Automated evaluation of masseter muscle volume: deep learning prognostic approach in oral cancer

Abstract

Background: Sarcopenia has been identified as a potential negative prognostic factor in cancer patients. In this study, our objective was to investigate the relationship between the assessment method for sarcopenia using the masseter muscle volume measured on computed tomography (CT) images and the life expectancy of patients with oral cancer. We also developed a learning model using deep learning to automatically extract the masseter muscle volume and investigated its association with the life expectancy of oral cancer patients.

Methods: To develop the learning model for masseter muscle volume, we used manually extracted data from CT images of 277 patients. We established the association between manually extracted masseter muscle volume and the life expectancy of oral cancer patients. Additionally, we compared the correlation between the groups of manual and automatic extraction in the masseter muscle volume learning model.

Results: Our findings revealed a significant association between manually extracted masseter muscle volume on CT images and the life expectancy of patients with oral cancer. Notably, the manual and automatic extraction groups in the masseter muscle volume learning model showed a high correlation. Furthermore, the masseter muscle volume automatically extracted using the developed learning model exhibited a strong association with life expectancy.

Conclusions: The sarcopenia assessment method is useful for predicting the life expectancy of patients with oral cancer. In the future, it is crucial to validate and analyze various factors within the oral surgery field, extending beyond cancer patients.

Keywords: Deep learning; Oral cancer; Sarcopenia.

Fig. 1

Schematic representation of SegResNet Adapted and partially modified from Andriy M. Springer: 311–320, 2018. Schematic overview of the deep learning architecture. The encoder part consists of normalization by group, a rectified linear unit, and 3 × 3 × 3 convolution, and the initial number of filters is 16. The decoder part consists of an upsizing and 1 × 1 × 1 convolution. The segmentation map is output with the same spatial size as the input image, and the input image is reconstructed. The input images are compressed to 128 × 128 × 128 voxel and are used as the network input

Fig. 2

Overall survival by masseter muscle volume (MMV) Both men and women had significantly lower overall survival rate in the low MMV group (men: hazard ratio [HR] = 0.598; 95% confidence interval [CI], 0.438–0.726; p < 0.001; women: HR = 0.616; 95% CI, 0.433–0.755; p < 0.001)

Fig. 3

Comparison between masseter muscle volume (MMV) and artificial intelligence MMV (AIMMV)

Fig. 4

Overall survival based on the artificial intelligence masseter muscle volume In the low AIMMV group, both men and women had significantly lower overall survival rate (men: HR = 0.690; 95% CI, 0.547–0.795; p < 0.001; women: HR = 0.746; 95% CI, 0.611–0.840; p = 0.013)