Predicting renal cryoablation complications: New risk score based on tumor size and location and patient history

Grant D. Schmit, Louis A. Schenck, R. Houston Thompson, Stephen A. Boorjian, A. Nicholas Kurup, Adam J. Weisbrod, Daryl J. Kor, Matthew R. Callstrom, Thomas D. Atwell, Rickey E. Carter

Research output: Contribution to journalArticle

34 Scopus citations

Abstract

Purpose: To identify tumor and patient-related risk factors for major complications following renal cryoablation and to develop a model for predicting these adverse events. Materials and Methods: Institutional review board approval and informed patient consent were obtained for this HIPAA-compliant retrospective study. All 398 renal cryoablation procedures performed from 2003 through 2011 were reviewed to identify tumor and patient-related risk factors associated with major complications (Clavien-Dindo classification, ≥ grade III). A scoring system for predicting these adverse events was then developed using risk factor weighting obtained from a multivariate logistic regression model. To internally validate this model, the scoring system was then applied to all 73 renal cryoablation procedures performed during 2012. Results: Among tumor-related factors evaluated, Maximal tumor diameter (P = .0006) and Central tumor location (P = .02) were significantly associated with major complications. Among patient-related factors evaluated, prior Myocardial infarction (MI) (P = .002) and Complicated diabetes mellitus (P = .01) were significantly associated with major complications. This resulted in the (MC)2 risk scoring system, with (MC)2 risk score = 2.5 points (for tumors ≤ 2.5 cm in maximal diameter) or 0.1 points for each millimeter of maximal tumor diameter (for tumors > 2.5 cm) + 1.5 points (if central tumor location) + 2.5 points (if patient history of prior MI) + 3.0 points (if patient history of complicated diabetes). Mean (MC)2 risk score for all renal cryoablations was 4.7 (standard deviation, 1.9; range, 2.5-15.3). The observed major complication rates were 2.0% (95% confidence interval [CI]: 0.6%, 4.6%) in the low-risk group (score < 5.0), 12.8% (95% CI: 7.5%, 19.9%) in the moderate-risk group (score of 5.0-8.0), and 39.1% (95% CI: 19.7%, 61.5%) in the high-risk group (score > 8.0). Application of the (MC)2 scoring system to the validation group yielded a concordance index of 0.82 (95% CI: 0.62, 1.00). Conclusion: The results of this study suggest that the (MC)2 risk score is a valuable tool for predicting major complications in patients undergoing renal cryoablation. However, external validation is warranted.

Original languageEnglish (US)
Pages (from-to)903-910
Number of pages8
JournalRadiology
Volume272
Issue number3
DOIs
StatePublished - Sep 2014

    Fingerprint

ASJC Scopus subject areas

  • Radiology Nuclear Medicine and imaging

Cite this

Schmit, G. D., Schenck, L. A., Thompson, R. H., Boorjian, S. A., Kurup, A. N., Weisbrod, A. J., Kor, D. J., Callstrom, M. R., Atwell, T. D., & Carter, R. E. (2014). Predicting renal cryoablation complications: New risk score based on tumor size and location and patient history. Radiology, 272(3), 903-910. https://doi.org/10.1148/radiol.14132548