CT-assessed pectoralis major muscle measurements as risk factors for low neoadjuvant chemotherapy efficacy in osteosarcoma: a retrospective study.

BACKGROUND

Neoadjuvant chemotherapy (NAC) response in terms of tumor necrosis for osteosarcoma is of clinical significance for subsequent treatment, CT-acquired parameters indicating sarcopenia were reported to be the risk factors of low NAC efficacy in a variety of malignant tumors, yet the association between chest CT-derived pectoralis major muscle measurements and tumor necrosis of osteosarcoma have not been revealed.

METHODS

The pectoralis major muscle area, normalized pectoralis major muscle index (PMI) and CT attenuation acquired from preoperative chest CT as well as clinical data of 131 osteosarcoma patients were retrospectively analyzed and compared between favorable NAC response (tumor necrosis ≥ 90%) and poor NAC response group (tumor necrosis < 90%). For the evaluation of interobserver reproducibility, interclass correlation coefficient (ICC) with 95% confidence interval (CI) was used. The receiver operating characteristic (ROC) analysis was performed to determine the efficiency of each measurement and to calculate the optimal cut-off values of significant parameters. Multivariable binary logistic regression analysis was performed to determine the risk factors of low NAC efficacy.

RESULTS

A total of 79 lesions from 78 patients were eligible for this study (mean age, 17.6 ± 0.9 years; 45 male), 34 lesions constituted the favorable response (FR) group and 45 lesions constituted the poor response (PR) group. Pectoralis major muscle measurements were significantly higher in FR group than in PR group (muscle area, 11.56 ± 4.48 VS 9.42 ± 3.57, P = 0.021; PMI, 7.32 ± 2.23 VS 5.64 ± 1.54, P < 0.001; CT attenuation, 38.78 ± 8.44 VS 31.96 ± 11.37, P = 0.004). Excellent interobserver agreement was found in these parameters with ICC values ranging 0.925-0.944. In ROC analysis, PMI showed an area under curve (AUC) of 0.727, 95% CI: 0.612, 0.842, the threshold value for differentiating FR from PR was 7.147. PMI were risk factors of low NAC efficacy with the P value of 0.007.

CONCLUSIONS

Chest CT-accessed pectoralis major muscle area, PMI and CT attenuation were significantly correlated with tumor necrosis in osteosarcoma, these can be potentially helpful imaging auxiliary biomarkers for predicting NAC response of osteosarcoma.