Sarcoma tumor size (T) staging: Are radiology or pathology measurements more appropriate?
e22522 Background: In the AJCC version 7 TNMG staging of soft tissue sarcomas (STS), the longest dimension (1D) of the primary tumor at pathology analysis is the gold standard for tumor size (T) staging. However, measurements may differ between scans and actual tissue measurement, due to tissue elasticity, deformation, and/or formalin fixation. Thus, tumor size may change compared to preoperative imaging. Should STS T stage be defined by imaging or by direct tumor measurement? We examined the variability of the measurements between radiology and pathology data, examining 1D, cross sectional area (2D), and tumor volume (3D) to assign a T stage. Methods: We reviewed all patients (pts) with extremity STS who had surgery at Mt Sinai Hospital New York (2010-2015). MRI or CT and resected gross tumor measurements were available for 79 pts. After eliminating 10 samples with grossly irregular shapes and 11 samples with incomplete data, 58 tumors had complete 3D measurements. We calculated Pearson correlation coefficients for paired variables (radiology vs pathology size in 1D, 2D and 3D). Results: Imaging measures correlated well with direct tumor measurements. Pearson correlation coefficients for 1D, 2D and 3D measurements were 0.93, 0.72 and 0.78, respectively (all p < 0.01). The SEM (radiology/pathology size) = 0.13, i.e. 13% for 1D measurements. Thus, T stage could be incorrectly assigned in up to 13% of samples near 5 cm in size; this proportion decreases the further the tumor is from the 5 cm cutoff. Conclusions: As shown previously in the rationale for RECIST, 1D measures provide the smallest variance between radiology and pathology. The situation is made more complex in AJCC version 8, with four T categories. It remains unclear how to stage the 10-15% of primary STS with irregular shapes. These data support the use of nomograms for risk assessment rather than using bins created by the AJCC tumor staging system. 3D tumor volumes, if used at all, may play a greater role when assessing therapy responses or contending with tumors of irregular shape, rather than for routine STS staging.