Does size reliably predict malignancy in soft tissue tumours?

PURPOSE: In this retrospective study we examined whether size is a viable marker of tumour malignancy in soft tissue masses (STM) and if the ratio of width and length (RALD) of an STM reflects tumour biology more accurately. METHODS: Measurements of maximal lesion size and perpendicular diameter were performed in available MRI and ultrasonography studies of 212 patients (mean age 54.4 ± 17.2 years, male:female 1:1.12) with a histologically verified diagnosis. RESULTS: Overall, 28.2 % of lesions were malignant, 11.1 % intermediate, and 58.8 % benign. Size alone was a weak predictor of malignancy in STMs (sensitivity 68.8 %, specificity 50.3 %, positive predictive value [PPV] 44.0 %, negative predictive value [NPV] 80.4 %). RALD showed better discriminatory power with greater separation between benign and malignant entities and higher values for sensitivity (83.6 %), specificity (53.6 %), and NPV (89.0 %). A weighted combination of size, age and RALD improved diagnostic power, demonstrating higher values for sensitivity (77.0 %), specificity (80.1 %), PPV (61.0 %), and NPV (89.6 %). CONCLUSIONS: Size should not be used alone to estimate an STM's malignancy. RALD better reflects a lesion's growth pattern and a combination of age, size, and RALD helps to discriminate more accurately between benign, intermediate, and malignant entities. These findings should help to estimate easily whether a newly found STM is benign or malignant prior to further workup. KEY POINTS: •Size does not reliably differentiate between benign, intermediate, and malignant tumours •The R ALD (ratio of lateral to axial diameter) improves diagnostic confidence •When combined with age and size, STM differentiation was further enhanced •These measurements can aid in earlier detection of sarcomas.