Can Renal Mass Biopsy Assessment of Tumor Grade be Safely Substituted for by a Predictive Model?

Claudio Jeldres, Maxine Sun, Daniel Liberman, Giovanni Lughezzani, Alexandre de la Taille, Jacques Tostain, Antoine Valeri, Luca Cindolo, Vincenzo Ficarra, Walter Artibani, Richard Zigeuner, Arnaud Mejean, Jean Luc Descotes, Eric Lechevallier, Peter F. Mulders, Paul Perrotte, Jean Jacques Patard, Pierre I. Karakiewicz

Research output: Contribution to journalArticlepeer-review

50 Scopus citations

Abstract

Purpose: Fuhrman grade represents a key determinant of the natural history of small renal masses that represent renal cell carcinoma. We tested whether renal mass biopsy prediction of Fuhrman grade in the nephrectomy specimen could be safely substituted for by an accurate statistical model. To date the best available model has shown poor accuracy (55.6%), which is close to flipping a coin (50%) and clearly inadequate for use in clinical practice. Materials and Methods: We identified 1,139 patients with T1aN0M0 renal cell carcinoma treated with partial or radical nephrectomy at 11 participating institutions from 1989 to 2004. This cohort was used in univariate and multivariate logistic regression models predicting high Fuhrman grade (III-IV) at nephrectomy. Predictors included age at diagnosis, gender, tumor size and symptom classification. Multivariate logistic regression coefficients were used to generate a nomogram. Results: The rate of Fuhrman grade III-IV in patients with T1aN0M0 renal cell carcinoma was 12.3%. Stratifying patients with Fuhrman grade III-IV by age, gender, histological subtypes and sample size failed to reveal statistically significant differences. On univariate analysis predicting Fuhrman grade III-IV at nephrectomy only tumor size was a statistically significant predictor (p = 0.05). The most accurate multivariate nomogram for Fuhrman grade III-IV prediction was 58.3% (95% CI 57.8-58.9) accurate. Of all tested predictors only tumor size achieved independent predictor status (p = 0.009). Conclusions: Our analysis derived in European patients shows that statistical models cannot safely replace renal mass biopsy based prediction of Fuhrman grade III-IV at nephrectomy. Our findings corroborate a report from the United States in which a similar model had 55.6% accuracy. Jointly the studies indicate that statistical models are unreliable and cannot safely be substituted for renal mass biopsy in North American or European patients.

Original languageEnglish (US)
Pages (from-to)2585-2589
Number of pages5
JournalJournal of Urology
Volume182
Issue number6
DOIs
StatePublished - Dec 2009
Externally publishedYes

Keywords

  • Europe
  • biopsy
  • carcinoma
  • kidney
  • models
  • renal cell
  • statistical

Fingerprint

Dive into the research topics of 'Can Renal Mass Biopsy Assessment of Tumor Grade be Safely Substituted for by a Predictive Model?'. Together they form a unique fingerprint.

Cite this