Development and prospective validation of a novel weighted quantitative scoring system aimed at predicting the pathological features of cystic renal masses

Eur Radiol. 2019 Apr;29(4):1809-1819. doi: 10.1007/s00330-018-5722-6. Epub 2018 Oct 11.

Abstract

Objectives: To develop and prospectively validate a novel weighted quantitative scoring system based on CT findings, namely, the renal cyst index (RCI), aimed at preoperatively predicting the pathological features of cystic renal masses (CRMs).

Methods: The RCI was based on four critical features of CRMs: the cyst wall, septal, nodule, and cyst contents. These parameters were scored with 1, 2, or 3 points. Weight coefficients for these parameters were determined by the multivariable logistic regression. The odds ratio (OR) and 95% confidence interval (95% CI) were used to summarise the results. The RCI was defined as the sum of these four weight coefficients. Malignancy risk prediction models were built based on the retrospective evaluation of 441 patients. We also compared the prediction ability of the RCI with the Bosniak classification in the 441 patients and applied these novel models to 152 masses resected in our institution to prospectively validate the efficiency of the RCI.

Results: The wall point (OR = 5.71 [95% CI = 1.734-18.808, p = 0.004, point = 2], OR = 12.665 [95% CI = 3.750-42.770, p < 0.001, point = 3]), septal point (OR = 3.325 [95% CI = 1.272-8.692, p = 0.014, point = 3]), nodule point (OR = 4.588 [95% CI = 1.429-14.729, p < 0.001, point = 2], OR = 17.032 [95% CI = 5.017-57.820, p = 0.010, point = 3]), content point (OR = 22.822 [95% CI = 1.041-495.995, p = 0.047, point = 2], OR = 2.723 [95% CI = 1.296-10.696, p = 0.015, point = 3]), and RCI (OR = 1.247 [95% CI = 1.197-1.299, p < 0.001]) were significantly associated with malignancy. Masses with an RCI < 6 were regarded as benign masses; masses with an RCI ≥ 10 were regarded as malignant masses. The malignancy risk of masses with an RCI > 6 but < 10 were determined by a nomogram. The prediction ability of the RCI was significantly superior to the Bosniak classification for Bosniak IIF and III masses (AUC: 0.912 vs. 0.753, p = 0.001). The RCI also accurately predicted the pathological features of 152 masses.

Conclusion: The RCI is a reliable quantitative scoring system in predicting the malignancy risk of CRMs, and it outperformed the Bosniak classification system in some ways.

Key points: • The renal cyst index (RCI) is a useful weighted quantitative classification system based on CT findings for diagnosing cystic renal masses. • The RCI outperforms the Bosniak classification system in some ways, especially for Bosniak IIF and III masses. • Masses with an RCI < 6 can be regarded as a simple cyst, while those with an RCI > 10 can be regarded as malignant masses.

Keywords: Cystic kidney diseases; ROC curve; Renal cell carcinoma; Tomography, x-ray computed.

Publication types

  • Validation Study

MeSH terms

  • Adult
  • Aged
  • Carcinoma, Renal Cell / classification
  • Carcinoma, Renal Cell / diagnostic imaging*
  • Carcinoma, Renal Cell / pathology*
  • Female
  • Humans
  • Kidney Diseases, Cystic / classification
  • Kidney Diseases, Cystic / diagnostic imaging*
  • Kidney Diseases, Cystic / pathology*
  • Kidney Neoplasms / classification
  • Kidney Neoplasms / diagnostic imaging*
  • Kidney Neoplasms / pathology*
  • Logistic Models
  • Male
  • Middle Aged
  • Nomograms*
  • Odds Ratio
  • Prospective Studies
  • Retrospective Studies
  • Tomography, X-Ray Computed*