High-grade soft-tissue sarcomas: tumor response assessment--pilot study to assess the correlation between radiologic and pathologic response by using RECIST and Choi criteria

Radiology. 2009 May;251(2):447-56. doi: 10.1148/radiol.2512081403. Epub 2009 Mar 4.


Purpose: To compare radiologic response as defined according to both Response Evaluation Criteria in Solid Tumors (RECIST) and the new Choi criteria recently proposed for gastrointestinal stromal tumors with pathologic response in high-grade soft-tissue sarcomas (STSs) treated with preoperative chemotherapy and radiation therapy.

Materials and methods: The institutional ethical committee approved the trial in which patients were enrolled. Signed informed consent was obtained. Thirty-seven patients (21 men, 16 women; mean age, 44.2 years) enrolled in a collaborative randomized trial on preoperative chemotherapy and radiation therapy in localized high-risk STS at a single institution were selected for this retrospective analysis. Tumor response to preoperative treatment was assessed by using both RECIST and Choi criteria at computed tomography (CT) and was adapted to be used at magnetic resonance (MR) imaging. Pathologic response was assessed as either good or very good. Sensitivity, specificity, and predictive value of RECIST and Choi criteria were calculated with pathologic response as the reference standard and were reported with 95% confidence intervals.

Results: For 28 patients without synovial sarcomas, sensitivity of RECIST versus adapted Choi criteria was 32.0% versus 88.0% for good response and 41.2% versus 82.4% for very good response, respectively; specificity for pathologic response was 100% versus 100% for not a good response and 90.9% versus 27.3% for not a very good response, respectively. In synovial sarcoma, the nontreatment-related neoplastic cystic component of the tumor was a major obstacle for both RECIST and Choi criteria.

Conclusion: In STS treated with chemotherapy and radiation therapy, tumor size may be insufficient to render actual tumor response. Tumor attenuation at CT or tumor contrast material enhancement at MR imaging may complement tumor size, thus making Choi criteria more predictive of pathologic response.

Publication types

  • Comparative Study
  • Randomized Controlled Trial

MeSH terms

  • Adolescent
  • Adult
  • Aged
  • Biopsy
  • Female
  • Gastrointestinal Stromal Tumors / diagnosis*
  • Gastrointestinal Stromal Tumors / therapy*
  • Humans
  • Image Interpretation, Computer-Assisted / methods*
  • Magnetic Resonance Imaging / methods*
  • Male
  • Middle Aged
  • Outcome Assessment, Health Care / methods
  • Pilot Projects
  • Prognosis
  • Reproducibility of Results
  • Sarcoma / diagnosis*
  • Sarcoma / therapy*
  • Sensitivity and Specificity
  • Statistics as Topic
  • Tomography, X-Ray Computed / methods*
  • Treatment Outcome
  • Young Adult