Purpose: Unplanned excisions of soft tissue sarcomas (STS) occur due to inadequate suspicion for malignancy, resulting in higher local recurrence rates despite subsequent tumor bed excision (TBE) and oncologic management. MRI findings may influence the need for adjuvant therapy and aid definitive surgical planning. We sought to identify MRI features associated with presence of residual disease after unplanned excision, focusing on disease that was measurable but small (less than 5 cubic cm).
Materials and Methods: IRB approved review of medical records identified patients referred to our institution following unplanned excision of STS from January 2013 to August 2015. The study group consisted of those patients who underwent subsequent TBE at our institution and had MRI of the excision site prior to TBE. Patients with gross residual tumor (greater than 5 cubic cm) were excluded. MRI studies were evaluated to determine depth of tissue involvement, presence of nodular enhancement, and maximum length of soft tissue edema. In order to reduce the influence of the body part size on the measure of edema, a normalized measure of edema was obtained by dividing the maximum length of edema by the sum of the anterior-posterior and transverse dimensions of the nearest long bone. Results were correlated with histology from unplanned excision and TBE.
Results: 30 patients underwent tumor bed excision with MRI prior to TBE. Of these patients, 8 had gross residual tumor and were excluded, leaving 22 patients (13 males and 9 females; mean age 57). 18/22 tumors were grade 2 or 3. STS were mainly located in the extremities (12 in lower extremities, 8 in upper extremities and 2 in the lower back). Involvement of deep compartments in MRI was present in 4/22 cases. Original margins were positive in 17/22 unplanned excisions with inadequate margin assessment in the remaining 5 cases. Pathology proven residual tumor was present at TBE in 10/22 cases, 3 of which involved deep compartments. Residual tumor was present in 4/7 cases with nodular enhancement and 6/15 cases with absent nodular enhancement (sensitivity=40%; specificity=75%; PPV=57%; NPV=60%). Mean length of soft tissue edema was greater among patients with residual tumor than those without, although this difference was not considered significant (two sample t-test; n=22; mean difference=4.3 cm; p=0.07). However, the presence of measurable tumor was associated with significantly greater amounts of edema when normalized to body part size (two sample t-test; n=18; mean difference=0.11; p=0.01). There was no association between extent of soft tissue edema and neoadjuvant therapy or time interval to MRI from unplanned excision.
Conclusion: Low positive predictive value of nodular enhancement for residual tumor suggests that postsurgical inflammatory changes or scar tissue frequently mimic residual tumor following unplanned excision. Extent of soft tissue edema may be a predictor of residual tumor. Greater soft tissue edema on MRI after unplanned excision can potentially demonstrate the need for adjuvant therapy and definitive tumor bed excision given an increased likelihood of residual disease presence.