Association of Skeletal Muscle Area with Toxicity and Treatment Tolerance in Patients Receiving Chemoradiation for Advanced Lung Cancer

Abstract

<h3>Purpose/Objective(s)</h3> Significant toxicity is common in the treatment of advanced non-small cell lung cancer (NSCLC) and small cell lung cancer (SCLC) and can be associated with adverse events, such as chemotherapy delay or dose reduction, and worse clinical outcomes. Baseline low skeletal muscle (SM) area is a marker of sarcopenia and has been associated with worse survival in other malignancies, but the association of SM area and toxicity in lung cancer is less studied. <h3>Materials/Methods</h3> Patients with locally advanced or oligo-metastatic NSCLC and extensive-stage SCLC treated with concurrent chemoradiation (CRT) from 2002-2013 at a single institution were reviewed. A deep-learning pipeline utilized existing pre-treatment computed tomography scans to calculate SM area at the 3^rd lumbar vertebral level. Gold standard SM index (SMI) was calculated, adjusting for height, sex, and dichotomized per previously validated cutoff values. Grade 3 or higher (G3+) toxicity, per NCI CTCAE v5.0, was assessed within 21-days of first chemotherapy cycle. Multivariate analysis (MVA) of toxicity endpoints with SMI and baseline characteristics were analyzed by logistic regression analysis, and with overall survival (OS) using Cox regression analysis. <h3>Results</h3> A total of 455 patients met inclusion criteria, with median follow-up of 23.0mo (range 1.0-193.0mo) and median age of 63y (range 29-88y). Among 387 patients with NSCLC, most were clinical stage (AJCC 7th edition) IIIA (43%), IIIB (29%), or IV (11%), while 11% had upfront surgery and adjuvant CRT. Most common chemotherapy regimen was cisplatin-based (48%). Patients with low SMI were more likely to be older (median age 70y vs 62y), ECOG performance status (PS) >0 (75% vs 60%), lower BMI (median BMI 23.3 vs 27.7), and not receive cisplatin-based regimen (39% vs 56%). There was no difference in histology, stage, surgery, or every 3-week (q3w) chemotherapy dosing. On MVA, low SMI was associated with increased risk of G3+ toxicity (OR 1.72, p<0.01) and chemotherapy delay or dose reduction (OR 1.99, p = 0.029). G3+ toxicity was also associated with surgery, but not age, PS, q3w dosing, or regimen. Chemotherapy delay or dose reduction was not associated with age, PS, surgery, and q3w dosing, or regimen. In patients with NSCLC, median OS was 26.2m. G3+ toxicity (HR 1.56, p=0.016), stage 4 (HR 1.280, p=0.003), and cisplatin (HR 1.65, p=0.013) were associated with worse OS on Cox regression, but not low SMI (HR 1.25, p = 0.11), PS, surgery, or regimen. In patients with SCLC, SMI or toxicity were not associated with OS. <h3>Conclusion</h3> Low SMI predicted higher risk of G3+ toxicity during first cycle of chemotherapy in lung cancer. High-risk patients with low SMI experienced significant adverse events and should be considered for more aggressive symptom management or alternative treatment strategies.