Study Design
Eligibility criteria for E2399, a phase II function preservation trial, included patients with previously untreated resectable, biopsy proven squamous cell carcinoma or the OP and L, stageT2N+M0 or T3-T4N0-3M0, age ≥ 18,ECOG PS of 0-2, with good organ system function. Patients received induction chemotherapy of paclitaxel175 mg/m2 IV over 3 hours followed by carboplatin AUC 6 IV over 30 to 60 minutes for 2 cycles 21 days apart. At the completion of induction therapy, patients with a tumor response or stable disease went on to receive radiation therapy 70Gy in 35 fractions over seven weeks with concurrent weekly paclitaxel 30mg/m2 IV. The protocol was approved by the Institutional Review Boards at all Cooperative Group sites. Written informed consent was obtained from all participants. Treatment results and biologic correlative data havepreviously been published.6,7
QOL and Functional Assessment
Questionnaires:
Functional Assessment Cancer Therapy-Head Neck (FACT-HN).The FACT-HN consists of 27 items in the FACT-G (version 4) and an 11-item HN module that measures concerns/symptoms associated with head and neck cancer.8,9,10A subscale score is computed by summing across all items, with higher scores reflecting better QOL. The FACT-HN has been used in HNC trials9 and has demonstrated sensitivity to change in disease status. Validity has been supported in numerous studies.10,11Internal consistency of the subscale is adequate.12The Trial Outcome Index (TOI) is a composite score which includes physical, functional, and a HN cancer specific subscale of the FACT. The questionnaire was completed by the patient at baseline, immediately following induction, and at 3, 6, 12, and 24 months post-chemoradiation.
Performance Status Scale Head and Neck Cancer Patients (PSS-HN). This 3-domain scale evaluates normalcy of diet, eating in public, and speech and has been used in numerous studies of head and neck cancer patients.13,14Validity and reliability has been previously documented.15 The PSS-HN was administered by the study staff at the same time points as the FACT-HN.
Voice Handicap Index (VHI).This 30 item patient-reported instrument evaluates function, emotion, and physical areas that may be affected by voice disorders such as those disorders caused by CCR therapy for head and neck cancer.16 A Likert-like response format is used with 7 possible responses reflecting degree of impairment. Items are summed to provide a total score. Internal consistency of the 30-item scale (r=0.95) is adequate. The VHI was completed at time points similar to the FACT-HN.
Voice Desirability Assessment (VDA). Physicians rated ability to communicate, not aesthetic quality of voice, on a linear scale from zero (aphonic) to fifty (perfect voice).17The VDA was also completed at baseline, post-induction, and at 3, 6, 12, and 24 months post treatment.
Swallowing Assessment:Patients were scheduled for swallow assessment by a Speech and Language Pathologist (SLP) and a modified barium swallow (MBS) and at baseline, 3, 12 and 24 months post treatment. The procedure for conducting the MBS was delineated in the protocol in order to ensure constancy. Prior to enrolling patients on the study each site was required to identify a Speech and Language Pathologist (SLP) who would conduct study related swallow assessment and the MBS studies. The designated SLP was required to review the protocol and pass a brief quiz aimed at ensuring familiarity with the guidelines. The SLP who conducted the assessment and the MBS scored the patients swallowing function using the American Speech-Language Hearing Association’s Functional Communication Measure (FCM), a Medicare standard for scoring swallowing function.The FCM is a seven-point rating scale designed to describe the change in an individual's swallowing ability over time. The FCM reports global functioning taking into account subjective issues including: diet level, need for compensatory strategies or cueing, and alternate mode of feeding. In addition, videos of the MBS were scored centrally by the study SLP using the Dysphagia Outcome Severity Scale (DOSS).The DOSS uses only data obtained directly from the MBS and concentrates on the identification of physiologic and anatomic abnormalities.
Statistical Considerations: Analyses were conducted using SPSS (Version 26). Descriptive and graphical methods were used to summarize the data distributions. Frequency distributions were used for nominal and ordinal variables, with the exception of age (mean, SD), all continuous data distributions were heavily skewed. Medians and 25th-75th inter-quartile (IQR) were used to summarize those data. Comparisons of baseline patient demographic and clinical characteristics were conducted using the likelihood Chi-Square statistic (nominal and ordinal data) and Mann-Whitney tests (skewed continuous data). Longitudinal tests of change were conducted using generalized linear modeling that adjusted the standard errors for the lack of independence of measurement. This approach also allowed for missing assessments; however, because missing 12-month data was not assumed to be random (i.e., likely survival related) for inclusion in the longitudinal analysis, a patient must have completed both the baseline and the 12-month assessment. Measures of concordance between the two ordinal FCM and DOSS measures were conducted using the symmetric Somer’sdstatistic. Bivariate associations were assessed using Spearman’s Rho. Maintaining a maximum alpha of .05 was used for making determinations of statistical significance.