related outcomes.
This study benefits from a large sample size of patients with similar clinical characteristics. We have demonstrated independent, positive associations for 2 swallowing activities during RT and CRT—eat and exercise—in approximately 500 cases of patients who had complete response to definitive RT or CRT for pharyngeal cancers. Both early (duration of gastrostomy dependence) and long-term (diet) swallowing-related end points were significantly better as swallowing behaviors during RT or CRT increased between subgroups of patients. On the basis of these data, future investigations are encouraged to consider the patient’s swallowing function during RT or CRT based on PO status and exercise adherence as possible important covariables when reporting post-RT swallowing outcomes.
Eat and exercise, 2 goals commonly prescribed by speech pathologists in proactive swallowing therapy, were independently associated with favorable outcomes after adjustment for confounding markers of tumor and treatment burden. It is not likely that the referral alone to a speech pathologist confers benefit, but rather it is the provision of and adherence to evidence-based swallowing therapy that are key to optimizing functional outcomes. In this study, swallowing outcomes did not significantly differ between the 380 patients who did and the 117 patients who did not follow up with appointments with the speech pathologist. Referral bias inherent to observational studies and patient characteristics that influence attendance at appointments are important variables that likely influence these outcomes. These variables were not available in the present study, and future evaluations are needed to elucidate the influence of consultation patterns and patient attendance on swallowing recovery. More important, adherence to evidence-based swallowing therapy prescribed by the speech pathologist improves outcomes and was significantly associated with better swallowing end points in this study. Similar to the case in any discipline, it is not the appointment that makes the difference but the adherence to appropriate selected interventions because the patient who sees a specialist but fails to take the prescribed medication will benefit less than one who adheres.
The observational, retrospective methods used in this study inherently limit our ability to explore detailed dose-response relationships regarding the amount of exercise or minimal PO necessary to confer benefit. These are important considerations for future studies, and ongoing dose-response trials may begin to answer some of these questions.24 Other important considerations are baseline performance status and acute toxic effects (eg, mucositis, odynophagia) that may affect patients’ ability to eat and exercise during treatment. These factors were not available in this retrospective data set but should be examined in prospective studies to ensure that the effects observed in this analysis are not merely a reflection of poorer performance status or greater acute toxic effects that preclude swallowing activity during RT or CRT. In addition, adherence data were not prospectively measured and relied on abstraction from the medical record. If adherence details were not completely documented in the medical record, a loss of statistical power may have resulted in overly conservative estimates of associations in our analyses. As such, it is possible that future investigations that prospectively ascertain functional status during RT or CRT might show larger differences in outcomes associated with eat and exercise. Finally, additional study is also needed to examine the long-term physiologic effects of these swallowing goals during RT and CRT using gold-standard modified barium swallow studies, and the influence of these swallowing practices on outcomes in other sites of head and neck cancer, such as the larynx.
Long-term swallowing outcomes were best in patients who both maintained full PO throughout RT or CRT and reported adherence to swallowing exercises and uniformly worst in those who were NPO status at the end of treatment and nonadherent to the exercise regimen. Multivariable models show independent effects of 2 swallowing goals—eat and exercise—and subgroup analyses suggest additive effects of eat and exercise. Proactive swallowing therapy that facilitates safe PO throughout RT and CRT and swallowing exercise adherence should be considered an essential component of modern, multidisciplinary head and neck care. Our findings, in concert with those of prior rigorous trials, offer support for early referral to the speech pathologist to begin proactive swallowing therapy before definitive RT or CRT.
Corresponding Author: Katherine A. Hutcheson, PhD, Department of Head and Neck Surgery, University of Texas MD Anderson Cancer Center, 1515 Holcombe Blvd, Unit 1445, Houston, TX 77030 (firstname.lastname@example.org).
Submitted for Publication: February 22, 2013; final revision received July 3, 2013; accepted July 28, 2013.
Published Online: September 19, 2013. doi:10.1001/jamaoto.2013.4715.
Author Contributions: Drs Lai and Lewin served as co–senior authors, each with equal contribution to the manuscript. Drs Hutcheson and Lai had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: Lai.
Acquisition of data: Bhayani, Gold.
Analysis and interpretation of data: Hutcheson, Bhayani, Beadle, Shinn, Lai, Lewin.
Drafting of the manuscript: Hutcheson.
Critical revision of the manuscript for important intellectual content: All authors.
Statistical analysis: Hutcheson.
Administrative, technical, or material support: Bhayani, Gold, Shinn.
Study supervision: Beadle, Lai, Lewin.
Conflict of Interest Disclosures: None reported.
Funding/Support: Support was provided by the UT Health Innovation for Cancer Prevention Research Fellowship, TheUniversity of Texas School of Public Health–Cancer Prevention, and Research Institute of Texas (CPRIT) grant RP101503 (to Dr Hutcheson).
Role of the Sponsor: The funding sources had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; and preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.
Previous Presentation: This study was presented at the American Head and Neck Society 2013 Annual Meeting; April 10, 2013; Orlando, Florida.
Additional Contributions: Janet Hampton, University of Texas MD Anderson Cancer Center, helped prepare the manuscript, and Clare P. Alvarez, MS, CCC-SLP, Denise A. Barringer, MS, CCC-SLP, and Asher Lisec, BA, University of Texas MD Anderson Cancer Center, contributed to data collection.