Radiological assessment of muscle mass and quality (sarcopenia) in women with breast cancer

Purpose Sarcopenia, the loss of muscle mass, is divided in “primary” or “age-related” and “secondary” when causal factors other than ageing are evident. Cancer is one of the major causes of secondary sarcopenia, associated with negative clinical outcome, even in breast cancer (BC). BC is the most frequently diagnosed cancer in women worldwide. Computer Tomography (CT) is considered the gold standard to evaluate sarcopenia, also in BC patients. Other radiological techniques (Magnetic Resonance Imaging, MRI and Ultrasound, US) can be easily used to assess body composition, especially when CT is not available. The aim of my study was to evaluate with radiological techniques the muscle mass variation in women with BC recruited both prospectively and retrospectively at our University Hospital. A new method to assess muscle mass on breast MRI was developed. The first step was to find if there was a correlation between psoas muscle area (PMA) assessed on CT images and pectoralis muscle area (TPA) assessed on breast MRI. Material and Methods A total of 26 women included in the study was evaluated with both body CT and breast MRI. Reconstructed axial images with both a 1.25-mm and a 5-mm slice thickness and axial T1-weighted images were evaluated by two radiologists to calculate TPA and PMA. Descriptive statistical analysis included inter- and intra-reader agreement and the correlation between TPA on CT and PMA on MRI. Results Comparing axial 5-mm-slice-thickness body CT images and T1-weighted fat-saturated MR images, the Pearson r correlation coefficient was 0.52. Comparing axial 1.25-mm slice thickness body CT images and T1-weighted MR images, the Pearson r (– 1 < r < + 1) correlation coefficient was 0.70 and the coefficient of determination was 0.49, p < 0.05. The inter-reader agreement was almost perfect (0.81–1) for axial 1.25-mm and 5-mm CT images, respectively. The intra-reader agreement of reader 1 was k = 0.98 and k = 0.94 for 1.25-mm and 5-mm CT images, respectively. The intra-reader agreement of reader 2 was 0.95 and 0.94 for 1.25-mm and 5-mm, respectively. On axial pre-contrast T1-weighted images, the inter-reader agreement was 0.61, p < 0.05, considered good (0.61–0.8). Intra-observer agreement of reader 1 and reader 2 for PMA estimation were good (0.62 and 0.64). Conclusion The results demonstrated a strong correlation between PMA assessed on breast MRI and TPA assessed on body CT images. In addition, the technique for measurement of PMA has also been shown to be highly reproducible between different readers.