Introduction:

Non-muscle invasive bladder cancer (NMIBC) has high rates of recurrence and low risk of progression. However, the natural history of NMIBC is heterogeneous with significant variation depending on stage and grade at diagnosis. Currently available risk calculators provide broad estimates of any recurrence and progression to muscle-invasive disease, but are not able to predict the stage and grade of the recurrence. Understanding the patterns of recurrence on a more granular level may enable risk-stratified NMIBC treatment and surveillance. The objective of our study was to predict the risk of multiple levels of recurrence and progression outcomes across NMIBC stage and grade in a large cohort from two community-based health systems.

Methods:

The analysis cohort consisted of 2956 NMIBC (stage <T2) patients from Kaiser Permanente Northwest (KPNW) and Geisinger diagnosed between 1994-2015. Data were derived from tumor registries, electronic health records, membership files, and pathology records.  Recurrences (n=1062) were identified using a computer algorithm and all pathology reports were annotated for stage and grade by two urologic oncologists (TG and MN).  We defined initial NMIBC diagnosis as a composite stage and grade variable as follows: papillary urothelial neoplasm of low malignant potential (PUNLMP), Ta low grade, Ta high grade, Tis or Ta with carcinoma in situ (CIS), T1 low grade, and T1 high grade.  Recurrences were defined as low risk (PUNLMP or Ta low grade), intermediate risk (Ta high grade, CIS, T1 low grade), or high risk (T1 high grade), or progression to muscle invasion. We defined four outcomes as follows: any recurrence, intermediate risk recurrence or higher, high risk recurrence or higher, and progression to muscle invasion. Separate Cox proportional hazards regressions were performed to determine the association between each initial stage/grade category and time to each ordered outcome.  Multivariable models were adjusted for age, sex, health system (KPNW or Geisinger), initial tumor size (<3cm or >=3cm), initial number of tumors (unifocal vs. multifocal), perioperative intravesical chemotherapy (yes/no), and induction intravesical therapy (yes/no). We also computed 1- and 5-year risk estimates and associated 95% confidence intervals (CI) based on these models.

Results:

Mean age at diagnosis was 69 years and the cohort was predominantly white (98.5%), male (79%), and from KPNW (66%).  At initial diagnosis, nearly half (1420) of tumors were Ta low grade, 78% were unifocal, and 77% were less than 3cm in size.  Over a median follow up of 29.4 months, there were 1062 recurrences (35.9% of the cohort) and a total of 111 patients had progression to muscle-invasion (3.8%). PUNLMP was associated with a decreased risk of any recurrence as compared to Ta low grade (adjusted HR 0.72, 95% CI 0.56-0.92, Table 1).  The adjusted hazard of high risk recurrence (T1 high grade) or progression increased with increasing diagnosis stage/grade from Ta high grade (adjusted HR 2.60, 95% CI 1.62-4.15), Tis or Ta with CIS (adjusted HR 4.74, 95% CI 3.01-7.47), and T1 high grade (adjusted HR 7.14, 95% CI 4.97-10.26). Table 2 describes the predicted 1- and 5-year risks of each outcome by stage and grade at diagnosis.  For example, the predicted risk of a high risk recurrence (T1 high grade) or progression to muscle-invasive disease for PUNLMP was 0.3% at 1 year and 2% at 5 years.  For Ta high grade, the predicted risk of a high risk recurrence or progression was 4% at 1 year and 10% at 5 years.

Conclusion:

In this large cohort of NMIBC patients from two community-based health systems, we assessed the predicted risks of new clinically relevant definitions of recurrence that incorporate stage and grade of the recurrence. To our knowledge, this is also one of the largest existing series of long-term PUNLMP outcomes. We found that the 1- and 5- year predicted risk of high risk recurrences (T1 high grade) and progression to muscle-invasive disease increased with higher composite stage and grade at diagnosis.  The predicted risks of multilevel NMIBC recurrence and progression outcomes may be useful for counseling patients and for designing risk-stratified surveillance schedules.

Funding: NIH NCI 1R21CA191610-01 (PI: McMullen & Nielsen)