P1.06-004 Evaluating the Non-Small Cell Lung Cancer Symptom Assessment Questionnaire (NSCLC-SAQ): Preliminary Results from the Quantitative Pilot Study

Topic: Advanced General


      In collaboration with the US Food and Drug Administration (FDA), the Patient-Reported Outcome (PRO) Consortium’s NSCLC Working Group has developed the 7-item NSCLC-SAQ. Content includes cough, pain (2 items), shortness of breath, fatigue (2 items), and appetite. A quantitative pilot study is underway to evaluate the NSCLC-SAQ’s item-level and scale-level performance.


      Eligible subjects with clinically-diagnosed advanced NSCLC from US-based clinical sites completed a questionnaire battery (demographics, NSCLC-SAQ, NCCN/FACT Lung Symptom Index-17 [FLSI-17], Patient Global Impression of Severity [PGIS]) using a tablet computer. For this interim analysis, items were evaluated for response distribution, ceiling/floor effects, missing data, and item-to-item correlations. Exploratory factor and Rasch analyses were examined. Internal consistency reliability was estimated using Cronbach’s coefficient alpha. Construct validity was assessed with Pearson correlations between the NSCLC-SAQ and FLSI-17 Disease-Related Symptom (DRS) subscale. The PGIS was used to assess the NSCLC-SAQ’s known-groups validity.


      For this interim analysis, 117 (of the anticipated 150) subjects from nine sites were included. Subjects’ mean age was 64 years old (range 40-85), 55% were female, and 84% were white (non-Hispanic), ECOG performance status at enrollment was: 0 (33%), 1 (50%), and 2 (17%). NSCLC staging was: Stage IIIB (15%) and IV (85%). A total of 34% were treatment-naïve, 32% had received first-line treatment only, and 34% had received second- or third-line treatment. Mean scores for the 7 items of the NSCLC-SAQ ranged from 0.9 to 2.2 using a response scale between 0 (“No <symptom> at all” or “Never”) to 4 (“Very Severe <symptom>” or “Always”). Subjects used the full range of responses (i.e., 0, 1, 2, 3, and 4) and provided answers for all NSCLC-SAQ items. Rasch analyses showed the items were ordered and the person-to-item distribution was acceptable. Factor analysis indicated a single component accounting for 47% of the variance, which supports a unidimensional scale structure and the ability to calculate a total symptom score. Cronbach’s alpha was 0.80. The NSCLC-SAQ total symptom score correlated highly (r=0.87) with the FLSI-17 DRS and was able to discriminate levels of overall symptom severity as assessed by the PGIS (p<0.001).


      The NSCLC-SAQ has been developed in accordance with the FDA’s PRO Guidance. This study provides quantitative evidence of adequate item and scale performance. These data will be submitted to the FDA to support qualification of the NSCLC-SAQ as a measure to assess a symptom endpoint for efficacy evaluation and product labeling.


      symptom assessment, NSCLC-SAQ, non-small cell lung cancer