Background: Internal hemipelvectomy is a challenging procedure often complicated by wound healing issues. To guide surgical planning and reduce risk of wound complications, we created and validated a preoperative risk stratification model. Study Design: We reviewed clinical, surgical, and pathological data from 411 patients who underwent en bloc internal hemipelvectomy for malignant pelvic tumors between 1996 and 2018. The sample included a training set (N=123) from a U.S. tertiary hospital and a validation set (N=288) from centers in Canada and the Netherlands. Using logistic regression, we identified key predictors of wound complications requiring surgical intervention and developed a risk scoring system incorporating Enneking-Dunham resection type and tumor anatomic compartment extension. Patients were stratified into three risk categories: low (0–8 points), intermediate (9–13 points), and high (≥14 points). Results: Wound complications requiring surgery occurred in 34% of the training cohort and 35% of the validation cohort. In the training cohort, rates of wound complications were 7.4%, 40%, and 67% in the low-, intermediate-, and high-risk groups, respectively (p<0.001); rates were similar in the validation cohort (7.1%, 32%, and 49%; p<0.001). The risk model demonstrated fair-to-good discriminative performance; the areas under the receiver operating characteristic curves were 0.83 (training cohort) and 0.71 (validation cohort) despite heterogeneity in tumor histology, reconstruction techniques, and perioperative management. Conclusions: We present the first externally validated prognostic staging system for predicting wound complications after internal hemipelvectomy. The model is simple, reproducible, and clinically applicable, allowing surgeons to preoperatively identify high-risk patients who may benefit from modified surgical approaches or prophylactic soft-tissue reconstruction. Further prospective studies are warranted to optimize management strategies for intermediate- and high- risk patients. Level of evidence: Prognostic studies, Level III
Ogura et al. (Fri,) studied this question.