Predictive value of magnetic resonance imaging in determining presence of residual disease after marginal excision of unsuspected soft tissue sarcomas of the hand.

PURPOSE

Diagnosis of a soft tissue sarcoma of the hand is at times made only in retrospect after marginal excision of a presumed benign soft tissue mass. Magnetic resonance imaging (MRI) helps determine the presence of residual disease and the extent of contamination caused by marginal excision of unsuspected soft tissue sarcomas and assists in the planning of definitive treatment with surgery and radiotherapy when required. We sought to determine the accuracy of MRI in detecting residual sarcoma after marginal excision.

METHODS

We retrospectively studied case records and imaging studies for all patients with soft tissue sarcomas of the hand evaluated and treated at our institution from 1996 to 2005. We included in this study 33 patients who underwent definitive surgery at our center after prior marginal excision. MRI scans done before definitive wide resection were classified as positive or negative for residual tumor. A musculoskeletal radiologist, blinded to the sarcoma type, findings on histopathology, and surgery, reviewed 19 scans to see whether the accuracy could be improved.

RESULTS

There were 8 epithelioid sarcomas, 6 synovial sarcomas, 4 malignant fibrous histiocytomas, 2 leiomyosarcomas, 2 liposarcomas, 2 myxofibrosarcomas, and one each of 9 other diagnoses. A total of 11 were low-grade and 22 were high-grade tumors, with 4 superficial and 29 deep tumors. Pathology examination after definitive wide resection or partial hand amputation showed that 15 patients had residual tumor, 9 gross and 6 microscopic. The sensitivity of MRI in detecting residual soft tissue sarcoma of the hand was 60%, specificity was 78%, positive predictive value was 69%, and negative predictive value was 70%. The sensitivity of MRI in detecting gross residual soft tissue sarcoma of the hand was 89%, specificity was 79%, positive predictive value was 62%, and negative predictive value was 95%. Even when an experienced musculoskeletal radiologist reassessed 19 MRI scans, the accuracy did not improve.

CONCLUSIONS

Magnetic resonance imaging does not reliably detect residual gross or microscopic soft tissue sarcoma after marginal excision of unsuspected soft tissue sarcomas of the hand, with residual tumor not readily distinguished from postoperative change. The absence of disease on MRI should not be used as the sole criterion in determining whether a repeat resection should be performed.

TYPE OF STUDY/LEVEL OF EVIDENCE: Diagnostic III.