Recurrence and progression in patients with non-muscle invasive bladder cancer: prognostic models including multicolor fluorescence in situ hybridization molecular grading.

OBJECTIVE

To test the prognostic value of multicolor fluorescence in situ hybridization analyses of tumor cells in urine for prediction of the recurrence and progression of tumor in patients with intermediate risk non-muscle invasive bladder cancer.

METHODS

A total of 168 patients with non-muscle invasive bladder cancer were included in the study. Fluorescence in situ hybridization was carried out on the bladder wash urine collected before resection. Tumors were classified as low molecular grading if they had a diploid chromosomal pattern or only a loss of p16 or ch3 aneuploidy, and as high molecular grading if they showed aneuploidy of ch7 or 17. Cox regression models assessed the added prognostic value of fluorescence in situ hybridization for primary tumor recurrence or progression, respectively.

RESULTS

Median follow up was 67 months. A total of 57% of tumors were classified as low molecular grading. The 2- and 5-year recurrence-free survival was 68% and 49% for low molecular grading, and 47% and 30% for high molecular grading, respectively. The 2- and 5-year progression-free survival was 95% and 84% for low molecular grading, and 79% and 58% for high molecular grading tumor patients, respectively. Molecular grading (hazard ratio 1.60; P = 0.03) was associated with recurrence, when also accounting for histopathology and a patient's characteristics. Both cancer severity score (hazard ratio 1.51; P < 0.01) and molecular grading (hazard ratio 2.53; P < 0.01) independently and positively predicted progression in multivariable models. The C-index for predicting recurrence increased from 0.58 to 0.61 when molecular grading fluorescence in situ hybridization was included in the model, and from 0.68 to 0.72 when predicting progression.

CONCLUSIONS

Fluorescence in situ hybridization-based molecular grading increases the accuracy of a prognostic model, predicting both recurrence and progression in patients with intermediate risk non-muscle invasive bladder cancer.