(P 462) EXTERNAL VALIDATION OF A NOMOGRAM PREDICTING THE RISK OF POSTOPERATIVE MAJOR WOUND COMPLICATION IN SOFT TISSUE SARCOMA FOLLOWING PREOPERATIVE RADIOTHERAPY AND RESECTION

Objective: Standard of care for soft tissue sarcoma (STS) is wide surgical resection combined with external beam radiation therapy (EBRT) and occasional chemotherapy. Unfortunately, postoperative surgical site complications and major wound complications (MWCs) are common in the setting of STS and are exacerbated by EBRT (particularly preoperative EBRT). While MWC have not been shown to affect long term recurrence or overall survival, these complications do contribute to overall morbidity.

Significant work has been done to identify risk factors associated with the development of MWC. Nomograms are useful disease-specific prognostic models that can be utilized to predict clinical outcomes. In 2022, Ouyang and colleagues developed a nomogram designed to predict the risk of MWC following preoperative EBRT and resection of STS of the trunk and extremity.1 In this cohort, 126 patients at a single institution met inclusion criteria and 19% experienced MWC. Age, tumor size, and metastasis at presentation were identified as independent risk factors for MWC. A nomogram was generated with a receiver operating characteristic curve (ROC) area under the curve (AUC) value of 0.86, suggesting that this model effectively predicts MWC. To date, this nomogram has not been externally validated.

The purpose of this study was to externally validate the nomogram generated by Ouyang and colleagues.

Methods: Institutional Review Board approval was obtained for this study. We retrospectively reviewed all patients at a single sarcoma center who underwent preoperative radiation followed by surgical resection for STS of the extremity or trunk between 2008 and 2022. Model variables were those proposed by the Ouyang nomogram: sex, age, body mass index, diagnosis of diabetes, presence of metastases on presentation, type of surgical closure, and tumor characteristics (site, size, volume, depth, grade, histological subtype). MWC were identified and were defined as any of the following within four months of initial resection: formal wound debridement in the operating room, procedural interventions such as percutaneous drainage, treatment with intravenous antibiotics, or deep wound packing for more than 120 days. Model discrimination (ROC and AUC), calibration, and overall performance (Brier score) were used to assess the model’s performance.

Results: Two hundred seventy-nine patients were treated with preoperative radiation followed by surgical resection for management of STS of the extremity or trunk. Of these patients, 104 had MWC (37.3%).

The AUC was 0.633, suggesting overall poor discriminatory ability. The calibration curve demonstrated poor overlap between the model predictions and actual patient outcomes. Figures 1 and 2

A Brier score is a measure of how accurate predictions are compared to actual outcomes. A Brier score is on a scale of 0 to 1, and lower values are indicative of better model prediction performance. The Brier score for this model was 0.348, indicating poor model performance.

Conclusion: In external validation by a larger, independent cohort, the nomogram proposed by Ouyang performed significantly worse than on the training dataset. Additional work is needed to develop models predicting MWCs following STS resection.

Works Cited:
Ouyang, Z. et al. Nomogram Predicting the Risk of Postoperative Major Wound Complication in Soft Tissue Sarcoma of the Trunk and Extremities after Preoperative Radiotherapy. Cancers 14, 4096 (2022).